Fix GitHub issue #3716: gcc-8 warnings

Summary:
Fix the following gcc-8 warnings:
- conflicting C language linkage declaration [-Werror]
- writing to an object with no trivial copy-assignment [-Werror=class-memaccess]
- array subscript -1 is below array bounds [-Werror=array-bounds]

Solves https://github.com/facebook/rocksdb/issues/3716
Closes https://github.com/facebook/rocksdb/pull/3736

Differential Revision: D7684161

Pulled By: yiwu-arbug

fbshipit-source-id: 47c0423d26b74add251f1d3595211eee1e41e54a

CMakeLists.txt

@@ -586,6 +586,7 @@ set(SOURCES
         utilities/leveldb_options/leveldb_options.cc
         utilities/lua/rocks_lua_compaction_filter.cc
         utilities/memory/memory_util.cc
+        utilities/merge_operators/bytesxor.cc
         utilities/merge_operators/max.cc
         utilities/merge_operators/put.cc
         utilities/merge_operators/string_append/stringappend.cc

HISTORY.md

@@ -1,5 +1,32 @@
 # Rocksdb Change Log
-## Unreleased
+## 5.11.4 (03/16/2018)
+### Bug Fixes
+* Fix WAL corruption caused by race condition between user write thread and backup/checkpoint thread.
+
+## 5.11.2 (02/24/2018)
+### Bug Fixes
+* Fix bug in iterator readahead causing blocks to incorrectly be considered truncated (corrupted).
+
+## 5.11.1 (02/22/2018)
+### New Features
+* Follow rsync-style naming convention for BackupEngine tempfiles. This enables some optimizations when run on GlusterFS.
+
+## 5.11.0 (01/08/2018)
+### Public API Change
+* Add `autoTune` and `getBytesPerSecond()` to RocksJava RateLimiter
+
+### New Features
+* Add a new histogram stat called rocksdb.db.flush.micros for memtable flush.
+* Add "--use_txn" option to use transactional API in db_stress.
+* Disable onboard cache for compaction output in Windows platform.
+* Improve the performance of iterators doing long range scans by using readahead.
+
+### Bug Fixes
+* Fix a stack-use-after-scope bug in ForwardIterator.
+* Fix builds on platforms including Linux, Windows, and PowerPC.
+* Fix buffer overrun in backup engine for DBs with huge number of files.
+* Fix a mislabel bug for bottom-pri compaction threads.
+* Fix DB::Flush() keep waiting after flush finish under certain condition.
 
 ## 5.10.0 (12/11/2017)
 ### Public API Change
@@ -9,6 +36,8 @@
 * Provide lifetime hints when writing files on Linux. This reduces hardware write-amp on storage devices supporting multiple streams.
 * Add a DB stat, `NUMBER_ITER_SKIP`, which returns how many internal keys were skipped during iterations (e.g., due to being tombstones or duplicate versions of a key).
 * Add PerfContext counters, `key_lock_wait_count` and `key_lock_wait_time`, which measure the number of times transactions wait on key locks and total amount of time waiting.
+* Support dynamically changing `ColumnFamilyOptions::compaction_options_universal`.
+* Batch update stats at the end of each `Get`, rather than for each block cache access.
 
 ### Bug Fixes
 * Fix IOError on WAL write doesn't propagate to write group follower

INSTALL.md

@@ -107,6 +107,41 @@ to build a portable binary, add `PORTABLE=1` before your make commands, like thi
             * run `brew tap homebrew/versions; brew install gcc48 --use-llvm` to install gcc 4.8 (or higher).
     * run `brew install rocksdb`
 
+* **FreeBSD** (11.01):
+
+    * You can either install RocksDB from the Ports system using `cd /usr/ports/databases/rocksdb && make install`, or you can follow the details below to install dependencies and compile from source code:
+
+    * Install the dependencies for RocksDB:
+
+        export BATCH=YES
+        cd /usr/ports/devel/gmake && make install
+        cd /usr/ports/devel/gflags && make install
+
+        cd /usr/ports/archivers/snappy && make install
+        cd /usr/ports/archivers/bzip2 && make install
+        cd /usr/ports/archivers/liblz4 && make install
+        cd /usr/ports/archivesrs/zstd && make install
+
+        cd /usr/ports/devel/git && make install
+
+
+    * Install the dependencies for RocksJava (optional):
+
+        export BATCH=yes
+        cd /usr/ports/java/openjdk7 && make install
+
+    * Build RocksDB from source:
+        cd ~
+        git clone https://github.com/facebook/rocksdb.git
+        cd rocksdb
+        gmake static_lib
+
+    * Build RocksJava from source (optional):
+        cd rocksdb
+        export JAVA_HOME=/usr/local/openjdk7
+        gmake rocksdbjava
+
+
 * **iOS**:
   * Run: `TARGET_OS=IOS make static_lib`. When building the project which uses rocksdb iOS library, make sure to define two important pre-processing macros: `ROCKSDB_LITE` and `IOS_CROSS_COMPILE`.
 

Makefile

@@ -1553,8 +1553,8 @@ ZLIB_DOWNLOAD_BASE ?= http://zlib.net
 BZIP2_VER ?= 1.0.6
 BZIP2_SHA256 ?= a2848f34fcd5d6cf47def00461fcb528a0484d8edef8208d6d2e2909dc61d9cd
 BZIP2_DOWNLOAD_BASE ?= http://www.bzip.org
-SNAPPY_VER ?= 1.1.7
-SNAPPY_SHA256 ?= 3dfa02e873ff51a11ee02b9ca391807f0c8ea0529a4924afa645fbf97163f9d4
+SNAPPY_VER ?= 1.1.4
+SNAPPY_SHA256 ?= 134bfe122fd25599bb807bb8130e7ba6d9bdb851e0b16efcb83ac4f5d0b70057
 SNAPPY_DOWNLOAD_BASE ?= https://github.com/google/snappy/releases/download
 LZ4_VER ?= 1.8.0
 LZ4_SHA256 ?= 2ca482ea7a9bb103603108b5a7510b7592b90158c151ff50a28f1ca8389fccf6
@@ -1574,7 +1574,7 @@ else
 endif
 endif
 ifeq ($(PLATFORM), OS_FREEBSD)
-	JAVA_INCLUDE += -I$(JAVA_HOME)/include/freebsd
+	JAVA_INCLUDE = -I$(JAVA_HOME)/include -I$(JAVA_HOME)/include/freebsd
 	ROCKSDBJNILIB = librocksdbjni-freebsd$(ARCH).so
 	ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-freebsd$(ARCH).jar
 endif
@@ -1600,7 +1600,7 @@ libz.a:
 		exit 1; \
 	fi
 	tar xvzf zlib-$(ZLIB_VER).tar.gz
-	cd zlib-$(ZLIB_VER) && CFLAGS='-fPIC ${EXTRA_CFLAGS}' LDFLAGS='${EXTRA_LDFLAGS}' ./configure --static && make
+	cd zlib-$(ZLIB_VER) && CFLAGS='-fPIC ${EXTRA_CFLAGS}' LDFLAGS='${EXTRA_LDFLAGS}' ./configure --static && $(MAKE)
 	cp zlib-$(ZLIB_VER)/libz.a .
 
 libbz2.a:
@@ -1612,7 +1612,7 @@ libbz2.a:
 		exit 1; \
 	fi
 	tar xvzf bzip2-$(BZIP2_VER).tar.gz
-	cd bzip2-$(BZIP2_VER) && make CFLAGS='-fPIC -O2 -g -D_FILE_OFFSET_BITS=64 ${EXTRA_CFLAGS}' AR='ar ${EXTRA_ARFLAGS}'
+	cd bzip2-$(BZIP2_VER) && $(MAKE) CFLAGS='-fPIC -O2 -g -D_FILE_OFFSET_BITS=64 ${EXTRA_CFLAGS}' AR='ar ${EXTRA_ARFLAGS}'
 	cp bzip2-$(BZIP2_VER)/libbz2.a .
 
 libsnappy.a:
@@ -1625,7 +1625,7 @@ libsnappy.a:
 	fi
 	tar xvzf snappy-$(SNAPPY_VER).tar.gz
 	cd snappy-$(SNAPPY_VER) && CFLAGS='${EXTRA_CFLAGS}' CXXFLAGS='${EXTRA_CXXFLAGS}' LDFLAGS='${EXTRA_LDFLAGS}' ./configure --with-pic --enable-static --disable-shared
-	cd snappy-$(SNAPPY_VER) && make ${SNAPPY_MAKE_TARGET}
+	cd snappy-$(SNAPPY_VER) && $(MAKE) ${SNAPPY_MAKE_TARGET}
 	cp snappy-$(SNAPPY_VER)/.libs/libsnappy.a .
 
 liblz4.a:
@@ -1638,7 +1638,7 @@ liblz4.a:
 		exit 1; \
 	fi
 	tar xvzf lz4-$(LZ4_VER).tar.gz
-	cd lz4-$(LZ4_VER)/lib && make CFLAGS='-fPIC -O2 ${EXTRA_CFLAGS}' all
+	cd lz4-$(LZ4_VER)/lib && $(MAKE) CFLAGS='-fPIC -O2 ${EXTRA_CFLAGS}' all
 	cp lz4-$(LZ4_VER)/lib/liblz4.a .
 
 libzstd.a:
@@ -1651,7 +1651,7 @@ libzstd.a:
 		exit 1; \
 	fi
 	tar xvzf zstd-$(ZSTD_VER).tar.gz
-	cd zstd-$(ZSTD_VER)/lib && DESTDIR=. PREFIX= make CFLAGS='-fPIC -O2 ${EXTRA_CFLAGS}' install
+	cd zstd-$(ZSTD_VER)/lib && DESTDIR=. PREFIX= $(MAKE) CFLAGS='-fPIC -O2 ${EXTRA_CFLAGS}' install
 	cp zstd-$(ZSTD_VER)/lib/libzstd.a .
 
 # A version of each $(LIBOBJECTS) compiled with -fPIC and a fixed set of static compression libraries
@@ -1750,7 +1750,7 @@ JAVA_C_LIBOBJECTS = $(patsubst %.c.o,jl/%.c.o,$(JAVA_C_OBJECTS))
 JAVA_ASM_LIBOBJECTS = $(patsubst %.S.o,jl/%.S.o,$(JAVA_ASM_OBJECTS))
 endif
 
-java_libobjects = $(patsubst %,jl/%,$(LIBOBJECTS))
+java_libobjects = $(patsubst %,jl/%,$(LIB_CC_OBJECTS))
 CLEAN_FILES += jl
 java_all_libobjects = $(java_libobjects)
 

TARGETS

@@ -242,6 +242,7 @@ cpp_library(
         "utilities/leveldb_options/leveldb_options.cc",
         "utilities/lua/rocks_lua_compaction_filter.cc",
         "utilities/memory/memory_util.cc",
+        "utilities/merge_operators/bytesxor.cc",
         "utilities/merge_operators/max.cc",
         "utilities/merge_operators/put.cc",
         "utilities/merge_operators/string_append/stringappend.cc",

buckifier/buckify_rocksdb.py

@@ -36,7 +36,7 @@ def parse_src_mk(repo_path):
 # get all .cc / .c files
 def get_cc_files(repo_path):
     cc_files = []
-    for root, dirnames, filenames in os.walk(repo_path):
+    for root, dirnames, filenames in os.walk(repo_path):  # noqa: B007 T25377293 Grandfathered in
         root = root[(len(repo_path) + 1):]
         if "java" in root:
             # Skip java

buckifier/rocks_test_runner.sh

@@ -1,4 +1,5 @@
 #!/usr/bin/env bash
 # Create a tmp directory for the test to use
 TEST_DIR=$(mktemp -d /dev/shm/fbcode_rocksdb_XXXXXXX)
+# shellcheck disable=SC2068
 TEST_TMPDIR="$TEST_DIR" $@ && rm -rf "$TEST_DIR"

build_tools/build_detect_platform

@@ -51,11 +51,13 @@ if [ -z "$ROCKSDB_NO_FBCODE" -a -d /mnt/gvfs/third-party ]; then
     FBCODE_BUILD="true"
     # If we're compiling with TSAN we need pic build
     PIC_BUILD=$COMPILE_WITH_TSAN
-    if [ -z "$ROCKSDB_FBCODE_BUILD_WITH_481" ]; then
-      source "$PWD/build_tools/fbcode_config.sh"
-    else
+    if [ -n "$ROCKSDB_FBCODE_BUILD_WITH_481" ]; then
       # we need this to build with MySQL. Don't use for other purposes.
       source "$PWD/build_tools/fbcode_config4.8.1.sh"
+    elif [ -n "$ROCKSDB_FBCODE_BUILD_WITH_5xx" ]; then
+      source "$PWD/build_tools/fbcode_config.sh"
+    else
+      source "$PWD/build_tools/fbcode_config_platform007.sh"
     fi
 fi
 
@@ -141,6 +143,7 @@ case "$TARGET_OS" in
         ;;
     FreeBSD)
         PLATFORM=OS_FREEBSD
+        CXX=clang++
         COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp -D_REENTRANT -DOS_FREEBSD"
         PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lpthread"
         # PORT_FILES=port/freebsd/freebsd_specific.cc
@@ -481,6 +484,9 @@ if test -z "$PORTABLE"; then
     COMMON_FLAGS="$COMMON_FLAGS -mcpu=$POWER -mtune=$POWER "
   elif test -n "`echo $TARGET_ARCHITECTURE | grep ^s390x`"; then
     COMMON_FLAGS="$COMMON_FLAGS -march=z10 "
+  elif test -n "`echo $TARGET_ARCHITECTURE | grep ^arm`"; then
+    # TODO: Handle this with approprite options.
+    COMMON_FLAGS="$COMMON_FLAGS"
   elif [ "$TARGET_OS" != AIX ] && [ "$TARGET_OS" != SunOS ]; then
     COMMON_FLAGS="$COMMON_FLAGS -march=native "
   elif test "$USE_SSE"; then

build_tools/cont_integration.sh

@@ -13,10 +13,12 @@ error=0
 
 function log {
   DATE=`date +%Y-%m-%d:%H:%M:%S`
+  # shellcheck disable=SC2068
   echo $DATE $@
 }
 
 function log_err {
+  # shellcheck disable=SC2145
   log "ERROR: $@ Error code: $error."
 }
 

build_tools/dependencies.sh

@@ -1,3 +1,4 @@
+# shellcheck disable=SC2148
 GCC_BASE=/mnt/gvfs/third-party2/gcc/8219ec1bcedf8ad9da05e121e193364de2cc4f61/5.x/centos6-native/c447969
 CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/64d8d58e3d84f8bde7a029763d4f5baf39d0d5b9/stable/centos6-native/6aaf4de
 LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/ba9be983c81de7299b59fe71950c664a84dcb5f8/5.x/gcc-5-glibc-2.23/339d858

build_tools/dependencies_4.8.1.sh

@@ -1,3 +1,4 @@
+# shellcheck disable=SC2148
 GCC_BASE=/mnt/gvfs/third-party2/gcc/cf7d14c625ce30bae1a4661c2319c5a283e4dd22/4.8.1/centos6-native/cc6c9dc
 CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/8598c375b0e94e1448182eb3df034704144a838d/stable/centos6-native/3f16ddd
 LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/d6e0a7da6faba45f5e5b1638f9edd7afc2f34e7d/4.8.1/gcc-4.8.1-glibc-2.17/8aac7fc

build_tools/dependencies_platform007.sh

@@ -0,0 +1,18 @@
+GCC_BASE=/mnt/gvfs/third-party2/gcc/6e8e715624fd15256a7970073387793dfcf79b46/7.x/centos7-native/b2ef2b6
+CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/ef37e1faa1c29782abfac1ae65a291b9b7966f6d/stable/centos7-native/c9f9104
+LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/c67031f0f739ac61575a061518d6ef5038f99f90/7.x/platform007/5620abc
+GLIBC_BASE=/mnt/gvfs/third-party2/glibc/60d6f124a78798b73944f5ba87c2306ae3460153/2.26/platform007/f259413
+SNAPPY_BASE=/mnt/gvfs/third-party2/snappy/7f9bdaada18f59bc27ec2b0871eb8a6144343aef/1.1.3/platform007/ca4da3d
+ZLIB_BASE=/mnt/gvfs/third-party2/zlib/22c2d65676fb7c23cfa797c4f6937f38b026f3cf/1.2.8/platform007/ca4da3d
+BZIP2_BASE=/mnt/gvfs/third-party2/bzip2/dc49a21c5fceec6456a7a28a94dcd16690af1337/1.0.6/platform007/ca4da3d
+LZ4_BASE=/mnt/gvfs/third-party2/lz4/907b498203d297947f3bb70b9466f47e100f1873/r131/platform007/ca4da3d
+ZSTD_BASE=/mnt/gvfs/third-party2/zstd/3ee276cbacfad3074e3f07bf826ac47f06970f4e/1.3.5/platform007/15a3614
+GFLAGS_BASE=/mnt/gvfs/third-party2/gflags/0b9929d2588991c65a57168bf88aff2db87c5d48/2.2.0/platform007/ca4da3d
+JEMALLOC_BASE=/mnt/gvfs/third-party2/jemalloc/9c910d36d6235cc40e8ff559358f1833452300ca/master/platform007/5b0f53e
+NUMA_BASE=/mnt/gvfs/third-party2/numa/9cbf2460284c669ed19c3ccb200a71f7dd7e53c7/2.0.11/platform007/ca4da3d
+LIBUNWIND_BASE=/mnt/gvfs/third-party2/libunwind/bf3d7497fe4e6d007354f0adffa16ce3003f8338/1.3/platform007/6f3e0a9
+TBB_BASE=/mnt/gvfs/third-party2/tbb/ff4e0b093534704d8abab678a4fd7f5ea7b094c7/2018_U5/platform007/ca4da3d
+KERNEL_HEADERS_BASE=/mnt/gvfs/third-party2/kernel-headers/b5c4a61a5c483ba24722005ae07895971a2ac707/fb/platform007/da39a3e
+BINUTILS_BASE=/mnt/gvfs/third-party2/binutils/92ff90349e2f43ea0a8246d8b1cf17b6869013e3/2.29.1/centos7-native/da39a3e
+VALGRIND_BASE=/mnt/gvfs/third-party2/valgrind/f3f697a28122e6bcd513273dd9c1ff23852fc59f/3.13.0/platform007/ca4da3d
+LUA_BASE=/mnt/gvfs/third-party2/lua/f0cd714433206d5139df61659eb7b28b1dea6683/5.3.4/platform007/5007832

build_tools/fbcode_config_platform007.sh

@@ -0,0 +1,157 @@
+#!/bin/sh
+#
+# Set environment variables so that we can compile rocksdb using
+# fbcode settings.  It uses the latest g++ and clang compilers and also
+# uses jemalloc
+# Environment variables that change the behavior of this script:
+# PIC_BUILD -- if true, it will only take pic versions of libraries from fbcode. libraries that don't have pic variant will not be included
+
+
+BASEDIR=`dirname $BASH_SOURCE`
+source "$BASEDIR/dependencies_platform007.sh"
+
+CFLAGS=""
+
+# libgcc
+LIBGCC_INCLUDE="$LIBGCC_BASE/include/c++/7.3.0"
+LIBGCC_LIBS=" -L $LIBGCC_BASE/lib"
+
+# glibc
+GLIBC_INCLUDE="$GLIBC_BASE/include"
+GLIBC_LIBS=" -L $GLIBC_BASE/lib"
+
+# snappy
+SNAPPY_INCLUDE=" -I $SNAPPY_BASE/include/"
+if test -z $PIC_BUILD; then
+  SNAPPY_LIBS=" $SNAPPY_BASE/lib/libsnappy.a"
+else
+  SNAPPY_LIBS=" $SNAPPY_BASE/lib/libsnappy_pic.a"
+fi
+CFLAGS+=" -DSNAPPY"
+
+if test -z $PIC_BUILD; then
+  # location of zlib headers and libraries
+  ZLIB_INCLUDE=" -I $ZLIB_BASE/include/"
+  ZLIB_LIBS=" $ZLIB_BASE/lib/libz.a"
+  CFLAGS+=" -DZLIB"
+
+  # location of bzip headers and libraries
+  BZIP_INCLUDE=" -I $BZIP2_BASE/include/"
+  BZIP_LIBS=" $BZIP2_BASE/lib/libbz2.a"
+  CFLAGS+=" -DBZIP2"
+
+  LZ4_INCLUDE=" -I $LZ4_BASE/include/"
+  LZ4_LIBS=" $LZ4_BASE/lib/liblz4.a"
+  CFLAGS+=" -DLZ4"
+fi
+
+ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
+if test -z $PIC_BUILD; then
+  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd.a"
+else
+  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd_pic.a"
+fi
+CFLAGS+=" -DZSTD"
+
+# location of gflags headers and libraries
+GFLAGS_INCLUDE=" -I $GFLAGS_BASE/include/"
+if test -z $PIC_BUILD; then
+  GFLAGS_LIBS=" $GFLAGS_BASE/lib/libgflags.a"
+else
+  GFLAGS_LIBS=" $GFLAGS_BASE/lib/libgflags_pic.a"
+fi
+CFLAGS+=" -DGFLAGS=gflags"
+
+# location of jemalloc
+JEMALLOC_INCLUDE=" -I $JEMALLOC_BASE/include/"
+JEMALLOC_LIB=" $JEMALLOC_BASE/lib/libjemalloc.a"
+
+if test -z $PIC_BUILD; then
+  # location of numa
+  NUMA_INCLUDE=" -I $NUMA_BASE/include/"
+  NUMA_LIB=" $NUMA_BASE/lib/libnuma.a"
+  CFLAGS+=" -DNUMA"
+
+  # location of libunwind
+  LIBUNWIND="$LIBUNWIND_BASE/lib/libunwind.a"
+fi
+
+# location of TBB
+TBB_INCLUDE=" -isystem $TBB_BASE/include/"
+if test -z $PIC_BUILD; then
+  TBB_LIBS="$TBB_BASE/lib/libtbb.a"
+else
+  TBB_LIBS="$TBB_BASE/lib/libtbb_pic.a"
+fi
+CFLAGS+=" -DTBB"
+
+# use Intel SSE support for checksum calculations
+export USE_SSE=1
+export PORTABLE=1
+
+BINUTILS="$BINUTILS_BASE/bin"
+AR="$BINUTILS/ar"
+
+DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP_INCLUDE $LZ4_INCLUDE $ZSTD_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE $TBB_INCLUDE"
+
+STDLIBS="-L $GCC_BASE/lib64"
+
+CLANG_BIN="$CLANG_BASE/bin"
+CLANG_LIB="$CLANG_BASE/lib"
+CLANG_SRC="$CLANG_BASE/../../src"
+
+CLANG_ANALYZER="$CLANG_BIN/clang++"
+CLANG_SCAN_BUILD="$CLANG_SRC/llvm/tools/clang/tools/scan-build/bin/scan-build"
+
+if [ -z "$USE_CLANG" ]; then
+  # gcc
+  CC="$GCC_BASE/bin/gcc"
+  CXX="$GCC_BASE/bin/g++"
+
+  CFLAGS+=" -B$BINUTILS/gold"
+  CFLAGS+=" -isystem $LIBGCC_INCLUDE"
+  CFLAGS+=" -isystem $GLIBC_INCLUDE"
+  JEMALLOC=1
+else
+  # clang
+  CLANG_INCLUDE="$CLANG_LIB/clang/stable/include"
+  CC="$CLANG_BIN/clang"
+  CXX="$CLANG_BIN/clang++"
+
+  KERNEL_HEADERS_INCLUDE="$KERNEL_HEADERS_BASE/include"
+
+  CFLAGS+=" -B$BINUTILS/gold -nostdinc -nostdlib"
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/7.x "
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/7.x/x86_64-facebook-linux "
+  CFLAGS+=" -isystem $GLIBC_INCLUDE"
+  CFLAGS+=" -isystem $LIBGCC_INCLUDE"
+  CFLAGS+=" -isystem $CLANG_INCLUDE"
+  CFLAGS+=" -isystem $KERNEL_HEADERS_INCLUDE/linux "
+  CFLAGS+=" -isystem $KERNEL_HEADERS_INCLUDE "
+  CFLAGS+=" -Wno-expansion-to-defined "
+  CXXFLAGS="-nostdinc++"
+fi
+
+CFLAGS+=" $DEPS_INCLUDE"
+CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE -DROCKSDB_RANGESYNC_PRESENT -DROCKSDB_SCHED_GETCPU_PRESENT -DROCKSDB_SUPPORT_THREAD_LOCAL -DHAVE_SSE42"
+CXXFLAGS+=" $CFLAGS"
+
+EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS"
+EXEC_LDFLAGS+=" -B$BINUTILS/gold"
+EXEC_LDFLAGS+=" -Wl,--dynamic-linker,/usr/local/fbcode/platform007/lib/ld.so"
+EXEC_LDFLAGS+=" $LIBUNWIND"
+EXEC_LDFLAGS+=" -Wl,-rpath=/usr/local/fbcode/platform007/lib"
+# required by libtbb
+EXEC_LDFLAGS+=" -ldl"
+
+PLATFORM_LDFLAGS="$LIBGCC_LIBS $GLIBC_LIBS $STDLIBS -lgcc -lstdc++"
+
+EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $TBB_LIBS"
+
+VALGRIND_VER="$VALGRIND_BASE/bin/"
+
+# lua not supported because it's on track for deprecation, I think
+LUA_PATH=
+LUA_LIB=
+
+export CC CXX AR CFLAGS CXXFLAGS EXEC_LDFLAGS EXEC_LDFLAGS_SHARED VALGRIND_VER JEMALLOC_LIB JEMALLOC_INCLUDE CLANG_ANALYZER CLANG_SCAN_BUILD LUA_PATH LUA_LIB

build_tools/make_package.sh

@@ -1,3 +1,4 @@
+# shellcheck disable=SC1113
 #/usr/bin/env bash
 
 set -e
@@ -28,12 +29,14 @@ function package() {
     if dpkg --get-selections | grep --quiet $1; then
       log "$1 is already installed. skipping."
     else
+      # shellcheck disable=SC2068
       apt-get install $@ -y
     fi
   elif [[ $OS = "centos" ]]; then
     if rpm -qa | grep --quiet $1; then
       log "$1 is already installed. skipping."
     else
+      # shellcheck disable=SC2068
       yum install $@ -y
     fi
   fi
@@ -52,6 +55,7 @@ function gem_install() {
   if gem list | grep --quiet $1; then
     log "$1 is already installed. skipping."
   else
+    # shellcheck disable=SC2068
     gem install $@
   fi
 }
@@ -125,4 +129,5 @@ function main() {
     include $LIB_DIR
 }
 
+# shellcheck disable=SC2068
 main $@

build_tools/rocksdb-lego-determinator

@@ -85,8 +85,9 @@ NON_SHM="TMPD=/tmp/rocksdb_test_tmp"
 GCC_481="ROCKSDB_FBCODE_BUILD_WITH_481=1"
 ASAN="COMPILE_WITH_ASAN=1"
 CLANG="USE_CLANG=1"
-LITE="OPT=\"-DROCKSDB_LITE -g\""
-TSAN="COMPILE_WITH_TSAN=1"
+# in gcc-5 there are known problems with TSAN like https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71090.
+# using platform007 gives us gcc-8 or higher which has that bug fixed.
+TSAN="ROCKSDB_FBCODE_BUILD_WITH_PLATFORM007=1 COMPILE_WITH_TSAN=1"
 UBSAN="COMPILE_WITH_UBSAN=1"
 DISABLE_JEMALLOC="DISABLE_JEMALLOC=1"
 HTTP_PROXY="https_proxy=http://fwdproxy.29.prn1:8080 http_proxy=http://fwdproxy.29.prn1:8080 ftp_proxy=http://fwdproxy.29.prn1:8080"

build_tools/update_dependencies.sh

@@ -53,6 +53,45 @@ function get_lib_base()
   log_variable $__res_var
 }
 
+###########################################################
+#                platform007 dependencies                 #
+###########################################################
+
+OUTPUT="$BASEDIR/dependencies_platform007.sh"
+
+rm -f "$OUTPUT"
+touch "$OUTPUT"
+
+echo "Writing dependencies to $OUTPUT"
+
+# Compilers locations
+GCC_BASE=`readlink -f $TP2_LATEST/gcc/7.x/centos7-native/*/`
+CLANG_BASE=`readlink -f $TP2_LATEST/llvm-fb/stable/centos7-native/*/`
+
+log_variable GCC_BASE
+log_variable CLANG_BASE
+
+# Libraries locations
+get_lib_base libgcc     7.x     platform007
+get_lib_base glibc      2.26    platform007
+get_lib_base snappy     LATEST  platform007
+get_lib_base zlib       LATEST  platform007
+get_lib_base bzip2      LATEST  platform007
+get_lib_base lz4        LATEST  platform007
+get_lib_base zstd       LATEST  platform007
+get_lib_base gflags     LATEST  platform007
+get_lib_base jemalloc   LATEST  platform007
+get_lib_base numa       LATEST  platform007
+get_lib_base libunwind  LATEST  platform007
+get_lib_base tbb        LATEST  platform007
+
+get_lib_base kernel-headers fb platform007
+get_lib_base binutils   LATEST centos7-native
+get_lib_base valgrind   LATEST platform007
+get_lib_base lua        5.3.4  platform007
+
+git diff $OUTPUT
+
 ###########################################################
 #                   5.x dependencies                      #
 ###########################################################

coverage/parse_gcov_output.py

@@ -72,7 +72,7 @@ def display_file_coverage(per_file_coverage, total_coverage):
     header_template = \
         "%" + str(max_file_name_length) + "s\t%s\t%s"
     separator = "-" * (max_file_name_length + 10 + 20)
-    print header_template % ("Filename", "Coverage", "Lines")
+    print header_template % ("Filename", "Coverage", "Lines")  # noqa: E999 T25377293 Grandfathered in
     print separator
 
     # -- Print body

db/c.cc

@@ -1388,23 +1388,24 @@ void rocksdb_writebatch_put_log_data(
   b->rep.PutLogData(Slice(blob, len));
 }
 
+class H : public WriteBatch::Handler {
+ public:
+  void* state_;
+  void (*put_)(void*, const char* k, size_t klen, const char* v, size_t vlen);
+  void (*deleted_)(void*, const char* k, size_t klen);
+  virtual void Put(const Slice& key, const Slice& value) override {
+    (*put_)(state_, key.data(), key.size(), value.data(), value.size());
+  }
+  virtual void Delete(const Slice& key) override {
+    (*deleted_)(state_, key.data(), key.size());
+  }
+};
+
 void rocksdb_writebatch_iterate(
     rocksdb_writebatch_t* b,
     void* state,
     void (*put)(void*, const char* k, size_t klen, const char* v, size_t vlen),
     void (*deleted)(void*, const char* k, size_t klen)) {
-  class H : public WriteBatch::Handler {
-   public:
-    void* state_;
-    void (*put_)(void*, const char* k, size_t klen, const char* v, size_t vlen);
-    void (*deleted_)(void*, const char* k, size_t klen);
-    virtual void Put(const Slice& key, const Slice& value) override {
-      (*put_)(state_, key.data(), key.size(), value.data(), value.size());
-    }
-    virtual void Delete(const Slice& key) override {
-      (*deleted_)(state_, key.data(), key.size());
-    }
-  };
   H handler;
   handler.state_ = state;
   handler.put_ = put;
@@ -1649,18 +1650,6 @@ void rocksdb_writebatch_wi_iterate(
     void* state,
     void (*put)(void*, const char* k, size_t klen, const char* v, size_t vlen),
     void (*deleted)(void*, const char* k, size_t klen)) {
-  class H : public WriteBatch::Handler {
-   public:
-    void* state_;
-    void (*put_)(void*, const char* k, size_t klen, const char* v, size_t vlen);
-    void (*deleted_)(void*, const char* k, size_t klen);
-    virtual void Put(const Slice& key, const Slice& value) override {
-      (*put_)(state_, key.data(), key.size(), value.data(), value.size());
-    }
-    virtual void Delete(const Slice& key) override {
-      (*deleted_)(state_, key.data(), key.size());
-    }
-  };
   H handler;
   handler.state_ = state;
   handler.put_ = put;
@@ -3109,20 +3098,21 @@ void rocksdb_slicetransform_destroy(rocksdb_slicetransform_t* st) {
   delete st;
 }
 
+struct Wrapper : public rocksdb_slicetransform_t {
+  const SliceTransform* rep_;
+  ~Wrapper() { delete rep_; }
+  const char* Name() const override { return rep_->Name(); }
+  Slice Transform(const Slice& src) const override {
+    return rep_->Transform(src);
+  }
+  bool InDomain(const Slice& src) const override {
+    return rep_->InDomain(src);
+  }
+  bool InRange(const Slice& src) const override { return rep_->InRange(src); }
+  static void DoNothing(void*) { }
+};
+
 rocksdb_slicetransform_t* rocksdb_slicetransform_create_fixed_prefix(size_t prefixLen) {
-  struct Wrapper : public rocksdb_slicetransform_t {
-    const SliceTransform* rep_;
-    ~Wrapper() { delete rep_; }
-    const char* Name() const override { return rep_->Name(); }
-    Slice Transform(const Slice& src) const override {
-      return rep_->Transform(src);
-    }
-    bool InDomain(const Slice& src) const override {
-      return rep_->InDomain(src);
-    }
-    bool InRange(const Slice& src) const override { return rep_->InRange(src); }
-    static void DoNothing(void*) { }
-  };
   Wrapper* wrapper = new Wrapper;
   wrapper->rep_ = rocksdb::NewFixedPrefixTransform(prefixLen);
   wrapper->state_ = nullptr;
@@ -3131,19 +3121,6 @@ rocksdb_slicetransform_t* rocksdb_slicetransform_create_fixed_prefix(size_t pref
 }
 
 rocksdb_slicetransform_t* rocksdb_slicetransform_create_noop() {
-  struct Wrapper : public rocksdb_slicetransform_t {
-    const SliceTransform* rep_;
-    ~Wrapper() { delete rep_; }
-    const char* Name() const override { return rep_->Name(); }
-    Slice Transform(const Slice& src) const override {
-      return rep_->Transform(src);
-    }
-    bool InDomain(const Slice& src) const override {
-      return rep_->InDomain(src);
-    }
-    bool InRange(const Slice& src) const override { return rep_->InRange(src); }
-    static void DoNothing(void*) { }
-  };
   Wrapper* wrapper = new Wrapper;
   wrapper->rep_ = rocksdb::NewNoopTransform();
   wrapper->state_ = nullptr;

db/column_family.cc

@@ -344,12 +344,13 @@ void SuperVersionUnrefHandle(void* ptr) {
   // When latter happens, we are in ~ColumnFamilyData(), no get should happen as
   // well.
   SuperVersion* sv = static_cast<SuperVersion*>(ptr);
-  if (sv->Unref()) {
-    sv->db_mutex->Lock();
-    sv->Cleanup();
-    sv->db_mutex->Unlock();
-    delete sv;
-  }
+  bool was_last_ref __attribute__((__unused__));
+  was_last_ref = sv->Unref();
+  // Thread-local SuperVersions can't outlive ColumnFamilyData::super_version_.
+  // This is important because we can't do SuperVersion cleanup here.
+  // That would require locking DB mutex, which would deadlock because
+  // SuperVersionUnrefHandle is called with locked ThreadLocalPtr mutex.
+  assert(!was_last_ref);
 }
 }  // anonymous namespace
 
@@ -385,7 +386,8 @@ ColumnFamilyData::ColumnFamilyData(
       pending_flush_(false),
       pending_compaction_(false),
       prev_compaction_needed_bytes_(0),
-      allow_2pc_(db_options.allow_2pc) {
+      allow_2pc_(db_options.allow_2pc),
+      last_memtable_id_(0) {
   Ref();
 
   // Convert user defined table properties collector factories to internal ones.
@@ -965,6 +967,12 @@ void ColumnFamilyData::InstallSuperVersion(
       RecalculateWriteStallConditions(mutable_cf_options);
 
   if (old_superversion != nullptr) {
+    // Reset SuperVersions cached in thread local storage.
+    // This should be done before old_superversion->Unref(). That's to ensure
+    // that local_sv_ never holds the last reference to SuperVersion, since
+    // it has no means to safely do SuperVersion cleanup.
+    ResetThreadLocalSuperVersions();
+
     if (old_superversion->mutable_cf_options.write_buffer_size !=
         mutable_cf_options.write_buffer_size) {
       mem_->UpdateWriteBufferSize(mutable_cf_options.write_buffer_size);
@@ -980,9 +988,6 @@ void ColumnFamilyData::InstallSuperVersion(
       sv_context->superversions_to_free.push_back(old_superversion);
     }
   }
-
-  // Reset SuperVersions cached in thread local storage
-  ResetThreadLocalSuperVersions();
 }
 
 void ColumnFamilyData::ResetThreadLocalSuperVersions() {
@@ -994,10 +999,12 @@ void ColumnFamilyData::ResetThreadLocalSuperVersions() {
       continue;
     }
     auto sv = static_cast<SuperVersion*>(ptr);
-    if (sv->Unref()) {
-      sv->Cleanup();
-      delete sv;
-    }
+    bool was_last_ref __attribute__((__unused__));
+    was_last_ref = sv->Unref();
+    // sv couldn't have been the last reference because
+    // ResetThreadLocalSuperVersions() is called before
+    // unref'ing super_version_.
+    assert(!was_last_ref);
   }
 }
 

db/column_family.h

@@ -239,7 +239,11 @@ class ColumnFamilyData {
   void SetCurrent(Version* _current);
   uint64_t GetNumLiveVersions() const;  // REQUIRE: DB mutex held
   uint64_t GetTotalSstFilesSize() const;  // REQUIRE: DB mutex held
-  void SetMemtable(MemTable* new_mem) { mem_ = new_mem; }
+  void SetMemtable(MemTable* new_mem) {
+    uint64_t memtable_id = last_memtable_id_.fetch_add(1) + 1;
+    new_mem->SetID(memtable_id);
+    mem_ = new_mem;
+  }
 
   // calculate the oldest log needed for the durability of this column family
   uint64_t OldestLogToKeep();
@@ -419,6 +423,9 @@ class ColumnFamilyData {
 
   // if the database was opened with 2pc enabled
   bool allow_2pc_;
+
+  // Memtable id to track flush.
+  std::atomic<uint64_t> last_memtable_id_;
 };
 
 // ColumnFamilySet has interesting thread-safety requirements

db/db_flush_test.cc

@@ -126,6 +126,41 @@ TEST_F(DBFlushTest, FlushInLowPriThreadPool) {
   ASSERT_EQ(1, num_compactions);
 }
 
+TEST_F(DBFlushTest, ManualFlushWithMinWriteBufferNumberToMerge) {
+  Options options = CurrentOptions();
+  options.write_buffer_size = 100;
+  options.max_write_buffer_number = 4;
+  options.min_write_buffer_number_to_merge = 3;
+  Reopen(options);
+
+  SyncPoint::GetInstance()->LoadDependency(
+      {{"DBImpl::BGWorkFlush",
+        "DBFlushTest::ManualFlushWithMinWriteBufferNumberToMerge:1"},
+       {"DBFlushTest::ManualFlushWithMinWriteBufferNumberToMerge:2",
+        "FlushJob::WriteLevel0Table"}});
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  ASSERT_OK(Put("key1", "value1"));
+
+  port::Thread t([&]() {
+    // The call wait for flush to finish, i.e. with flush_options.wait = true.
+    ASSERT_OK(Flush());
+  });
+
+  // Wait for flush start.
+  TEST_SYNC_POINT("DBFlushTest::ManualFlushWithMinWriteBufferNumberToMerge:1");
+  // Insert a second memtable before the manual flush finish.
+  // At the end of the manual flush job, it will check if further flush
+  // is needed, but it will not trigger flush of the second memtable because
+  // min_write_buffer_number_to_merge is not reached.
+  ASSERT_OK(Put("key2", "value2"));
+  ASSERT_OK(dbfull()->TEST_SwitchMemtable());
+  TEST_SYNC_POINT("DBFlushTest::ManualFlushWithMinWriteBufferNumberToMerge:2");
+
+  // Manual flush should return, without waiting for flush indefinitely.
+  t.join();
+}
+
 TEST_P(DBFlushDirectIOTest, DirectIO) {
   Options options;
   options.create_if_missing = true;

db/db_impl.h

@@ -807,8 +807,12 @@ class DBImpl : public DB {
   Status FlushMemTable(ColumnFamilyData* cfd, const FlushOptions& options,
                        bool writes_stopped = false);
 
-  // Wait for memtable flushed
-  Status WaitForFlushMemTable(ColumnFamilyData* cfd);
+  // Wait for memtable flushed.
+  // If flush_memtable_id is non-null, wait until the memtable with the ID
+  // gets flush. Otherwise, wait until the column family don't have any
+  // memtable pending flush.
+  Status WaitForFlushMemTable(ColumnFamilyData* cfd,
+                              const uint64_t* flush_memtable_id = nullptr);
 
   // REQUIRES: mutex locked
   Status SwitchWAL(WriteContext* write_context);

db/db_impl_compaction_flush.cc

@@ -809,7 +809,13 @@ int DBImpl::Level0StopWriteTrigger(ColumnFamilyHandle* column_family) {
 Status DBImpl::Flush(const FlushOptions& flush_options,
                      ColumnFamilyHandle* column_family) {
   auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
-  return FlushMemTable(cfh->cfd(), flush_options);
+  ROCKS_LOG_INFO(immutable_db_options_.info_log, "[%s] Manual flush start.",
+                 cfh->GetName().c_str());
+  Status s = FlushMemTable(cfh->cfd(), flush_options);
+  ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                 "[%s] Manual flush finished, status: %s\n",
+                 cfh->GetName().c_str(), s.ToString().c_str());
+  return s;
 }
 
 Status DBImpl::RunManualCompaction(ColumnFamilyData* cfd, int input_level,
@@ -944,6 +950,7 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
                              const FlushOptions& flush_options,
                              bool writes_stopped) {
   Status s;
+  uint64_t flush_memtable_id = 0;
   {
     WriteContext context;
     InstrumentedMutexLock guard_lock(&mutex_);
@@ -961,6 +968,7 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
 
     // SwitchMemtable() will release and reacquire mutex during execution
     s = SwitchMemtable(cfd, &context);
+    flush_memtable_id = cfd->imm()->GetLatestMemTableID();
 
     if (!writes_stopped) {
       write_thread_.ExitUnbatched(&w);
@@ -975,16 +983,19 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
 
   if (s.ok() && flush_options.wait) {
     // Wait until the compaction completes
-    s = WaitForFlushMemTable(cfd);
+    s = WaitForFlushMemTable(cfd, &flush_memtable_id);
   }
   return s;
 }
 
-Status DBImpl::WaitForFlushMemTable(ColumnFamilyData* cfd) {
+Status DBImpl::WaitForFlushMemTable(ColumnFamilyData* cfd,
+                                    const uint64_t* flush_memtable_id) {
   Status s;
   // Wait until the compaction completes
   InstrumentedMutexLock l(&mutex_);
-  while (cfd->imm()->NumNotFlushed() > 0 && bg_error_.ok()) {
+  while (cfd->imm()->NumNotFlushed() > 0 && bg_error_.ok() &&
+         (flush_memtable_id == nullptr ||
+          cfd->imm()->GetEarliestMemTableID() <= *flush_memtable_id)) {
     if (shutting_down_.load(std::memory_order_acquire)) {
       return Status::ShutdownInProgress();
     }

db/db_iter.cc

@@ -217,6 +217,9 @@ class DBIter final: public Iterator {
         *prop = "Iterator is not valid.";
       }
       return Status::OK();
+    } else if (prop_name == "rocksdb.iterator.internal-key") {
+      *prop = saved_key_.GetUserKey().ToString();
+      return Status::OK();
     }
     return Status::InvalidArgument("Undentified property.");
   }

db/db_iterator_test.cc

@@ -55,6 +55,7 @@ TEST_F(DBIteratorTest, IteratorProperty) {
   Options options = CurrentOptions();
   CreateAndReopenWithCF({"pikachu"}, options);
   Put(1, "1", "2");
+  Delete(1, "2");
   ReadOptions ropt;
   ropt.pin_data = false;
   {
@@ -64,9 +65,15 @@ TEST_F(DBIteratorTest, IteratorProperty) {
     ASSERT_NOK(iter->GetProperty("non_existing.value", &prop_value));
     ASSERT_OK(iter->GetProperty("rocksdb.iterator.is-key-pinned", &prop_value));
     ASSERT_EQ("0", prop_value);
+    ASSERT_OK(iter->GetProperty("rocksdb.iterator.internal-key", &prop_value));
+    ASSERT_EQ("1", prop_value);
     iter->Next();
     ASSERT_OK(iter->GetProperty("rocksdb.iterator.is-key-pinned", &prop_value));
     ASSERT_EQ("Iterator is not valid.", prop_value);
+
+    // Get internal key at which the iteration stopped (tombstone in this case).
+    ASSERT_OK(iter->GetProperty("rocksdb.iterator.internal-key", &prop_value));
+    ASSERT_EQ("2", prop_value);
   }
   Close();
 }
@@ -2157,6 +2164,48 @@ TEST_F(DBIteratorTest, SkipStatistics) {
   ASSERT_EQ(skip_count, TestGetTickerCount(options, NUMBER_ITER_SKIP));
 }
 
+TEST_F(DBIteratorTest, SeekAfterHittingManyInternalKeys) {
+  Options options = CurrentOptions();
+  DestroyAndReopen(options);
+  ReadOptions ropts;
+  ropts.max_skippable_internal_keys = 2;
+
+  Put("1", "val_1");
+  // Add more tombstones than max_skippable_internal_keys so that Next() fails.
+  Delete("2");
+  Delete("3");
+  Delete("4");
+  Delete("5");
+  Put("6", "val_6");
+
+  unique_ptr<Iterator> iter(db_->NewIterator(ropts));
+  iter->SeekToFirst();
+
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().ToString(), "1");
+  ASSERT_EQ(iter->value().ToString(), "val_1");
+
+  // This should fail as incomplete due to too many non-visible internal keys on
+  // the way to the next valid user key.
+  iter->Next();
+  ASSERT_TRUE(!iter->Valid());
+  ASSERT_TRUE(iter->status().IsIncomplete());
+
+  // Get the internal key at which Next() failed.
+  std::string prop_value;
+  ASSERT_OK(iter->GetProperty("rocksdb.iterator.internal-key", &prop_value));
+  ASSERT_EQ("4", prop_value);
+
+  // Create a new iterator to seek to the internal key.
+  unique_ptr<Iterator> iter2(db_->NewIterator(ropts));
+  iter2->Seek(prop_value);
+  ASSERT_TRUE(iter2->Valid());
+  ASSERT_OK(iter2->status());
+
+  ASSERT_EQ(iter2->key().ToString(), "6");
+  ASSERT_EQ(iter2->value().ToString(), "val_6");
+}
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/db_test.cc

@@ -5656,6 +5656,50 @@ TEST_F(DBTest, PauseBackgroundWorkTest) {
   // now it's done
   ASSERT_TRUE(done.load());
 }
+
+// Keep spawning short-living threads that create an iterator and quit.
+// Meanwhile in another thread keep flushing memtables.
+// This used to cause a deadlock.
+TEST_F(DBTest, ThreadLocalPtrDeadlock) {
+  std::atomic<int> flushes_done{0};
+  std::atomic<int> threads_destroyed{0};
+  auto done = [&] {
+    return flushes_done.load() > 10;
+  };
+
+  std::thread flushing_thread([&] {
+    for (int i = 0; !done(); ++i) {
+      ASSERT_OK(db_->Put(WriteOptions(), Slice("hi"),
+                         Slice(std::to_string(i).c_str())));
+      ASSERT_OK(db_->Flush(FlushOptions()));
+      int cnt = ++flushes_done;
+      fprintf(stderr, "Flushed %d times\n", cnt);
+    }
+  });
+
+  std::vector<std::thread> thread_spawning_threads(10);
+  for (auto& t: thread_spawning_threads) {
+    t = std::thread([&] {
+      while (!done()) {
+        {
+          std::thread tmp_thread([&] {
+            auto it = db_->NewIterator(ReadOptions());
+            delete it;
+          });
+          tmp_thread.join();
+        }
+        ++threads_destroyed;
+      }
+    });
+  }
+
+  for (auto& t: thread_spawning_threads) {
+    t.join();
+  }
+  flushing_thread.join();
+  fprintf(stderr, "Done. Flushed %d times, destroyed %d threads\n",
+          flushes_done.load(), threads_destroyed.load());
+}
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/log_reader.h

@@ -54,6 +54,7 @@ class Reader {
   // The Reader will start reading at the first record located at physical
   // position >= initial_offset within the file.
   Reader(std::shared_ptr<Logger> info_log,
+  // @lint-ignore TXT2 T25377293 Grandfathered in
 	 unique_ptr<SequentialFileReader>&& file,
          Reporter* reporter, bool checksum, uint64_t initial_offset,
          uint64_t log_num);

db/memtable.h

@@ -368,6 +368,11 @@ class MemTable {
     return oldest_key_time_.load(std::memory_order_relaxed);
   }
 
+  // REQUIRES: db_mutex held.
+  void SetID(uint64_t id) { id_ = id; }
+
+  uint64_t GetID() const { return id_; }
+
  private:
   enum FlushStateEnum { FLUSH_NOT_REQUESTED, FLUSH_REQUESTED, FLUSH_SCHEDULED };
 
@@ -437,6 +442,9 @@ class MemTable {
   // Timestamp of oldest key
   std::atomic<uint64_t> oldest_key_time_;
 
+  // Memtable id to track flush.
+  uint64_t id_ = 0;
+
   // Returns a heuristic flush decision
   bool ShouldFlushNow() const;
 

db/memtable_list.h

@@ -5,11 +5,12 @@
 //
 #pragma once
 
-#include <string>
-#include <list>
-#include <vector>
-#include <set>
 #include <deque>
+#include <limits>
+#include <list>
+#include <set>
+#include <string>
+#include <vector>
 
 #include "db/dbformat.h"
 #include "db/memtable.h"
@@ -244,6 +245,22 @@ class MemTableList {
 
   uint64_t GetMinLogContainingPrepSection();
 
+  uint64_t GetEarliestMemTableID() const {
+    auto& memlist = current_->memlist_;
+    if (memlist.empty()) {
+      return std::numeric_limits<uint64_t>::max();
+    }
+    return memlist.back()->GetID();
+  }
+
+  uint64_t GetLatestMemTableID() const {
+    auto& memlist = current_->memlist_;
+    if (memlist.empty()) {
+      return 0;
+    }
+    return memlist.front()->GetID();
+  }
+
  private:
   // DB mutex held
   void InstallNewVersion();

hdfs/setup.sh

@@ -1,3 +1,4 @@
+# shellcheck disable=SC2148
 export USE_HDFS=1
 export LD_LIBRARY_PATH=$JAVA_HOME/jre/lib/amd64/server:$JAVA_HOME/jre/lib/amd64:/usr/lib/hadoop/lib/native
 

include/rocksdb/iterator.h

@@ -97,6 +97,9 @@ class Iterator : public Cleanable {
   // Property "rocksdb.iterator.super-version-number":
   //   LSM version used by the iterator. The same format as DB Property
   //   kCurrentSuperVersionNumber. See its comment for more information.
+  // Property "rocksdb.iterator.internal-key":
+  //   Get the user-key portion of the internal key at which the iteration
+  //   stopped.
   virtual Status GetProperty(std::string prop_name, std::string* prop);
 
  private:

include/rocksdb/utilities/stackable_db.h

@@ -4,6 +4,7 @@
 
 #pragma once
 #include <map>
+#include <memory>
 #include <string>
 #include "rocksdb/db.h"
 
@@ -18,11 +19,20 @@ namespace rocksdb {
 // This class contains APIs to stack rocksdb wrappers.Eg. Stack TTL over base d
 class StackableDB : public DB {
  public:
-  // StackableDB is the owner of db now!
+  // StackableDB take sole ownership of the underlying db.
   explicit StackableDB(DB* db) : db_(db) {}
 
+  // StackableDB take shared ownership of the underlying db.
+  explicit StackableDB(std::shared_ptr<DB> db)
+      : db_(db.get()), shared_db_ptr_(db) {}
+
   ~StackableDB() {
-    delete db_;
+    if (shared_db_ptr_ == nullptr) {
+      delete db_;
+    } else {
+      assert(shared_db_ptr_.get() == db_);
+    }
+    db_ = nullptr;
   }
 
   virtual DB* GetBaseDB() {
@@ -373,6 +383,7 @@ class StackableDB : public DB {
 
  protected:
   DB* db_;
+  std::shared_ptr<DB> shared_db_ptr_;
 };
 
 } //  namespace rocksdb

include/rocksdb/version.h

@@ -5,8 +5,8 @@
 #pragma once
 
 #define ROCKSDB_MAJOR 5
-#define ROCKSDB_MINOR 10
-#define ROCKSDB_PATCH 0
+#define ROCKSDB_MINOR 11
+#define ROCKSDB_PATCH 4
 
 // Do not use these. We made the mistake of declaring macros starting with
 // double underscore. Now we have to live with our choice. We'll deprecate these

java/CMakeLists.txt

@@ -24,6 +24,7 @@ set(JNI_NATIVE_SOURCES
         rocksjni/lru_cache.cc
         rocksjni/memtablejni.cc
         rocksjni/merge_operator.cc
+        rocksjni/native_comparator_wrapper_test.cc
         rocksjni/options.cc
         rocksjni/options_util.cc
         rocksjni/ratelimiterjni.cc
@@ -87,6 +88,8 @@ set(NATIVE_JAVA_CLASSES
         org.rocksdb.LRUCache
         org.rocksdb.MemTableConfig
         org.rocksdb.MergeOperator
+        org.rocksdb.NativeComparatorWrapper
+        org.rocksdb.NativeComparatorWrapperTest.NativeStringComparatorWrapper
         org.rocksdb.NativeLibraryLoader
         org.rocksdb.Options
         org.rocksdb.OptionsUtil
@@ -122,6 +125,8 @@ set(NATIVE_JAVA_CLASSES
         org.rocksdb.WriteBatchTestInternalHelper
         org.rocksdb.WriteBatchWithIndex
         org.rocksdb.WriteOptions
+        org.rocksdb.util.CapturingWriteBatchHandler
+        org.rocksdb.util.WriteBatchGetter
 )
 
 include(FindJava)
@@ -177,6 +182,7 @@ add_jar(
   src/main/java/org/rocksdb/CompactionStyle.java
   src/main/java/org/rocksdb/Comparator.java
   src/main/java/org/rocksdb/ComparatorOptions.java
+  src/main/java/org/rocksdb/ComparatorType.java
   src/main/java/org/rocksdb/CompressionOptions.java
   src/main/java/org/rocksdb/CompressionType.java
   src/main/java/org/rocksdb/DBOptions.java
@@ -203,6 +209,7 @@ add_jar(
   src/main/java/org/rocksdb/MergeOperator.java
   src/main/java/org/rocksdb/MutableColumnFamilyOptions.java
   src/main/java/org/rocksdb/MutableColumnFamilyOptionsInterface.java
+  src/main/java/org/rocksdb/NativeComparatorWrapper.java
   src/main/java/org/rocksdb/NativeLibraryLoader.java
   src/main/java/org/rocksdb/Options.java
   src/main/java/org/rocksdb/OptionsUtil.java
@@ -244,11 +251,14 @@ add_jar(
   src/main/java/org/rocksdb/WriteOptions.java
   src/test/java/org/rocksdb/BackupEngineTest.java
   src/test/java/org/rocksdb/IngestExternalFileOptionsTest.java
+  src/test/java/org/rocksdb/NativeComparatorWrapperTest.java
   src/test/java/org/rocksdb/PlatformRandomHelper.java
   src/test/java/org/rocksdb/RocksDBExceptionTest.java
   src/test/java/org/rocksdb/RocksMemoryResource.java
   src/test/java/org/rocksdb/SnapshotTest.java
   src/test/java/org/rocksdb/WriteBatchTest.java
+  src/test/java/org/rocksdb/util/CapturingWriteBatchHandler.java
+  src/test/java/org/rocksdb/util/WriteBatchGetter.java
   INCLUDE_JARS ${JAVA_TESTCLASSPATH}
 )
 

java/Makefile

@@ -1,6 +1,7 @@
 NATIVE_JAVA_CLASSES = org.rocksdb.AbstractCompactionFilter\
 	org.rocksdb.AbstractCompactionFilterFactory\
 	org.rocksdb.AbstractSlice\
+	org.rocksdb.AbstractTransactionNotifier\
 	org.rocksdb.BackupEngine\
 	org.rocksdb.BackupableDBOptions\
 	org.rocksdb.BlockBasedTableConfig\
@@ -29,6 +30,9 @@ NATIVE_JAVA_CLASSES = org.rocksdb.AbstractCompactionFilter\
 	org.rocksdb.Logger\
 	org.rocksdb.LRUCache\
 	org.rocksdb.MergeOperator\
+	org.rocksdb.NativeComparatorWrapper\
+	org.rocksdb.OptimisticTransactionDB\
+	org.rocksdb.OptimisticTransactionOptions\
 	org.rocksdb.Options\
 	org.rocksdb.OptionsUtil\
 	org.rocksdb.PlainTableConfig\
@@ -45,6 +49,10 @@ NATIVE_JAVA_CLASSES = org.rocksdb.AbstractCompactionFilter\
 	org.rocksdb.Slice\
 	org.rocksdb.SstFileWriter\
 	org.rocksdb.Statistics\
+	org.rocksdb.Transaction\
+	org.rocksdb.TransactionDB\
+	org.rocksdb.TransactionDBOptions\
+	org.rocksdb.TransactionOptions\
 	org.rocksdb.TransactionLogIterator\
 	org.rocksdb.TtlDB\
 	org.rocksdb.VectorMemTableConfig\
@@ -57,6 +65,7 @@ NATIVE_JAVA_CLASSES = org.rocksdb.AbstractCompactionFilter\
 	org.rocksdb.WBWIRocksIterator
 
 NATIVE_JAVA_TEST_CLASSES = org.rocksdb.RocksDBExceptionTest\
+    org.rocksdb.NativeComparatorWrapperTest.NativeStringComparatorWrapper\
     org.rocksdb.WriteBatchTest\
     org.rocksdb.WriteBatchTestInternalHelper
 
@@ -104,7 +113,11 @@ JAVA_TESTS = org.rocksdb.BackupableDBOptionsTest\
 	org.rocksdb.MergeTest\
 	org.rocksdb.MixedOptionsTest\
 	org.rocksdb.MutableColumnFamilyOptionsTest\
+	org.rocksdb.NativeComparatorWrapperTest\
 	org.rocksdb.NativeLibraryLoaderTest\
+	org.rocksdb.OptimisticTransactionTest\
+	org.rocksdb.OptimisticTransactionDBTest\
+	org.rocksdb.OptimisticTransactionOptionsTest\
 	org.rocksdb.OptionsUtilTest\
 	org.rocksdb.OptionsTest\
 	org.rocksdb.PlainTableConfigTest\
@@ -120,6 +133,10 @@ JAVA_TESTS = org.rocksdb.BackupableDBOptionsTest\
 	org.rocksdb.SliceTest\
 	org.rocksdb.SnapshotTest\
 	org.rocksdb.SstFileWriterTest\
+	org.rocksdb.TransactionTest\
+	org.rocksdb.TransactionDBTest\
+	org.rocksdb.TransactionOptionsTest\
+	org.rocksdb.TransactionDBOptionsTest\
 	org.rocksdb.TransactionLogIteratorTest\
 	org.rocksdb.TtlDBTest\
 	org.rocksdb.StatisticsTest\
@@ -209,6 +226,20 @@ column_family_sample: java
 	java $(JAVA_ARGS) -Djava.library.path=target -cp $(MAIN_CLASSES):$(SAMPLES_MAIN_CLASSES) RocksDBColumnFamilySample /tmp/rocksdbjni
 	$(AM_V_at)@rm -rf /tmp/rocksdbjni
 
+transaction_sample: java
+	$(AM_V_GEN)mkdir -p $(SAMPLES_MAIN_CLASSES)
+	$(AM_V_at)javac -cp $(MAIN_CLASSES) -d $(SAMPLES_MAIN_CLASSES) $(SAMPLES_MAIN_SRC)/TransactionSample.java
+	$(AM_V_at)@rm -rf /tmp/rocksdbjni
+	java -ea -Xcheck:jni -Djava.library.path=target -cp $(MAIN_CLASSES):$(SAMPLES_MAIN_CLASSES) TransactionSample /tmp/rocksdbjni
+	$(AM_V_at)@rm -rf /tmp/rocksdbjni
+
+optimistic_transaction_sample: java
+	$(AM_V_GEN)mkdir -p $(SAMPLES_MAIN_CLASSES)
+	$(AM_V_at)javac -cp $(MAIN_CLASSES) -d $(SAMPLES_MAIN_CLASSES) $(SAMPLES_MAIN_SRC)/OptimisticTransactionSample.java
+	$(AM_V_at)@rm -rf /tmp/rocksdbjni
+	java -ea -Xcheck:jni -Djava.library.path=target -cp $(MAIN_CLASSES):$(SAMPLES_MAIN_CLASSES) OptimisticTransactionSample /tmp/rocksdbjni
+	$(AM_V_at)@rm -rf /tmp/rocksdbjni
+
 resolve_test_deps:
 	test -d "$(JAVA_TEST_LIBDIR)" || mkdir -p "$(JAVA_TEST_LIBDIR)"
 	test -s "$(JAVA_JUNIT_JAR)" || cp $(MVN_LOCAL)/junit/junit/4.12/junit-4.12.jar $(JAVA_TEST_LIBDIR) || curl -k -L -o $(JAVA_JUNIT_JAR) $(SEARCH_REPO_URL)junit/junit/4.12/junit-4.12.jar

java/jdb_bench.sh

@@ -1,3 +1,4 @@
+# shellcheck disable=SC2148
 PLATFORM=64
 if [ `getconf LONG_BIT` != "64" ]
 then
@@ -7,4 +8,5 @@ fi
 ROCKS_JAR=`find target -name rocksdbjni*.jar`
 
 echo "Running benchmark in $PLATFORM-Bit mode."
+# shellcheck disable=SC2068
 java -server -d$PLATFORM -XX:NewSize=4m -XX:+AggressiveOpts -Djava.library.path=target -cp "${ROCKS_JAR}:benchmark/target/classes" org.rocksdb.benchmark.DbBenchmark $@

java/rocksjni/columnfamilyhandle.cc

@@ -3,8 +3,8 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
 //
-// This file implements the "bridge" between Java and C++ and enables
-// calling c++ rocksdb::Iterator methods from Java side.
+// This file implements the "bridge" between Java and C++ for
+// rocksdb::ColumnFamilyHandle.
 
 #include <stdio.h>
 #include <stdlib.h>
@@ -13,14 +13,56 @@
 #include "include/org_rocksdb_ColumnFamilyHandle.h"
 #include "rocksjni/portal.h"
 
+/*
+ * Class:     org_rocksdb_ColumnFamilyHandle
+ * Method:    getName
+ * Signature: (J)[B
+ */
+jbyteArray Java_org_rocksdb_ColumnFamilyHandle_getName(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* cfh = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jhandle);
+  std::string cf_name = cfh->GetName();
+  return rocksdb::JniUtil::copyBytes(env, cf_name);
+}
+
+/*
+* Class:     org_rocksdb_ColumnFamilyHandle
+* Method:    getID
+* Signature: (J)I
+*/
+jint Java_org_rocksdb_ColumnFamilyHandle_getID(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* cfh = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jhandle);
+  const int32_t id = cfh->GetID();
+  return static_cast<jint>(id);
+}
+
+/*
+ * Class:     org_rocksdb_ColumnFamilyHandle
+ * Method:    getDescriptor
+ * Signature: (J)Lorg/rocksdb/ColumnFamilyDescriptor;
+ */
+jobject Java_org_rocksdb_ColumnFamilyHandle_getDescriptor(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* cfh = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jhandle);
+  rocksdb::ColumnFamilyDescriptor desc;
+  rocksdb::Status s = cfh->GetDescriptor(&desc);
+  if (s.ok()) {
+    return rocksdb::ColumnFamilyDescriptorJni::construct(env, &desc);
+  } else {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
+    return nullptr;
+  }
+}
+
 /*
  * Class:     org_rocksdb_ColumnFamilyHandle
  * Method:    disposeInternal
  * Signature: (J)V
  */
 void Java_org_rocksdb_ColumnFamilyHandle_disposeInternal(
-    JNIEnv* env, jobject jobj, jlong handle) {
-  auto* cfh = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(handle);
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* cfh = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jhandle);
   assert(cfh != nullptr);
   delete cfh;
 }

java/rocksjni/compaction_filter_factory.cc

@@ -34,4 +34,5 @@ void Java_org_rocksdb_AbstractCompactionFilterFactory_disposeInternal(
   auto* ptr_sptr_cff =
       reinterpret_cast<std::shared_ptr<rocksdb::CompactionFilterFactoryJniCallback> *>(jhandle);
   delete ptr_sptr_cff;
+// @lint-ignore TXT4 T25377293 Grandfathered in
 }

java/rocksjni/comparator.cc

@@ -14,6 +14,7 @@
 
 #include "include/org_rocksdb_Comparator.h"
 #include "include/org_rocksdb_DirectComparator.h"
+#include "include/org_rocksdb_NativeComparatorWrapper.h"
 #include "rocksjni/comparatorjnicallback.h"
 #include "rocksjni/portal.h"
 
@@ -49,4 +50,16 @@ jlong Java_org_rocksdb_DirectComparator_createNewDirectComparator0(
       new rocksdb::DirectComparatorJniCallback(env, jobj, copt);
   return reinterpret_cast<jlong>(c);
 }
+
+/*
+ * Class:     org_rocksdb_NativeComparatorWrapper
+ * Method:    disposeInternal
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_NativeComparatorWrapper_disposeInternal(
+    JNIEnv* env, jobject jobj, jlong jcomparator_handle) {
+  auto* comparator =
+      reinterpret_cast<rocksdb::Comparator*>(jcomparator_handle);
+  delete comparator;
+}
 // </editor-fold>

java/rocksjni/ingest_external_file_options.cc

@@ -146,4 +146,5 @@ void Java_org_rocksdb_IngestExternalFileOptions_disposeInternal(
   auto* options =
       reinterpret_cast<rocksdb::IngestExternalFileOptions*>(jhandle);
   delete options;
+// @lint-ignore TXT4 T25377293 Grandfathered in
 }

java/rocksjni/jnicallback.cc

@@ -49,4 +49,5 @@ JniCallback::~JniCallback() {
 
   releaseJniEnv(attached_thread);
 }
+// @lint-ignore TXT4 T25377293 Grandfathered in
 }  // namespace rocksdb

java/rocksjni/jnicallback.h

@@ -25,4 +25,5 @@ namespace rocksdb {
   };
 }
 
+// @lint-ignore TXT4 T25377293 Grandfathered in
 #endif  // JAVA_ROCKSJNI_JNICALLBACK_H_

java/rocksjni/native_comparator_wrapper_test.cc

@@ -0,0 +1,50 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+#include <jni.h>
+#include <string>
+
+#include "rocksdb/comparator.h"
+#include "rocksdb/slice.h"
+
+#include "include/org_rocksdb_NativeComparatorWrapperTest_NativeStringComparatorWrapper.h"
+
+namespace rocksdb {
+
+class NativeComparatorWrapperTestStringComparator
+    : public Comparator {
+
+  const char* Name() const {
+    return "NativeComparatorWrapperTestStringComparator";
+  }
+
+  int Compare(
+      const Slice& a, const Slice& b) const {
+    return a.ToString().compare(b.ToString());
+  }
+
+  void FindShortestSeparator(
+      std::string* start, const Slice& limit) const {
+    return;
+  }
+
+  void FindShortSuccessor(
+      std::string* key) const {
+    return;
+  }
+};
+}  // end of rocksdb namespace
+
+/*
+ * Class:     org_rocksdb_NativeComparatorWrapperTest_NativeStringComparatorWrapper
+ * Method:    newStringComparator
+ * Signature: ()J
+ */
+jlong Java_org_rocksdb_NativeComparatorWrapperTest_00024NativeStringComparatorWrapper_newStringComparator(
+    JNIEnv* env , jobject jobj) {
+  auto* comparator =
+      new rocksdb::NativeComparatorWrapperTestStringComparator();
+  return reinterpret_cast<jlong>(comparator);
+}

java/rocksjni/optimistic_transaction_db.cc

@@ -0,0 +1,267 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// This file implements the "bridge" between Java and C++
+// for rocksdb::TransactionDB.
+
+#include <jni.h>
+
+#include "include/org_rocksdb_OptimisticTransactionDB.h"
+
+#include "rocksdb/options.h"
+#include "rocksdb/utilities/optimistic_transaction_db.h"
+#include "rocksdb/utilities/transaction.h"
+
+#include "rocksjni/portal.h"
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionDB
+ * Method:    open
+ * Signature: (JLjava/lang/String;)J
+ */
+jlong Java_org_rocksdb_OptimisticTransactionDB_open__JLjava_lang_String_2(
+    JNIEnv* env, jclass jcls, jlong joptions_handle, jstring jdb_path) {
+  const char* db_path = env->GetStringUTFChars(jdb_path, nullptr);
+  if (db_path == nullptr) {
+      // exception thrown: OutOfMemoryError
+      return 0;
+  }
+
+  auto* options = reinterpret_cast<rocksdb::Options*>(joptions_handle);
+  rocksdb::OptimisticTransactionDB* otdb = nullptr;
+  rocksdb::Status s =
+      rocksdb::OptimisticTransactionDB::Open(*options, db_path, &otdb);
+  env->ReleaseStringUTFChars(jdb_path, db_path);
+
+  if (s.ok()) {
+    return reinterpret_cast<jlong>(otdb);
+  } else {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
+    return 0;
+  }
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionDB
+ * Method:    open
+ * Signature: (JLjava/lang/String;[[B[J)[J
+ */
+jlongArray Java_org_rocksdb_OptimisticTransactionDB_open__JLjava_lang_String_2_3_3B_3J(
+    JNIEnv* env, jclass jcls, jlong jdb_options_handle, jstring jdb_path,
+    jobjectArray jcolumn_names, jlongArray jcolumn_options_handles) {
+  const char* db_path = env->GetStringUTFChars(jdb_path, nullptr);
+  if (db_path == nullptr) {
+      // exception thrown: OutOfMemoryError
+      return nullptr;
+  }
+
+  std::vector<rocksdb::ColumnFamilyDescriptor> column_families;
+  const jsize len_cols = env->GetArrayLength(jcolumn_names);
+  if (len_cols > 0) {
+    if (env->EnsureLocalCapacity(len_cols) != 0) {
+      // out of memory
+      env->ReleaseStringUTFChars(jdb_path, db_path);
+      return nullptr;
+    }
+
+    jlong* jco =
+        env->GetLongArrayElements(jcolumn_options_handles, nullptr);
+    if(jco == nullptr) {
+        // exception thrown: OutOfMemoryError
+        env->ReleaseStringUTFChars(jdb_path, db_path);
+        return nullptr;
+    }
+
+    for (int i = 0; i < len_cols; i++) {
+        const jobject jcn = env->GetObjectArrayElement(jcolumn_names, i);
+        if (env->ExceptionCheck()) {
+            // exception thrown: ArrayIndexOutOfBoundsException
+            env->ReleaseLongArrayElements(jcolumn_options_handles, jco, JNI_ABORT);
+            env->ReleaseStringUTFChars(jdb_path, db_path);
+            return nullptr;
+        }
+
+        const jbyteArray jcn_ba = reinterpret_cast<jbyteArray>(jcn);
+        const jsize jcf_name_len = env->GetArrayLength(jcn_ba);
+        if (env->EnsureLocalCapacity(jcf_name_len) != 0) {
+          // out of memory
+          env->DeleteLocalRef(jcn);
+          env->ReleaseLongArrayElements(jcolumn_options_handles, jco, JNI_ABORT);
+          env->ReleaseStringUTFChars(jdb_path, db_path);
+          return nullptr;
+        }
+
+        jbyte* jcf_name = env->GetByteArrayElements(jcn_ba, nullptr);
+        if (jcf_name == nullptr) {
+            // exception thrown: OutOfMemoryError
+            env->DeleteLocalRef(jcn);
+            env->ReleaseLongArrayElements(jcolumn_options_handles, jco, JNI_ABORT);
+            env->ReleaseStringUTFChars(jdb_path, db_path);
+            return nullptr;
+        }
+
+        const std::string cf_name(reinterpret_cast<char *>(jcf_name), jcf_name_len);
+        const rocksdb::ColumnFamilyOptions* cf_options =
+            reinterpret_cast<rocksdb::ColumnFamilyOptions*>(jco[i]);
+        column_families.push_back(
+            rocksdb::ColumnFamilyDescriptor(cf_name, *cf_options));
+
+        env->ReleaseByteArrayElements(jcn_ba, jcf_name, JNI_ABORT);
+        env->DeleteLocalRef(jcn);
+    }
+    env->ReleaseLongArrayElements(jcolumn_options_handles, jco, JNI_ABORT);
+  }
+
+  auto* db_options = reinterpret_cast<rocksdb::DBOptions*>(jdb_options_handle);
+  std::vector<rocksdb::ColumnFamilyHandle*> handles;
+  rocksdb::OptimisticTransactionDB* otdb = nullptr;
+  const rocksdb::Status s = rocksdb::OptimisticTransactionDB::Open(*db_options,
+      db_path, column_families, &handles, &otdb);
+
+  env->ReleaseStringUTFChars(jdb_path, db_path);
+
+  // check if open operation was successful
+  if (s.ok()) {
+    const jsize resultsLen = 1 + len_cols;  // db handle + column family handles
+    std::unique_ptr<jlong[]> results =
+        std::unique_ptr<jlong[]>(new jlong[resultsLen]);
+    results[0] = reinterpret_cast<jlong>(otdb);
+    for (int i = 1; i <= len_cols; i++) {
+      results[i] = reinterpret_cast<jlong>(handles[i - 1]);
+    }
+
+    jlongArray jresults = env->NewLongArray(resultsLen);
+    if (jresults == nullptr) {
+        // exception thrown: OutOfMemoryError
+        return nullptr;
+    }
+    env->SetLongArrayRegion(jresults, 0, resultsLen, results.get());
+    if (env->ExceptionCheck()) {
+        // exception thrown: ArrayIndexOutOfBoundsException
+        return nullptr;
+    }
+    return jresults;
+  }
+
+  rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
+  return nullptr;
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionDB
+ * Method:    beginTransaction
+ * Signature: (JJ)J
+ */
+jlong Java_org_rocksdb_OptimisticTransactionDB_beginTransaction__JJ(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jwrite_options_handle) {
+  auto* optimistic_txn_db =
+      reinterpret_cast<rocksdb::OptimisticTransactionDB*>(jhandle);
+  auto* write_options =
+      reinterpret_cast<rocksdb::WriteOptions*>(jwrite_options_handle);
+  rocksdb::Transaction* txn =
+      optimistic_txn_db->BeginTransaction(*write_options);
+  return reinterpret_cast<jlong>(txn);
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionDB
+ * Method:    beginTransaction
+ * Signature: (JJJ)J
+ */
+jlong Java_org_rocksdb_OptimisticTransactionDB_beginTransaction__JJJ(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jwrite_options_handle,
+    jlong joptimistic_txn_options_handle) {
+  auto* optimistic_txn_db =
+      reinterpret_cast<rocksdb::OptimisticTransactionDB*>(jhandle);
+  auto* write_options =
+      reinterpret_cast<rocksdb::WriteOptions*>(jwrite_options_handle);
+  auto* optimistic_txn_options =
+      reinterpret_cast<rocksdb::OptimisticTransactionOptions*>(
+      joptimistic_txn_options_handle);
+  rocksdb::Transaction* txn =
+      optimistic_txn_db->BeginTransaction(*write_options,
+      *optimistic_txn_options);
+  return reinterpret_cast<jlong>(txn);
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionDB
+ * Method:    beginTransaction_withOld
+ * Signature: (JJJ)J
+ */
+jlong Java_org_rocksdb_OptimisticTransactionDB_beginTransaction_1withOld__JJJ(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jwrite_options_handle,
+    jlong jold_txn_handle) {
+  auto* optimistic_txn_db =
+      reinterpret_cast<rocksdb::OptimisticTransactionDB*>(jhandle);
+  auto* write_options =
+      reinterpret_cast<rocksdb::WriteOptions*>(jwrite_options_handle);
+  auto* old_txn =
+      reinterpret_cast<rocksdb::Transaction*>(
+      jold_txn_handle);
+  rocksdb::OptimisticTransactionOptions optimistic_txn_options;
+  rocksdb::Transaction* txn =
+      optimistic_txn_db->BeginTransaction(*write_options,
+          optimistic_txn_options, old_txn);
+
+  // RocksJava relies on the assumption that
+  // we do not allocate a new Transaction object
+  // when providing an old_optimistic_txn
+  assert(txn == old_txn);
+
+  return reinterpret_cast<jlong>(txn);
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionDB
+ * Method:    beginTransaction_withOld
+ * Signature: (JJJJ)J
+ */
+jlong Java_org_rocksdb_OptimisticTransactionDB_beginTransaction_1withOld__JJJJ(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jwrite_options_handle,
+    jlong joptimistic_txn_options_handle, jlong jold_txn_handle) {
+  auto* optimistic_txn_db =
+      reinterpret_cast<rocksdb::OptimisticTransactionDB*>(jhandle);
+  auto* write_options =
+      reinterpret_cast<rocksdb::WriteOptions*>(jwrite_options_handle);
+  auto* optimistic_txn_options =
+      reinterpret_cast<rocksdb::OptimisticTransactionOptions*>(
+        joptimistic_txn_options_handle);
+  auto* old_txn =
+      reinterpret_cast<rocksdb::Transaction*>(
+      jold_txn_handle);
+  rocksdb::Transaction* txn =
+      optimistic_txn_db->BeginTransaction(*write_options,
+      *optimistic_txn_options, old_txn);
+
+  // RocksJava relies on the assumption that
+  // we do not allocate a new Transaction object
+  // when providing an old_optimisic_txn
+  assert(txn == old_txn);
+
+  return reinterpret_cast<jlong>(txn);
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionDB
+ * Method:    getBaseDB
+ * Signature: (J)J
+ */
+jlong Java_org_rocksdb_OptimisticTransactionDB_getBaseDB(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* optimistic_txn_db =
+      reinterpret_cast<rocksdb::OptimisticTransactionDB*>(jhandle);
+  return reinterpret_cast<jlong>(optimistic_txn_db->GetBaseDB());
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionDB
+ * Method:    disposeInternal
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_OptimisticTransactionDB_disposeInternal(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  delete reinterpret_cast<rocksdb::OptimisticTransactionDB*>(jhandle);
+}

java/rocksjni/optimistic_transaction_options.cc

@@ -0,0 +1,72 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// This file implements the "bridge" between Java and C++
+// for rocksdb::OptimisticTransactionOptions.
+
+#include <jni.h>
+
+#include "include/org_rocksdb_OptimisticTransactionOptions.h"
+
+#include "rocksdb/comparator.h"
+#include "rocksdb/utilities/optimistic_transaction_db.h"
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionOptions
+ * Method:    newOptimisticTransactionOptions
+ * Signature: ()J
+ */
+jlong Java_org_rocksdb_OptimisticTransactionOptions_newOptimisticTransactionOptions(
+    JNIEnv* env, jclass jcls) {
+  rocksdb::OptimisticTransactionOptions* opts =
+      new rocksdb::OptimisticTransactionOptions();
+  return reinterpret_cast<jlong>(opts);
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionOptions
+ * Method:    isSetSnapshot
+ * Signature: (J)Z
+ */
+jboolean Java_org_rocksdb_OptimisticTransactionOptions_isSetSnapshot(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* opts =
+      reinterpret_cast<rocksdb::OptimisticTransactionOptions*>(jhandle);
+  return opts->set_snapshot;
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionOptions
+ * Method:    setSetSnapshot
+ * Signature: (JZ)V
+ */
+void Java_org_rocksdb_OptimisticTransactionOptions_setSetSnapshot(JNIEnv* env,
+    jobject jobj, jlong jhandle, jboolean jset_snapshot) {
+  auto* opts =
+      reinterpret_cast<rocksdb::OptimisticTransactionOptions*>(jhandle);
+  opts->set_snapshot = jset_snapshot;
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionOptions
+ * Method:    setComparator
+ * Signature: (JJ)V
+ */
+void Java_org_rocksdb_OptimisticTransactionOptions_setComparator(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jcomparator_handle) {
+  auto* opts =
+      reinterpret_cast<rocksdb::OptimisticTransactionOptions*>(jhandle);
+  opts->cmp = reinterpret_cast<rocksdb::Comparator*>(jcomparator_handle);
+}
+
+/*
+ * Class:     org_rocksdb_OptimisticTransactionOptions
+ * Method:    disposeInternal
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_OptimisticTransactionOptions_disposeInternal(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  delete reinterpret_cast<rocksdb::OptimisticTransactionOptions*>(jhandle);
+}

java/rocksjni/options.cc

@@ -146,19 +146,33 @@ void Java_org_rocksdb_Options_setComparatorHandle__JI(
 /*
  * Class:     org_rocksdb_Options
  * Method:    setComparatorHandle
- * Signature: (JJZ)V
+ * Signature: (JJB)V
  */
-void Java_org_rocksdb_Options_setComparatorHandle__JJZ(
+void Java_org_rocksdb_Options_setComparatorHandle__JJB(
     JNIEnv* env, jobject jobj, jlong jopt_handle, jlong jcomparator_handle,
-    jboolean is_direct) {
-  auto* opt = reinterpret_cast<rocksdb::Options*>(jopt_handle);
-  if(is_direct) {
-    opt->comparator =
-        reinterpret_cast<rocksdb::DirectComparatorJniCallback*>(jcomparator_handle);
-  } else {
-    opt->comparator =
-        reinterpret_cast<rocksdb::ComparatorJniCallback*>(jcomparator_handle);
+    jbyte jcomparator_type) {
+  rocksdb::Comparator *comparator = nullptr;
+  switch(jcomparator_type) {
+      // JAVA_COMPARATOR
+      case 0x0:
+        comparator =
+            reinterpret_cast<rocksdb::ComparatorJniCallback*>(jcomparator_handle);
+        break;
+
+      // JAVA_DIRECT_COMPARATOR
+      case 0x1:
+        comparator =
+            reinterpret_cast<rocksdb::DirectComparatorJniCallback*>(jcomparator_handle);
+        break;
+
+      // JAVA_NATIVE_COMPARATOR_WRAPPER
+      case 0x2:
+        comparator =
+            reinterpret_cast<rocksdb::Comparator*>(jcomparator_handle);
+        break;
   }
+  auto* opt = reinterpret_cast<rocksdb::Options*>(jopt_handle);
+  opt->comparator = comparator;
 }
 
 /*
@@ -2960,19 +2974,33 @@ void Java_org_rocksdb_ColumnFamilyOptions_setComparatorHandle__JI(
 /*
  * Class:     org_rocksdb_ColumnFamilyOptions
  * Method:    setComparatorHandle
- * Signature: (JJZ)V
+ * Signature: (JJB)V
  */
-void Java_org_rocksdb_ColumnFamilyOptions_setComparatorHandle__JJZ(
+void Java_org_rocksdb_ColumnFamilyOptions_setComparatorHandle__JJB(
     JNIEnv* env, jobject jobj, jlong jopt_handle, jlong jcomparator_handle,
-    jboolean is_direct) {
-  auto* opt = reinterpret_cast<rocksdb::ColumnFamilyOptions*>(jopt_handle);
-  if(is_direct) {
-    opt->comparator =
-        reinterpret_cast<rocksdb::DirectComparatorJniCallback*>(jcomparator_handle);
-  } else {
-    opt->comparator =
-        reinterpret_cast<rocksdb::ComparatorJniCallback*>(jcomparator_handle);
+    jbyte jcomparator_type) {
+  rocksdb::Comparator *comparator = nullptr;
+  switch(jcomparator_type) {
+      // JAVA_COMPARATOR
+      case 0x0:
+        comparator =
+            reinterpret_cast<rocksdb::ComparatorJniCallback*>(jcomparator_handle);
+        break;
+
+      // JAVA_DIRECT_COMPARATOR
+      case 0x1:
+        comparator =
+            reinterpret_cast<rocksdb::DirectComparatorJniCallback*>(jcomparator_handle);
+        break;
+
+      // JAVA_NATIVE_COMPARATOR_WRAPPER
+      case 0x2:
+        comparator =
+            reinterpret_cast<rocksdb::Comparator*>(jcomparator_handle);
+        break;
   }
+  auto* opt = reinterpret_cast<rocksdb::ColumnFamilyOptions*>(jopt_handle);
+  opt->comparator = comparator;
 }
 
 /*

java/rocksjni/rocks_callback_object.cc

@@ -24,4 +24,5 @@ void Java_org_rocksdb_RocksCallbackObject_disposeInternal(
       //   2) Comparator -> BaseComparatorJniCallback + JniCallback -> ComparatorJniCallback
       // I think this is okay, as Comparator and JniCallback both have virtual destructors...
       delete reinterpret_cast<rocksdb::JniCallback*>(handle);
+// @lint-ignore TXT4 T25377293 Grandfathered in
 }

java/rocksjni/sst_file_writerjni.cc

@@ -20,24 +20,34 @@
 /*
  * Class:     org_rocksdb_SstFileWriter
  * Method:    newSstFileWriter
- * Signature: (JJJZ)J
+ * Signature: (JJJB)J
  */
-jlong Java_org_rocksdb_SstFileWriter_newSstFileWriter__JJJZ(JNIEnv *env,
-    jclass jcls, jlong jenvoptions,  jlong joptions, jlong jcomparator,
-    jboolean is_direct) {
+jlong Java_org_rocksdb_SstFileWriter_newSstFileWriter__JJJB(JNIEnv *env,
+    jclass jcls, jlong jenvoptions,  jlong joptions, jlong jcomparator_handle,
+    jbyte jcomparator_type) {
+  rocksdb::Comparator *comparator = nullptr;
+  switch(jcomparator_type) {
+      // JAVA_COMPARATOR
+      case 0x0:
+        comparator =
+            reinterpret_cast<rocksdb::ComparatorJniCallback*>(jcomparator_handle);
+        break;
+
+      // JAVA_DIRECT_COMPARATOR
+      case 0x1:
+        comparator =
+            reinterpret_cast<rocksdb::DirectComparatorJniCallback*>(jcomparator_handle);
+        break;
+
+      // JAVA_NATIVE_COMPARATOR_WRAPPER
+      case 0x2:
+        comparator =
+            reinterpret_cast<rocksdb::Comparator*>(jcomparator_handle);
+        break;
+  }
   auto *env_options =
       reinterpret_cast<const rocksdb::EnvOptions *>(jenvoptions);
   auto *options = reinterpret_cast<const rocksdb::Options *>(joptions);
-
-  rocksdb::Comparator *comparator = nullptr;
-  if(is_direct) {
-    comparator =
-        reinterpret_cast<rocksdb::DirectComparatorJniCallback*>(jcomparator);
-  } else {
-    comparator =
-        reinterpret_cast<rocksdb::ComparatorJniCallback*>(jcomparator);
-  }
-
   rocksdb::SstFileWriter *sst_file_writer =
       new rocksdb::SstFileWriter(*env_options, *options, comparator);
   return reinterpret_cast<jlong>(sst_file_writer);

java/rocksjni/statisticsjni.cc

@@ -30,4 +30,5 @@ namespace rocksdb {
 
     return true;
   }
+// @lint-ignore TXT4 T25377293 Grandfathered in
 };

java/rocksjni/statisticsjni.h

@@ -30,4 +30,5 @@ namespace rocksdb {
 
 }  // namespace rocksdb
 
+// @lint-ignore TXT4 T25377293 Grandfathered in
 #endif  // JAVA_ROCKSJNI_STATISTICSJNI_H_

java/rocksjni/transaction_db.cc

@@ -0,0 +1,431 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// This file implements the "bridge" between Java and C++
+// for rocksdb::TransactionDB.
+
+#include <jni.h>
+#include <functional>
+#include <memory>
+#include <utility>
+
+
+#include "include/org_rocksdb_TransactionDB.h"
+
+#include "rocksdb/options.h"
+#include "rocksdb/utilities/transaction.h"
+#include "rocksdb/utilities/transaction_db.h"
+
+#include "rocksjni/portal.h"
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    open
+ * Signature: (JJLjava/lang/String;)J
+ */
+jlong Java_org_rocksdb_TransactionDB_open__JJLjava_lang_String_2(JNIEnv* env,
+    jclass jcls, jlong joptions_handle, jlong jtxn_db_options_handle,
+    jstring jdb_path) {
+  auto* options = reinterpret_cast<rocksdb::Options*>(joptions_handle);
+  auto* txn_db_options =
+      reinterpret_cast<rocksdb::TransactionDBOptions*>(jtxn_db_options_handle);
+  rocksdb::TransactionDB* tdb = nullptr;
+  const char* db_path = env->GetStringUTFChars(jdb_path, nullptr);
+  if (db_path == nullptr) {
+    // exception thrown: OutOfMemoryError
+    return 0;
+  }
+  rocksdb::Status s =
+      rocksdb::TransactionDB::Open(*options, *txn_db_options, db_path, &tdb);
+  env->ReleaseStringUTFChars(jdb_path, db_path);
+
+  if (s.ok()) {
+    return reinterpret_cast<jlong>(tdb);
+  } else {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
+    return 0;
+  }
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    open
+ * Signature: (JJLjava/lang/String;[[B[J)[J
+ */
+jlongArray Java_org_rocksdb_TransactionDB_open__JJLjava_lang_String_2_3_3B_3J(
+    JNIEnv* env, jclass jcls, jlong jdb_options_handle,
+    jlong jtxn_db_options_handle, jstring jdb_path,
+    jobjectArray jcolumn_names,
+    jlongArray jcolumn_options_handles) {
+  const char* db_path = env->GetStringUTFChars(jdb_path, nullptr);
+  if (db_path == nullptr) {
+    // exception thrown: OutOfMemoryError
+    return nullptr;
+  }
+
+  const jsize len_cols = env->GetArrayLength(jcolumn_names);
+  if (env->EnsureLocalCapacity(len_cols) != 0) {
+    // out of memory
+    env->ReleaseStringUTFChars(jdb_path, db_path);
+    return nullptr;
+  }
+
+  jlong* jco = env->GetLongArrayElements(jcolumn_options_handles, nullptr);
+  if (jco == nullptr) {
+      // exception thrown: OutOfMemoryError
+      env->ReleaseStringUTFChars(jdb_path, db_path);
+      return nullptr;
+  }
+  std::vector<rocksdb::ColumnFamilyDescriptor> column_families;
+  for (int i = 0; i < len_cols; i++) {
+    const jobject jcn = env->GetObjectArrayElement(jcolumn_names, i);
+    if (env->ExceptionCheck()) {
+        // exception thrown: ArrayIndexOutOfBoundsException
+        env->ReleaseLongArrayElements(jcolumn_options_handles, jco, JNI_ABORT);
+        env->ReleaseStringUTFChars(jdb_path, db_path);
+        return nullptr;
+    }
+    const jbyteArray jcn_ba = reinterpret_cast<jbyteArray>(jcn);
+    jbyte* jcf_name = env->GetByteArrayElements(jcn_ba, nullptr);
+    if (jcf_name == nullptr) {
+      // exception thrown: OutOfMemoryError
+      env->DeleteLocalRef(jcn);
+      env->ReleaseLongArrayElements(jcolumn_options_handles, jco, JNI_ABORT);
+      env->ReleaseStringUTFChars(jdb_path, db_path);
+      return nullptr;
+    }
+
+    const int jcf_name_len = env->GetArrayLength(jcn_ba);
+    if (env->EnsureLocalCapacity(jcf_name_len) != 0) {
+        // out of memory
+        env->ReleaseByteArrayElements(jcn_ba, jcf_name, JNI_ABORT);
+        env->DeleteLocalRef(jcn);
+        env->ReleaseLongArrayElements(jcolumn_options_handles, jco, JNI_ABORT);
+        env->ReleaseStringUTFChars(jdb_path, db_path);
+        return nullptr;
+    }
+    const std::string cf_name(reinterpret_cast<char *>(jcf_name), jcf_name_len);
+    const rocksdb::ColumnFamilyOptions* cf_options =
+      reinterpret_cast<rocksdb::ColumnFamilyOptions*>(jco[i]);
+    column_families.push_back(
+      rocksdb::ColumnFamilyDescriptor(cf_name, *cf_options));
+
+    env->ReleaseByteArrayElements(jcn_ba, jcf_name, JNI_ABORT);
+    env->DeleteLocalRef(jcn);
+  }
+  env->ReleaseLongArrayElements(jcolumn_options_handles, jco, JNI_ABORT);
+
+  auto* db_options = reinterpret_cast<rocksdb::DBOptions*>(jdb_options_handle);
+  auto* txn_db_options =
+      reinterpret_cast<rocksdb::TransactionDBOptions*>(jtxn_db_options_handle);
+  std::vector<rocksdb::ColumnFamilyHandle*> handles;
+  rocksdb::TransactionDB* tdb = nullptr;
+  const rocksdb::Status s = rocksdb::TransactionDB::Open(*db_options, *txn_db_options,
+      db_path, column_families, &handles, &tdb);
+
+  // check if open operation was successful
+  if (s.ok()) {
+    const jsize resultsLen = 1 + len_cols;  // db handle + column family handles
+    std::unique_ptr<jlong[]> results =
+        std::unique_ptr<jlong[]>(new jlong[resultsLen]);
+    results[0] = reinterpret_cast<jlong>(tdb);
+    for (int i = 1; i <= len_cols; i++) {
+      results[i] = reinterpret_cast<jlong>(handles[i - 1]);
+    }
+
+    jlongArray jresults = env->NewLongArray(resultsLen);
+    if (jresults == nullptr) {
+        // exception thrown: OutOfMemoryError
+        return nullptr;
+    }
+    env->SetLongArrayRegion(jresults, 0, resultsLen, results.get());
+    if (env->ExceptionCheck()) {
+        // exception thrown: ArrayIndexOutOfBoundsException
+        env->DeleteLocalRef(jresults);
+        return nullptr;
+    }
+    return jresults;
+  } else {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
+    return nullptr;
+  }
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    beginTransaction
+ * Signature: (JJ)J
+ */
+jlong Java_org_rocksdb_TransactionDB_beginTransaction__JJ(JNIEnv* env,
+    jobject jobj, jlong jhandle, jlong jwrite_options_handle) {
+  auto* txn_db = reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+  auto* write_options =
+      reinterpret_cast<rocksdb::WriteOptions*>(jwrite_options_handle);
+  rocksdb::Transaction* txn = txn_db->BeginTransaction(*write_options);
+  return reinterpret_cast<jlong>(txn);
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    beginTransaction
+ * Signature: (JJJ)J
+ */
+jlong Java_org_rocksdb_TransactionDB_beginTransaction__JJJ(JNIEnv* env,
+    jobject jobj, jlong jhandle, jlong jwrite_options_handle,
+    jlong jtxn_options_handle) {
+  auto* txn_db = reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+  auto* write_options =
+      reinterpret_cast<rocksdb::WriteOptions*>(jwrite_options_handle);
+  auto* txn_options =
+      reinterpret_cast<rocksdb::TransactionOptions*>(jtxn_options_handle);
+  rocksdb::Transaction* txn =
+      txn_db->BeginTransaction(*write_options, *txn_options);
+  return reinterpret_cast<jlong>(txn);
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    beginTransaction_withOld
+ * Signature: (JJJ)J
+ */
+jlong Java_org_rocksdb_TransactionDB_beginTransaction_1withOld__JJJ(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jwrite_options_handle,
+    jlong jold_txn_handle) {
+  auto* txn_db = reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+  auto* write_options =
+      reinterpret_cast<rocksdb::WriteOptions*>(jwrite_options_handle);
+  auto* old_txn = reinterpret_cast<rocksdb::Transaction*>(jold_txn_handle);
+  rocksdb::TransactionOptions txn_options;
+  rocksdb::Transaction* txn =
+      txn_db->BeginTransaction(*write_options, txn_options, old_txn);
+
+  // RocksJava relies on the assumption that
+  // we do not allocate a new Transaction object
+  // when providing an old_txn
+  assert(txn == old_txn);
+
+  return reinterpret_cast<jlong>(txn);
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    beginTransaction_withOld
+ * Signature: (JJJJ)J
+ */
+jlong Java_org_rocksdb_TransactionDB_beginTransaction_1withOld__JJJJ(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jwrite_options_handle,
+    jlong jtxn_options_handle, jlong jold_txn_handle) {
+  auto* txn_db = reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+  auto* write_options =
+      reinterpret_cast<rocksdb::WriteOptions*>(jwrite_options_handle);
+  auto* txn_options =
+      reinterpret_cast<rocksdb::TransactionOptions*>(jtxn_options_handle);
+  auto* old_txn = reinterpret_cast<rocksdb::Transaction*>(jold_txn_handle);
+  rocksdb::Transaction* txn = txn_db->BeginTransaction(*write_options,
+      *txn_options, old_txn);
+
+  // RocksJava relies on the assumption that
+  // we do not allocate a new Transaction object
+  // when providing an old_txn
+  assert(txn == old_txn);
+
+  return reinterpret_cast<jlong>(txn);
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    getTransactionByName
+ * Signature: (JLjava/lang/String;)J
+ */
+jlong Java_org_rocksdb_TransactionDB_getTransactionByName(JNIEnv* env,
+    jobject jobj, jlong jhandle, jstring jname) {
+  auto* txn_db = reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+  const char* name = env->GetStringUTFChars(jname, nullptr);
+  if (name == nullptr) {
+      // exception thrown: OutOfMemoryError
+      return 0;
+  }
+  rocksdb::Transaction* txn = txn_db->GetTransactionByName(name);
+  env->ReleaseStringUTFChars(jname, name);
+  return reinterpret_cast<jlong>(txn);
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    getAllPreparedTransactions
+ * Signature: (J)[J
+ */
+jlongArray Java_org_rocksdb_TransactionDB_getAllPreparedTransactions(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* txn_db = reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+  std::vector<rocksdb::Transaction*> txns;
+  txn_db->GetAllPreparedTransactions(&txns);
+
+  const size_t size = txns.size();
+  assert(size < UINT32_MAX);  // does it fit in a jint?
+
+  const jsize len = static_cast<jsize>(size);
+  jlong tmp[len];
+  for (jsize i = 0; i < len; ++i) {
+      tmp[i] = reinterpret_cast<jlong>(txns[i]);
+  }
+
+  jlongArray jtxns = env->NewLongArray(len);
+  if (jtxns == nullptr) {
+      // exception thrown: OutOfMemoryError
+      return nullptr;
+  }
+  env->SetLongArrayRegion(jtxns, 0, len, tmp);
+  if (env->ExceptionCheck()) {
+      // exception thrown: ArrayIndexOutOfBoundsException
+      env->DeleteLocalRef(jtxns);
+      return nullptr;
+  }
+
+  return jtxns;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    getLockStatusData
+ * Signature: (J)Ljava/util/Map;
+ */
+jobject Java_org_rocksdb_TransactionDB_getLockStatusData(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* txn_db = reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+  const std::unordered_multimap<uint32_t, rocksdb::KeyLockInfo> lock_status_data =
+      txn_db->GetLockStatusData();
+  const jobject jlock_status_data = rocksdb::HashMapJni::construct(env,
+        static_cast<uint32_t>(lock_status_data.size()));
+  if (jlock_status_data == nullptr) {
+      // exception occurred
+      return nullptr;
+  }
+
+  const rocksdb::HashMapJni::FnMapKV<const int32_t, const rocksdb::KeyLockInfo> fn_map_kv =
+      [env, txn_db, &lock_status_data](const std::pair<const int32_t, const rocksdb::KeyLockInfo>& pair) {
+          const jobject jlong_column_family_id =
+              rocksdb::LongJni::valueOf(env, pair.first);
+          if (jlong_column_family_id == nullptr) {
+              // an error occurred
+              return std::unique_ptr<std::pair<jobject, jobject>>(nullptr);
+          }
+          const jobject jkey_lock_info =
+              rocksdb::KeyLockInfoJni::construct(env, pair.second);
+          if (jkey_lock_info == nullptr) {
+             // an error occurred
+             return std::unique_ptr<std::pair<jobject, jobject>>(nullptr);
+          }
+          return std::unique_ptr<std::pair<jobject, jobject>>(new std::pair<jobject, jobject>(jlong_column_family_id,
+              jkey_lock_info));
+      };
+
+  if(!rocksdb::HashMapJni::putAll(env, jlock_status_data,
+      lock_status_data.begin(), lock_status_data.end(), fn_map_kv)) {
+    // exception occcurred
+    return nullptr;
+  }
+
+  return jlock_status_data;
+}
+
+/*
+* Class:     org_rocksdb_TransactionDB
+* Method:    getDeadlockInfoBuffer
+* Signature: (J)[Lorg/rocksdb/TransactionDB/DeadlockPath;
+*/
+jobjectArray Java_org_rocksdb_TransactionDB_getDeadlockInfoBuffer(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* txn_db = reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+  const std::vector<rocksdb::DeadlockPath> deadlock_info_buffer =
+      txn_db->GetDeadlockInfoBuffer();
+
+  const jsize deadlock_info_buffer_len =
+      static_cast<jsize>(deadlock_info_buffer.size());
+  jobjectArray jdeadlock_info_buffer =
+      env->NewObjectArray(deadlock_info_buffer_len,
+        rocksdb::DeadlockPathJni::getJClass(env), nullptr);
+  if (jdeadlock_info_buffer == nullptr) {
+      // exception thrown: OutOfMemoryError
+      return nullptr;
+  }
+  jsize jdeadlock_info_buffer_offset = 0;
+
+  auto buf_end = deadlock_info_buffer.end();
+  for (auto buf_it = deadlock_info_buffer.begin(); buf_it != buf_end; ++buf_it) {
+    const rocksdb::DeadlockPath deadlock_path = *buf_it;
+    const std::vector<rocksdb::DeadlockInfo> deadlock_infos
+        = deadlock_path.path;
+    const jsize deadlock_infos_len =
+        static_cast<jsize>(deadlock_info_buffer.size());
+    jobjectArray jdeadlock_infos = env->NewObjectArray(deadlock_infos_len,
+        rocksdb::DeadlockInfoJni::getJClass(env), nullptr);
+    if (jdeadlock_infos == nullptr) {
+        // exception thrown: OutOfMemoryError
+        env->DeleteLocalRef(jdeadlock_info_buffer);
+        return nullptr;
+    }
+    jsize jdeadlock_infos_offset = 0;
+
+    auto infos_end = deadlock_infos.end();
+    for (auto infos_it = deadlock_infos.begin(); infos_it != infos_end; ++infos_it) {
+        const rocksdb::DeadlockInfo deadlock_info = *infos_it;
+        const jobject jdeadlock_info = rocksdb::TransactionDBJni::newDeadlockInfo(
+            env, jobj, deadlock_info.m_txn_id, deadlock_info.m_cf_id,
+            deadlock_info.m_waiting_key, deadlock_info.m_exclusive);
+        if (jdeadlock_info == nullptr) {
+            // exception occcurred
+            env->DeleteLocalRef(jdeadlock_info_buffer);
+            return nullptr;
+        }
+        env->SetObjectArrayElement(jdeadlock_infos, jdeadlock_infos_offset++, jdeadlock_info);
+        if (env->ExceptionCheck()) {
+            // exception thrown: ArrayIndexOutOfBoundsException or ArrayStoreException
+            env->DeleteLocalRef(jdeadlock_info);
+            env->DeleteLocalRef(jdeadlock_info_buffer);
+            return nullptr;
+        }
+    }
+
+    const jobject jdeadlock_path =
+        rocksdb::DeadlockPathJni::construct(env, jdeadlock_infos,
+            deadlock_path.limit_exceeded);
+    if(jdeadlock_path == nullptr) {
+        // exception occcurred
+        env->DeleteLocalRef(jdeadlock_info_buffer);
+        return nullptr;
+    }
+    env->SetObjectArrayElement(jdeadlock_info_buffer, jdeadlock_info_buffer_offset++, jdeadlock_path);
+    if (env->ExceptionCheck()) {
+        // exception thrown: ArrayIndexOutOfBoundsException or ArrayStoreException
+        env->DeleteLocalRef(jdeadlock_path);
+        env->DeleteLocalRef(jdeadlock_info_buffer);
+        return nullptr;
+    }
+  }
+
+  return jdeadlock_info_buffer;
+}
+
+/*
+* Class:     org_rocksdb_TransactionDB
+* Method:    setDeadlockInfoBufferSize
+* Signature: (JI)V
+*/
+void Java_org_rocksdb_TransactionDB_setDeadlockInfoBufferSize(
+    JNIEnv* env, jobject jobj, jlong jhandle, jint jdeadlock_info_buffer_size) {
+  auto* txn_db = reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+  txn_db->SetDeadlockInfoBufferSize(jdeadlock_info_buffer_size);
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDB
+ * Method:    disposeInternal
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_TransactionDB_disposeInternal(JNIEnv* env, jobject jobj,
+    jlong jhandle) {
+  delete reinterpret_cast<rocksdb::TransactionDB*>(jhandle);
+}

java/rocksjni/transaction_db_options.cc

@@ -0,0 +1,147 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// This file implements the "bridge" between Java and C++
+// for rocksdb::TransactionDBOptions.
+
+#include <jni.h>
+
+#include "include/org_rocksdb_TransactionDBOptions.h"
+
+#include "rocksdb/utilities/transaction_db.h"
+
+#include "rocksjni/portal.h"
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    newTransactionDBOptions
+ * Signature: ()J
+ */
+jlong Java_org_rocksdb_TransactionDBOptions_newTransactionDBOptions(
+    JNIEnv* env, jclass jcls) {
+  rocksdb::TransactionDBOptions* opts = new rocksdb::TransactionDBOptions();
+  return reinterpret_cast<jlong>(opts);
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    getMaxNumLocks
+ * Signature: (J)J
+ */
+jlong Java_org_rocksdb_TransactionDBOptions_getMaxNumLocks(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  return opts->max_num_locks;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    setMaxNumLocks
+ * Signature: (JJ)V
+ */
+void Java_org_rocksdb_TransactionDBOptions_setMaxNumLocks(JNIEnv* env,
+    jobject jobj, jlong jhandle, jlong jmax_num_locks) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  opts->max_num_locks = jmax_num_locks;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    getNumStripes
+ * Signature: (J)J
+ */
+jlong Java_org_rocksdb_TransactionDBOptions_getNumStripes(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  return opts->num_stripes;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    setNumStripes
+ * Signature: (JJ)V
+ */
+void Java_org_rocksdb_TransactionDBOptions_setNumStripes(JNIEnv* env,
+    jobject jobj, jlong jhandle, jlong jnum_stripes) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  opts->num_stripes = jnum_stripes;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    getTransactionLockTimeout
+ * Signature: (J)J
+ */
+jlong Java_org_rocksdb_TransactionDBOptions_getTransactionLockTimeout(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  return opts->transaction_lock_timeout;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    setTransactionLockTimeout
+ * Signature: (JJ)V
+ */
+void Java_org_rocksdb_TransactionDBOptions_setTransactionLockTimeout(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jtransaction_lock_timeout) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  opts->transaction_lock_timeout = jtransaction_lock_timeout;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    getDefaultLockTimeout
+ * Signature: (J)J
+ */
+jlong Java_org_rocksdb_TransactionDBOptions_getDefaultLockTimeout(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  return opts->default_lock_timeout;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    setDefaultLockTimeout
+ * Signature: (JJ)V
+ */
+void Java_org_rocksdb_TransactionDBOptions_setDefaultLockTimeout(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jdefault_lock_timeout) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  opts->default_lock_timeout = jdefault_lock_timeout;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    getWritePolicy
+ * Signature: (J)B
+ */
+jbyte Java_org_rocksdb_TransactionDBOptions_getWritePolicy(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  return rocksdb::TxnDBWritePolicyJni::toJavaTxnDBWritePolicy(opts->write_policy);
+}
+
+/*
+* Class:     org_rocksdb_TransactionDBOptions
+* Method:    setWritePolicy
+* Signature: (JB)V
+*/
+void Java_org_rocksdb_TransactionDBOptions_setWritePolicy(
+    JNIEnv* env, jobject jobj, jlong jhandle, jbyte jwrite_policy) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+  opts->write_policy =
+      rocksdb::TxnDBWritePolicyJni::toCppTxnDBWritePolicy(jwrite_policy);
+}
+
+/*
+ * Class:     org_rocksdb_TransactionDBOptions
+ * Method:    disposeInternal
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_TransactionDBOptions_disposeInternal(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  delete reinterpret_cast<rocksdb::TransactionDBOptions*>(jhandle);
+}

java/rocksjni/transaction_notifier.cc

@@ -0,0 +1,42 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// This file implements the "bridge" between Java and C++
+// for rocksdb::TransactionNotifier.
+
+#include <jni.h>
+
+#include "include/org_rocksdb_AbstractTransactionNotifier.h"
+#include "rocksjni/transaction_notifier_jnicallback.h"
+
+/*
+ * Class:     org_rocksdb_AbstractTransactionNotifier
+ * Method:    createNewTransactionNotifier
+ * Signature: ()J
+ */
+jlong Java_org_rocksdb_AbstractTransactionNotifier_createNewTransactionNotifier(
+    JNIEnv* env, jobject jobj) {
+  auto* transaction_notifier =
+      new rocksdb::TransactionNotifierJniCallback(env, jobj);
+  auto* sptr_transaction_notifier =
+      new std::shared_ptr<rocksdb::TransactionNotifierJniCallback>(
+          transaction_notifier);
+  return reinterpret_cast<jlong>(sptr_transaction_notifier);
+}
+
+/*
+ * Class:     org_rocksdb_AbstractTransactionNotifier
+ * Method:    disposeInternal
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_AbstractTransactionNotifier_disposeInternal(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  // TODO(AR) refactor to use JniCallback::JniCallback
+  // when https://github.com/facebook/rocksdb/pull/1241/ is merged
+  std::shared_ptr<rocksdb::TransactionNotifierJniCallback>* handle =
+      reinterpret_cast<std::shared_ptr<
+          rocksdb::TransactionNotifierJniCallback>*>(jhandle);
+  delete handle;
+}

java/rocksjni/transaction_notifier_jnicallback.cc

@@ -0,0 +1,39 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// This file implements the callback "bridge" between Java and C++ for
+// rocksdb::TransactionNotifier.
+
+#include "rocksjni/transaction_notifier_jnicallback.h"
+#include "rocksjni/portal.h"
+
+namespace rocksdb {
+
+TransactionNotifierJniCallback::TransactionNotifierJniCallback(JNIEnv* env,
+    jobject jtransaction_notifier) : JniCallback(env, jtransaction_notifier) {
+  // we cache the method id for the JNI callback
+  m_jsnapshot_created_methodID =
+      AbstractTransactionNotifierJni::getSnapshotCreatedMethodId(env);
+}
+
+void TransactionNotifierJniCallback::SnapshotCreated(
+    const Snapshot* newSnapshot) {
+  jboolean attached_thread = JNI_FALSE;
+  JNIEnv* env = getJniEnv(&attached_thread);
+  assert(env != nullptr);
+
+  env->CallVoidMethod(m_jcallback_obj,
+      m_jsnapshot_created_methodID, reinterpret_cast<jlong>(newSnapshot));
+
+  if(env->ExceptionCheck()) {
+    // exception thrown from CallVoidMethod
+    env->ExceptionDescribe();  // print out exception to stderr
+    releaseJniEnv(attached_thread);
+    return;
+  }
+
+  releaseJniEnv(attached_thread);
+}
+}  // namespace rocksdb

java/rocksjni/transaction_notifier_jnicallback.h

@@ -0,0 +1,42 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// This file implements the callback "bridge" between Java and C++ for
+// rocksdb::TransactionNotifier.
+
+#ifndef JAVA_ROCKSJNI_TRANSACTION_NOTIFIER_JNICALLBACK_H_
+#define JAVA_ROCKSJNI_TRANSACTION_NOTIFIER_JNICALLBACK_H_
+
+#include <jni.h>
+
+#include "rocksdb/utilities/transaction.h"
+#include "rocksjni/jnicallback.h"
+
+namespace rocksdb {
+
+/**
+ * This class acts as a bridge between C++
+ * and Java. The methods in this class will be
+ * called back from the RocksDB TransactionDB or OptimisticTransactionDB (C++),
+ * we then callback to the appropriate Java method
+ * this enables TransactionNotifier to be implemented in Java.
+ *
+ * Unlike RocksJava's Comparator JNI Callback, we do not attempt
+ * to reduce Java object allocations by caching the Snapshot object
+ * presented to the callback. This could be revisited in future
+ * if performance is lacking.
+ */
+class TransactionNotifierJniCallback: public JniCallback,
+    public TransactionNotifier {
+ public:
+  TransactionNotifierJniCallback(JNIEnv* env, jobject jtransaction_notifier);
+  virtual void SnapshotCreated(const Snapshot* newSnapshot);
+
+ private:
+  jmethodID m_jsnapshot_created_methodID;
+};
+}  // namespace rocksdb
+
+#endif  // JAVA_ROCKSJNI_TRANSACTION_NOTIFIER_JNICALLBACK_H_

java/rocksjni/transaction_options.cc

@@ -0,0 +1,166 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// This file implements the "bridge" between Java and C++
+// for rocksdb::TransactionOptions.
+
+#include <jni.h>
+
+#include "include/org_rocksdb_TransactionOptions.h"
+
+#include "rocksdb/utilities/transaction_db.h"
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    newTransactionOptions
+ * Signature: ()J
+ */
+jlong Java_org_rocksdb_TransactionOptions_newTransactionOptions(JNIEnv* env,
+    jclass jcls) {
+  auto* opts = new rocksdb::TransactionOptions();
+  return reinterpret_cast<jlong>(opts);
+}
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    isSetSnapshot
+ * Signature: (J)Z
+ */
+jboolean Java_org_rocksdb_TransactionOptions_isSetSnapshot(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  return opts->set_snapshot;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    setSetSnapshot
+ * Signature: (JZ)V
+ */
+void Java_org_rocksdb_TransactionOptions_setSetSnapshot(JNIEnv* env,
+    jobject jobj, jlong jhandle, jboolean jset_snapshot) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  opts->set_snapshot = jset_snapshot;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    isDeadlockDetect
+ * Signature: (J)Z
+ */
+jboolean Java_org_rocksdb_TransactionOptions_isDeadlockDetect(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  return opts->deadlock_detect;
+}
+
+/*
+* Class:     org_rocksdb_TransactionOptions
+* Method:    setDeadlockDetect
+* Signature: (JZ)V
+*/
+void Java_org_rocksdb_TransactionOptions_setDeadlockDetect(
+    JNIEnv* env, jobject jobj, jlong jhandle, jboolean jdeadlock_detect) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  opts->deadlock_detect = jdeadlock_detect;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    getLockTimeout
+ * Signature: (J)J
+ */
+jlong Java_org_rocksdb_TransactionOptions_getLockTimeout(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  return opts->lock_timeout;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    setLockTimeout
+ * Signature: (JJ)V
+ */
+void Java_org_rocksdb_TransactionOptions_setLockTimeout(JNIEnv* env,
+    jobject jobj, jlong jhandle, jlong jlock_timeout) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  opts->lock_timeout = jlock_timeout;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    getExpiration
+ * Signature: (J)J
+ */
+jlong Java_org_rocksdb_TransactionOptions_getExpiration(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  return opts->expiration;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    setExpiration
+ * Signature: (JJ)V
+ */
+void Java_org_rocksdb_TransactionOptions_setExpiration(JNIEnv* env,
+    jobject jobj, jlong jhandle, jlong jexpiration) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  opts->expiration = jexpiration;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    getDeadlockDetectDepth
+ * Signature: (J)J
+ */
+jlong Java_org_rocksdb_TransactionOptions_getDeadlockDetectDepth(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  return opts->deadlock_detect_depth;
+}
+
+/*
+* Class:     org_rocksdb_TransactionOptions
+* Method:    setDeadlockDetectDepth
+* Signature: (JJ)V
+*/
+void Java_org_rocksdb_TransactionOptions_setDeadlockDetectDepth(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jdeadlock_detect_depth) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  opts->deadlock_detect_depth = jdeadlock_detect_depth;
+}
+
+/*
+* Class:     org_rocksdb_TransactionOptions
+* Method:    getMaxWriteBatchSize
+* Signature: (J)J
+*/
+jlong Java_org_rocksdb_TransactionOptions_getMaxWriteBatchSize(
+    JNIEnv* env, jobject jobj, jlong jhandle) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  return opts->max_write_batch_size;
+}
+
+/*
+* Class:     org_rocksdb_TransactionOptions
+* Method:    setMaxWriteBatchSize
+* Signature: (JJ)V
+*/
+void Java_org_rocksdb_TransactionOptions_setMaxWriteBatchSize(
+    JNIEnv* env, jobject jobj, jlong jhandle, jlong jmax_write_batch_size) {
+  auto* opts = reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+  opts->max_write_batch_size = jmax_write_batch_size;
+}
+
+/*
+ * Class:     org_rocksdb_TransactionOptions
+ * Method:    disposeInternal
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_TransactionOptions_disposeInternal(JNIEnv* env,
+    jobject jobj, jlong jhandle) {
+  delete reinterpret_cast<rocksdb::TransactionOptions*>(jhandle);
+}

java/rocksjni/write_batch.cc

@@ -27,12 +27,34 @@
  * Method:    newWriteBatch
  * Signature: (I)J
  */
-jlong Java_org_rocksdb_WriteBatch_newWriteBatch(
+jlong Java_org_rocksdb_WriteBatch_newWriteBatch__I(
     JNIEnv* env, jclass jcls, jint jreserved_bytes) {
   auto* wb = new rocksdb::WriteBatch(static_cast<size_t>(jreserved_bytes));
   return reinterpret_cast<jlong>(wb);
 }
 
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    newWriteBatch
+ * Signature: ([BI)J
+ */
+jlong Java_org_rocksdb_WriteBatch_newWriteBatch___3BI(
+    JNIEnv* env, jclass jcls, jbyteArray jserialized,
+    jint jserialized_length) {
+  jboolean has_exception = JNI_FALSE;
+  std::string serialized = rocksdb::JniUtil::byteString<std::string>(env,
+    jserialized, jserialized_length,
+    [](const char* str, const size_t len) { return std::string(str, len); },
+    &has_exception);
+  if(has_exception == JNI_TRUE) {
+    // exception occurred
+    return 0;
+  }
+
+  auto* wb = new rocksdb::WriteBatch(serialized);
+  return reinterpret_cast<jlong>(wb);
+}
+
 /*
  * Class:     org_rocksdb_WriteBatch
  * Method:    count0
@@ -90,6 +112,37 @@ void Java_org_rocksdb_WriteBatch_rollbackToSavePoint0(
   rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
 }
 
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    popSavePoint
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_WriteBatch_popSavePoint(
+    JNIEnv* env, jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  auto s = wb->PopSavePoint();
+
+  if (s.ok()) {
+    return;
+  }
+  rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    setMaxBytes
+ * Signature: (JJ)V
+ */
+void Java_org_rocksdb_WriteBatch_setMaxBytes(
+    JNIEnv* env, jobject jobj, jlong jwb_handle, jlong jmax_bytes) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  wb->SetMaxBytes(static_cast<size_t>(jmax_bytes));
+}
+
 /*
  * Class:     org_rocksdb_WriteBatch
  * Method:    put
@@ -102,10 +155,13 @@ void Java_org_rocksdb_WriteBatch_put__J_3BI_3BI(
   auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
   assert(wb != nullptr);
   auto put = [&wb] (rocksdb::Slice key, rocksdb::Slice value) {
-    wb->Put(key, value);
+    return wb->Put(key, value);
   };
-  rocksdb::JniUtil::kv_op(put, env, jobj, jkey, jkey_len, jentry_value,
-      jentry_value_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(put, env,
+      jobj, jkey, jkey_len, jentry_value, jentry_value_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -122,10 +178,13 @@ void Java_org_rocksdb_WriteBatch_put__J_3BI_3BIJ(
   auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
   assert(cf_handle != nullptr);
   auto put = [&wb, &cf_handle] (rocksdb::Slice key, rocksdb::Slice value) {
-    wb->Put(cf_handle, key, value);
+    return wb->Put(cf_handle, key, value);
   };
-  rocksdb::JniUtil::kv_op(put, env, jobj, jkey, jkey_len, jentry_value,
-      jentry_value_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(put, env,
+      jobj, jkey, jkey_len, jentry_value, jentry_value_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -140,10 +199,13 @@ void Java_org_rocksdb_WriteBatch_merge__J_3BI_3BI(
   auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
   assert(wb != nullptr);
   auto merge = [&wb] (rocksdb::Slice key, rocksdb::Slice value) {
-    wb->Merge(key, value);
+    return wb->Merge(key, value);
   };
-  rocksdb::JniUtil::kv_op(merge, env, jobj, jkey, jkey_len, jentry_value,
-      jentry_value_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(merge, env,
+      jobj, jkey, jkey_len, jentry_value, jentry_value_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -160,34 +222,41 @@ void Java_org_rocksdb_WriteBatch_merge__J_3BI_3BIJ(
   auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
   assert(cf_handle != nullptr);
   auto merge = [&wb, &cf_handle] (rocksdb::Slice key, rocksdb::Slice value) {
-    wb->Merge(cf_handle, key, value);
+    return wb->Merge(cf_handle, key, value);
   };
-  rocksdb::JniUtil::kv_op(merge, env, jobj, jkey, jkey_len, jentry_value,
-      jentry_value_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(merge, env,
+      jobj, jkey, jkey_len, jentry_value, jentry_value_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
  * Class:     org_rocksdb_WriteBatch
- * Method:    remove
+ * Method:    delete
  * Signature: (J[BI)V
  */
-void Java_org_rocksdb_WriteBatch_remove__J_3BI(
+void Java_org_rocksdb_WriteBatch_delete__J_3BI(
     JNIEnv* env, jobject jobj, jlong jwb_handle,
     jbyteArray jkey, jint jkey_len) {
   auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
   assert(wb != nullptr);
   auto remove = [&wb] (rocksdb::Slice key) {
-    wb->Delete(key);
+    return wb->Delete(key);
   };
-  rocksdb::JniUtil::k_op(remove, env, jobj, jkey, jkey_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(remove, env,
+      jobj, jkey, jkey_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
  * Class:     org_rocksdb_WriteBatch
- * Method:    remove
+ * Method:    delete
  * Signature: (J[BIJ)V
  */
-void Java_org_rocksdb_WriteBatch_remove__J_3BIJ(
+void Java_org_rocksdb_WriteBatch_delete__J_3BIJ(
     JNIEnv* env, jobject jobj, jlong jwb_handle,
     jbyteArray jkey, jint jkey_len, jlong jcf_handle) {
   auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
@@ -195,9 +264,55 @@ void Java_org_rocksdb_WriteBatch_remove__J_3BIJ(
   auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
   assert(cf_handle != nullptr);
   auto remove = [&wb, &cf_handle] (rocksdb::Slice key) {
-    wb->Delete(cf_handle, key);
+    return wb->Delete(cf_handle, key);
   };
-  rocksdb::JniUtil::k_op(remove, env, jobj, jkey, jkey_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(remove, env,
+      jobj, jkey, jkey_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    singleDelete
+ * Signature: (J[BI)V
+ */
+void Java_org_rocksdb_WriteBatch_singleDelete__J_3BI(
+    JNIEnv* env, jobject jobj, jlong jwb_handle, jbyteArray jkey,
+    jint jkey_len) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+  auto single_delete = [&wb] (rocksdb::Slice key) {
+    return wb->SingleDelete(key);
+  };
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(single_delete,
+      env, jobj, jkey, jkey_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    singleDelete
+ * Signature: (J[BIJ)V
+ */
+void Java_org_rocksdb_WriteBatch_singleDelete__J_3BIJ(
+    JNIEnv* env, jobject jobj, jlong jwb_handle, jbyteArray jkey,
+    jint jkey_len, jlong jcf_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+  auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
+  assert(cf_handle != nullptr);
+  auto single_delete = [&wb, &cf_handle] (rocksdb::Slice key) {
+    return wb->SingleDelete(cf_handle, key);
+  };
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(single_delete,
+      env, jobj, jkey, jkey_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -205,19 +320,20 @@ void Java_org_rocksdb_WriteBatch_remove__J_3BIJ(
  * Method:    deleteRange
  * Signature: (J[BI[BI)V
  */
-JNIEXPORT void JNICALL Java_org_rocksdb_WriteBatch_deleteRange__J_3BI_3BI(
-    JNIEnv*, jobject, jlong, jbyteArray, jint, jbyteArray, jint);
-
 void Java_org_rocksdb_WriteBatch_deleteRange__J_3BI_3BI(
     JNIEnv* env, jobject jobj, jlong jwb_handle, jbyteArray jbegin_key,
     jint jbegin_key_len, jbyteArray jend_key, jint jend_key_len) {
   auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
   assert(wb != nullptr);
   auto deleteRange = [&wb](rocksdb::Slice beginKey, rocksdb::Slice endKey) {
-    wb->DeleteRange(beginKey, endKey);
+    return wb->DeleteRange(beginKey, endKey);
   };
-  rocksdb::JniUtil::kv_op(deleteRange, env, jobj, jbegin_key, jbegin_key_len,
-                          jend_key, jend_key_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(
+      deleteRange, env, jobj, jbegin_key, jbegin_key_len, jend_key,
+      jend_key_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -234,11 +350,15 @@ void Java_org_rocksdb_WriteBatch_deleteRange__J_3BI_3BIJ(
   auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
   assert(cf_handle != nullptr);
   auto deleteRange = [&wb, &cf_handle](rocksdb::Slice beginKey,
-                                       rocksdb::Slice endKey) {
-    wb->DeleteRange(cf_handle, beginKey, endKey);
+      rocksdb::Slice endKey) {
+    return wb->DeleteRange(cf_handle, beginKey, endKey);
   };
-  rocksdb::JniUtil::kv_op(deleteRange, env, jobj, jbegin_key, jbegin_key_len,
-                          jend_key, jend_key_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(
+      deleteRange, env, jobj, jbegin_key, jbegin_key_len, jend_key,
+      jend_key_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -252,9 +372,13 @@ void Java_org_rocksdb_WriteBatch_putLogData(
   auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
   assert(wb != nullptr);
   auto putLogData = [&wb] (rocksdb::Slice blob) {
-    wb->PutLogData(blob);
+    return wb->PutLogData(blob);
   };
-  rocksdb::JniUtil::k_op(putLogData, env, jobj, jblob, jblob_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(putLogData,
+      env, jobj, jblob, jblob_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -263,7 +387,7 @@ void Java_org_rocksdb_WriteBatch_putLogData(
  * Signature: (JJ)V
  */
 void Java_org_rocksdb_WriteBatch_iterate(
-    JNIEnv* env , jobject jobj, jlong jwb_handle, jlong handlerHandle) {
+    JNIEnv* env, jobject jobj, jlong jwb_handle, jlong handlerHandle) {
   auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
   assert(wb != nullptr);
 
@@ -276,6 +400,178 @@ void Java_org_rocksdb_WriteBatch_iterate(
   rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
 }
 
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    data
+ * Signature: (J)[B
+ */
+jbyteArray Java_org_rocksdb_WriteBatch_data(
+    JNIEnv* env, jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  auto data = wb->Data();
+  return rocksdb::JniUtil::copyBytes(env, data);
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    getDataSize
+ * Signature: (J)J
+ */
+jlong Java_org_rocksdb_WriteBatch_getDataSize(
+    JNIEnv* env, jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  auto data_size = wb->GetDataSize();
+  return static_cast<jlong>(data_size);
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    hasPut
+ * Signature: (J)Z
+ */
+jboolean Java_org_rocksdb_WriteBatch_hasPut(
+    JNIEnv* env, jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  return wb->HasPut();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    hasDelete
+ * Signature: (J)Z
+ */
+jboolean Java_org_rocksdb_WriteBatch_hasDelete(
+    JNIEnv* env, jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  return wb->HasDelete();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    hasSingleDelete
+ * Signature: (J)Z
+ */
+JNIEXPORT jboolean JNICALL Java_org_rocksdb_WriteBatch_hasSingleDelete(
+    JNIEnv* env , jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  return wb->HasSingleDelete();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    hasDeleteRange
+ * Signature: (J)Z
+ */
+JNIEXPORT jboolean JNICALL Java_org_rocksdb_WriteBatch_hasDeleteRange(
+    JNIEnv* env , jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  return wb->HasDeleteRange();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    hasMerge
+ * Signature: (J)Z
+ */
+JNIEXPORT jboolean JNICALL Java_org_rocksdb_WriteBatch_hasMerge(
+    JNIEnv* env , jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  return wb->HasMerge();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    hasBeginPrepare
+ * Signature: (J)Z
+ */
+JNIEXPORT jboolean JNICALL Java_org_rocksdb_WriteBatch_hasBeginPrepare(
+    JNIEnv* env , jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  return wb->HasBeginPrepare();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    hasEndPrepare
+ * Signature: (J)Z
+ */
+JNIEXPORT jboolean JNICALL Java_org_rocksdb_WriteBatch_hasEndPrepare(
+    JNIEnv* env , jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  return wb->HasEndPrepare();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    hasCommit
+ * Signature: (J)Z
+ */
+JNIEXPORT jboolean JNICALL Java_org_rocksdb_WriteBatch_hasCommit(
+    JNIEnv* env , jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  return wb->HasCommit();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    hasRollback
+ * Signature: (J)Z
+ */
+JNIEXPORT jboolean JNICALL Java_org_rocksdb_WriteBatch_hasRollback(
+    JNIEnv* env , jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  return wb->HasRollback();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    markWalTerminationPoint
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_WriteBatch_markWalTerminationPoint(
+    JNIEnv* env, jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  wb->MarkWalTerminationPoint();
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatch
+ * Method:    getWalTerminationPoint
+ * Signature: (J)Lorg/rocksdb/WriteBatch/SavePoint;
+ */
+jobject Java_org_rocksdb_WriteBatch_getWalTerminationPoint(
+    JNIEnv* env, jobject jobj, jlong jwb_handle) {
+  auto* wb = reinterpret_cast<rocksdb::WriteBatch*>(jwb_handle);
+  assert(wb != nullptr);
+
+  auto save_point = wb->GetWalTerminationPoint();
+  return rocksdb::WriteBatchSavePointJni::construct(env, save_point);
+}
+
 /*
  * Class:     org_rocksdb_WriteBatch
  * Method:    disposeInternal

java/rocksjni/write_batch_test.cc

@@ -87,8 +87,32 @@ jbyteArray Java_org_rocksdb_WriteBatchTest_getContents(
         state.append(")");
         count++;
         break;
+      case rocksdb::kTypeSingleDeletion:
+        state.append("SingleDelete(");
+        state.append(ikey.user_key.ToString());
+        state.append(")");
+        count++;
+        break;
+      case rocksdb::kTypeRangeDeletion:
+        state.append("DeleteRange(");
+        state.append(ikey.user_key.ToString());
+        state.append(", ");
+        state.append(iter->value().ToString());
+        state.append(")");
+        count++;
+        break;
+      case rocksdb::kTypeLogData:
+        state.append("LogData(");
+        state.append(ikey.user_key.ToString());
+        state.append(")");
+        count++;
+        break;
       default:
         assert(false);
+        state.append("Err:Expected(");
+        state.append(std::to_string(ikey.type));
+        state.append(")");
+        count++;
         break;
     }
     state.append("@");
@@ -96,8 +120,12 @@ jbyteArray Java_org_rocksdb_WriteBatchTest_getContents(
   }
   if (!s.ok()) {
     state.append(s.ToString());
-  } else if (count != rocksdb::WriteBatchInternal::Count(b)) {
-    state.append("CountMismatch()");
+  } else if (rocksdb::WriteBatchInternal::Count(b) != count) {
+    state.append("Err:CountMismatch(expected=");
+    state.append(std::to_string(rocksdb::WriteBatchInternal::Count(b)));
+    state.append(", actual=");
+    state.append(std::to_string(count));
+    state.append(")");
   }
   delete mem->Unref();
 

java/rocksjni/write_batch_with_index.cc

@@ -39,19 +39,31 @@ jlong Java_org_rocksdb_WriteBatchWithIndex_newWriteBatchWithIndex__Z(
 /*
  * Class:     org_rocksdb_WriteBatchWithIndex
  * Method:    newWriteBatchWithIndex
- * Signature: (JZIZ)J
+ * Signature: (JBIZ)J
  */
-jlong Java_org_rocksdb_WriteBatchWithIndex_newWriteBatchWithIndex__JZIZ(
+jlong Java_org_rocksdb_WriteBatchWithIndex_newWriteBatchWithIndex__JBIZ(
     JNIEnv* env, jclass jcls, jlong jfallback_index_comparator_handle,
-    jboolean is_direct, jint jreserved_bytes, jboolean joverwrite_key) {
-rocksdb::Comparator *fallback_comparator = nullptr;
-if(is_direct) {
-  fallback_comparator =
-      reinterpret_cast<rocksdb::DirectComparatorJniCallback*>(jfallback_index_comparator_handle);
-} else {
-  fallback_comparator =
-      reinterpret_cast<rocksdb::ComparatorJniCallback*>(jfallback_index_comparator_handle);
-}
+    jbyte jcomparator_type, jint jreserved_bytes, jboolean joverwrite_key) {
+  rocksdb::Comparator *fallback_comparator = nullptr;
+  switch(jcomparator_type) {
+      // JAVA_COMPARATOR
+      case 0x0:
+        fallback_comparator =
+            reinterpret_cast<rocksdb::ComparatorJniCallback*>(jfallback_index_comparator_handle);
+        break;
+
+      // JAVA_DIRECT_COMPARATOR
+      case 0x1:
+        fallback_comparator =
+            reinterpret_cast<rocksdb::DirectComparatorJniCallback*>(jfallback_index_comparator_handle);
+        break;
+
+      // JAVA_NATIVE_COMPARATOR_WRAPPER
+      case 0x2:
+        fallback_comparator =
+            reinterpret_cast<rocksdb::Comparator*>(jfallback_index_comparator_handle);
+        break;
+  }
   auto* wbwi =
       new rocksdb::WriteBatchWithIndex(
           fallback_comparator,
@@ -83,10 +95,13 @@ void Java_org_rocksdb_WriteBatchWithIndex_put__J_3BI_3BI(
   auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
   assert(wbwi != nullptr);
   auto put = [&wbwi] (rocksdb::Slice key, rocksdb::Slice value) {
-    wbwi->Put(key, value);
+    return wbwi->Put(key, value);
   };
-  rocksdb::JniUtil::kv_op(put, env, jobj, jkey, jkey_len, jentry_value,
-      jentry_value_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(put, env,
+      jobj, jkey, jkey_len, jentry_value, jentry_value_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -103,10 +118,13 @@ void Java_org_rocksdb_WriteBatchWithIndex_put__J_3BI_3BIJ(
   auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
   assert(cf_handle != nullptr);
   auto put = [&wbwi, &cf_handle] (rocksdb::Slice key, rocksdb::Slice value) {
-    wbwi->Put(cf_handle, key, value);
+    return wbwi->Put(cf_handle, key, value);
   };
-  rocksdb::JniUtil::kv_op(put, env, jobj, jkey, jkey_len, jentry_value,
-      jentry_value_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(put, env,
+      jobj, jkey, jkey_len, jentry_value, jentry_value_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -120,10 +138,13 @@ void Java_org_rocksdb_WriteBatchWithIndex_merge__J_3BI_3BI(
   auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
   assert(wbwi != nullptr);
   auto merge = [&wbwi] (rocksdb::Slice key, rocksdb::Slice value) {
-    wbwi->Merge(key, value);
+    return wbwi->Merge(key, value);
   };
-  rocksdb::JniUtil::kv_op(merge, env, jobj, jkey, jkey_len, jentry_value,
-      jentry_value_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(merge, env,
+      jobj, jkey, jkey_len, jentry_value, jentry_value_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -140,34 +161,41 @@ void Java_org_rocksdb_WriteBatchWithIndex_merge__J_3BI_3BIJ(
   auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
   assert(cf_handle != nullptr);
   auto merge = [&wbwi, &cf_handle] (rocksdb::Slice key, rocksdb::Slice value) {
-    wbwi->Merge(cf_handle, key, value);
+    return wbwi->Merge(cf_handle, key, value);
   };
-  rocksdb::JniUtil::kv_op(merge, env, jobj, jkey, jkey_len, jentry_value,
-      jentry_value_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(merge, env,
+      jobj, jkey, jkey_len, jentry_value, jentry_value_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
  * Class:     org_rocksdb_WriteBatchWithIndex
- * Method:    remove
+ * Method:    delete
  * Signature: (J[BI)V
  */
-void Java_org_rocksdb_WriteBatchWithIndex_remove__J_3BI(
+void Java_org_rocksdb_WriteBatchWithIndex_delete__J_3BI(
     JNIEnv* env, jobject jobj, jlong jwbwi_handle, jbyteArray jkey,
     jint jkey_len) {
   auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
   assert(wbwi != nullptr);
   auto remove = [&wbwi] (rocksdb::Slice key) {
-    wbwi->Delete(key);
+    return wbwi->Delete(key);
   };
-  rocksdb::JniUtil::k_op(remove, env, jobj, jkey, jkey_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(remove, env,
+      jobj, jkey, jkey_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
  * Class:     org_rocksdb_WriteBatchWithIndex
- * Method:    remove
+ * Method:    delete
  * Signature: (J[BIJ)V
  */
-void Java_org_rocksdb_WriteBatchWithIndex_remove__J_3BIJ(
+void Java_org_rocksdb_WriteBatchWithIndex_delete__J_3BIJ(
     JNIEnv* env, jobject jobj, jlong jwbwi_handle, jbyteArray jkey,
     jint jkey_len, jlong jcf_handle) {
   auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
@@ -175,9 +203,55 @@ void Java_org_rocksdb_WriteBatchWithIndex_remove__J_3BIJ(
   auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
   assert(cf_handle != nullptr);
   auto remove = [&wbwi, &cf_handle] (rocksdb::Slice key) {
-    wbwi->Delete(cf_handle, key);
+    return wbwi->Delete(cf_handle, key);
   };
-  rocksdb::JniUtil::k_op(remove, env, jobj, jkey, jkey_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(remove, env,
+      jobj, jkey, jkey_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatchWithIndex
+ * Method:    singleDelete
+ * Signature: (J[BI)V
+ */
+void Java_org_rocksdb_WriteBatchWithIndex_singleDelete__J_3BI(
+    JNIEnv* env, jobject jobj, jlong jwbwi_handle, jbyteArray jkey,
+    jint jkey_len) {
+  auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
+  assert(wbwi != nullptr);
+  auto single_delete = [&wbwi] (rocksdb::Slice key) {
+    return wbwi->SingleDelete(key);
+  };
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(single_delete,
+      env, jobj, jkey, jkey_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatchWithIndex
+ * Method:    singleDelete
+ * Signature: (J[BIJ)V
+ */
+void Java_org_rocksdb_WriteBatchWithIndex_singleDelete__J_3BIJ(
+  JNIEnv* env, jobject jobj, jlong jwbwi_handle, jbyteArray jkey,
+  jint jkey_len, jlong jcf_handle) {
+  auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
+  assert(wbwi != nullptr);
+  auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
+  assert(cf_handle != nullptr);
+  auto single_delete = [&wbwi, &cf_handle] (rocksdb::Slice key) {
+    return wbwi->SingleDelete(cf_handle, key);
+  };
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(single_delete,
+      env, jobj, jkey, jkey_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -191,10 +265,14 @@ void Java_org_rocksdb_WriteBatchWithIndex_deleteRange__J_3BI_3BI(
   auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
   assert(wbwi != nullptr);
   auto deleteRange = [&wbwi](rocksdb::Slice beginKey, rocksdb::Slice endKey) {
-    wbwi->DeleteRange(beginKey, endKey);
+    return wbwi->DeleteRange(beginKey, endKey);
   };
-  rocksdb::JniUtil::kv_op(deleteRange, env, jobj, jbegin_key, jbegin_key_len,
-                          jend_key, jend_key_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(
+      deleteRange, env, jobj, jbegin_key, jbegin_key_len, jend_key,
+      jend_key_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -211,11 +289,15 @@ void Java_org_rocksdb_WriteBatchWithIndex_deleteRange__J_3BI_3BIJ(
   auto* cf_handle = reinterpret_cast<rocksdb::ColumnFamilyHandle*>(jcf_handle);
   assert(cf_handle != nullptr);
   auto deleteRange = [&wbwi, &cf_handle](rocksdb::Slice beginKey,
-                                         rocksdb::Slice endKey) {
-    wbwi->DeleteRange(cf_handle, beginKey, endKey);
+      rocksdb::Slice endKey) {
+    return wbwi->DeleteRange(cf_handle, beginKey, endKey);
   };
-  rocksdb::JniUtil::kv_op(deleteRange, env, jobj, jbegin_key, jbegin_key_len,
-                          jend_key, jend_key_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::kv_op(
+      deleteRange, env, jobj, jbegin_key, jbegin_key_len, jend_key,
+      jend_key_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -229,9 +311,13 @@ void Java_org_rocksdb_WriteBatchWithIndex_putLogData(
   auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
   assert(wbwi != nullptr);
   auto putLogData = [&wbwi] (rocksdb::Slice blob) {
-    wbwi->PutLogData(blob);
+    return wbwi->PutLogData(blob);
   };
-  rocksdb::JniUtil::k_op(putLogData, env, jobj, jblob, jblob_len);
+  std::unique_ptr<rocksdb::Status> status = rocksdb::JniUtil::k_op(putLogData,
+      env, jobj, jblob, jblob_len);
+  if (status != nullptr && !status->ok()) {
+    rocksdb::RocksDBExceptionJni::ThrowNew(env, status);
+  }
 }
 
 /*
@@ -279,6 +365,54 @@ void Java_org_rocksdb_WriteBatchWithIndex_rollbackToSavePoint0(
   rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
 }
 
+/*
+ * Class:     org_rocksdb_WriteBatchWithIndex
+ * Method:    popSavePoint
+ * Signature: (J)V
+ */
+void Java_org_rocksdb_WriteBatchWithIndex_popSavePoint(
+    JNIEnv* env, jobject jobj, jlong jwbwi_handle) {
+  auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
+  assert(wbwi != nullptr);
+
+  auto s = wbwi->PopSavePoint();
+
+  if (s.ok()) {
+    return;
+  }
+
+  rocksdb::RocksDBExceptionJni::ThrowNew(env, s);
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatchWithIndex
+ * Method:    setMaxBytes
+ * Signature: (JJ)V
+ */
+void Java_org_rocksdb_WriteBatchWithIndex_setMaxBytes(
+    JNIEnv* env, jobject jobj, jlong jwbwi_handle, jlong jmax_bytes) {
+  auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
+  assert(wbwi != nullptr);
+
+  wbwi->SetMaxBytes(static_cast<size_t>(jmax_bytes));
+}
+
+/*
+ * Class:     org_rocksdb_WriteBatchWithIndex
+ * Method:    getWriteBatch
+ * Signature: (J)Lorg/rocksdb/WriteBatch;
+ */
+jobject Java_org_rocksdb_WriteBatchWithIndex_getWriteBatch(
+    JNIEnv* env, jobject jobj, jlong jwbwi_handle) {
+  auto* wbwi = reinterpret_cast<rocksdb::WriteBatchWithIndex*>(jwbwi_handle);
+  assert(wbwi != nullptr);
+
+  auto* wb = wbwi->GetWriteBatch();
+
+  // TODO(AR) is the `wb` object owned by us?
+  return rocksdb::WriteBatchJni::construct(env, wb);
+}
+
 /*
  * Class:     org_rocksdb_WriteBatchWithIndex
  * Method:    iterator0
@@ -551,33 +685,15 @@ jlongArray Java_org_rocksdb_WBWIRocksIterator_entry1(
 
   jlong results[3];
 
-  //set the type of the write entry
-  switch (we.type) {
-    case rocksdb::kPutRecord:
-      results[0] = 0x1;
-      break;
+  // set the type of the write entry
+  results[0] = rocksdb::WriteTypeJni::toJavaWriteType(we.type);
 
-    case rocksdb::kMergeRecord:
-      results[0] = 0x2;
-      break;
-
-    case rocksdb::kDeleteRecord:
-      results[0] = 0x4;
-      break;
-
-    case rocksdb::kLogDataRecord:
-      results[0] = 0x8;
-      break;
-
-    default:
-      results[0] = 0x0;
-  }
-
-  // key_slice and value_slice will be freed by org.rocksdb.DirectSlice#close
+  // NOTE: key_slice and value_slice will be freed by org.rocksdb.DirectSlice#close
 
   auto* key_slice = new rocksdb::Slice(we.key.data(), we.key.size());
   results[1] = reinterpret_cast<jlong>(key_slice);
   if (we.type == rocksdb::kDeleteRecord
+      || we.type == rocksdb::kSingleDeleteRecord
       || we.type == rocksdb::kLogDataRecord) {
     // set native handle of value slice to null if no value available
     results[2] = 0;

java/rocksjni/writebatchhandlerjnicallback.cc

@@ -14,24 +14,62 @@ WriteBatchHandlerJniCallback::WriteBatchHandlerJniCallback(
     JNIEnv* env, jobject jWriteBatchHandler)
     : JniCallback(env, jWriteBatchHandler), m_env(env) {
 
+  m_jPutCfMethodId = WriteBatchHandlerJni::getPutCfMethodId(env);
+  if(m_jPutCfMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
   m_jPutMethodId = WriteBatchHandlerJni::getPutMethodId(env);
   if(m_jPutMethodId == nullptr) {
     // exception thrown
     return;
   }
 
+  m_jMergeCfMethodId = WriteBatchHandlerJni::getMergeCfMethodId(env);
+  if(m_jMergeCfMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
   m_jMergeMethodId = WriteBatchHandlerJni::getMergeMethodId(env);
   if(m_jMergeMethodId == nullptr) {
     // exception thrown
     return;
   }
 
+  m_jDeleteCfMethodId = WriteBatchHandlerJni::getDeleteCfMethodId(env);
+  if(m_jDeleteCfMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
   m_jDeleteMethodId = WriteBatchHandlerJni::getDeleteMethodId(env);
   if(m_jDeleteMethodId == nullptr) {
     // exception thrown
     return;
   }
 
+  m_jSingleDeleteCfMethodId =
+      WriteBatchHandlerJni::getSingleDeleteCfMethodId(env);
+  if(m_jSingleDeleteCfMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
+  m_jSingleDeleteMethodId = WriteBatchHandlerJni::getSingleDeleteMethodId(env);
+  if(m_jSingleDeleteMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
+  m_jDeleteRangeCfMethodId =
+      WriteBatchHandlerJni::getDeleteRangeCfMethodId(env);
+  if (m_jDeleteRangeCfMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
   m_jDeleteRangeMethodId = WriteBatchHandlerJni::getDeleteRangeMethodId(env);
   if (m_jDeleteRangeMethodId == nullptr) {
     // exception thrown
@@ -44,6 +82,45 @@ WriteBatchHandlerJniCallback::WriteBatchHandlerJniCallback(
     return;
   }
 
+  m_jPutBlobIndexCfMethodId =
+      WriteBatchHandlerJni::getPutBlobIndexCfMethodId(env);
+  if(m_jPutBlobIndexCfMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
+  m_jMarkBeginPrepareMethodId =
+      WriteBatchHandlerJni::getMarkBeginPrepareMethodId(env);
+  if(m_jMarkBeginPrepareMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
+  m_jMarkEndPrepareMethodId =
+      WriteBatchHandlerJni::getMarkEndPrepareMethodId(env);
+  if(m_jMarkEndPrepareMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
+  m_jMarkNoopMethodId = WriteBatchHandlerJni::getMarkNoopMethodId(env);
+  if(m_jMarkNoopMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+    
+  m_jMarkRollbackMethodId = WriteBatchHandlerJni::getMarkRollbackMethodId(env);
+  if(m_jMarkRollbackMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
+  m_jMarkCommitMethodId = WriteBatchHandlerJni::getMarkCommitMethodId(env);
+  if(m_jMarkCommitMethodId == nullptr) {
+    // exception thrown
+    return;
+  }
+
   m_jContinueMethodId = WriteBatchHandlerJni::getContinueMethodId(env);
   if(m_jContinueMethodId == nullptr) {
     // exception thrown
@@ -51,196 +128,272 @@ WriteBatchHandlerJniCallback::WriteBatchHandlerJniCallback(
   }
 }
 
+rocksdb::Status WriteBatchHandlerJniCallback::PutCF(uint32_t column_family_id,
+    const Slice& key, const Slice& value) {
+  auto put = [this, column_family_id] (
+      jbyteArray j_key, jbyteArray j_value) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jPutCfMethodId,
+      static_cast<jint>(column_family_id),
+      j_key,
+      j_value);
+  };
+  auto status = WriteBatchHandlerJniCallback::kv_op(key, value, put);
+  if(status == nullptr) {
+    return rocksdb::Status::OK();   // TODO(AR) what to do if there is an Exception but we don't know the rocksdb::Status?
+  } else {
+    return rocksdb::Status(*status);
+  }
+}
+
 void WriteBatchHandlerJniCallback::Put(const Slice& key, const Slice& value) {
-  const jbyteArray j_key = sliceToJArray(key);
-  if(j_key == nullptr) {
-    // exception thrown
-    if(m_env->ExceptionCheck()) {
-      m_env->ExceptionDescribe();
-    }
-    return;
-  }
-
-  const jbyteArray j_value = sliceToJArray(value);
-  if(j_value == nullptr) {
-    // exception thrown
-    if(m_env->ExceptionCheck()) {
-      m_env->ExceptionDescribe();
-    }
-    if(j_key != nullptr) {
-      m_env->DeleteLocalRef(j_key);
-    }
-    return;
-  }
-
-  m_env->CallVoidMethod(
+  auto put = [this] (
+        jbyteArray j_key, jbyteArray j_value) {
+    m_env->CallVoidMethod(
       m_jcallback_obj,
       m_jPutMethodId,
       j_key,
       j_value);
-  if(m_env->ExceptionCheck()) {
-    // exception thrown
-    m_env->ExceptionDescribe();
-    if(j_value != nullptr) {
-      m_env->DeleteLocalRef(j_value);
-    }
-    if(j_key != nullptr) {
-      m_env->DeleteLocalRef(j_key);
-    }
-    return;
-  }
+  };
+  WriteBatchHandlerJniCallback::kv_op(key, value, put);
+}
 
-  if(j_value != nullptr) {
-    m_env->DeleteLocalRef(j_value);
-  }
-  if(j_key != nullptr) {
-    m_env->DeleteLocalRef(j_key);
+rocksdb::Status WriteBatchHandlerJniCallback::MergeCF(uint32_t column_family_id,
+    const Slice& key, const Slice& value) {
+  auto merge = [this, column_family_id] (
+        jbyteArray j_key, jbyteArray j_value) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jMergeCfMethodId,
+      static_cast<jint>(column_family_id),
+      j_key,
+      j_value);
+  };
+  auto status = WriteBatchHandlerJniCallback::kv_op(key, value, merge);
+  if(status == nullptr) {
+    return rocksdb::Status::OK();   // TODO(AR) what to do if there is an Exception but we don't know the rocksdb::Status?
+  } else {
+    return rocksdb::Status(*status);
   }
 }
 
 void WriteBatchHandlerJniCallback::Merge(const Slice& key, const Slice& value) {
-  const jbyteArray j_key = sliceToJArray(key);
-  if(j_key == nullptr) {
-    // exception thrown
-    if(m_env->ExceptionCheck()) {
-      m_env->ExceptionDescribe();
-    }
-    return;
-  }
-
-  const jbyteArray j_value = sliceToJArray(value);
-  if(j_value == nullptr) {
-    // exception thrown
-    if(m_env->ExceptionCheck()) {
-      m_env->ExceptionDescribe();
-    }
-    if(j_key != nullptr) {
-      m_env->DeleteLocalRef(j_key);
-    }
-    return;
-  }
-
-  m_env->CallVoidMethod(
+  auto merge = [this] (
+        jbyteArray j_key, jbyteArray j_value) {
+    m_env->CallVoidMethod(
       m_jcallback_obj,
       m_jMergeMethodId,
       j_key,
       j_value);
-  if(m_env->ExceptionCheck()) {
-    // exception thrown
-    m_env->ExceptionDescribe();
-    if(j_value != nullptr) {
-      m_env->DeleteLocalRef(j_value);
-    }
-    if(j_key != nullptr) {
-      m_env->DeleteLocalRef(j_key);
-    }
-    return;
-  }
+  };
+  WriteBatchHandlerJniCallback::kv_op(key, value, merge);
+}
 
-  if(j_value != nullptr) {
-    m_env->DeleteLocalRef(j_value);
-  }
-  if(j_key != nullptr) {
-    m_env->DeleteLocalRef(j_key);
+rocksdb::Status WriteBatchHandlerJniCallback::DeleteCF(uint32_t column_family_id,
+    const Slice& key) {
+  auto remove = [this, column_family_id] (jbyteArray j_key) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jDeleteCfMethodId,
+      static_cast<jint>(column_family_id),
+      j_key);
+  };
+  auto status = WriteBatchHandlerJniCallback::k_op(key, remove);
+  if(status == nullptr) {
+    return rocksdb::Status::OK();   // TODO(AR) what to do if there is an Exception but we don't know the rocksdb::Status?
+  } else {
+    return rocksdb::Status(*status);
   }
 }
 
 void WriteBatchHandlerJniCallback::Delete(const Slice& key) {
-  const jbyteArray j_key = sliceToJArray(key);
-  if(j_key == nullptr) {
-    // exception thrown
-    if(m_env->ExceptionCheck()) {
-      m_env->ExceptionDescribe();
-    }
-    return;
-  }
-
-  m_env->CallVoidMethod(
+  auto remove = [this] (jbyteArray j_key) {
+    m_env->CallVoidMethod(
       m_jcallback_obj,
       m_jDeleteMethodId,
       j_key);
-  if(m_env->ExceptionCheck()) {
-    // exception thrown
-    m_env->ExceptionDescribe();
-    if(j_key != nullptr) {
-      m_env->DeleteLocalRef(j_key);
-    }
-    return;
-  }
+  };
+  WriteBatchHandlerJniCallback::k_op(key, remove);
+}
 
-  if(j_key != nullptr) {
-    m_env->DeleteLocalRef(j_key);
+rocksdb::Status WriteBatchHandlerJniCallback::SingleDeleteCF(uint32_t column_family_id,
+    const Slice& key) {
+  auto singleDelete = [this, column_family_id] (jbyteArray j_key) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jSingleDeleteCfMethodId,
+      static_cast<jint>(column_family_id),
+      j_key);
+  };
+  auto status = WriteBatchHandlerJniCallback::k_op(key, singleDelete);
+  if(status == nullptr) {
+    return rocksdb::Status::OK();   // TODO(AR) what to do if there is an Exception but we don't know the rocksdb::Status?
+  } else {
+    return rocksdb::Status(*status);
+  }
+}
+
+void WriteBatchHandlerJniCallback::SingleDelete(const Slice& key) {
+  auto singleDelete = [this] (jbyteArray j_key) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jSingleDeleteMethodId,
+      j_key);
+  };
+  WriteBatchHandlerJniCallback::k_op(key, singleDelete);
+}
+
+rocksdb::Status WriteBatchHandlerJniCallback::DeleteRangeCF(uint32_t column_family_id,
+    const Slice& beginKey, const Slice& endKey) {
+  auto deleteRange = [this, column_family_id] (
+        jbyteArray j_beginKey, jbyteArray j_endKey) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jDeleteRangeCfMethodId,
+      static_cast<jint>(column_family_id),
+      j_beginKey,
+      j_endKey);
+  };
+  auto status = WriteBatchHandlerJniCallback::kv_op(beginKey, endKey, deleteRange);
+  if(status == nullptr) {
+    return rocksdb::Status::OK();   // TODO(AR) what to do if there is an Exception but we don't know the rocksdb::Status?
+  } else {
+    return rocksdb::Status(*status);
   }
 }
 
 void WriteBatchHandlerJniCallback::DeleteRange(const Slice& beginKey,
-                                               const Slice& endKey) {
-  const jbyteArray j_beginKey = sliceToJArray(beginKey);
-  if (j_beginKey == nullptr) {
-    // exception thrown
-    if (m_env->ExceptionCheck()) {
-      m_env->ExceptionDescribe();
-    }
-    return;
-  }
-
-  const jbyteArray j_endKey = sliceToJArray(beginKey);
-  if (j_endKey == nullptr) {
-    // exception thrown
-    if (m_env->ExceptionCheck()) {
-      m_env->ExceptionDescribe();
-    }
-    return;
-  }
-
-  m_env->CallVoidMethod(m_jcallback_obj, m_jDeleteRangeMethodId,
-                        j_beginKey, j_endKey);
-  if (m_env->ExceptionCheck()) {
-    // exception thrown
-    m_env->ExceptionDescribe();
-    if (j_beginKey != nullptr) {
-      m_env->DeleteLocalRef(j_beginKey);
-    }
-    if (j_endKey != nullptr) {
-      m_env->DeleteLocalRef(j_endKey);
-    }
-    return;
-  }
-
-  if (j_beginKey != nullptr) {
-    m_env->DeleteLocalRef(j_beginKey);
-  }
-
-  if (j_endKey != nullptr) {
-    m_env->DeleteLocalRef(j_endKey);
-  }
+    const Slice& endKey) {
+  auto deleteRange = [this] (
+        jbyteArray j_beginKey, jbyteArray j_endKey) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jDeleteRangeMethodId,
+      j_beginKey,
+      j_endKey);
+  };
+  WriteBatchHandlerJniCallback::kv_op(beginKey, endKey, deleteRange);
 }
 
 void WriteBatchHandlerJniCallback::LogData(const Slice& blob) {
-  const jbyteArray j_blob = sliceToJArray(blob);
-  if(j_blob == nullptr) {
-    // exception thrown
-    if(m_env->ExceptionCheck()) {
-      m_env->ExceptionDescribe();
-    }
-    return;
-  }
-
-  m_env->CallVoidMethod(
+  auto logData = [this] (jbyteArray j_blob) {
+    m_env->CallVoidMethod(
       m_jcallback_obj,
       m_jLogDataMethodId,
       j_blob);
-  if(m_env->ExceptionCheck()) {
+  };
+  WriteBatchHandlerJniCallback::k_op(blob, logData);
+}
+
+rocksdb::Status WriteBatchHandlerJniCallback::PutBlobIndexCF(uint32_t column_family_id,
+    const Slice& key, const Slice& value) {
+  auto putBlobIndex = [this, column_family_id] (
+      jbyteArray j_key, jbyteArray j_value) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jPutBlobIndexCfMethodId,
+      static_cast<jint>(column_family_id),
+      j_key,
+      j_value);
+  };
+  auto status = WriteBatchHandlerJniCallback::kv_op(key, value, putBlobIndex);
+  if(status == nullptr) {
+    return rocksdb::Status::OK();   // TODO(AR) what to do if there is an Exception but we don't know the rocksdb::Status?
+  } else {
+    return rocksdb::Status(*status);
+  }
+}
+
+rocksdb::Status WriteBatchHandlerJniCallback::MarkBeginPrepare() {
+  m_env->CallVoidMethod(m_jcallback_obj, m_jMarkBeginPrepareMethodId);
+
+  // check for Exception, in-particular RocksDBException
+  if (m_env->ExceptionCheck()) {
     // exception thrown
-    m_env->ExceptionDescribe();
-    if(j_blob != nullptr) {
-      m_env->DeleteLocalRef(j_blob);
+    jthrowable exception = m_env->ExceptionOccurred();
+    std::unique_ptr<rocksdb::Status> status = rocksdb::RocksDBExceptionJni::toCppStatus(m_env, exception);
+    if (status == nullptr) {
+      // unkown status or exception occurred extracting status
+      m_env->ExceptionDescribe();
+      return rocksdb::Status::OK();  // TODO(AR) probably need a better error code here
+
+    } else {
+      m_env->ExceptionClear();  // clear the exception, as we have extracted the status
+      return rocksdb::Status(*status);
     }
-    return;
   }
 
-  if(j_blob != nullptr) {
-    m_env->DeleteLocalRef(j_blob);
+  return rocksdb::Status::OK();
+}
+
+rocksdb::Status WriteBatchHandlerJniCallback::MarkEndPrepare(const Slice& xid) {
+  auto markEndPrepare = [this] (
+      jbyteArray j_xid) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jMarkEndPrepareMethodId,
+      j_xid);
+  };
+  auto status = WriteBatchHandlerJniCallback::k_op(xid, markEndPrepare);
+  if(status == nullptr) {
+    return rocksdb::Status::OK();   // TODO(AR) what to do if there is an Exception but we don't know the rocksdb::Status?
+  } else {
+    return rocksdb::Status(*status);
+  }
+}
+
+rocksdb::Status WriteBatchHandlerJniCallback::MarkNoop(bool empty_batch) {
+  m_env->CallVoidMethod(m_jcallback_obj, m_jMarkNoopMethodId, static_cast<jboolean>(empty_batch));
+
+  // check for Exception, in-particular RocksDBException
+  if (m_env->ExceptionCheck()) {
+    // exception thrown
+    jthrowable exception = m_env->ExceptionOccurred();
+    std::unique_ptr<rocksdb::Status> status = rocksdb::RocksDBExceptionJni::toCppStatus(m_env, exception);
+    if (status == nullptr) {
+      // unkown status or exception occurred extracting status
+      m_env->ExceptionDescribe();
+      return rocksdb::Status::OK();  // TODO(AR) probably need a better error code here
+
+    } else {
+      m_env->ExceptionClear();  // clear the exception, as we have extracted the status
+      return rocksdb::Status(*status);
+    }
+  }
+
+  return rocksdb::Status::OK();
+}
+
+rocksdb::Status WriteBatchHandlerJniCallback::MarkRollback(const Slice& xid) {
+  auto markRollback = [this] (
+      jbyteArray j_xid) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jMarkRollbackMethodId,
+      j_xid);
+  };
+  auto status = WriteBatchHandlerJniCallback::k_op(xid, markRollback);
+  if(status == nullptr) {
+    return rocksdb::Status::OK();   // TODO(AR) what to do if there is an Exception but we don't know the rocksdb::Status?
+  } else {
+    return rocksdb::Status(*status);
+  }
+}
+
+rocksdb::Status WriteBatchHandlerJniCallback::MarkCommit(const Slice& xid) {
+  auto markCommit = [this] (
+      jbyteArray j_xid) {
+    m_env->CallVoidMethod(
+      m_jcallback_obj,
+      m_jMarkCommitMethodId,
+      j_xid);
+  };
+  auto status = WriteBatchHandlerJniCallback::k_op(xid, markCommit);
+  if(status == nullptr) {
+    return rocksdb::Status::OK();   // TODO(AR) what to do if there is an Exception but we don't know the rocksdb::Status?
+  } else {
+    return rocksdb::Status(*status);
   }
 }
 
@@ -256,39 +409,101 @@ bool WriteBatchHandlerJniCallback::Continue() {
   return static_cast<bool>(jContinue == JNI_TRUE);
 }
 
-/*
- * Creates a Java Byte Array from the data in a Slice
- *
- * When calling this function
- * you must remember to call env->DeleteLocalRef
- * on the result after you have finished with it
- *
- * @param s A Slice to convery to a Java byte array
- *
- * @return A reference to a Java byte array, or a nullptr if an
- *     exception occurs
- */
-jbyteArray WriteBatchHandlerJniCallback::sliceToJArray(const Slice& s) {
-
-  // TODO(AR) move to JniUtil
-
-  jbyteArray ja = m_env->NewByteArray(static_cast<jsize>(s.size()));
-  if(ja == nullptr) {
-    // exception thrown: OutOfMemoryError
-    return nullptr;
-  }
-
-  m_env->SetByteArrayRegion(
-      ja, 0, static_cast<jsize>(s.size()),
-      const_cast<jbyte*>(reinterpret_cast<const jbyte*>(s.data())));
-  if(m_env->ExceptionCheck()) {
-    if(ja != nullptr) {
-      m_env->DeleteLocalRef(ja);
+std::unique_ptr<rocksdb::Status> WriteBatchHandlerJniCallback::kv_op(const Slice& key, const Slice& value, std::function<void(jbyteArray, jbyteArray)> kvFn) {
+    const jbyteArray j_key = JniUtil::copyBytes(m_env, key);
+  if (j_key == nullptr) {
+    // exception thrown
+    if (m_env->ExceptionCheck()) {
+      m_env->ExceptionDescribe();
     }
-    // exception thrown: ArrayIndexOutOfBoundsException
     return nullptr;
   }
 
-  return ja;
+  const jbyteArray j_value = JniUtil::copyBytes(m_env, value);
+  if (j_value == nullptr) {
+    // exception thrown
+    if (m_env->ExceptionCheck()) {
+      m_env->ExceptionDescribe();
+    }
+    if (j_key != nullptr) {
+      m_env->DeleteLocalRef(j_key);
+    }
+    return nullptr;
+  }
+
+  kvFn(j_key, j_value);
+
+  // check for Exception, in-particular RocksDBException
+  if (m_env->ExceptionCheck()) {
+    if (j_value != nullptr) {
+      m_env->DeleteLocalRef(j_value);
+    }
+    if (j_key != nullptr) {
+      m_env->DeleteLocalRef(j_key);
+    }
+
+    // exception thrown
+    jthrowable exception = m_env->ExceptionOccurred();
+    std::unique_ptr<rocksdb::Status> status = rocksdb::RocksDBExceptionJni::toCppStatus(m_env, exception);
+    if (status == nullptr) {
+      // unkown status or exception occurred extracting status
+      m_env->ExceptionDescribe();
+      return nullptr;
+
+    } else {
+      m_env->ExceptionClear();  // clear the exception, as we have extracted the status
+      return status;
+    }
+  }
+
+  if (j_value != nullptr) {
+    m_env->DeleteLocalRef(j_value);
+  }
+  if (j_key != nullptr) {
+    m_env->DeleteLocalRef(j_key);
+  }
+
+  // all OK
+  return std::unique_ptr<rocksdb::Status>(new rocksdb::Status(rocksdb::Status::OK()));
+}
+
+std::unique_ptr<rocksdb::Status> WriteBatchHandlerJniCallback::k_op(const Slice& key, std::function<void(jbyteArray)> kFn) {
+    const jbyteArray j_key = JniUtil::copyBytes(m_env, key);
+  if (j_key == nullptr) {
+    // exception thrown
+    if (m_env->ExceptionCheck()) {
+      m_env->ExceptionDescribe();
+    }
+    return nullptr;
+  }
+
+  kFn(j_key);
+
+  // check for Exception, in-particular RocksDBException
+  if (m_env->ExceptionCheck()) {
+    if (j_key != nullptr) {
+      m_env->DeleteLocalRef(j_key);
+    }
+
+    // exception thrown
+    jthrowable exception = m_env->ExceptionOccurred();
+    std::unique_ptr<rocksdb::Status> status = rocksdb::RocksDBExceptionJni::toCppStatus(m_env, exception);
+    if (status == nullptr) {
+      // unkown status or exception occurred extracting status
+      m_env->ExceptionDescribe();
+      return nullptr;
+
+    } else {
+      m_env->ExceptionClear();  // clear the exception, as we have extracted the status
+      return status;
+    }
+  }
+
+  if (j_key != nullptr) {
+    m_env->DeleteLocalRef(j_key);
+  }
+
+  // all OK
+  return std::unique_ptr<rocksdb::Status>(new rocksdb::Status(rocksdb::Status::OK()));
 }
 }  // namespace rocksdb

java/rocksjni/writebatchhandlerjnicallback.h

@@ -9,7 +9,9 @@
 #ifndef JAVA_ROCKSJNI_WRITEBATCHHANDLERJNICALLBACK_H_
 #define JAVA_ROCKSJNI_WRITEBATCHHANDLERJNICALLBACK_H_
 
+#include <functional>
 #include <jni.h>
+#include <memory>
 #include "rocksjni/jnicallback.h"
 #include "rocksdb/write_batch.h"
 
@@ -25,22 +27,57 @@ class WriteBatchHandlerJniCallback : public JniCallback, public WriteBatch::Hand
  public:
     WriteBatchHandlerJniCallback(
       JNIEnv* env, jobject jWriteBackHandler);
+    Status PutCF(uint32_t column_family_id, const Slice& key,
+        const Slice& value);
     void Put(const Slice& key, const Slice& value);
+    Status MergeCF(uint32_t column_family_id, const Slice& key,
+        const Slice& value);
     void Merge(const Slice& key, const Slice& value);
+    Status DeleteCF(uint32_t column_family_id, const Slice& key);
     void Delete(const Slice& key);
+    Status SingleDeleteCF(uint32_t column_family_id, const Slice& key);
+    void SingleDelete(const Slice& key);
+    Status DeleteRangeCF(uint32_t column_family_id, const Slice& beginKey,
+        const Slice& endKey);
     void DeleteRange(const Slice& beginKey, const Slice& endKey);
     void LogData(const Slice& blob);
+    Status PutBlobIndexCF(uint32_t column_family_id, const Slice& key,
+        const Slice& value);
+    Status MarkBeginPrepare();
+    Status MarkEndPrepare(const Slice& xid);
+    Status MarkNoop(bool empty_batch);
+    Status MarkRollback(const Slice& xid);
+    Status MarkCommit(const Slice& xid);
     bool Continue();
 
  private:
     JNIEnv* m_env;
-    jbyteArray sliceToJArray(const Slice& s);
+    jmethodID m_jPutCfMethodId;
     jmethodID m_jPutMethodId;
+    jmethodID m_jMergeCfMethodId;
     jmethodID m_jMergeMethodId;
+    jmethodID m_jDeleteCfMethodId;
     jmethodID m_jDeleteMethodId;
+    jmethodID m_jSingleDeleteCfMethodId;
+    jmethodID m_jSingleDeleteMethodId;
+    jmethodID m_jDeleteRangeCfMethodId;
     jmethodID m_jDeleteRangeMethodId;
     jmethodID m_jLogDataMethodId;
+    jmethodID m_jPutBlobIndexCfMethodId;
+    jmethodID m_jMarkBeginPrepareMethodId;
+    jmethodID m_jMarkEndPrepareMethodId;
+    jmethodID m_jMarkNoopMethodId;
+    jmethodID m_jMarkRollbackMethodId;
+    jmethodID m_jMarkCommitMethodId;
     jmethodID m_jContinueMethodId;
+    /**
+     * @return A pointer to a rocksdb::Status or nullptr if an unexpected exception occurred
+     */
+    std::unique_ptr<rocksdb::Status> kv_op(const Slice& key, const Slice& value, std::function<void(jbyteArray, jbyteArray)> kvFn);
+    /**
+     * @return A pointer to a rocksdb::Status or nullptr if an unexpected exception occurred
+     */
+    std::unique_ptr<rocksdb::Status> k_op(const Slice& key, std::function<void(jbyteArray)> kFn);
 };
 }  // namespace rocksdb
 

java/samples/src/main/java/OptimisticTransactionSample.java

@@ -0,0 +1,184 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+import org.rocksdb.*;
+
+import static java.nio.charset.StandardCharsets.UTF_8;
+
+/**
+ * Demonstrates using Transactions on an OptimisticTransactionDB with
+ * varying isolation guarantees
+ */
+public class OptimisticTransactionSample {
+  private static final String dbPath = "/tmp/rocksdb_optimistic_transaction_example";
+
+  public static final void main(final String args[]) throws RocksDBException {
+
+    try(final Options options = new Options()
+        .setCreateIfMissing(true);
+        final OptimisticTransactionDB txnDb =
+            OptimisticTransactionDB.open(options, dbPath)) {
+
+      try (final WriteOptions writeOptions = new WriteOptions();
+           final ReadOptions readOptions = new ReadOptions()) {
+
+        ////////////////////////////////////////////////////////
+        //
+        // Simple OptimisticTransaction Example ("Read Committed")
+        //
+        ////////////////////////////////////////////////////////
+        readCommitted(txnDb, writeOptions, readOptions);
+
+
+        ////////////////////////////////////////////////////////
+        //
+        // "Repeatable Read" (Snapshot Isolation) Example
+        //   -- Using a single Snapshot
+        //
+        ////////////////////////////////////////////////////////
+        repeatableRead(txnDb, writeOptions, readOptions);
+
+
+        ////////////////////////////////////////////////////////
+        //
+        // "Read Committed" (Monotonic Atomic Views) Example
+        //   --Using multiple Snapshots
+        //
+        ////////////////////////////////////////////////////////
+        readCommitted_monotonicAtomicViews(txnDb, writeOptions, readOptions);
+      }
+    }
+  }
+
+  /**
+   * Demonstrates "Read Committed" isolation
+   */
+  private static void readCommitted(final OptimisticTransactionDB txnDb,
+      final WriteOptions writeOptions, final ReadOptions readOptions)
+      throws RocksDBException {
+    final byte key1[] = "abc".getBytes(UTF_8);
+    final byte value1[] = "def".getBytes(UTF_8);
+
+    final byte key2[] = "xyz".getBytes(UTF_8);
+    final byte value2[] = "zzz".getBytes(UTF_8);
+
+    // Start a transaction
+    try(final Transaction txn = txnDb.beginTransaction(writeOptions)) {
+      // Read a key in this transaction
+      byte[] value = txn.get(readOptions, key1);
+      assert(value == null);
+
+      // Write a key in this transaction
+      txn.put(key1, value1);
+
+      // Read a key OUTSIDE this transaction. Does not affect txn.
+      value = txnDb.get(readOptions, key1);
+      assert(value == null);
+
+      // Write a key OUTSIDE of this transaction.
+      // Does not affect txn since this is an unrelated key.
+      // If we wrote key 'abc' here, the transaction would fail to commit.
+      txnDb.put(writeOptions, key2, value2);
+
+      // Commit transaction
+      txn.commit();
+    }
+  }
+
+  /**
+   * Demonstrates "Repeatable Read" (Snapshot Isolation) isolation
+   */
+  private static void repeatableRead(final OptimisticTransactionDB txnDb,
+      final WriteOptions writeOptions, final ReadOptions readOptions)
+      throws RocksDBException {
+
+    final byte key1[] = "ghi".getBytes(UTF_8);
+    final byte value1[] = "jkl".getBytes(UTF_8);
+
+    // Set a snapshot at start of transaction by setting setSnapshot(true)
+    try(final OptimisticTransactionOptions txnOptions =
+            new OptimisticTransactionOptions().setSetSnapshot(true);
+        final Transaction txn =
+            txnDb.beginTransaction(writeOptions, txnOptions)) {
+
+      final Snapshot snapshot = txn.getSnapshot();
+
+      // Write a key OUTSIDE of transaction
+      txnDb.put(writeOptions, key1, value1);
+
+      // Read a key using the snapshot.
+      readOptions.setSnapshot(snapshot);
+      final byte[] value = txn.getForUpdate(readOptions, key1, true);
+      assert(value == value1);
+
+      try {
+        // Attempt to commit transaction
+        txn.commit();
+        throw new IllegalStateException();
+      } catch(final RocksDBException e) {
+        // Transaction could not commit since the write outside of the txn
+        // conflicted with the read!
+        assert(e.getStatus().getCode() == Status.Code.Busy);
+      }
+
+      txn.rollback();
+    } finally {
+      // Clear snapshot from read options since it is no longer valid
+      readOptions.setSnapshot(null);
+    }
+  }
+
+  /**
+   * Demonstrates "Read Committed" (Monotonic Atomic Views) isolation
+   *
+   * In this example, we set the snapshot multiple times.  This is probably
+   * only necessary if you have very strict isolation requirements to
+   * implement.
+   */
+  private static void readCommitted_monotonicAtomicViews(
+      final OptimisticTransactionDB txnDb, final WriteOptions writeOptions,
+      final ReadOptions readOptions) throws RocksDBException {
+
+    final byte keyX[] = "x".getBytes(UTF_8);
+    final byte valueX[] = "x".getBytes(UTF_8);
+
+    final byte keyY[] = "y".getBytes(UTF_8);
+    final byte valueY[] = "y".getBytes(UTF_8);
+
+    try (final OptimisticTransactionOptions txnOptions =
+             new OptimisticTransactionOptions().setSetSnapshot(true);
+         final Transaction txn =
+             txnDb.beginTransaction(writeOptions, txnOptions)) {
+
+      // Do some reads and writes to key "x"
+      Snapshot snapshot = txnDb.getSnapshot();
+      readOptions.setSnapshot(snapshot);
+      byte[] value = txn.get(readOptions, keyX);
+      txn.put(valueX, valueX);
+
+      // Do a write outside of the transaction to key "y"
+      txnDb.put(writeOptions, keyY, valueY);
+
+      // Set a new snapshot in the transaction
+      txn.setSnapshot();
+      snapshot = txnDb.getSnapshot();
+      readOptions.setSnapshot(snapshot);
+
+      // Do some reads and writes to key "y"
+      // Since the snapshot was advanced, the write done outside of the
+      // transaction does not conflict.
+      value = txn.getForUpdate(readOptions, keyY, true);
+      txn.put(keyY, valueY);
+
+      // Commit.  Since the snapshot was advanced, the write done outside of the
+      // transaction does not prevent this transaction from Committing.
+      txn.commit();
+
+    } finally {
+      // Clear snapshot from read options since it is no longer valid
+      readOptions.setSnapshot(null);
+    }
+  }
+}

java/samples/src/main/java/TransactionSample.java

@@ -0,0 +1,183 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+import org.rocksdb.*;
+
+import static java.nio.charset.StandardCharsets.UTF_8;
+
+/**
+ * Demonstrates using Transactions on a TransactionDB with
+ * varying isolation guarantees
+ */
+public class TransactionSample {
+  private static final String dbPath = "/tmp/rocksdb_transaction_example";
+
+  public static final void main(final String args[]) throws RocksDBException {
+
+    try(final Options options = new Options()
+        .setCreateIfMissing(true);
+        final TransactionDBOptions txnDbOptions = new TransactionDBOptions();
+        final TransactionDB txnDb =
+            TransactionDB.open(options, txnDbOptions, dbPath)) {
+
+      try (final WriteOptions writeOptions = new WriteOptions();
+           final ReadOptions readOptions = new ReadOptions()) {
+
+        ////////////////////////////////////////////////////////
+        //
+        // Simple Transaction Example ("Read Committed")
+        //
+        ////////////////////////////////////////////////////////
+        readCommitted(txnDb, writeOptions, readOptions);
+
+
+        ////////////////////////////////////////////////////////
+        //
+        // "Repeatable Read" (Snapshot Isolation) Example
+        //   -- Using a single Snapshot
+        //
+        ////////////////////////////////////////////////////////
+        repeatableRead(txnDb, writeOptions, readOptions);
+
+
+        ////////////////////////////////////////////////////////
+        //
+        // "Read Committed" (Monotonic Atomic Views) Example
+        //   --Using multiple Snapshots
+        //
+        ////////////////////////////////////////////////////////
+        readCommitted_monotonicAtomicViews(txnDb, writeOptions, readOptions);
+      }
+    }
+  }
+
+  /**
+   * Demonstrates "Read Committed" isolation
+   */
+  private static void readCommitted(final TransactionDB txnDb,
+      final WriteOptions writeOptions, final ReadOptions readOptions)
+      throws RocksDBException {
+    final byte key1[] = "abc".getBytes(UTF_8);
+    final byte value1[] = "def".getBytes(UTF_8);
+
+    final byte key2[] = "xyz".getBytes(UTF_8);
+    final byte value2[] = "zzz".getBytes(UTF_8);
+
+    // Start a transaction
+    try(final Transaction txn = txnDb.beginTransaction(writeOptions)) {
+      // Read a key in this transaction
+      byte[] value = txn.get(readOptions, key1);
+      assert(value == null);
+
+      // Write a key in this transaction
+      txn.put(key1, value1);
+
+      // Read a key OUTSIDE this transaction. Does not affect txn.
+      value = txnDb.get(readOptions, key1);
+      assert(value == null);
+
+      // Write a key OUTSIDE of this transaction.
+      // Does not affect txn since this is an unrelated key.
+      // If we wrote key 'abc' here, the transaction would fail to commit.
+      txnDb.put(writeOptions, key2, value2);
+
+      // Commit transaction
+      txn.commit();
+    }
+  }
+
+  /**
+   * Demonstrates "Repeatable Read" (Snapshot Isolation) isolation
+   */
+  private static void repeatableRead(final TransactionDB txnDb,
+      final WriteOptions writeOptions, final ReadOptions readOptions)
+      throws RocksDBException {
+
+    final byte key1[] = "ghi".getBytes(UTF_8);
+    final byte value1[] = "jkl".getBytes(UTF_8);
+
+    // Set a snapshot at start of transaction by setting setSnapshot(true)
+    try(final TransactionOptions txnOptions = new TransactionOptions()
+          .setSetSnapshot(true);
+        final Transaction txn =
+            txnDb.beginTransaction(writeOptions, txnOptions)) {
+
+      final Snapshot snapshot = txn.getSnapshot();
+
+      // Write a key OUTSIDE of transaction
+      txnDb.put(writeOptions, key1, value1);
+
+      // Attempt to read a key using the snapshot.  This will fail since
+      // the previous write outside this txn conflicts with this read.
+      readOptions.setSnapshot(snapshot);
+
+      try {
+        final byte[] value = txn.getForUpdate(readOptions, key1, true);
+        throw new IllegalStateException();
+      } catch(final RocksDBException e) {
+        assert(e.getStatus().getCode() == Status.Code.Busy);
+      }
+
+      txn.rollback();
+    } finally {
+      // Clear snapshot from read options since it is no longer valid
+      readOptions.setSnapshot(null);
+    }
+  }
+
+  /**
+   * Demonstrates "Read Committed" (Monotonic Atomic Views) isolation
+   *
+   * In this example, we set the snapshot multiple times.  This is probably
+   * only necessary if you have very strict isolation requirements to
+   * implement.
+   */
+  private static void readCommitted_monotonicAtomicViews(
+      final TransactionDB txnDb, final WriteOptions writeOptions,
+      final ReadOptions readOptions) throws RocksDBException {
+
+    final byte keyX[] = "x".getBytes(UTF_8);
+    final byte valueX[] = "x".getBytes(UTF_8);
+
+    final byte keyY[] = "y".getBytes(UTF_8);
+    final byte valueY[] = "y".getBytes(UTF_8);
+
+    try (final TransactionOptions txnOptions = new TransactionOptions()
+        .setSetSnapshot(true);
+         final Transaction txn =
+             txnDb.beginTransaction(writeOptions, txnOptions)) {
+
+      // Do some reads and writes to key "x"
+      Snapshot snapshot = txnDb.getSnapshot();
+      readOptions.setSnapshot(snapshot);
+      byte[] value = txn.get(readOptions, keyX);
+      txn.put(valueX, valueX);
+
+      // Do a write outside of the transaction to key "y"
+      txnDb.put(writeOptions, keyY, valueY);
+
+      // Set a new snapshot in the transaction
+      txn.setSnapshot();
+      txn.setSavePoint();
+      snapshot = txnDb.getSnapshot();
+      readOptions.setSnapshot(snapshot);
+
+      // Do some reads and writes to key "y"
+      // Since the snapshot was advanced, the write done outside of the
+      // transaction does not conflict.
+      value = txn.getForUpdate(readOptions, keyY, true);
+      txn.put(keyY, valueY);
+
+      // Decide we want to revert the last write from this transaction.
+      txn.rollbackToSavePoint();
+
+      // Commit.
+      txn.commit();
+    } finally {
+      // Clear snapshot from read options since it is no longer valid
+      readOptions.setSnapshot(null);
+    }
+  }
+}

java/src/main/java/org/rocksdb/AbstractComparator.java

@@ -17,10 +17,23 @@ package org.rocksdb;
 public abstract class AbstractComparator<T extends AbstractSlice<?>>
     extends RocksCallbackObject {
 
+  protected AbstractComparator() {
+    super();
+  }
+
   protected AbstractComparator(final ComparatorOptions copt) {
     super(copt.nativeHandle_);
   }
 
+  /**
+   * Get the type of this comparator.
+   *
+   * Used for determining the correct C++ cast in native code.
+   *
+   * @return The type of the comparator.
+   */
+  abstract ComparatorType getComparatorType();
+
   /**
    * The name of the comparator.  Used to check for comparator
    * mismatches (i.e., a DB created with one comparator is

java/src/main/java/org/rocksdb/AbstractTransactionNotifier.java

@@ -0,0 +1,54 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+/**
+ * Provides notification to the caller of SetSnapshotOnNextOperation when
+ * the actual snapshot gets created
+ */
+public abstract class AbstractTransactionNotifier
+    extends RocksCallbackObject {
+
+  protected AbstractTransactionNotifier() {
+    super();
+  }
+
+  /**
+   * Implement this method to receive notification when a snapshot is
+   * requested via {@link Transaction#setSnapshotOnNextOperation()}.
+   *
+   * @param newSnapshot the snapshot that has been created.
+   */
+  public abstract void snapshotCreated(final Snapshot newSnapshot);
+
+  /**
+   * This is intentionally private as it is the callback hook
+   * from JNI
+   */
+  private void snapshotCreated(final long snapshotHandle) {
+    snapshotCreated(new Snapshot(snapshotHandle));
+  }
+
+  @Override
+  protected long initializeNative(final long... nativeParameterHandles) {
+    return createNewTransactionNotifier();
+  }
+
+  private native long createNewTransactionNotifier();
+
+  /**
+   * Deletes underlying C++ TransactionNotifier pointer.
+   *
+   * Note that this function should be called only after all
+   * Transactions referencing the comparator are closed.
+   * Otherwise an undefined behavior will occur.
+   */
+  @Override
+  protected void disposeInternal() {
+    disposeInternal(nativeHandle_);
+  }
+  protected final native void disposeInternal(final long handle);
+}

java/src/main/java/org/rocksdb/AbstractWriteBatch.java

@@ -18,52 +18,80 @@ public abstract class AbstractWriteBatch extends RocksObject
   }
 
   @Override
-  public void put(byte[] key, byte[] value) {
+  public void put(byte[] key, byte[] value) throws RocksDBException {
     put(nativeHandle_, key, key.length, value, value.length);
   }
 
   @Override
   public void put(ColumnFamilyHandle columnFamilyHandle, byte[] key,
-      byte[] value) {
+      byte[] value) throws RocksDBException {
     put(nativeHandle_, key, key.length, value, value.length,
         columnFamilyHandle.nativeHandle_);
   }
 
   @Override
-  public void merge(byte[] key, byte[] value) {
+  public void merge(byte[] key, byte[] value) throws RocksDBException {
     merge(nativeHandle_, key, key.length, value, value.length);
   }
 
   @Override
   public void merge(ColumnFamilyHandle columnFamilyHandle, byte[] key,
-      byte[] value) {
+      byte[] value) throws RocksDBException {
     merge(nativeHandle_, key, key.length, value, value.length,
         columnFamilyHandle.nativeHandle_);
   }
 
   @Override
-  public void remove(byte[] key) {
-    remove(nativeHandle_, key, key.length);
+  @Deprecated
+  public void remove(byte[] key) throws RocksDBException {
+    delete(nativeHandle_, key, key.length);
   }
 
   @Override
-  public void remove(ColumnFamilyHandle columnFamilyHandle, byte[] key) {
-    remove(nativeHandle_, key, key.length, columnFamilyHandle.nativeHandle_);
+  @Deprecated
+  public void remove(ColumnFamilyHandle columnFamilyHandle, byte[] key)
+      throws RocksDBException {
+    delete(nativeHandle_, key, key.length, columnFamilyHandle.nativeHandle_);
   }
 
   @Override
-  public void deleteRange(byte[] beginKey, byte[] endKey) {
+  public void delete(byte[] key) throws RocksDBException {
+    delete(nativeHandle_, key, key.length);
+  }
+
+  @Override
+  public void delete(ColumnFamilyHandle columnFamilyHandle, byte[] key)
+      throws RocksDBException {
+    delete(nativeHandle_, key, key.length, columnFamilyHandle.nativeHandle_);
+  }
+
+
+  @Override
+  public void singleDelete(byte[] key) throws RocksDBException {
+    singleDelete(nativeHandle_, key, key.length);
+  }
+
+  @Override
+  public void singleDelete(ColumnFamilyHandle columnFamilyHandle, byte[] key)
+      throws RocksDBException {
+    singleDelete(nativeHandle_, key, key.length, columnFamilyHandle.nativeHandle_);
+  }
+
+  @Override
+  public void deleteRange(byte[] beginKey, byte[] endKey)
+      throws RocksDBException {
     deleteRange(nativeHandle_, beginKey, beginKey.length, endKey, endKey.length);
   }
 
   @Override
-  public void deleteRange(ColumnFamilyHandle columnFamilyHandle, byte[] beginKey, byte[] endKey) {
+  public void deleteRange(ColumnFamilyHandle columnFamilyHandle,
+      byte[] beginKey, byte[] endKey) throws RocksDBException {
     deleteRange(nativeHandle_, beginKey, beginKey.length, endKey, endKey.length,
         columnFamilyHandle.nativeHandle_);
   }
 
   @Override
-  public void putLogData(byte[] blob) {
+  public void putLogData(byte[] blob) throws RocksDBException {
     putLogData(nativeHandle_, blob, blob.length);
   }
 
@@ -82,38 +110,67 @@ public abstract class AbstractWriteBatch extends RocksObject
     rollbackToSavePoint0(nativeHandle_);
   }
 
+  @Override
+  public void popSavePoint() throws RocksDBException {
+    popSavePoint(nativeHandle_);
+  }
+
+  @Override
+  public void setMaxBytes(final long maxBytes) {
+    setMaxBytes(nativeHandle_, maxBytes);
+  }
+
+  @Override
+  public WriteBatch getWriteBatch() {
+    return getWriteBatch(nativeHandle_);
+  }
+
   abstract int count0(final long handle);
 
   abstract void put(final long handle, final byte[] key, final int keyLen,
-      final byte[] value, final int valueLen);
+      final byte[] value, final int valueLen) throws RocksDBException;
 
   abstract void put(final long handle, final byte[] key, final int keyLen,
-      final byte[] value, final int valueLen, final long cfHandle);
+      final byte[] value, final int valueLen, final long cfHandle)
+      throws RocksDBException;
 
   abstract void merge(final long handle, final byte[] key, final int keyLen,
-      final byte[] value, final int valueLen);
+      final byte[] value, final int valueLen) throws RocksDBException;
 
   abstract void merge(final long handle, final byte[] key, final int keyLen,
-      final byte[] value, final int valueLen, final long cfHandle);
+      final byte[] value, final int valueLen, final long cfHandle)
+      throws RocksDBException;
 
-  abstract void remove(final long handle, final byte[] key,
-      final int keyLen);
+  abstract void delete(final long handle, final byte[] key,
+      final int keyLen) throws RocksDBException;
 
-  abstract void remove(final long handle, final byte[] key,
-      final int keyLen, final long cfHandle);
+  abstract void delete(final long handle, final byte[] key,
+      final int keyLen, final long cfHandle) throws RocksDBException;
+
+  abstract void singleDelete(final long handle, final byte[] key,
+                       final int keyLen) throws RocksDBException;
+
+  abstract void singleDelete(final long handle, final byte[] key,
+                       final int keyLen, final long cfHandle) throws RocksDBException;
 
   abstract void deleteRange(final long handle, final byte[] beginKey, final int beginKeyLen,
-      final byte[] endKey, final int endKeyLen);
+      final byte[] endKey, final int endKeyLen) throws RocksDBException;
 
   abstract void deleteRange(final long handle, final byte[] beginKey, final int beginKeyLen,
-      final byte[] endKey, final int endKeyLen, final long cfHandle);
+      final byte[] endKey, final int endKeyLen, final long cfHandle) throws RocksDBException;
 
   abstract void putLogData(final long handle, final byte[] blob,
-      final int blobLen);
+      final int blobLen) throws RocksDBException;
 
   abstract void clear0(final long handle);
 
   abstract void setSavePoint0(final long handle);
 
   abstract void rollbackToSavePoint0(final long handle);
+
+  abstract void popSavePoint(final long handle) throws RocksDBException;
+
+  abstract void setMaxBytes(final long handle, long maxBytes);
+
+  abstract WriteBatch getWriteBatch(final long handle);
 }

java/src/main/java/org/rocksdb/ColumnFamilyDescriptor.java

@@ -5,6 +5,8 @@
 
 package org.rocksdb;
 
+import java.util.Arrays;
+
 /**
  * <p>Describes a column family with a
  * name and respective Options.</p>
@@ -32,7 +34,7 @@ public class ColumnFamilyDescriptor {
    * @since 3.10.0
    */
   public ColumnFamilyDescriptor(final byte[] columnFamilyName,
-      final ColumnFamilyOptions columnFamilyOptions) {
+                                final ColumnFamilyOptions columnFamilyOptions) {
     columnFamilyName_ = columnFamilyName;
     columnFamilyOptions_ = columnFamilyOptions;
   }
@@ -43,19 +45,65 @@ public class ColumnFamilyDescriptor {
    * @return column family name.
    * @since 3.10.0
    */
-  public byte[] columnFamilyName() {
+  public byte[] getName() {
     return columnFamilyName_;
   }
 
+  /**
+   * Retrieve name of column family.
+   *
+   * @return column family name.
+   * @since 3.10.0
+   *
+   * @deprecated Use {@link #getName()} instead.
+   */
+  @Deprecated
+  public byte[] columnFamilyName() {
+    return getName();
+  }
+
   /**
    * Retrieve assigned options instance.
    *
    * @return Options instance assigned to this instance.
    */
-  public ColumnFamilyOptions columnFamilyOptions() {
+  public ColumnFamilyOptions getOptions() {
     return columnFamilyOptions_;
   }
 
+  /**
+   * Retrieve assigned options instance.
+   *
+   * @return Options instance assigned to this instance.
+   *
+   * @deprecated Use {@link #getOptions()} instead.
+   */
+  @Deprecated
+  public ColumnFamilyOptions columnFamilyOptions() {
+    return getOptions();
+  }
+
+  @Override
+  public boolean equals(final Object o) {
+    if (this == o) {
+      return true;
+    }
+    if (o == null || getClass() != o.getClass()) {
+      return false;
+    }
+
+    final ColumnFamilyDescriptor that = (ColumnFamilyDescriptor) o;
+    return Arrays.equals(columnFamilyName_, that.columnFamilyName_)
+            && columnFamilyOptions_.nativeHandle_ == that.columnFamilyOptions_.nativeHandle_;
+  }
+
+  @Override
+  public int hashCode() {
+    int result = (int) (columnFamilyOptions_.nativeHandle_ ^ (columnFamilyOptions_.nativeHandle_ >>> 32));
+    result = 31 * result + Arrays.hashCode(columnFamilyName_);
+    return result;
+  }
+
   private final byte[] columnFamilyName_;
   private final ColumnFamilyOptions columnFamilyOptions_;
 }

java/src/main/java/org/rocksdb/ColumnFamilyHandle.java

@@ -5,6 +5,9 @@
 
 package org.rocksdb;
 
+import java.util.Arrays;
+import java.util.Objects;
+
 /**
  * ColumnFamilyHandle class to hold handles to underlying rocksdb
  * ColumnFamily Pointers.
@@ -21,6 +24,63 @@ public class ColumnFamilyHandle extends RocksObject {
     this.rocksDB_ = rocksDB;
   }
 
+  /**
+   * Gets the name of the Column Family.
+   *
+   * @return The name of the Column Family.
+   */
+  public byte[] getName() {
+    return getName(nativeHandle_);
+  }
+
+  /**
+   * Gets the ID of the Column Family.
+   *
+   * @return the ID of the Column Family.
+   */
+  public int getID() {
+    return getID(nativeHandle_);
+  }
+
+  /**
+   * Gets the up-to-date descriptor of the column family
+   * associated with this handle. Since it fills "*desc" with the up-to-date
+   * information, this call might internally lock and release DB mutex to
+   * access the up-to-date CF options. In addition, all the pointer-typed
+   * options cannot be referenced any longer than the original options exist.
+   *
+   * Note that this function is not supported in RocksDBLite.
+   *
+   * @return the up-to-date descriptor.
+   *
+   * @throws RocksDBException if an error occurs whilst retrieving the
+   *     descriptor.
+   */
+  public ColumnFamilyDescriptor getDescriptor() throws RocksDBException {
+    assert(isOwningHandle());
+    return getDescriptor(nativeHandle_);
+  }
+
+  @Override
+  public boolean equals(final Object o) {
+    if (this == o) {
+      return true;
+    }
+    if (o == null || getClass() != o.getClass()) {
+      return false;
+    }
+
+    final ColumnFamilyHandle that = (ColumnFamilyHandle) o;
+    return rocksDB_.nativeHandle_ == that.rocksDB_.nativeHandle_ &&
+        getID() == that.getID() &&
+        Arrays.equals(getName(), that.getName());
+  }
+
+  @Override
+  public int hashCode() {
+    return Objects.hash(getName(), getID(), rocksDB_.nativeHandle_);
+  }
+
   /**
    * <p>Deletes underlying C++ iterator pointer.</p>
    *
@@ -36,6 +96,9 @@ public class ColumnFamilyHandle extends RocksObject {
     }
   }
 
+  private native byte[] getName(final long handle);
+  private native int getID(final long handle);
+  private native ColumnFamilyDescriptor getDescriptor(final long handle) throws RocksDBException;
   @Override protected final native void disposeInternal(final long handle);
 
   private final RocksDB rocksDB_;

java/src/main/java/org/rocksdb/ColumnFamilyOptions.java

@@ -33,6 +33,18 @@ public class ColumnFamilyOptions extends RocksObject
     super(newColumnFamilyOptions());
   }
 
+  /**
+   * <p>Constructor to be used by
+   * {@link #getColumnFamilyOptionsFromProps(java.util.Properties)},
+   * {@link ColumnFamilyDescriptor#columnFamilyOptions()}
+   * and also called via JNI.</p>
+   *
+   * @param handle native handle to ColumnFamilyOptions instance.
+   */
+  ColumnFamilyOptions(final long handle) {
+    super(handle);
+  }
+
   /**
    * <p>Method to get a options instance by using pre-configured
    * property values. If one or many values are undefined in
@@ -131,7 +143,7 @@ public class ColumnFamilyOptions extends RocksObject
       final AbstractComparator<? extends AbstractSlice<?>> comparator) {
     assert (isOwningHandle());
     setComparatorHandle(nativeHandle_, comparator.nativeHandle_,
-            comparator instanceof DirectComparator);
+            comparator.getComparatorType().getValue());
     comparator_ = comparator;
     return this;
   }
@@ -768,17 +780,6 @@ public class ColumnFamilyOptions extends RocksObject
     return forceConsistencyChecks(nativeHandle_);
   }
 
-  /**
-   * <p>Constructor to be used by
-   * {@link #getColumnFamilyOptionsFromProps(java.util.Properties)}</p>
-   * and also called via JNI.
-   *
-   * @param handle native handle to ColumnFamilyOptions instance.
-   */
-  public ColumnFamilyOptions(final long handle) {
-    super(handle);
-  }
-
   private static native long getColumnFamilyOptionsFromProps(
       String optString);
 
@@ -794,7 +795,7 @@ public class ColumnFamilyOptions extends RocksObject
       long memtableMemoryBudget);
   private native void setComparatorHandle(long handle, int builtinComparator);
   private native void setComparatorHandle(long optHandle,
-      long comparatorHandle, boolean isDirect);
+      long comparatorHandle, byte comparatorType);
   private native void setMergeOperatorName(long handle, String name);
   private native void setMergeOperator(long handle, long mergeOperatorHandle);
   private native void setCompactionFilterHandle(long handle,

java/src/main/java/org/rocksdb/Comparator.java

@@ -25,5 +25,10 @@ public abstract class Comparator extends AbstractComparator<Slice> {
     return createNewComparator0(nativeParameterHandles[0]);
   }
 
+  @Override
+  final ComparatorType getComparatorType() {
+    return ComparatorType.JAVA_COMPARATOR;
+  }
+
   private native long createNewComparator0(final long comparatorOptionsHandle);
 }

java/src/main/java/org/rocksdb/ComparatorType.java

@@ -0,0 +1,49 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+enum ComparatorType {
+  JAVA_COMPARATOR((byte)0x0),
+  JAVA_DIRECT_COMPARATOR((byte)0x1),
+  JAVA_NATIVE_COMPARATOR_WRAPPER((byte)0x2);
+
+  private final byte value;
+
+  ComparatorType(final byte value) {
+    this.value = value;
+  }
+
+  /**
+   * <p>Returns the byte value of the enumerations value.</p>
+   *
+   * @return byte representation
+   */
+  byte getValue() {
+    return value;
+  }
+
+  /**
+   * <p>Get the ComparatorType enumeration value by
+   * passing the byte identifier to this method.</p>
+   *
+   * @param byteIdentifier of ComparatorType.
+   *
+   * @return ComparatorType instance.
+   *
+   * @throws IllegalArgumentException if the comparator type for the byteIdentifier
+   *     cannot be found
+   */
+  static ComparatorType getComparatorType(final byte byteIdentifier) {
+    for (final ComparatorType comparatorType : ComparatorType.values()) {
+      if (comparatorType.getValue() == byteIdentifier) {
+        return comparatorType;
+      }
+    }
+
+    throw new IllegalArgumentException(
+        "Illegal value provided for ComparatorType.");
+  }
+}

java/src/main/java/org/rocksdb/DBOptionsInterface.java

@@ -269,7 +269,10 @@ public interface DBOptionsInterface<T extends DBOptionsInterface> {
    * Statistics objects should not be shared between DB instances as
    * it does not use any locks to prevent concurrent updates.</p>
    *
+   * @param statistics The statistics to set
+   *
    * @return the instance of the current object.
+   *
    * @see RocksDB#open(org.rocksdb.Options, String)
    */
   T setStatistics(final Statistics statistics);
@@ -277,7 +280,9 @@ public interface DBOptionsInterface<T extends DBOptionsInterface> {
   /**
    * <p>Returns statistics object.</p>
    *
-   * @return the instance of the statistics object or null if there is no statistics object.
+   * @return the instance of the statistics object or null if there is no
+   * statistics object.
+   *
    * @see #setStatistics(Statistics)
    */
   Statistics statistics();

java/src/main/java/org/rocksdb/DirectComparator.java

@@ -25,6 +25,11 @@ public abstract class DirectComparator extends AbstractComparator<DirectSlice> {
     return createNewDirectComparator0(nativeParameterHandles[0]);
   }
 
+  @Override
+  final ComparatorType getComparatorType() {
+    return ComparatorType.JAVA_DIRECT_COMPARATOR;
+  }
+
   private native long createNewDirectComparator0(
       final long comparatorOptionsHandle);
 }

java/src/main/java/org/rocksdb/NativeComparatorWrapper.java

@@ -0,0 +1,57 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+/**
+ * A simple abstraction to allow a Java class to wrap a custom comparator
+ * implemented in C++.
+ *
+ * The native comparator must directly extend rocksdb::Comparator.
+ */
+public abstract class NativeComparatorWrapper
+    extends AbstractComparator<Slice> {
+
+  @Override
+  final ComparatorType getComparatorType() {
+    return ComparatorType.JAVA_NATIVE_COMPARATOR_WRAPPER;
+  }
+
+  @Override
+  public final String name() {
+    throw new IllegalStateException("This should not be called. " +
+        "Implementation is in Native code");
+  }
+
+  @Override
+  public final int compare(final Slice s1, final Slice s2) {
+    throw new IllegalStateException("This should not be called. " +
+        "Implementation is in Native code");
+  }
+
+  @Override
+  public final String findShortestSeparator(final String start, final Slice limit) {
+    throw new IllegalStateException("This should not be called. " +
+        "Implementation is in Native code");
+  }
+
+  @Override
+  public final String findShortSuccessor(final String key) {
+    throw new IllegalStateException("This should not be called. " +
+        "Implementation is in Native code");
+  }
+
+  /**
+   * We override {@link RocksCallbackObject#disposeInternal()}
+   * as disposing of a native rocksd::Comparator extension requires
+   * a slightly different approach as it is not really a RocksCallbackObject
+   */
+  @Override
+  protected void disposeInternal() {
+    disposeInternal(nativeHandle_);
+  }
+
+  private native void disposeInternal(final long handle);
+}

java/src/main/java/org/rocksdb/OptimisticTransactionDB.java

@@ -0,0 +1,175 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+import java.util.List;
+
+/**
+ * Database with Transaction support.
+ */
+public class OptimisticTransactionDB extends RocksDB
+    implements TransactionalDB<OptimisticTransactionOptions> {
+
+  /**
+   * Private constructor.
+   *
+   * @param nativeHandle The native handle of the C++ OptimisticTransactionDB
+   *     object
+   */
+  private OptimisticTransactionDB(final long nativeHandle) {
+    super(nativeHandle);
+  }
+
+  /**
+   * Open an OptimisticTransactionDB similar to
+   * {@link RocksDB#open(Options, String)}.
+   *
+   * @param options {@link org.rocksdb.Options} instance.
+   * @param path the path to the rocksdb.
+   *
+   * @return a {@link OptimisticTransactionDB} instance on success, null if the
+   * specified {@link OptimisticTransactionDB} can not be opened.
+   *
+   * @throws RocksDBException if an error occurs whilst opening the database.
+   */
+  public static OptimisticTransactionDB open(final Options options,
+      final String path) throws RocksDBException {
+    final OptimisticTransactionDB otdb = new OptimisticTransactionDB(open(
+        options.nativeHandle_, path));
+
+    // when non-default Options is used, keeping an Options reference
+    // in RocksDB can prevent Java to GC during the life-time of
+    // the currently-created RocksDB.
+    otdb.storeOptionsInstance(options);
+
+    return otdb;
+  }
+
+  /**
+   * Open an OptimisticTransactionDB similar to
+   * {@link RocksDB#open(DBOptions, String, List, List)}.
+   *
+   * @param dbOptions {@link org.rocksdb.DBOptions} instance.
+   * @param path the path to the rocksdb.
+   * @param columnFamilyDescriptors list of column family descriptors
+   * @param columnFamilyHandles will be filled with ColumnFamilyHandle instances
+   *
+   * @return a {@link OptimisticTransactionDB} instance on success, null if the
+   *     specified {@link OptimisticTransactionDB} can not be opened.
+   *
+   * @throws RocksDBException if an error occurs whilst opening the database.
+   */
+  public static OptimisticTransactionDB open(final DBOptions dbOptions,
+      final String path,
+      final List<ColumnFamilyDescriptor> columnFamilyDescriptors,
+      final List<ColumnFamilyHandle> columnFamilyHandles)
+      throws RocksDBException {
+
+    final byte[][] cfNames = new byte[columnFamilyDescriptors.size()][];
+    final long[] cfOptionHandles = new long[columnFamilyDescriptors.size()];
+    for (int i = 0; i < columnFamilyDescriptors.size(); i++) {
+      final ColumnFamilyDescriptor cfDescriptor = columnFamilyDescriptors
+          .get(i);
+      cfNames[i] = cfDescriptor.columnFamilyName();
+      cfOptionHandles[i] = cfDescriptor.columnFamilyOptions().nativeHandle_;
+    }
+
+    final long[] handles = open(dbOptions.nativeHandle_, path, cfNames,
+        cfOptionHandles);
+    final OptimisticTransactionDB otdb =
+        new OptimisticTransactionDB(handles[0]);
+
+    // when non-default Options is used, keeping an Options reference
+    // in RocksDB can prevent Java to GC during the life-time of
+    // the currently-created RocksDB.
+    otdb.storeOptionsInstance(dbOptions);
+
+    for (int i = 1; i < handles.length; i++) {
+      columnFamilyHandles.add(new ColumnFamilyHandle(otdb, handles[i]));
+    }
+
+    return otdb;
+  }
+
+  @Override
+  public Transaction beginTransaction(final WriteOptions writeOptions) {
+    return new Transaction(this, beginTransaction(nativeHandle_,
+        writeOptions.nativeHandle_));
+  }
+
+  @Override
+  public Transaction beginTransaction(final WriteOptions writeOptions,
+      final OptimisticTransactionOptions optimisticTransactionOptions) {
+    return new Transaction(this, beginTransaction(nativeHandle_,
+        writeOptions.nativeHandle_,
+        optimisticTransactionOptions.nativeHandle_));
+  }
+
+  // TODO(AR) consider having beingTransaction(... oldTransaction) set a
+  // reference count inside Transaction, so that we can always call
+  // Transaction#close but the object is only disposed when there are as many
+  // closes as beginTransaction. Makes the try-with-resources paradigm easier for
+  // java developers
+
+  @Override
+  public Transaction beginTransaction(final WriteOptions writeOptions,
+      final Transaction oldTransaction) {
+    final long jtxn_handle = beginTransaction_withOld(nativeHandle_,
+        writeOptions.nativeHandle_, oldTransaction.nativeHandle_);
+
+    // RocksJava relies on the assumption that
+    // we do not allocate a new Transaction object
+    // when providing an old_txn
+    assert(jtxn_handle == oldTransaction.nativeHandle_);
+
+    return oldTransaction;
+  }
+
+  @Override
+  public Transaction beginTransaction(final WriteOptions writeOptions,
+      final OptimisticTransactionOptions optimisticTransactionOptions,
+      final Transaction oldTransaction) {
+    final long jtxn_handle = beginTransaction_withOld(nativeHandle_,
+        writeOptions.nativeHandle_, optimisticTransactionOptions.nativeHandle_,
+        oldTransaction.nativeHandle_);
+
+    // RocksJava relies on the assumption that
+    // we do not allocate a new Transaction object
+    // when providing an old_txn
+    assert(jtxn_handle == oldTransaction.nativeHandle_);
+
+    return oldTransaction;
+  }
+
+  /**
+   * Get the underlying database that was opened.
+   *
+   * @return The underlying database that was opened.
+   */
+  public RocksDB getBaseDB() {
+    final RocksDB db = new RocksDB(getBaseDB(nativeHandle_));
+    db.disOwnNativeHandle();
+    return db;
+  }
+
+  protected static native long open(final long optionsHandle,
+      final String path) throws RocksDBException;
+  protected static native long[] open(final long handle, final String path,
+      final byte[][] columnFamilyNames, final long[] columnFamilyOptions);
+  private native long beginTransaction(final long handle,
+      final long writeOptionsHandle);
+  private native long beginTransaction(final long handle,
+      final long writeOptionsHandle,
+      final long optimisticTransactionOptionsHandle);
+  private native long beginTransaction_withOld(final long handle,
+      final long writeOptionsHandle, final long oldTransactionHandle);
+  private native long beginTransaction_withOld(final long handle,
+      final long writeOptionsHandle,
+      final long optimisticTransactionOptionsHandle,
+      final long oldTransactionHandle);
+  private native long getBaseDB(final long handle);
+  @Override protected final native void disposeInternal(final long handle);
+}

java/src/main/java/org/rocksdb/OptimisticTransactionOptions.java

@@ -0,0 +1,53 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+public class OptimisticTransactionOptions extends RocksObject
+    implements TransactionalOptions {
+
+  public OptimisticTransactionOptions() {
+    super(newOptimisticTransactionOptions());
+  }
+
+  @Override
+  public boolean isSetSnapshot() {
+    assert(isOwningHandle());
+    return isSetSnapshot(nativeHandle_);
+  }
+
+  @Override
+  public OptimisticTransactionOptions setSetSnapshot(
+      final boolean setSnapshot) {
+    assert(isOwningHandle());
+    setSetSnapshot(nativeHandle_, setSnapshot);
+    return this;
+  }
+
+  /**
+   * Should be set if the DB has a non-default comparator.
+   * See comment in
+   * {@link WriteBatchWithIndex#WriteBatchWithIndex(AbstractComparator, int, boolean)}
+   * constructor.
+   *
+   * @param comparator The comparator to use for the transaction.
+   *
+   * @return this OptimisticTransactionOptions instance
+   */
+  public OptimisticTransactionOptions setComparator(
+      final AbstractComparator<? extends AbstractSlice<?>> comparator) {
+    assert(isOwningHandle());
+    setComparator(nativeHandle_, comparator.nativeHandle_);
+    return this;
+  }
+
+  private native static long newOptimisticTransactionOptions();
+  private native boolean isSetSnapshot(final long handle);
+  private native void setSetSnapshot(final long handle,
+      final boolean setSnapshot);
+  private native void setComparator(final long handle,
+      final long comparatorHandle);
+  @Override protected final native void disposeInternal(final long handle);
+}

java/src/main/java/org/rocksdb/Options.java

@@ -170,7 +170,7 @@ public class Options extends RocksObject
       final AbstractComparator<? extends AbstractSlice<?>> comparator) {
     assert(isOwningHandle());
     setComparatorHandle(nativeHandle_, comparator.nativeHandle_,
-            comparator instanceof DirectComparator);
+            comparator.getComparatorType().getValue());
     comparator_ = comparator;
     return this;
   }
@@ -1734,7 +1734,7 @@ public class Options extends RocksObject
       long memtableMemoryBudget);
   private native void setComparatorHandle(long handle, int builtinComparator);
   private native void setComparatorHandle(long optHandle,
-      long comparatorHandle, boolean isDirect);
+      long comparatorHandle, byte comparatorType);
   private native void setMergeOperatorName(
       long handle, String name);
   private native void setMergeOperator(

java/src/main/java/org/rocksdb/RocksDB.java

@@ -435,7 +435,7 @@ public class RocksDB extends RocksObject {
         path));
   }
 
-  private void storeOptionsInstance(DBOptionsInterface options) {
+  protected void storeOptionsInstance(DBOptionsInterface options) {
     options_ = options;
   }
 
@@ -1683,7 +1683,7 @@ public class RocksDB extends RocksObject {
    * @return The handle of the default column family
    */
   public ColumnFamilyHandle getDefaultColumnFamily() {
-    ColumnFamilyHandle cfHandle = new ColumnFamilyHandle(this,
+    final ColumnFamilyHandle cfHandle = new ColumnFamilyHandle(this,
         getDefaultColumnFamily(nativeHandle_));
     cfHandle.disOwnNativeHandle();
     return cfHandle;
@@ -2359,8 +2359,9 @@ public class RocksDB extends RocksObject {
       final long[] columnFamilyHandles, final long readOptHandle)
       throws RocksDBException;
   protected native long getSnapshot(long nativeHandle);
-  protected native void releaseSnapshot(long nativeHandle, long snapshotHandle);
-  @Override protected final native void disposeInternal(final long handle);
+  protected native void releaseSnapshot(
+      long nativeHandle, long snapshotHandle);
+  @Override protected native void disposeInternal(final long handle);
   private native long getDefaultColumnFamily(long handle);
   private native long createColumnFamily(final long handle,
       final byte[] columnFamilyName, final long columnFamilyOptions)

java/src/main/java/org/rocksdb/Snapshot.java

@@ -11,6 +11,10 @@ package org.rocksdb;
 public class Snapshot extends RocksObject {
   Snapshot(final long nativeHandle) {
     super(nativeHandle);
+
+    // The pointer to the snapshot is always released
+    // by the database instance.
+    disOwnNativeHandle();
   }
 
   /**
@@ -20,17 +24,17 @@ public class Snapshot extends RocksObject {
    *     this snapshot.
    */
   public long getSequenceNumber() {
-    assert(isOwningHandle());
     return getSequenceNumber(nativeHandle_);
   }
 
-  /**
-   * Dont release C++ Snapshot pointer. The pointer
-   * to the snapshot is released by the database
-   * instance.
-   */
   @Override
   protected final void disposeInternal(final long handle) {
+    /**
+     * Nothing to release, we never own the pointer for a
+     * Snapshot. The pointer
+     * to the snapshot is released by the database
+     * instance.
+     */
   }
 
   private native long getSequenceNumber(long handle);

java/src/main/java/org/rocksdb/SstFileWriter.java

@@ -31,7 +31,7 @@ public class SstFileWriter extends RocksObject {
       final AbstractComparator<? extends AbstractSlice<?>> comparator) {
     super(newSstFileWriter(
         envOptions.nativeHandle_, options.nativeHandle_, comparator.nativeHandle_,
-        comparator instanceof DirectComparator));
+        comparator.getComparatorType().getValue()));
   }
 
   /**
@@ -225,7 +225,7 @@ public void put(final byte[] key, final byte[] value)
 
   private native static long newSstFileWriter(
       final long envOptionsHandle, final long optionsHandle,
-      final long userComparatorHandle, final boolean isDirect);
+      final long userComparatorHandle, final byte comparatorType);
 
   private native static long newSstFileWriter(final long envOptionsHandle,
       final long optionsHandle);

java/src/main/java/org/rocksdb/Statistics.java

@@ -117,6 +117,8 @@ public class Statistics extends RocksObject {
 
   /**
    * Resets all ticker and histogram stats.
+   *
+   * @throws RocksDBException if an error occurs when resetting the statistics.
    */
   public void reset() throws RocksDBException {
     assert(isOwningHandle());
@@ -126,6 +128,7 @@ public class Statistics extends RocksObject {
   /**
    * String representation of the statistic object.
    */
+  @Override
   public String toString() {
     assert(isOwningHandle());
     return toString(nativeHandle_);

java/src/main/java/org/rocksdb/Status.java

@@ -87,6 +87,15 @@ public class Status {
       throw new IllegalArgumentException(
           "Illegal value provided for Code (" + value + ").");
     }
+
+    /**
+     * Returns the byte value of the enumerations value.
+     *
+     * @return byte representation
+     */
+    public byte getValue() {
+      return value;
+    }
   }
 
   // should stay in sync with /include/rocksdb/status.h:SubCode and /java/rocksjni/portal.h:toJavaStatusSubCode
@@ -116,5 +125,14 @@ public class Status {
       throw new IllegalArgumentException(
           "Illegal value provided for SubCode (" + value + ").");
     }
+
+    /**
+     * Returns the byte value of the enumerations value.
+     *
+     * @return byte representation
+     */
+    public byte getValue() {
+      return value;
+    }
   }
 }

java/src/main/java/org/rocksdb/TransactionDB.java

@@ -0,0 +1,354 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+import java.util.ArrayList;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+
+/**
+ * Database with Transaction support
+ */
+public class TransactionDB extends RocksDB
+    implements TransactionalDB<TransactionOptions> {
+
+  private TransactionDBOptions transactionDbOptions_;
+
+  /**
+   * Private constructor.
+   *
+   * @param nativeHandle The native handle of the C++ TransactionDB object
+   */
+  private TransactionDB(final long nativeHandle) {
+    super(nativeHandle);
+  }
+
+  /**
+   * Open a TransactionDB, similar to {@link RocksDB#open(Options, String)}.
+   *
+   * @param options {@link org.rocksdb.Options} instance.
+   * @param transactionDbOptions {@link org.rocksdb.TransactionDBOptions}
+   *     instance.
+   * @param path the path to the rocksdb.
+   *
+   * @return a {@link TransactionDB} instance on success, null if the specified
+   *     {@link TransactionDB} can not be opened.
+   *
+   * @throws RocksDBException if an error occurs whilst opening the database.
+   */
+  public static TransactionDB open(final Options options,
+      final TransactionDBOptions transactionDbOptions, final String path)
+      throws RocksDBException {
+    final TransactionDB tdb = new TransactionDB(open(options.nativeHandle_,
+        transactionDbOptions.nativeHandle_, path));
+
+    // when non-default Options is used, keeping an Options reference
+    // in RocksDB can prevent Java to GC during the life-time of
+    // the currently-created RocksDB.
+    tdb.storeOptionsInstance(options);
+    tdb.storeTransactionDbOptions(transactionDbOptions);
+
+    return tdb;
+  }
+
+  /**
+   * Open a TransactionDB, similar to
+   * {@link RocksDB#open(DBOptions, String, List, List)}.
+   *
+   * @param dbOptions {@link org.rocksdb.DBOptions} instance.
+   * @param transactionDbOptions {@link org.rocksdb.TransactionDBOptions}
+   *     instance.
+   * @param path the path to the rocksdb.
+   * @param columnFamilyDescriptors list of column family descriptors
+   * @param columnFamilyHandles will be filled with ColumnFamilyHandle instances
+   *
+   * @return a {@link TransactionDB} instance on success, null if the specified
+   *     {@link TransactionDB} can not be opened.
+   *
+   * @throws RocksDBException if an error occurs whilst opening the database.
+   */
+  public static TransactionDB open(final DBOptions dbOptions,
+      final TransactionDBOptions transactionDbOptions,
+      final String path,
+      final List<ColumnFamilyDescriptor> columnFamilyDescriptors,
+      final List<ColumnFamilyHandle> columnFamilyHandles)
+      throws RocksDBException {
+
+    final byte[][] cfNames = new byte[columnFamilyDescriptors.size()][];
+    final long[] cfOptionHandles = new long[columnFamilyDescriptors.size()];
+    for (int i = 0; i < columnFamilyDescriptors.size(); i++) {
+      final ColumnFamilyDescriptor cfDescriptor = columnFamilyDescriptors
+          .get(i);
+      cfNames[i] = cfDescriptor.columnFamilyName();
+      cfOptionHandles[i] = cfDescriptor.columnFamilyOptions().nativeHandle_;
+    }
+
+    final long[] handles = open(dbOptions.nativeHandle_,
+        transactionDbOptions.nativeHandle_, path, cfNames, cfOptionHandles);
+    final TransactionDB tdb = new TransactionDB(handles[0]);
+
+    // when non-default Options is used, keeping an Options reference
+    // in RocksDB can prevent Java to GC during the life-time of
+    // the currently-created RocksDB.
+    tdb.storeOptionsInstance(dbOptions);
+    tdb.storeTransactionDbOptions(transactionDbOptions);
+
+    for (int i = 1; i < handles.length; i++) {
+      columnFamilyHandles.add(new ColumnFamilyHandle(tdb, handles[i]));
+    }
+
+    return tdb;
+  }
+
+  @Override
+  public Transaction beginTransaction(final WriteOptions writeOptions) {
+    return new Transaction(this, beginTransaction(nativeHandle_,
+        writeOptions.nativeHandle_));
+  }
+
+  @Override
+  public Transaction beginTransaction(final WriteOptions writeOptions,
+      final TransactionOptions transactionOptions) {
+    return new Transaction(this, beginTransaction(nativeHandle_,
+        writeOptions.nativeHandle_, transactionOptions.nativeHandle_));
+  }
+
+  // TODO(AR) consider having beingTransaction(... oldTransaction) set a
+  // reference count inside Transaction, so that we can always call
+  // Transaction#close but the object is only disposed when there are as many
+  // closes as beginTransaction. Makes the try-with-resources paradigm easier for
+  // java developers
+
+  @Override
+  public Transaction beginTransaction(final WriteOptions writeOptions,
+      final Transaction oldTransaction) {
+    final long jtxnHandle = beginTransaction_withOld(nativeHandle_,
+        writeOptions.nativeHandle_, oldTransaction.nativeHandle_);
+
+    // RocksJava relies on the assumption that
+    // we do not allocate a new Transaction object
+    // when providing an old_txn
+    assert(jtxnHandle == oldTransaction.nativeHandle_);
+
+    return oldTransaction;
+  }
+
+  @Override
+  public Transaction beginTransaction(final WriteOptions writeOptions,
+      final TransactionOptions transactionOptions,
+      final Transaction oldTransaction) {
+    final long jtxn_handle = beginTransaction_withOld(nativeHandle_,
+        writeOptions.nativeHandle_, transactionOptions.nativeHandle_,
+        oldTransaction.nativeHandle_);
+
+    // RocksJava relies on the assumption that
+    // we do not allocate a new Transaction object
+    // when providing an old_txn
+    assert(jtxn_handle == oldTransaction.nativeHandle_);
+
+    return oldTransaction;
+  }
+
+  public Transaction getTransactionByName(final String transactionName) {
+    final long jtxnHandle = getTransactionByName(nativeHandle_, transactionName);
+    if(jtxnHandle == 0) {
+      return null;
+    }
+
+    final Transaction txn = new Transaction(this, jtxnHandle);
+
+    // this instance doesn't own the underlying C++ object
+    txn.disOwnNativeHandle();
+
+    return txn;
+  }
+
+  public List<Transaction> getAllPreparedTransactions() {
+    final long[] jtxnHandles = getAllPreparedTransactions(nativeHandle_);
+
+    final List<Transaction> txns = new ArrayList<>();
+    for(final long jtxnHandle : jtxnHandles) {
+      final Transaction txn = new Transaction(this, jtxnHandle);
+
+      // this instance doesn't own the underlying C++ object
+      txn.disOwnNativeHandle();
+
+      txns.add(txn);
+    }
+    return txns;
+  }
+
+  public static class KeyLockInfo {
+    private final String key;
+    private final long[] transactionIDs;
+    private final boolean exclusive;
+
+    public KeyLockInfo(final String key, final long transactionIDs[],
+        final boolean exclusive) {
+      this.key = key;
+      this.transactionIDs = transactionIDs;
+      this.exclusive = exclusive;
+    }
+
+    /**
+     * Get the key.
+     *
+     * @return the key
+     */
+    public String getKey() {
+      return key;
+    }
+
+    /**
+     * Get the Transaction IDs.
+     *
+     * @return the Transaction IDs.
+     */
+    public long[] getTransactionIDs() {
+      return transactionIDs;
+    }
+
+    /**
+     * Get the Lock status.
+     *
+     * @return true if the lock is exclusive, false if the lock is shared.
+     */
+    public boolean isExclusive() {
+      return exclusive;
+    }
+  }
+
+  /**
+   * Returns map of all locks held.
+   *
+   * @return a map of all the locks held.
+   */
+  public Map<Long, KeyLockInfo> getLockStatusData() {
+    return getLockStatusData(nativeHandle_);
+  }
+
+  /**
+   * Called from C++ native method {@link #getDeadlockInfoBuffer(long)}
+   * to construct a DeadlockInfo object.
+   *
+   * @param transactionID The transaction id
+   * @param columnFamilyId The id of the {@link ColumnFamilyHandle}
+   * @param waitingKey the key that we are waiting on
+   * @param exclusive true if the lock is exclusive, false if the lock is shared
+   *
+   * @return The waiting transactions
+   */
+  private DeadlockInfo newDeadlockInfo(
+      final long transactionID, final long columnFamilyId,
+      final String waitingKey, final boolean exclusive) {
+    return new DeadlockInfo(transactionID, columnFamilyId,
+        waitingKey, exclusive);
+  }
+
+  public static class DeadlockInfo {
+    private final long transactionID;
+    private final long columnFamilyId;
+    private final String waitingKey;
+    private final boolean exclusive;
+
+    private DeadlockInfo(final long transactionID, final long columnFamilyId,
+      final String waitingKey, final boolean exclusive) {
+      this.transactionID = transactionID;
+      this.columnFamilyId = columnFamilyId;
+      this.waitingKey = waitingKey;
+      this.exclusive = exclusive;
+    }
+
+    /**
+     * Get the Transaction ID.
+     *
+     * @return the transaction ID
+     */
+    public long getTransactionID() {
+      return transactionID;
+    }
+
+    /**
+     * Get the Column Family ID.
+     *
+     * @return The column family ID
+     */
+    public long getColumnFamilyId() {
+      return columnFamilyId;
+    }
+
+    /**
+     * Get the key that we are waiting on.
+     *
+     * @return the key that we are waiting on
+     */
+    public String getWaitingKey() {
+      return waitingKey;
+    }
+
+    /**
+     * Get the Lock status.
+     *
+     * @return true if the lock is exclusive, false if the lock is shared.
+     */
+    public boolean isExclusive() {
+      return exclusive;
+    }
+  }
+
+  public static class DeadlockPath {
+    final DeadlockInfo[] path;
+    final boolean limitExceeded;
+
+    public DeadlockPath(final DeadlockInfo[] path, final boolean limitExceeded) {
+      this.path = path;
+      this.limitExceeded = limitExceeded;
+    }
+
+    public boolean isEmpty() {
+      return path.length == 0 && !limitExceeded;
+    }
+  }
+
+  public DeadlockPath[] getDeadlockInfoBuffer() {
+    return getDeadlockInfoBuffer(nativeHandle_);
+  }
+
+  public void setDeadlockInfoBufferSize(final int targetSize) {
+    setDeadlockInfoBufferSize(nativeHandle_, targetSize);
+  }
+
+  private void storeTransactionDbOptions(
+      final TransactionDBOptions transactionDbOptions) {
+    this.transactionDbOptions_ = transactionDbOptions;
+  }
+
+  private static native long open(final long optionsHandle,
+      final long transactionDbOptionsHandle, final String path)
+      throws RocksDBException;
+  private static native long[] open(final long dbOptionsHandle,
+      final long transactionDbOptionsHandle, final String path,
+      final byte[][] columnFamilyNames, final long[] columnFamilyOptions);
+  private native long beginTransaction(final long handle,
+      final long writeOptionsHandle);
+  private native long beginTransaction(final long handle,
+      final long writeOptionsHandle, final long transactionOptionsHandle);
+  private native long beginTransaction_withOld(final long handle,
+      final long writeOptionsHandle, final long oldTransactionHandle);
+  private native long beginTransaction_withOld(final long handle,
+      final long writeOptionsHandle, final long transactionOptionsHandle,
+      final long oldTransactionHandle);
+  private native long getTransactionByName(final long handle,
+      final String name);
+  private native long[] getAllPreparedTransactions(final long handle);
+  private native Map<Long, KeyLockInfo> getLockStatusData(
+      final long handle);
+  private native DeadlockPath[] getDeadlockInfoBuffer(final long handle);
+  private native void setDeadlockInfoBufferSize(final long handle,
+      final int targetSize);
+  @Override protected final native void disposeInternal(final long handle);
+}

java/src/main/java/org/rocksdb/TransactionDBOptions.java

@@ -0,0 +1,217 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+public class TransactionDBOptions extends RocksObject {
+
+  public TransactionDBOptions() {
+    super(newTransactionDBOptions());
+  }
+
+  /**
+   * Specifies the maximum number of keys that can be locked at the same time
+   * per column family.
+   *
+   * If the number of locked keys is greater than {@link #getMaxNumLocks()},
+   * transaction writes (or GetForUpdate) will return an error.
+   *
+   * @return The maximum number of keys that can be locked
+   */
+  public long getMaxNumLocks() {
+    assert(isOwningHandle());
+    return getMaxNumLocks(nativeHandle_);
+  }
+
+  /**
+   * Specifies the maximum number of keys that can be locked at the same time
+   * per column family.
+   *
+   * If the number of locked keys is greater than {@link #getMaxNumLocks()},
+   * transaction writes (or GetForUpdate) will return an error.
+   *
+   * @param maxNumLocks The maximum number of keys that can be locked;
+   *     If this value is not positive, no limit will be enforced.
+   *
+   * @return this TransactionDBOptions instance
+   */
+  public TransactionDBOptions setMaxNumLocks(final long maxNumLocks) {
+    assert(isOwningHandle());
+    setMaxNumLocks(nativeHandle_, maxNumLocks);
+    return this;
+  }
+
+  /**
+   * The number of sub-tables per lock table (per column family)
+   *
+   * @return The number of sub-tables
+   */
+  public long getNumStripes() {
+    assert(isOwningHandle());
+    return getNumStripes(nativeHandle_);
+  }
+
+  /**
+   * Increasing this value will increase the concurrency by dividing the lock
+   * table (per column family) into more sub-tables, each with their own
+   * separate mutex.
+   *
+   * Default: 16
+   *
+   * @param numStripes The number of sub-tables
+   *
+   * @return this TransactionDBOptions instance
+   */
+  public TransactionDBOptions setNumStripes(final long numStripes) {
+    assert(isOwningHandle());
+    setNumStripes(nativeHandle_, numStripes);
+    return this;
+  }
+
+  /**
+   * The default wait timeout in milliseconds when
+   * a transaction attempts to lock a key if not specified by
+   * {@link TransactionOptions#setLockTimeout(long)}
+   *
+   * If 0, no waiting is done if a lock cannot instantly be acquired.
+   * If negative, there is no timeout.
+   *
+   * @return the default wait timeout in milliseconds
+   */
+  public long getTransactionLockTimeout() {
+    assert(isOwningHandle());
+    return getTransactionLockTimeout(nativeHandle_);
+  }
+
+  /**
+   * If positive, specifies the default wait timeout in milliseconds when
+   * a transaction attempts to lock a key if not specified by
+   * {@link TransactionOptions#setLockTimeout(long)}
+   *
+   * If 0, no waiting is done if a lock cannot instantly be acquired.
+   * If negative, there is no timeout. Not using a timeout is not recommended
+   * as it can lead to deadlocks.  Currently, there is no deadlock-detection to
+   * recover from a deadlock.
+   *
+   * Default: 1000
+   *
+   * @param transactionLockTimeout the default wait timeout in milliseconds
+   *
+   * @return this TransactionDBOptions instance
+   */
+  public TransactionDBOptions setTransactionLockTimeout(
+      final long transactionLockTimeout) {
+    assert(isOwningHandle());
+    setTransactionLockTimeout(nativeHandle_, transactionLockTimeout);
+    return this;
+  }
+
+  /**
+   * The wait timeout in milliseconds when writing a key
+   * OUTSIDE of a transaction (ie by calling {@link RocksDB#put},
+   * {@link RocksDB#merge}, {@link RocksDB#remove} or {@link RocksDB#write}
+   * directly).
+   *
+   * If 0, no waiting is done if a lock cannot instantly be acquired.
+   * If negative, there is no timeout and will block indefinitely when acquiring
+   * a lock.
+   *
+   * @return the timeout in milliseconds when writing a key OUTSIDE of a
+   *     transaction
+   */
+  public long getDefaultLockTimeout() {
+    assert(isOwningHandle());
+    return getDefaultLockTimeout(nativeHandle_);
+  }
+
+  /**
+   * If positive, specifies the wait timeout in milliseconds when writing a key
+   * OUTSIDE of a transaction (ie by calling {@link RocksDB#put},
+   * {@link RocksDB#merge}, {@link RocksDB#remove} or {@link RocksDB#write}
+   * directly).
+   *
+   * If 0, no waiting is done if a lock cannot instantly be acquired.
+   * If negative, there is no timeout and will block indefinitely when acquiring
+   * a lock.
+   *
+   * Not using a timeout can lead to deadlocks. Currently, there
+   * is no deadlock-detection to recover from a deadlock.  While DB writes
+   * cannot deadlock with other DB writes, they can deadlock with a transaction.
+   * A negative timeout should only be used if all transactions have a small
+   * expiration set.
+   *
+   * Default: 1000
+   *
+   * @param defaultLockTimeout the timeout in milliseconds when writing a key
+   *     OUTSIDE of a transaction
+   * @return this TransactionDBOptions instance
+   */
+   public TransactionDBOptions setDefaultLockTimeout(
+       final long defaultLockTimeout) {
+     assert(isOwningHandle());
+     setDefaultLockTimeout(nativeHandle_, defaultLockTimeout);
+     return this;
+   }
+
+//  /**
+//   * If set, the {@link TransactionDB} will use this implementation of a mutex
+//   * and condition variable for all transaction locking instead of the default
+//   * mutex/condvar implementation.
+//   *
+//   * @param transactionDbMutexFactory the mutex factory for the transactions
+//   *
+//   * @return this TransactionDBOptions instance
+//   */
+//  public TransactionDBOptions setCustomMutexFactory(
+//      final TransactionDBMutexFactory transactionDbMutexFactory) {
+//
+//  }
+
+  /**
+   * The policy for when to write the data into the DB. The default policy is to
+   * write only the committed data {@link TxnDBWritePolicy#WRITE_COMMITTED}.
+   * The data could be written before the commit phase. The DB then needs to
+   * provide the mechanisms to tell apart committed from uncommitted data.
+   *
+   * @return The write policy.
+   */
+  public TxnDBWritePolicy getWritePolicy() {
+    assert(isOwningHandle());
+    return TxnDBWritePolicy.getTxnDBWritePolicy(getWritePolicy(nativeHandle_));
+  }
+
+  /**
+   * The policy for when to write the data into the DB. The default policy is to
+   * write only the committed data {@link TxnDBWritePolicy#WRITE_COMMITTED}.
+   * The data could be written before the commit phase. The DB then needs to
+   * provide the mechanisms to tell apart committed from uncommitted data.
+   *
+   * @param writePolicy The write policy.
+   *
+   * @return this TransactionDBOptions instance
+   */
+  public TransactionDBOptions setWritePolicy(
+      final TxnDBWritePolicy writePolicy) {
+    assert(isOwningHandle());
+    setWritePolicy(nativeHandle_, writePolicy.getValue());
+    return this;
+  }
+
+  private native static long newTransactionDBOptions();
+  private native long getMaxNumLocks(final long handle);
+  private native void setMaxNumLocks(final long handle,
+      final long maxNumLocks);
+  private native long getNumStripes(final long handle);
+  private native void setNumStripes(final long handle, final long numStripes);
+  private native long getTransactionLockTimeout(final long handle);
+  private native void setTransactionLockTimeout(final long handle,
+      final long transactionLockTimeout);
+  private native long getDefaultLockTimeout(final long handle);
+  private native void setDefaultLockTimeout(final long handle,
+      final long transactionLockTimeout);
+  private native byte getWritePolicy(final long handle);
+  private native void setWritePolicy(final long handle, final byte writePolicy);
+  @Override protected final native void disposeInternal(final long handle);
+}

java/src/main/java/org/rocksdb/TransactionOptions.java

@@ -0,0 +1,189 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+public class TransactionOptions extends RocksObject
+    implements TransactionalOptions {
+
+  public TransactionOptions() {
+    super(newTransactionOptions());
+  }
+
+  @Override
+  public boolean isSetSnapshot() {
+    assert(isOwningHandle());
+    return isSetSnapshot(nativeHandle_);
+  }
+
+  @Override
+  public TransactionOptions setSetSnapshot(final boolean setSnapshot) {
+    assert(isOwningHandle());
+    setSetSnapshot(nativeHandle_, setSnapshot);
+    return this;
+  }
+
+  /**
+   * True means that before acquiring locks, this transaction will
+   * check if doing so will cause a deadlock. If so, it will return with
+   * {@link Status.Code#Busy}. The user should retry their transaction.
+   *
+   * @return true if a deadlock is detected.
+   */
+  public boolean isDeadlockDetect() {
+    assert(isOwningHandle());
+    return isDeadlockDetect(nativeHandle_);
+  }
+
+  /**
+   * Setting to true means that before acquiring locks, this transaction will
+   * check if doing so will cause a deadlock. If so, it will return with
+   * {@link Status.Code#Busy}. The user should retry their transaction.
+   *
+   * @param deadlockDetect true if we should detect deadlocks.
+   *
+   * @return this TransactionOptions instance
+   */
+  public TransactionOptions setDeadlockDetect(final boolean deadlockDetect) {
+    assert(isOwningHandle());
+    setDeadlockDetect(nativeHandle_, deadlockDetect);
+    return this;
+  }
+
+  /**
+   * The wait timeout in milliseconds when a transaction attempts to lock a key.
+   *
+   * If 0, no waiting is done if a lock cannot instantly be acquired.
+   * If negative, {@link TransactionDBOptions#getTransactionLockTimeout(long)}
+   * will be used
+   *
+   * @return the lock tiemout in milliseconds
+   */
+  public long getLockTimeout() {
+    assert(isOwningHandle());
+    return getLockTimeout(nativeHandle_);
+  }
+
+  /**
+   * If positive, specifies the wait timeout in milliseconds when
+   * a transaction attempts to lock a key.
+   *
+   * If 0, no waiting is done if a lock cannot instantly be acquired.
+   * If negative, {@link TransactionDBOptions#getTransactionLockTimeout(long)}
+   * will be used
+   *
+   * Default: -1
+   *
+   * @param lockTimeout the lock tiemout in milliseconds
+   *
+   * @return this TransactionOptions instance
+   */
+  public TransactionOptions setLockTimeout(final long lockTimeout) {
+    assert(isOwningHandle());
+    setLockTimeout(nativeHandle_, lockTimeout);
+    return this;
+  }
+
+  /**
+   * Expiration duration in milliseconds.
+   *
+   * If non-negative, transactions that last longer than this many milliseconds
+   * will fail to commit. If not set, a forgotten transaction that is never
+   * committed, rolled back, or deleted will never relinquish any locks it
+   * holds. This could prevent keys from being written by other writers.
+   *
+   * @return expiration the expiration duration in milliseconds
+   */
+  public long getExpiration() {
+    assert(isOwningHandle());
+    return getExpiration(nativeHandle_);
+  }
+
+  /**
+   * Expiration duration in milliseconds.
+   *
+   * If non-negative, transactions that last longer than this many milliseconds
+   * will fail to commit. If not set, a forgotten transaction that is never
+   * committed, rolled back, or deleted will never relinquish any locks it
+   * holds. This could prevent keys from being written by other writers.
+   *
+   * Default: -1
+   *
+   * @param expiration the expiration duration in milliseconds
+   *
+   * @return this TransactionOptions instance
+   */
+  public TransactionOptions setExpiration(final long expiration) {
+    assert(isOwningHandle());
+    setExpiration(nativeHandle_, expiration);
+    return this;
+  }
+
+  /**
+   * Gets the number of traversals to make during deadlock detection.
+   *
+   * @return the number of traversals to make during
+   *     deadlock detection
+   */
+  public long getDeadlockDetectDepth() {
+    return getDeadlockDetectDepth(nativeHandle_);
+  }
+
+  /**
+   * Sets the number of traversals to make during deadlock detection.
+   *
+   * Default: 50
+   *
+   * @param deadlockDetectDepth the the number of traversals to make during
+   *     deadlock detection
+   *
+   * @return this TransactionOptions instance
+   */
+  public TransactionOptions setDeadlockDetectDepth(
+      final long deadlockDetectDepth) {
+    setDeadlockDetectDepth(nativeHandle_, deadlockDetectDepth);
+    return this;
+  }
+
+  /**
+   * Get the maximum number of bytes that may be used for the write batch.
+   *
+   * @return the maximum number of bytes, 0 means no limit.
+   */
+  public long getMaxWriteBatchSize() {
+    return getMaxWriteBatchSize(nativeHandle_);
+  }
+
+  /**
+   * Set the maximum number of bytes that may be used for the write batch.
+   *
+   * @param maxWriteBatchSize the maximum number of bytes, 0 means no limit.
+   *
+   * @return this TransactionOptions instance
+   */
+  public TransactionOptions setMaxWriteBatchSize(final long maxWriteBatchSize) {
+    setMaxWriteBatchSize(nativeHandle_, maxWriteBatchSize);
+    return this;
+  }
+
+  private native static long newTransactionOptions();
+  private native boolean isSetSnapshot(final long handle);
+  private native void setSetSnapshot(final long handle,
+      final boolean setSnapshot);
+  private native boolean isDeadlockDetect(final long handle);
+  private native void setDeadlockDetect(final long handle,
+      final boolean deadlockDetect);
+  private native long getLockTimeout(final long handle);
+  private native void setLockTimeout(final long handle, final long lockTimeout);
+  private native long getExpiration(final long handle);
+  private native void setExpiration(final long handle, final long expiration);
+  private native long getDeadlockDetectDepth(final long handle);
+  private native void setDeadlockDetectDepth(final long handle,
+      final long deadlockDetectDepth);
+  private native long getMaxWriteBatchSize(final long handle);
+  private native void setMaxWriteBatchSize(final long handle,
+      final long maxWriteBatchSize);
+  @Override protected final native void disposeInternal(final long handle);
+}

java/src/main/java/org/rocksdb/TransactionalDB.java

@@ -0,0 +1,68 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+
+interface TransactionalDB<T extends TransactionalOptions>
+    extends AutoCloseable {
+
+  /**
+   * Starts a new Transaction.
+   *
+   * Caller is responsible for calling {@link #close()} on the returned
+   * transaction when it is no longer needed.
+   *
+   * @param writeOptions Any write options for the transaction
+   * @return a new transaction
+   */
+  Transaction beginTransaction(final WriteOptions writeOptions);
+
+  /**
+   * Starts a new Transaction.
+   *
+   * Caller is responsible for calling {@link #close()} on the returned
+   * transaction when it is no longer needed.
+   *
+   * @param writeOptions Any write options for the transaction
+   * @param transactionOptions Any options for the transaction
+   * @return a new transaction
+   */
+  Transaction beginTransaction(final WriteOptions writeOptions,
+      final T transactionOptions);
+
+  /**
+   * Starts a new Transaction.
+   *
+   * Caller is responsible for calling {@link #close()} on the returned
+   * transaction when it is no longer needed.
+   *
+   * @param writeOptions Any write options for the transaction
+   * @param oldTransaction this Transaction will be reused instead of allocating
+   *     a new one. This is an optimization to avoid extra allocations
+   *     when repeatedly creating transactions.
+   * @return The oldTransaction which has been reinitialized as a new
+   *     transaction
+   */
+  Transaction beginTransaction(final WriteOptions writeOptions,
+      final Transaction oldTransaction);
+
+  /**
+   * Starts a new Transaction.
+   *
+   * Caller is responsible for calling {@link #close()} on the returned
+   * transaction when it is no longer needed.
+   *
+   * @param writeOptions Any write options for the transaction
+   * @param transactionOptions Any options for the transaction
+   * @param oldTransaction this Transaction will be reused instead of allocating
+   *     a new one. This is an optimization to avoid extra allocations
+   *     when repeatedly creating transactions.
+   * @return The oldTransaction which has been reinitialized as a new
+   *     transaction
+   */
+  Transaction beginTransaction(final WriteOptions writeOptions,
+      final T transactionOptions, final Transaction oldTransaction);
+}

java/src/main/java/org/rocksdb/TransactionalOptions.java

@@ -0,0 +1,31 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+
+interface TransactionalOptions extends AutoCloseable {
+
+  /**
+   * True indicates snapshots will be set, just like if
+   * {@link Transaction#setSnapshot()} had been called
+   *
+   * @return whether a snapshot will be set
+   */
+  boolean isSetSnapshot();
+
+  /**
+   * Setting the setSnapshot to true is the same as calling
+   * {@link Transaction#setSnapshot()}.
+   *
+   * Default: false
+   *
+   * @param <T> The type of transactional options.
+   * @param setSnapshot Whether to set a snapshot
+   *
+   * @return this TransactionalOptions instance
+   */
+  <T extends TransactionalOptions> T setSetSnapshot(final boolean setSnapshot);
+}

java/src/main/java/org/rocksdb/TxnDBWritePolicy.java

@@ -0,0 +1,62 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+package org.rocksdb;
+
+/**
+ * The transaction db write policy.
+ */
+public enum TxnDBWritePolicy {
+  /**
+   * Write only the committed data.
+   */
+  WRITE_COMMITTED((byte)0x00),
+
+  /**
+   * Write data after the prepare phase of 2pc.
+   */
+  WRITE_PREPARED((byte)0x1),
+
+  /**
+   * Write data before the prepare phase of 2pc.
+   */
+  WRITE_UNPREPARED((byte)0x2);
+
+  private byte value;
+
+  TxnDBWritePolicy(final byte value) {
+    this.value = value;
+  }
+
+  /**
+   * <p>Returns the byte value of the enumerations value.</p>
+   *
+   * @return byte representation
+   */
+  public byte getValue() {
+    return value;
+  }
+
+  /**
+   * <p>Get the TxnDBWritePolicy enumeration value by
+   * passing the byte identifier to this method.</p>
+   *
+   * @param byteIdentifier of TxnDBWritePolicy.
+   *
+   * @return TxnDBWritePolicy instance.
+   *
+   * @throws IllegalArgumentException If TxnDBWritePolicy cannot be found for
+   *     the provided byteIdentifier
+   */
+  public static TxnDBWritePolicy getTxnDBWritePolicy(final byte byteIdentifier) {
+    for (final TxnDBWritePolicy txnDBWritePolicy : TxnDBWritePolicy.values()) {
+      if (txnDBWritePolicy.getValue() == byteIdentifier) {
+        return txnDBWritePolicy;
+      }
+    }
+
+    throw new IllegalArgumentException(
+        "Illegal value provided for TxnDBWritePolicy.");
+  }
+}

java/src/main/java/org/rocksdb/WALRecoveryMode.java

@@ -65,7 +65,7 @@ public enum WALRecoveryMode {
    *
    * @param byteIdentifier of WALRecoveryMode.
    *
-   * @return CompressionType instance.
+   * @return WALRecoveryMode instance.
    *
    * @throws IllegalArgumentException If WALRecoveryMode cannot be found for the
    *   provided byteIdentifier

java/src/main/java/org/rocksdb/WBWIRocksIterator.java

@@ -55,10 +55,13 @@ public class WBWIRocksIterator
    * that created the record in the Write Batch
    */
   public enum WriteType {
-    PUT((byte)0x1),
-    MERGE((byte)0x2),
-    DELETE((byte)0x4),
-    LOG((byte)0x8);
+    PUT((byte)0x0),
+    MERGE((byte)0x1),
+    DELETE((byte)0x2),
+    SINGLE_DELETE((byte)0x3),
+    DELETE_RANGE((byte)0x4),
+    LOG((byte)0x5),
+    XID((byte)0x6);
 
     final byte id;
     WriteType(final byte id) {

java/src/main/java/org/rocksdb/WriteBatch.java

@@ -39,6 +39,14 @@ public class WriteBatch extends AbstractWriteBatch {
     super(newWriteBatch(reserved_bytes));
   }
 
+  /**
+   * Constructs a WriteBatch instance from a serialized representation
+   * as returned by {@link #data()}.
+   */
+  public WriteBatch(final byte[] serialized) {
+    super(newWriteBatch(serialized, serialized.length));
+  }
+
   /**
    * Support for iterating over the contents of a batch.
    *
@@ -51,6 +59,134 @@ public class WriteBatch extends AbstractWriteBatch {
     iterate(nativeHandle_, handler.nativeHandle_);
   }
 
+  /**
+   * Retrieve the serialized version of this batch.
+   *
+   * @return the serialized representation of this write batch.
+   */
+  public byte[] data() {
+    return data(nativeHandle_);
+  }
+
+  /**
+   * Retrieve data size of the batch.
+   *
+   * @return the serialized data size of the batch.
+   */
+  public long getDataSize() {
+    return getDataSize(nativeHandle_);
+  }
+
+  /**
+   * Returns true if PutCF will be called during Iterate.
+   *
+   * Return true if PutCF will be called during Iterate.
+   */
+  public boolean hasPut() {
+    return hasPut(nativeHandle_);
+  }
+
+  /**
+   * Returns true if DeleteCF will be called during Iterate.
+   *
+   * Return true if DeleteCF will be called during Iterate.
+   */
+  public boolean hasDelete() {
+    return hasDelete(nativeHandle_);
+  }
+
+  /**
+   * Returns true if SingleDeleteCF will be called during Iterate.
+   *
+   * Return true if SingleDeleteCF will be called during Iterate.
+   */
+  public boolean hasSingleDelete() {
+    return hasSingleDelete(nativeHandle_);
+  }
+
+  /**
+   * Returns true if DeleteRangeCF will be called during Iterate.
+   *
+   * Return true if DeleteRangeCF will be called during Iterate.
+   */
+  public boolean hasDeleteRange() {
+    return hasDeleteRange(nativeHandle_);
+  }
+
+  /**
+   * Returns true if MergeCF will be called during Iterate.
+   *
+   * Return true if MergeCF will be called during Iterate.
+   */
+  public boolean hasMerge() {
+    return hasMerge(nativeHandle_);
+  }
+
+  /**
+   * Returns true if MarkBeginPrepare will be called during Iterate.
+   *
+   * Return true if MarkBeginPrepare will be called during Iterate.
+   */
+  public boolean hasBeginPrepare() {
+    return hasBeginPrepare(nativeHandle_);
+  }
+
+  /**
+   * Returns true if MarkEndPrepare will be called during Iterate.
+   *
+   * Return true if MarkEndPrepare will be called during Iterate.
+   */
+  public boolean hasEndPrepare() {
+    return hasEndPrepare(nativeHandle_);
+  }
+
+  /**
+   * Returns true if MarkCommit will be called during Iterate.
+   *
+   * Return true if MarkCommit will be called during Iterate.
+   */
+  public boolean hasCommit() {
+    return hasCommit(nativeHandle_);
+  }
+
+  /**
+   * Returns true if MarkRollback will be called during Iterate.
+   *
+   * Return true if MarkRollback will be called during Iterate.
+   */
+  public boolean hasRollback() {
+    return hasRollback(nativeHandle_);
+  }
+
+  @Override
+  public WriteBatch getWriteBatch() {
+    return this;
+  }
+
+  /**
+   * Marks this point in the WriteBatch as the last record to
+   * be inserted into the WAL, provided the WAL is enabled.
+   */
+  public void markWalTerminationPoint() {
+    markWalTerminationPoint(nativeHandle_);
+  }
+
+  /**
+   * Gets the WAL termination point.
+   *
+   * See {@link #markWalTerminationPoint()}
+   *
+   * @return the WAL termination point
+   */
+  public SavePoint getWalTerminationPoint() {
+    return getWalTerminationPoint(nativeHandle_);
+  }
+
+  @Override
+  WriteBatch getWriteBatch(final long handle) {
+    return this;
+  }
+
   /**
    * <p>Private WriteBatch constructor which is used to construct
    * WriteBatch instances from C++ side. As the reference to this
@@ -87,10 +223,14 @@ public class WriteBatch extends AbstractWriteBatch {
   @Override final native void merge(final long handle, final byte[] key,
       final int keyLen, final byte[] value, final int valueLen,
       final long cfHandle);
-  @Override final native void remove(final long handle, final byte[] key,
-      final int keyLen);
-  @Override final native void remove(final long handle, final byte[] key,
-      final int keyLen, final long cfHandle);
+  @Override final native void delete(final long handle, final byte[] key,
+      final int keyLen) throws RocksDBException;
+  @Override final native void delete(final long handle, final byte[] key,
+      final int keyLen, final long cfHandle) throws RocksDBException;
+  @Override final native void singleDelete(final long handle, final byte[] key,
+      final int keyLen) throws RocksDBException;
+  @Override final native void singleDelete(final long handle, final byte[] key,
+      final int keyLen, final long cfHandle) throws RocksDBException;
   @Override
   final native void deleteRange(final long handle, final byte[] beginKey, final int beginKeyLen,
       final byte[] endKey, final int endKeyLen);
@@ -98,15 +238,32 @@ public class WriteBatch extends AbstractWriteBatch {
   final native void deleteRange(final long handle, final byte[] beginKey, final int beginKeyLen,
       final byte[] endKey, final int endKeyLen, final long cfHandle);
   @Override final native void putLogData(final long handle,
-      final byte[] blob, final int blobLen);
+      final byte[] blob, final int blobLen) throws RocksDBException;
   @Override final native void clear0(final long handle);
   @Override final native void setSavePoint0(final long handle);
   @Override final native void rollbackToSavePoint0(final long handle);
+  @Override final native void popSavePoint(final long handle) throws RocksDBException;
+  @Override final native void setMaxBytes(final long nativeHandle,
+    final long maxBytes);
 
   private native static long newWriteBatch(final int reserved_bytes);
+  private native static long newWriteBatch(final byte[] serialized,
+      final int serializedLength);
   private native void iterate(final long handle, final long handlerHandle)
       throws RocksDBException;
-
+  private native byte[] data(final long nativeHandle);
+  private native long getDataSize(final long nativeHandle);
+  private native boolean hasPut(final long nativeHandle);
+  private native boolean hasDelete(final long nativeHandle);
+  private native boolean hasSingleDelete(final long nativeHandle);
+  private native boolean hasDeleteRange(final long nativeHandle);
+  private native boolean hasMerge(final long nativeHandle);
+  private native boolean hasBeginPrepare(final long nativeHandle);
+  private native boolean hasEndPrepare(final long nativeHandle);
+  private native boolean hasCommit(final long nativeHandle);
+  private native boolean hasRollback(final long nativeHandle);
+  private native void markWalTerminationPoint(final long nativeHandle);
+  private native SavePoint getWalTerminationPoint(final long nativeHandle);
 
   /**
    * Handler callback for iterating over the contents of a batch.
@@ -122,15 +279,38 @@ public class WriteBatch extends AbstractWriteBatch {
       return createNewHandler0();
     }
 
-    public abstract void put(byte[] key, byte[] value);
-    public abstract void merge(byte[] key, byte[] value);
-    public abstract void delete(byte[] key);
-    public abstract void deleteRange(byte[] beginKey, byte[] endKey);
-    public abstract void logData(byte[] blob);
+    public abstract void put(final int columnFamilyId, final byte[] key,
+        final byte[] value) throws RocksDBException;
+    public abstract void put(final byte[] key, final byte[] value);
+    public abstract void merge(final int columnFamilyId, final byte[] key,
+        final byte[] value) throws RocksDBException;
+    public abstract void merge(final byte[] key, final byte[] value);
+    public abstract void delete(final int columnFamilyId, final byte[] key)
+        throws RocksDBException;
+    public abstract void delete(final byte[] key);
+    public abstract void singleDelete(final int columnFamilyId,
+        final byte[] key) throws RocksDBException;
+    public abstract void singleDelete(final byte[] key);
+    public abstract void deleteRange(final int columnFamilyId,
+        final byte[] beginKey, final byte[] endKey) throws RocksDBException;
+    public abstract void deleteRange(final byte[] beginKey,
+        final byte[] endKey);
+    public abstract void logData(final byte[] blob);
+    public abstract void putBlobIndex(final int columnFamilyId,
+        final byte[] key, final byte[] value) throws RocksDBException;
+    public abstract void markBeginPrepare() throws RocksDBException;
+    public abstract void markEndPrepare(final byte[] xid)
+        throws RocksDBException;
+    public abstract void markNoop(final boolean emptyBatch)
+        throws RocksDBException;
+    public abstract void markRollback(final byte[] xid)
+        throws RocksDBException;
+    public abstract void markCommit(final byte[] xid)
+        throws RocksDBException;
 
     /**
      * shouldContinue is called by the underlying iterator
-     * WriteBatch::Iterate. If it returns false,
+     * {@link WriteBatch#iterate(Handler)}. If it returns false,
      * iteration is halted. Otherwise, it continues
      * iterating. The default implementation always
      * returns true.
@@ -144,4 +324,57 @@ public class WriteBatch extends AbstractWriteBatch {
 
     private native long createNewHandler0();
   }
+
+  /**
+   * A structure for describing the save point in the Write Batch.
+   */
+  public static class SavePoint {
+    private long size;
+    private long count;
+    private long contentFlags;
+
+    public SavePoint(final long size, final long count,
+        final long contentFlags) {
+      this.size = size;
+      this.count = count;
+      this.contentFlags = contentFlags;
+    }
+
+    public void clear() {
+      this.size = 0;
+      this.count = 0;
+      this.contentFlags = 0;
+    }
+
+    /**
+     * Get the size of the serialized representation.
+     *
+     * @return the size of the serialized representation.
+     */
+    public long getSize() {
+      return size;
+    }
+
+    /**
+     * Get the number of elements.
+     *
+     * @return the number of elements.
+     */
+    public long getCount() {
+      return count;
+    }
+
+    /**
+     * Get the content flags.
+     *
+     * @return the content flags.
+     */
+    public long getContentFlags() {
+      return contentFlags;
+    }
+
+    public boolean isCleared() {
+      return (size | count | contentFlags) == 0;
+    }
+  }
 }

java/src/main/java/org/rocksdb/WriteBatchInterface.java

@@ -24,7 +24,7 @@ public interface WriteBatchInterface {
      * @param key the specified key to be inserted.
      * @param value the value associated with the specified key.
      */
-    void put(byte[] key, byte[] value);
+    void put(byte[] key, byte[] value) throws RocksDBException;
 
     /**
      * <p>Store the mapping "key-&gt;value" within given column
@@ -36,7 +36,7 @@ public interface WriteBatchInterface {
      * @param value the value associated with the specified key.
      */
     void put(ColumnFamilyHandle columnFamilyHandle,
-                    byte[] key, byte[] value);
+                    byte[] key, byte[] value) throws RocksDBException;
 
     /**
      * <p>Merge "value" with the existing value of "key" in the database.
@@ -46,7 +46,7 @@ public interface WriteBatchInterface {
      * @param value the value to be merged with the current value for
      * the specified key.
      */
-    void merge(byte[] key, byte[] value);
+    void merge(byte[] key, byte[] value) throws RocksDBException;
 
     /**
      * <p>Merge "value" with the existing value of "key" in given column family.
@@ -58,14 +58,36 @@ public interface WriteBatchInterface {
      * the specified key.
      */
     void merge(ColumnFamilyHandle columnFamilyHandle,
-                      byte[] key, byte[] value);
+                      byte[] key, byte[] value) throws RocksDBException;
+
+    /**
+     * <p>If the database contains a mapping for "key", erase it.  Else do nothing.</p>
+     *
+     * @param key Key to delete within database
+     *
+     * @deprecated Use {@link #delete(byte[])}
+     */
+    @Deprecated
+    void remove(byte[] key) throws RocksDBException;
+
+    /**
+     * <p>If column family contains a mapping for "key", erase it.  Else do nothing.</p>
+     *
+     * @param columnFamilyHandle {@link ColumnFamilyHandle} instance
+     * @param key Key to delete within database
+     *
+     * @deprecated Use {@link #delete(ColumnFamilyHandle, byte[])}
+     */
+    @Deprecated
+    void remove(ColumnFamilyHandle columnFamilyHandle, byte[] key)
+        throws RocksDBException;
 
     /**
      * <p>If the database contains a mapping for "key", erase it.  Else do nothing.</p>
      *
      * @param key Key to delete within database
      */
-    void remove(byte[] key);
+    void delete(byte[] key) throws RocksDBException;
 
     /**
      * <p>If column family contains a mapping for "key", erase it.  Else do nothing.</p>
@@ -73,7 +95,58 @@ public interface WriteBatchInterface {
      * @param columnFamilyHandle {@link ColumnFamilyHandle} instance
      * @param key Key to delete within database
      */
-    void remove(ColumnFamilyHandle columnFamilyHandle, byte[] key);
+    void delete(ColumnFamilyHandle columnFamilyHandle, byte[] key)
+        throws RocksDBException;
+
+    /**
+     * Remove the database entry for {@code key}. Requires that the key exists
+     * and was not overwritten. It is not an error if the key did not exist
+     * in the database.
+     *
+     * If a key is overwritten (by calling {@link #put(byte[], byte[])} multiple
+     * times), then the result of calling SingleDelete() on this key is undefined.
+     * SingleDelete() only behaves correctly if there has been only one Put()
+     * for this key since the previous call to SingleDelete() for this key.
+     *
+     * This feature is currently an experimental performance optimization
+     * for a very specific workload. It is up to the caller to ensure that
+     * SingleDelete is only used for a key that is not deleted using Delete() or
+     * written using Merge(). Mixing SingleDelete operations with Deletes and
+     * Merges can result in undefined behavior.
+     *
+     * @param key Key to delete within database
+     *
+     * @throws RocksDBException thrown if error happens in underlying
+     *     native library.
+     */
+    @Experimental("Performance optimization for a very specific workload")
+    void singleDelete(final byte[] key) throws RocksDBException;
+
+    /**
+     * Remove the database entry for {@code key}. Requires that the key exists
+     * and was not overwritten. It is not an error if the key did not exist
+     * in the database.
+     *
+     * If a key is overwritten (by calling {@link #put(byte[], byte[])} multiple
+     * times), then the result of calling SingleDelete() on this key is undefined.
+     * SingleDelete() only behaves correctly if there has been only one Put()
+     * for this key since the previous call to SingleDelete() for this key.
+     *
+     * This feature is currently an experimental performance optimization
+     * for a very specific workload. It is up to the caller to ensure that
+     * SingleDelete is only used for a key that is not deleted using Delete() or
+     * written using Merge(). Mixing SingleDelete operations with Deletes and
+     * Merges can result in undefined behavior.
+     *
+     * @param columnFamilyHandle The column family to delete the key from
+     * @param key Key to delete within database
+     *
+     * @throws RocksDBException thrown if error happens in underlying
+     *     native library.
+     */
+    @Experimental("Performance optimization for a very specific workload")
+    void singleDelete(final ColumnFamilyHandle columnFamilyHandle,
+            final byte[] key) throws RocksDBException;
 
     /**
      * Removes the database entries in the range ["beginKey", "endKey"), i.e.,
@@ -89,7 +162,7 @@ public interface WriteBatchInterface {
      * @param endKey
      *          Last key to delete within database (excluded)
      */
-    void deleteRange(byte[] beginKey, byte[] endKey);
+    void deleteRange(byte[] beginKey, byte[] endKey) throws RocksDBException;
 
     /**
      * Removes the database entries in the range ["beginKey", "endKey"), i.e.,
@@ -106,7 +179,8 @@ public interface WriteBatchInterface {
      * @param endKey
      *          Last key to delete within database (excluded)
      */
-    void deleteRange(ColumnFamilyHandle columnFamilyHandle, byte[] beginKey, byte[] endKey);
+    void deleteRange(ColumnFamilyHandle columnFamilyHandle, byte[] beginKey,
+            byte[] endKey) throws RocksDBException;
 
     /**
      * Append a blob of arbitrary size to the records in this batch. The blob will
@@ -122,7 +196,7 @@ public interface WriteBatchInterface {
      *
      * @param blob binary object to be inserted
      */
-    void putLogData(byte[] blob);
+    void putLogData(byte[] blob) throws RocksDBException;
 
     /**
      * Clear all updates buffered in this batch
@@ -143,4 +217,30 @@ public interface WriteBatchInterface {
      * @throws RocksDBException if there is no previous call to SetSavePoint()
      */
     void rollbackToSavePoint() throws RocksDBException;
+
+    /**
+     * Pop the most recent save point.
+     *
+     * That is to say that it removes the last save point,
+     * which was set by {@link #setSavePoint()}.
+     *
+     * @throws RocksDBException If there is no previous call to
+     *     {@link #setSavePoint()}, an exception with
+     *     {@link Status.Code#NotFound} will be thrown.
+     */
+    void popSavePoint() throws RocksDBException;
+
+    /**
+     * Set the maximum size of the write batch.
+     *
+     * @param maxBytes the maximum size in bytes.
+     */
+    void setMaxBytes(long maxBytes);
+
+    /**
+     * Get the underlying Write Batch.
+     *
+     * @return the underlying WriteBatch.
+     */
+    WriteBatch getWriteBatch();
 }

java/src/main/java/org/rocksdb/WriteBatchWithIndex.java

@@ -61,7 +61,20 @@ public class WriteBatchWithIndex extends AbstractWriteBatch {
           fallbackIndexComparator, final int reservedBytes,
       final boolean overwriteKey) {
     super(newWriteBatchWithIndex(fallbackIndexComparator.nativeHandle_,
-        fallbackIndexComparator instanceof DirectComparator, reservedBytes, overwriteKey));
+        fallbackIndexComparator.getComparatorType().getValue(), reservedBytes,
+        overwriteKey));
+  }
+
+  /**
+   * <p>Private WriteBatchWithIndex constructor which is used to construct
+   * WriteBatchWithIndex instances from C++ side. As the reference to this
+   * object is also managed from C++ side the handle will be disowned.</p>
+   *
+   * @param nativeHandle address of native instance.
+   */
+  WriteBatchWithIndex(final long nativeHandle) {
+    super(nativeHandle);
+    disOwnNativeHandle();
   }
 
   /**
@@ -244,10 +257,14 @@ public class WriteBatchWithIndex extends AbstractWriteBatch {
   @Override final native void merge(final long handle, final byte[] key,
       final int keyLen, final byte[] value, final int valueLen,
       final long cfHandle);
-  @Override final native void remove(final long handle, final byte[] key,
-      final int keyLen);
-  @Override final native void remove(final long handle, final byte[] key,
-      final int keyLen, final long cfHandle);
+  @Override final native void delete(final long handle, final byte[] key,
+      final int keyLen) throws RocksDBException;
+  @Override final native void delete(final long handle, final byte[] key,
+      final int keyLen, final long cfHandle) throws RocksDBException;
+  @Override final native void singleDelete(final long handle, final byte[] key,
+      final int keyLen) throws RocksDBException;
+  @Override final native void singleDelete(final long handle, final byte[] key,
+      final int keyLen, final long cfHandle) throws RocksDBException;
   @Override
   final native void deleteRange(final long handle, final byte[] beginKey, final int beginKeyLen,
       final byte[] endKey, final int endKeyLen);
@@ -255,15 +272,20 @@ public class WriteBatchWithIndex extends AbstractWriteBatch {
   final native void deleteRange(final long handle, final byte[] beginKey, final int beginKeyLen,
       final byte[] endKey, final int endKeyLen, final long cfHandle);
   @Override final native void putLogData(final long handle, final byte[] blob,
-      final int blobLen);
+      final int blobLen) throws RocksDBException;
   @Override final native void clear0(final long handle);
   @Override final native void setSavePoint0(final long handle);
   @Override final native void rollbackToSavePoint0(final long handle);
+  @Override final native void popSavePoint(final long handle) throws RocksDBException;
+  @Override final native void setMaxBytes(final long nativeHandle,
+      final long maxBytes);
+  @Override final native WriteBatch getWriteBatch(final long handle);
 
   private native static long newWriteBatchWithIndex();
   private native static long newWriteBatchWithIndex(final boolean overwriteKey);
   private native static long newWriteBatchWithIndex(
-      final long fallbackIndexComparatorHandle, final boolean isDirect, final int reservedBytes,
+      final long fallbackIndexComparatorHandle,
+      final byte comparatorType, final int reservedBytes,
       final boolean overwriteKey);
   private native long iterator0(final long handle);
   private native long iterator1(final long handle, final long cfHandle);

java/src/main/java/org/rocksdb/WriteOptions.java

@@ -20,6 +20,12 @@ public class WriteOptions extends RocksObject {
 
   }
 
+  // TODO(AR) consider ownership
+  WriteOptions(final long nativeHandle) {
+    super(nativeHandle);
+    disOwnNativeHandle();
+  }
+
   /**
    * If true, the write will be flushed from the operating system
    * buffer cache (by calling WritableFile::Sync()) before the write

java/src/main/java/org/rocksdb/util/Environment.java

@@ -12,6 +12,10 @@ public class Environment {
     return (OS.contains("win"));
   }
 
+  public static boolean isFreeBSD() {
+    return (OS.contains("freebsd"));
+  }
+
   public static boolean isMac() {
     return (OS.contains("mac"));
   }
@@ -54,6 +58,8 @@ public class Environment {
       }
     } else if (isMac()) {
       return String.format("%sjni-osx", name);
+    } else if (isFreeBSD()) {
+      return String.format("%sjni-freebsd%s", name, is64Bit() ? "64" : "32");
     } else if (isAix() && is64Bit()) {
       return String.format("%sjni-aix64", name);
     } else if (isSolaris()) {
@@ -71,7 +77,7 @@ public class Environment {
   }
 
   private static String appendLibOsSuffix(final String libraryFileName, final boolean shared) {
-    if (isUnix() || isAix() || isSolaris()) {
+    if (isUnix() || isAix() || isSolaris() || isFreeBSD()) {
       return libraryFileName + ".so";
     } else if (isMac()) {
       return libraryFileName + (shared ? ".dylib" : ".jnilib");

java/src/test/java/org/rocksdb/AbstractTransactionTest.java

@@ -0,0 +1,903 @@
+// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+package org.rocksdb;
+
+import org.junit.Rule;
+import org.junit.Test;
+import org.junit.rules.TemporaryFolder;
+
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.List;
+import java.util.Random;
+
+import static java.nio.charset.StandardCharsets.UTF_8;
+import static org.assertj.core.api.Assertions.assertThat;
+import static org.assertj.core.api.Assertions.fail;
+
+/**
+ * Base class of {@link TransactionTest} and {@link OptimisticTransactionTest}
+ */
+public abstract class AbstractTransactionTest {
+
+  protected final static byte[] TXN_TEST_COLUMN_FAMILY = "txn_test_cf"
+      .getBytes();
+
+  protected static final Random rand = PlatformRandomHelper.
+      getPlatformSpecificRandomFactory();
+
+  @Rule
+  public TemporaryFolder dbFolder = new TemporaryFolder();
+
+  public abstract DBContainer startDb()
+      throws RocksDBException;
+
+  @Test
+  public void setSnapshot() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.setSnapshot();
+    }
+  }
+
+  @Test
+  public void setSnapshotOnNextOperation() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.setSnapshotOnNextOperation();
+      txn.put("key1".getBytes(), "value1".getBytes());
+    }
+  }
+
+  @Test
+  public void setSnapshotOnNextOperation_transactionNotifier() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+
+      try(final TestTransactionNotifier notifier = new TestTransactionNotifier()) {
+        txn.setSnapshotOnNextOperation(notifier);
+        txn.put("key1".getBytes(), "value1".getBytes());
+
+        txn.setSnapshotOnNextOperation(notifier);
+        txn.put("key2".getBytes(), "value2".getBytes());
+
+        assertThat(notifier.getCreatedSnapshots().size()).isEqualTo(2);
+      }
+    }
+  }
+
+  @Test
+  public void getSnapshot() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.setSnapshot();
+      final Snapshot snapshot = txn.getSnapshot();
+      assertThat(snapshot.isOwningHandle()).isFalse();
+    }
+  }
+
+  @Test
+  public void getSnapshot_null() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final Snapshot snapshot = txn.getSnapshot();
+      assertThat(snapshot).isNull();
+    }
+  }
+
+  @Test
+  public void clearSnapshot() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.setSnapshot();
+      txn.clearSnapshot();
+    }
+  }
+
+  @Test
+  public void clearSnapshot_none() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.clearSnapshot();
+    }
+  }
+
+  @Test
+  public void commit() throws RocksDBException {
+    final byte k1[] = "rollback-key1".getBytes(UTF_8);
+    final byte v1[] = "rollback-value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb()) {
+      try(final Transaction txn = dbContainer.beginTransaction()) {
+        txn.put(k1, v1);
+        txn.commit();
+      }
+
+      try(final ReadOptions readOptions = new ReadOptions();
+          final Transaction txn2 = dbContainer.beginTransaction()) {
+        assertThat(txn2.get(readOptions, k1)).isEqualTo(v1);
+      }
+    }
+  }
+
+  @Test
+  public void rollback() throws RocksDBException {
+    final byte k1[] = "rollback-key1".getBytes(UTF_8);
+    final byte v1[] = "rollback-value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb()) {
+      try(final Transaction txn = dbContainer.beginTransaction()) {
+        txn.put(k1, v1);
+        txn.rollback();
+      }
+
+      try(final ReadOptions readOptions = new ReadOptions();
+          final Transaction txn2 = dbContainer.beginTransaction()) {
+        assertThat(txn2.get(readOptions, k1)).isNull();
+      }
+    }
+  }
+
+  @Test
+  public void savePoint() throws RocksDBException {
+    final byte k1[] = "savePoint-key1".getBytes(UTF_8);
+    final byte v1[] = "savePoint-value1".getBytes(UTF_8);
+    final byte k2[] = "savePoint-key2".getBytes(UTF_8);
+    final byte v2[] = "savePoint-value2".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions()) {
+
+
+      try(final Transaction txn = dbContainer.beginTransaction()) {
+        txn.put(k1, v1);
+
+        assertThat(txn.get(readOptions, k1)).isEqualTo(v1);
+
+        txn.setSavePoint();
+
+        txn.put(k2, v2);
+
+        assertThat(txn.get(readOptions, k1)).isEqualTo(v1);
+        assertThat(txn.get(readOptions, k2)).isEqualTo(v2);
+
+        txn.rollbackToSavePoint();
+
+        assertThat(txn.get(readOptions, k1)).isEqualTo(v1);
+        assertThat(txn.get(readOptions, k2)).isNull();
+
+        txn.commit();
+      }
+
+      try(final Transaction txn2 = dbContainer.beginTransaction()) {
+        assertThat(txn2.get(readOptions, k1)).isEqualTo(v1);
+        assertThat(txn2.get(readOptions, k2)).isNull();
+      }
+    }
+  }
+
+  @Test
+  public void getPut_cf() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      assertThat(txn.get(testCf, readOptions, k1)).isNull();
+      txn.put(testCf, k1, v1);
+      assertThat(txn.get(testCf, readOptions, k1)).isEqualTo(v1);
+    }
+  }
+
+  @Test
+  public void getPut() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      assertThat(txn.get(readOptions, k1)).isNull();
+      txn.put(k1, v1);
+      assertThat(txn.get(readOptions, k1)).isEqualTo(v1);
+    }
+  }
+
+  @Test
+  public void multiGetPut_cf() throws RocksDBException {
+    final byte keys[][] = new byte[][] {
+        "key1".getBytes(UTF_8),
+        "key2".getBytes(UTF_8)};
+    final byte values[][] = new byte[][] {
+        "value1".getBytes(UTF_8),
+        "value2".getBytes(UTF_8)};
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      final List<ColumnFamilyHandle> cfList = Arrays.asList(testCf, testCf);
+
+      assertThat(txn.multiGet(readOptions, cfList, keys)).isEqualTo(new byte[][] { null, null });
+
+      txn.put(testCf, keys[0], values[0]);
+      txn.put(testCf, keys[1], values[1]);
+      assertThat(txn.multiGet(readOptions, cfList, keys)).isEqualTo(values);
+    }
+  }
+
+  @Test
+  public void multiGetPut() throws RocksDBException {
+    final byte keys[][] = new byte[][] {
+        "key1".getBytes(UTF_8),
+        "key2".getBytes(UTF_8)};
+    final byte values[][] = new byte[][] {
+        "value1".getBytes(UTF_8),
+        "value2".getBytes(UTF_8)};
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+
+      assertThat(txn.multiGet(readOptions, keys)).isEqualTo(new byte[][] { null, null });
+
+      txn.put(keys[0], values[0]);
+      txn.put(keys[1], values[1]);
+      assertThat(txn.multiGet(readOptions, keys)).isEqualTo(values);
+    }
+  }
+
+  @Test
+  public void getForUpdate_cf() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      assertThat(txn.getForUpdate(readOptions, testCf, k1, true)).isNull();
+      txn.put(testCf, k1, v1);
+      assertThat(txn.getForUpdate(readOptions, testCf, k1, true)).isEqualTo(v1);
+    }
+  }
+
+  @Test
+  public void getForUpdate() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      assertThat(txn.getForUpdate(readOptions, k1, true)).isNull();
+      txn.put(k1, v1);
+      assertThat(txn.getForUpdate(readOptions, k1, true)).isEqualTo(v1);
+    }
+  }
+
+  @Test
+  public void multiGetForUpdate_cf() throws RocksDBException {
+    final byte keys[][] = new byte[][] {
+        "key1".getBytes(UTF_8),
+        "key2".getBytes(UTF_8)};
+    final byte values[][] = new byte[][] {
+        "value1".getBytes(UTF_8),
+        "value2".getBytes(UTF_8)};
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      final List<ColumnFamilyHandle> cfList = Arrays.asList(testCf, testCf);
+
+      assertThat(txn.multiGetForUpdate(readOptions, cfList, keys))
+          .isEqualTo(new byte[][] { null, null });
+
+      txn.put(testCf, keys[0], values[0]);
+      txn.put(testCf, keys[1], values[1]);
+      assertThat(txn.multiGetForUpdate(readOptions, cfList, keys))
+          .isEqualTo(values);
+    }
+  }
+
+  @Test
+  public void multiGetForUpdate() throws RocksDBException {
+    final byte keys[][] = new byte[][]{
+        "key1".getBytes(UTF_8),
+        "key2".getBytes(UTF_8)};
+    final byte values[][] = new byte[][]{
+        "value1".getBytes(UTF_8),
+        "value2".getBytes(UTF_8)};
+
+    try (final DBContainer dbContainer = startDb();
+         final ReadOptions readOptions = new ReadOptions();
+         final Transaction txn = dbContainer.beginTransaction()) {
+      assertThat(txn.multiGetForUpdate(readOptions, keys)).isEqualTo(new byte[][]{null, null});
+
+      txn.put(keys[0], values[0]);
+      txn.put(keys[1], values[1]);
+      assertThat(txn.multiGetForUpdate(readOptions, keys)).isEqualTo(values);
+    }
+  }
+
+  @Test
+  public void getIterator() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+
+      final byte[] k1 = "key1".getBytes(UTF_8);
+      final byte[] v1 = "value1".getBytes(UTF_8);
+
+      txn.put(k1, v1);
+
+      try(final RocksIterator iterator = txn.getIterator(readOptions)) {
+        iterator.seek(k1);
+        assertThat(iterator.isValid()).isTrue();
+        assertThat(iterator.key()).isEqualTo(k1);
+        assertThat(iterator.value()).isEqualTo(v1);
+      }
+    }
+  }
+
+  @Test
+  public void getIterator_cf() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+
+      final byte[] k1 = "key1".getBytes(UTF_8);
+      final byte[] v1 = "value1".getBytes(UTF_8);
+
+      txn.put(testCf, k1, v1);
+
+      try(final RocksIterator iterator = txn.getIterator(readOptions, testCf)) {
+        iterator.seek(k1);
+        assertThat(iterator.isValid()).isTrue();
+        assertThat(iterator.key()).isEqualTo(k1);
+        assertThat(iterator.value()).isEqualTo(v1);
+      }
+    }
+  }
+
+  @Test
+  public void merge_cf() throws RocksDBException {
+    final byte[] k1 = "key1".getBytes(UTF_8);
+    final byte[] v1 = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      txn.merge(testCf, k1, v1);
+    }
+  }
+
+  @Test
+  public void merge() throws RocksDBException {
+    final byte[] k1 = "key1".getBytes(UTF_8);
+    final byte[] v1 = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.merge(k1, v1);
+    }
+  }
+
+
+  @Test
+  public void delete_cf() throws RocksDBException {
+    final byte[] k1 = "key1".getBytes(UTF_8);
+    final byte[] v1 = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      txn.put(testCf, k1, v1);
+      assertThat(txn.get(testCf, readOptions, k1)).isEqualTo(v1);
+
+      txn.delete(testCf, k1);
+      assertThat(txn.get(testCf, readOptions, k1)).isNull();
+    }
+  }
+
+  @Test
+  public void delete() throws RocksDBException {
+    final byte[] k1 = "key1".getBytes(UTF_8);
+    final byte[] v1 = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.put(k1, v1);
+      assertThat(txn.get(readOptions, k1)).isEqualTo(v1);
+
+      txn.delete(k1);
+      assertThat(txn.get(readOptions, k1)).isNull();
+    }
+  }
+
+  @Test
+  public void delete_parts_cf() throws RocksDBException {
+    final byte keyParts[][] = new byte[][] {
+        "ke".getBytes(UTF_8),
+        "y1".getBytes(UTF_8)};
+    final byte valueParts[][] = new byte[][] {
+        "val".getBytes(UTF_8),
+        "ue1".getBytes(UTF_8)};
+    final byte[] key = concat(keyParts);
+    final byte[] value = concat(valueParts);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      txn.put(testCf, keyParts, valueParts);
+      assertThat(txn.get(testCf, readOptions, key)).isEqualTo(value);
+
+      txn.delete(testCf, keyParts);
+
+      assertThat(txn.get(testCf, readOptions, key))
+          .isNull();
+    }
+  }
+
+  @Test
+  public void delete_parts() throws RocksDBException {
+    final byte keyParts[][] = new byte[][] {
+        "ke".getBytes(UTF_8),
+        "y1".getBytes(UTF_8)};
+    final byte valueParts[][] = new byte[][] {
+        "val".getBytes(UTF_8),
+        "ue1".getBytes(UTF_8)};
+    final byte[] key = concat(keyParts);
+    final byte[] value = concat(valueParts);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+
+      txn.put(keyParts, valueParts);
+
+      assertThat(txn.get(readOptions, key)).isEqualTo(value);
+
+      txn.delete(keyParts);
+
+      assertThat(txn.get(readOptions, key)).isNull();
+    }
+  }
+
+  @Test
+  public void getPutUntracked_cf() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      assertThat(txn.get(testCf, readOptions, k1)).isNull();
+      txn.putUntracked(testCf, k1, v1);
+      assertThat(txn.get(testCf, readOptions, k1)).isEqualTo(v1);
+    }
+  }
+
+  @Test
+  public void getPutUntracked() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      assertThat(txn.get(readOptions, k1)).isNull();
+      txn.putUntracked(k1, v1);
+      assertThat(txn.get(readOptions, k1)).isEqualTo(v1);
+    }
+  }
+
+  @Test
+  public void multiGetPutUntracked_cf() throws RocksDBException {
+    final byte keys[][] = new byte[][] {
+        "key1".getBytes(UTF_8),
+        "key2".getBytes(UTF_8)};
+    final byte values[][] = new byte[][] {
+        "value1".getBytes(UTF_8),
+        "value2".getBytes(UTF_8)};
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+
+      final List<ColumnFamilyHandle> cfList = Arrays.asList(testCf, testCf);
+
+      assertThat(txn.multiGet(readOptions, cfList, keys)).isEqualTo(new byte[][] { null, null });
+      txn.putUntracked(testCf, keys[0], values[0]);
+      txn.putUntracked(testCf, keys[1], values[1]);
+      assertThat(txn.multiGet(readOptions, cfList, keys)).isEqualTo(values);
+    }
+  }
+
+  @Test
+  public void multiGetPutUntracked() throws RocksDBException {
+    final byte keys[][] = new byte[][] {
+        "key1".getBytes(UTF_8),
+        "key2".getBytes(UTF_8)};
+    final byte values[][] = new byte[][] {
+        "value1".getBytes(UTF_8),
+        "value2".getBytes(UTF_8)};
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+
+      assertThat(txn.multiGet(readOptions, keys)).isEqualTo(new byte[][] { null, null });
+      txn.putUntracked(keys[0], values[0]);
+      txn.putUntracked(keys[1], values[1]);
+      assertThat(txn.multiGet(readOptions, keys)).isEqualTo(values);
+    }
+  }
+
+  @Test
+  public void mergeUntracked_cf() throws RocksDBException {
+    final byte[] k1 = "key1".getBytes(UTF_8);
+    final byte[] v1 = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      txn.mergeUntracked(testCf, k1, v1);
+    }
+  }
+
+  @Test
+  public void mergeUntracked() throws RocksDBException {
+    final byte[] k1 = "key1".getBytes(UTF_8);
+    final byte[] v1 = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.mergeUntracked(k1, v1);
+    }
+  }
+
+  @Test
+  public void deleteUntracked_cf() throws RocksDBException {
+    final byte[] k1 = "key1".getBytes(UTF_8);
+    final byte[] v1 = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      txn.put(testCf, k1, v1);
+      assertThat(txn.get(testCf, readOptions, k1)).isEqualTo(v1);
+
+      txn.deleteUntracked(testCf, k1);
+      assertThat(txn.get(testCf, readOptions, k1)).isNull();
+    }
+  }
+
+  @Test
+  public void deleteUntracked() throws RocksDBException {
+    final byte[] k1 = "key1".getBytes(UTF_8);
+    final byte[] v1 = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.put(k1, v1);
+      assertThat(txn.get(readOptions, k1)).isEqualTo(v1);
+
+      txn.deleteUntracked(k1);
+      assertThat(txn.get(readOptions, k1)).isNull();
+    }
+  }
+
+  @Test
+  public void deleteUntracked_parts_cf() throws RocksDBException {
+    final byte keyParts[][] = new byte[][] {
+        "ke".getBytes(UTF_8),
+        "y1".getBytes(UTF_8)};
+    final byte valueParts[][] = new byte[][] {
+        "val".getBytes(UTF_8),
+        "ue1".getBytes(UTF_8)};
+    final byte[] key = concat(keyParts);
+    final byte[] value = concat(valueParts);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      txn.put(testCf, keyParts, valueParts);
+      assertThat(txn.get(testCf, readOptions, key)).isEqualTo(value);
+
+      txn.deleteUntracked(testCf, keyParts);
+      assertThat(txn.get(testCf, readOptions, key)).isNull();
+    }
+  }
+
+  @Test
+  public void deleteUntracked_parts() throws RocksDBException {
+    final byte keyParts[][] = new byte[][] {
+        "ke".getBytes(UTF_8),
+        "y1".getBytes(UTF_8)};
+    final byte valueParts[][] = new byte[][] {
+        "val".getBytes(UTF_8),
+        "ue1".getBytes(UTF_8)};
+    final byte[] key = concat(keyParts);
+    final byte[] value = concat(valueParts);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.put(keyParts, valueParts);
+      assertThat(txn.get(readOptions, key)).isEqualTo(value);
+
+      txn.deleteUntracked(keyParts);
+      assertThat(txn.get(readOptions, key)).isNull();
+    }
+  }
+
+  @Test
+  public void putLogData() throws RocksDBException {
+    final byte[] blob = "blobby".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.putLogData(blob);
+    }
+  }
+
+  @Test
+  public void enabledDisableIndexing() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.disableIndexing();
+      txn.enableIndexing();
+      txn.disableIndexing();
+      txn.enableIndexing();
+    }
+  }
+
+  @Test
+  public void numKeys() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    final byte k2[] = "key2".getBytes(UTF_8);
+    final byte v2[] = "value2".getBytes(UTF_8);
+    final byte k3[] = "key3".getBytes(UTF_8);
+    final byte v3[] = "value3".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      txn.put(k1, v1);
+      txn.put(testCf, k2, v2);
+      txn.merge(k3, v3);
+      txn.delete(testCf, k2);
+
+      assertThat(txn.getNumKeys()).isEqualTo(3);
+      assertThat(txn.getNumPuts()).isEqualTo(2);
+      assertThat(txn.getNumMerges()).isEqualTo(1);
+      assertThat(txn.getNumDeletes()).isEqualTo(1);
+    }
+  }
+
+  @Test
+  public void elapsedTime() throws RocksDBException, InterruptedException {
+    final long preStartTxnTime = System.currentTimeMillis();
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      Thread.sleep(1);
+
+      final long txnElapsedTime = txn.getElapsedTime();
+      assertThat(txnElapsedTime).isLessThan(System.currentTimeMillis()
+          - preStartTxnTime);
+      assertThat(txnElapsedTime).isGreaterThan(0);
+    }
+  }
+
+  @Test
+  public void getWriteBatch() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+
+      txn.put(k1, v1);
+
+      final WriteBatchWithIndex writeBatch = txn.getWriteBatch();
+      assertThat(writeBatch).isNotNull();
+      assertThat(writeBatch.isOwningHandle()).isFalse();
+      assertThat(writeBatch.count()).isEqualTo(1);
+    }
+  }
+
+  @Test
+  public void setLockTimeout() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      txn.setLockTimeout(1000);
+    }
+  }
+
+  @Test
+  public void writeOptions() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final WriteOptions writeOptions = new WriteOptions()
+        .setDisableWAL(true)
+        .setSync(true);
+        final Transaction txn = dbContainer.beginTransaction(writeOptions)) {
+
+      txn.put(k1, v1);
+
+      WriteOptions txnWriteOptions = txn.getWriteOptions();
+      assertThat(txnWriteOptions).isNotNull();
+      assertThat(txnWriteOptions.isOwningHandle()).isFalse();
+      assertThat(txnWriteOptions).isNotSameAs(writeOptions);
+      assertThat(txnWriteOptions.disableWAL()).isTrue();
+      assertThat(txnWriteOptions.sync()).isTrue();
+
+      txn.setWriteOptions(txnWriteOptions.setSync(false));
+      txnWriteOptions = txn.getWriteOptions();
+      assertThat(txnWriteOptions).isNotNull();
+      assertThat(txnWriteOptions.isOwningHandle()).isFalse();
+      assertThat(txnWriteOptions).isNotSameAs(writeOptions);
+      assertThat(txnWriteOptions.disableWAL()).isTrue();
+      assertThat(txnWriteOptions.sync()).isFalse();
+    }
+  }
+
+  @Test
+  public void undoGetForUpdate_cf() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      final ColumnFamilyHandle testCf = dbContainer.getTestColumnFamily();
+      assertThat(txn.getForUpdate(readOptions, testCf, k1, true)).isNull();
+      txn.put(testCf, k1, v1);
+      assertThat(txn.getForUpdate(readOptions, testCf, k1, true)).isEqualTo(v1);
+      txn.undoGetForUpdate(testCf, k1);
+    }
+  }
+
+  @Test
+  public void undoGetForUpdate() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      assertThat(txn.getForUpdate(readOptions, k1, true)).isNull();
+      txn.put(k1, v1);
+      assertThat(txn.getForUpdate(readOptions, k1, true)).isEqualTo(v1);
+      txn.undoGetForUpdate(k1);
+    }
+  }
+
+  @Test
+  public void rebuildFromWriteBatch() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+    final byte k2[] = "key2".getBytes(UTF_8);
+    final byte v2[] = "value2".getBytes(UTF_8);
+    final byte k3[] = "key3".getBytes(UTF_8);
+    final byte v3[] = "value3".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final ReadOptions readOptions = new ReadOptions();
+        final Transaction txn = dbContainer.beginTransaction()) {
+
+      txn.put(k1, v1);
+
+      assertThat(txn.get(readOptions, k1)).isEqualTo(v1);
+      assertThat(txn.getNumKeys()).isEqualTo(1);
+
+      try(final WriteBatch writeBatch = new WriteBatch()) {
+        writeBatch.put(k2, v2);
+        writeBatch.put(k3, v3);
+        txn.rebuildFromWriteBatch(writeBatch);
+
+        assertThat(txn.get(readOptions, k1)).isEqualTo(v1);
+        assertThat(txn.get(readOptions, k2)).isEqualTo(v2);
+        assertThat(txn.get(readOptions, k3)).isEqualTo(v3);
+        assertThat(txn.getNumKeys()).isEqualTo(3);
+      }
+    }
+  }
+
+  @Test
+  public void getCommitTimeWriteBatch() throws RocksDBException {
+    final byte k1[] = "key1".getBytes(UTF_8);
+    final byte v1[] = "value1".getBytes(UTF_8);
+
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+
+      txn.put(k1, v1);
+      final WriteBatch writeBatch = txn.getCommitTimeWriteBatch();
+
+      assertThat(writeBatch).isNotNull();
+      assertThat(writeBatch.isOwningHandle()).isFalse();
+      assertThat(writeBatch.count()).isEqualTo(0);
+    }
+  }
+
+  @Test
+  public void logNumber() throws RocksDBException {
+    try(final DBContainer dbContainer = startDb();
+        final Transaction txn = dbContainer.beginTransaction()) {
+      assertThat(txn.getLogNumber()).isEqualTo(0);
+      final long logNumber = rand.nextLong();
+      txn.setLogNumber(logNumber);
+      assertThat(txn.getLogNumber()).isEqualTo(logNumber);
+    }
+  }
+
+  private static byte[] concat(final byte[][] bufs) {
+    int resultLength = 0;
+    for(final byte[] buf : bufs) {
+      resultLength += buf.length;
+    }
+
+    final byte[] result = new byte[resultLength];
+    int resultOffset = 0;
+    for(final byte[] buf : bufs) {
+      final int srcLength = buf.length;
+      System.arraycopy(buf, 0, result, resultOffset, srcLength);
+      resultOffset += srcLength;
+    }
+
+    return result;
+  }
+
+  private static class TestTransactionNotifier
+      extends AbstractTransactionNotifier {
+    private final List<Snapshot> createdSnapshots = new ArrayList<>();
+
+    @Override
+    public void snapshotCreated(final Snapshot newSnapshot) {
+      createdSnapshots.add(newSnapshot);
+    }
+
+    public List<Snapshot> getCreatedSnapshots() {
+      return createdSnapshots;
+    }
+  }
+
+  protected static abstract class DBContainer
+      implements AutoCloseable {
+    protected final WriteOptions writeOptions;
+    protected final List<ColumnFamilyHandle> columnFamilyHandles;
+    protected final ColumnFamilyOptions columnFamilyOptions;
+    protected final DBOptions options;
+
+    public DBContainer(final WriteOptions writeOptions,
+        final List<ColumnFamilyHandle> columnFamilyHandles,
+        final ColumnFamilyOptions columnFamilyOptions,
+        final DBOptions options) {
+      this.writeOptions = writeOptions;
+      this.columnFamilyHandles = columnFamilyHandles;
+      this.columnFamilyOptions = columnFamilyOptions;
+      this.options = options;
+    }
+
+    public abstract Transaction beginTransaction();
+
+    public abstract Transaction beginTransaction(
+        final WriteOptions writeOptions);
+
+    public ColumnFamilyHandle getTestColumnFamily() {
+      return columnFamilyHandles.get(1);
+    }
+
+    @Override
+    public abstract void close();
+  }
+}

java/src/test/java/org/rocksdb/ColumnFamilyTest.java

@@ -12,6 +12,7 @@ import org.junit.Rule;
 import org.junit.Test;
 import org.junit.rules.TemporaryFolder;
 
+import static java.nio.charset.StandardCharsets.UTF_8;
 import static org.assertj.core.api.Assertions.assertThat;
 
 public class ColumnFamilyTest {
@@ -23,6 +24,31 @@ public class ColumnFamilyTest {
   @Rule
   public TemporaryFolder dbFolder = new TemporaryFolder();
 
+  @Test
+  public void columnFamilyDescriptorName() throws RocksDBException {
+    final byte[] cfName = "some_name".getBytes(UTF_8);
+
+    try(final ColumnFamilyOptions cfOptions = new ColumnFamilyOptions()) {
+      final ColumnFamilyDescriptor cfDescriptor =
+              new ColumnFamilyDescriptor(cfName, cfOptions);
+      assertThat(cfDescriptor.getName()).isEqualTo(cfName);
+    }
+  }
+
+  @Test
+  public void columnFamilyDescriptorOptions() throws RocksDBException {
+    final byte[] cfName = "some_name".getBytes(UTF_8);
+
+    try(final ColumnFamilyOptions cfOptions = new ColumnFamilyOptions()
+            .setCompressionType(CompressionType.BZLIB2_COMPRESSION)) {
+      final ColumnFamilyDescriptor cfDescriptor =
+          new ColumnFamilyDescriptor(cfName, cfOptions);
+
+        assertThat(cfDescriptor.getOptions().compressionType())
+            .isEqualTo(CompressionType.BZLIB2_COMPRESSION);
+    }
+  }
+
   @Test
   public void listColumnFamilies() throws RocksDBException {
     try (final Options options = new Options().setCreateIfMissing(true);
@@ -47,6 +73,9 @@ public class ColumnFamilyTest {
       try {
         assertThat(cfh).isNotNull();
 
+        assertThat(cfh.getName()).isEqualTo("default".getBytes(UTF_8));
+        assertThat(cfh.getID()).isEqualTo(0);
+
         final byte[] key = "key".getBytes();
         final byte[] value = "value".getBytes();
 
@@ -64,15 +93,25 @@ public class ColumnFamilyTest {
 
   @Test
   public void createColumnFamily() throws RocksDBException {
+    final byte[] cfName = "new_cf".getBytes(UTF_8);
+    final ColumnFamilyDescriptor cfDescriptor = new ColumnFamilyDescriptor(cfName,
+            new ColumnFamilyOptions());
+
     try (final Options options = new Options().setCreateIfMissing(true);
          final RocksDB db = RocksDB.open(options,
-             dbFolder.getRoot().getAbsolutePath())) {
-      final ColumnFamilyHandle columnFamilyHandle = db.createColumnFamily(
-          new ColumnFamilyDescriptor("new_cf".getBytes(),
-              new ColumnFamilyOptions()));
+                 dbFolder.getRoot().getAbsolutePath())) {
+
+      final ColumnFamilyHandle columnFamilyHandle = db.createColumnFamily(cfDescriptor);
+
       try {
+        assertThat(columnFamilyHandle.getName()).isEqualTo(cfName);
+        assertThat(columnFamilyHandle.getID()).isEqualTo(1);
+
+        final ColumnFamilyDescriptor latestDescriptor = columnFamilyHandle.getDescriptor();
+        assertThat(latestDescriptor.getName()).isEqualTo(cfName);
+
         final List<byte[]> columnFamilyNames = RocksDB.listColumnFamilies(
-            options, dbFolder.getRoot().getAbsolutePath());
+                options, dbFolder.getRoot().getAbsolutePath());
         assertThat(columnFamilyNames).isNotNull();
         assertThat(columnFamilyNames.size()).isGreaterThan(0);
         assertThat(columnFamilyNames.size()).isEqualTo(2);