PARQUET-267: Detach thirdparty code from build configuration.

.travis.yml

@@ -20,8 +20,7 @@ addons:
     - g++-4.9
     - cmake
     - valgrind
-    - libboost-dev
-    #- libsnappy-dev currently handled by thirdparty scipts.
+    - libboost-dev                 #needed for thrift cpp compilation
     - libboost-program-options-dev #needed for thrift cpp compilation
     - libboost-test-dev            #needed for thrift cpp compilation
     - libssl-dev                   #needed for thrift cpp compilation
@@ -31,33 +30,24 @@ addons:
     - pkg-config                   #needed for thrift cpp compilation
 
 before_install:
-    - pushd thirdparty
-    # thrift cpp
+    - mkdir $HOME/build_dir
+    - cd $HOME/build_dir
     - >
       if [ $TRAVIS_OS_NAME == osx ]; then
         brew update &&
-        brew install thrift;
+        brew install thrift lz4 snappy;
       fi
     - >
       if [ $TRAVIS_OS_NAME == linux ]; then
-        wget http://archive.apache.org/dist/thrift/0.9.1/thrift-0.9.1.tar.gz &&
-        tar xfz thrift-0.9.1.tar.gz &&
-        pushd thrift-0.9.1 &&
-        ./configure CXXFLAGS='-fPIC' --without-qt4 --without-c_glib --without-csharp --without-java --without-erlang --without-nodejs --without-lua --without-python --without-perl --without-php --without-php_extension --without-ruby --without-haskell --without-go --without-d --with-cpp --prefix=$HOME/local &&
-        make clean &&
-        make install &&
-        popd;
+        cp -r $TRAVIS_BUILD_DIR/thirdparty . &&
+        ./thirdparty/download_thirdparty.sh &&
+        ./thirdparty/build_thirdparty.sh &&
+        export THRIFT_HOME=$HOME/build_dir/thirdparty/installed SNAPPY_HOME=$HOME/build_dir/thirdparty/installed LZ4_HOME=$HOME/build_dir/thirdparty/installed;
       fi
-    # snappy and lz4
-    - ./download_thirdparty.sh
-    - ./build_thirdparty.sh
-    - popd
 
 before_script:
     - export CC="gcc-4.9"
     - export CXX="g++-4.9"
-    - mkdir build
-    - cd build
-    - THRIFT_HOME=$HOME/local cmake ..
+    - cmake $TRAVIS_BUILD_DIR
 
 script: make

CMakeLists.txt

@@ -22,9 +22,6 @@ enable_testing()
 set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_SOURCE_DIR}/cmake_modules")
 set(BUILD_SUPPORT_DIR ${CMAKE_CURRENT_SOURCE_DIR}/build-support)
 
-set(THIRDPARTY_PREFIX ${CMAKE_SOURCE_DIR}/thirdparty/installed)
-set(CMAKE_PREFIX_PATH ${THIRDPARTY_PREFIX})
-
 if(APPLE)
   set(CMAKE_MACOSX_RPATH 1)
   set(CMAKE_OSX_DEPLOYMENT_TARGET 10.9)

README.md

@@ -2,32 +2,36 @@ Parquet-cpp [![Build Status](https://travis-ci.org/apache/parquet-cpp.svg)](http
 ===========
 A C++ library to read parquet files.
 
-To build you will need some version of boost installed and thrift 0.7+ installed.  
-(If you are building thrift from source, you will need to set the THRIFT_HOME env
-variable to the directory containing include/ and lib/.)
-
-Then run:
-<br>
-<code>
-thirdparty/download_thirdparty.sh
-</code>
-<br>
-<code>
-thirdparty/build_thirdparty.sh
-</code>
-<br>
-<code>
-cmake . 
-</code>
-<br>
-<code>
-make
-</code>
-
-The binaries will be built to ./bin which contains the libraries to link against as
+## Third Party Dependencies
+- snappy
+- lz4
+- thrift 0.7+ [install instructions](https://thrift.apache.org/docs/install/)
+
+Many package managers support some or all of these dependencies. E.g.:
+```shell
+ubuntu$ sudo apt-get install libboost-dev libsnappy-dev liblz4-dev
+```
+```shell
+mac$ brew install snappy lz4 thrift
+```
+
+./setup_build_env.sh tries to automate setting up a build environment for you with third party dependencies.  You use it by running `./setup_build_env.sh`.  By default, it will create a build directory `build/`.  You can override the build directory by setting the BUILD_DIR env variable to another location.
+
+Also feel free to take a look at our [.travis.yml](.travis.yml) to see how that build env is set up.
+
+
+## Build
+- `cmake .`
+  - You can customize dependent library locations through various environment variables:
+    - THRIFT_HOME customizes the thrift installed location.
+    - SNAPPY_HOME customizes the snappy installed location.
+    - LZ4_HOME customizes the lz4 installed location.
+- `make`
+
+The binaries will be built to ./debug which contains the libraries to link against as
 well as a few example executables.
 
-Incremental builds can be done afterwords with just <code> make </code>.
+Incremental builds can be done afterwords with just `make`.
 
 Design
 ========

cmake_modules/FindLz4.cmake

@@ -1,4 +1,3 @@
-# Copyright 2012 Cloudera Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -11,47 +10,71 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-
+#
+# Tries to find Lz4 headers and libraries.
+#
+# Usage of this module as follows:
+#
+#  find_package(Lz4)
+#
+# Variables used by this module, they can change the default behaviour and need
+# to be set before calling find_package:
+#
+#  Lz4_HOME - When set, this path is inspected instead of standard library
+#             locations as the root of the Lz4 installation.
+#             The environment variable LZ4_HOME overrides this veriable.
+#
 # - Find LZ4 (lz4.h, liblz4.a, liblz4.so, and liblz4.so.1)
 # This module defines
 #  LZ4_INCLUDE_DIR, directory containing headers
 #  LZ4_LIBS, directory containing lz4 libraries
 #  LZ4_STATIC_LIB, path to liblz4.a
+#  LZ4_SHARED_LIB, path to liblz4's shared library
 #  LZ4_FOUND, whether lz4 has been found
 
-set(LZ4_SEARCH_HEADER_PATHS
-  ${THIRDPARTY_PREFIX}/include
-)
-
-set(LZ4_SEARCH_LIB_PATH
-  ${THIRDPARTY_PREFIX}/lib
-)
+if( NOT "$ENV{LZ4_HOME}" STREQUAL "")
+    file( TO_CMAKE_PATH "$ENV{LZ4_HOME}" _native_path )
+    list( APPEND _lz4_roots ${_native_path} )
+elseif ( Lz4_HOME )
+    list( APPEND _lz4_roots ${Lz4_HOME} )
+endif()
 
-find_path(LZ4_INCLUDE_DIR lz4.h PATHS
-  ${LZ4_SEARCH_HEADER_PATHS}
-  # make sure we don't accidentally pick up a different version
-  NO_DEFAULT_PATH
-)
+# Try the parameterized roots, if they exist
+if ( _lz4_roots )
+    find_path( LZ4_INCLUDE_DIR NAMES lz4.h
+        PATHS ${_lz4_roots} NO_DEFAULT_PATH
+        PATH_SUFFIXES "include" )
+    find_library( LZ4_LIBRARIES NAMES lz4
+        PATHS ${_lz4_roots} NO_DEFAULT_PATH
+        PATH_SUFFIXES "lib" )
+else ()
+    find_path( LZ4_INCLUDE_DIR NAMES lz4.h )
+    find_library( LZ4_LIBRARIES NAMES lz4 )
+endif ()
 
-find_library(LZ4_LIB_PATH NAMES liblz4.a PATHS ${LZ4_SEARCH_LIB_PATH} NO_DEFAULT_PATH)
 
-if (LZ4_INCLUDE_DIR AND LZ4_LIB_PATH)
+if (LZ4_INCLUDE_DIR AND LZ4_LIBRARIES)
   set(LZ4_FOUND TRUE)
-  set(LZ4_LIBS ${LZ4_SEARCH_LIB_PATH})
-  set(LZ4_STATIC_LIB ${LZ4_SEARCH_LIB_PATH}/liblz4.a)
+  get_filename_component( LZ4_LIBS ${LZ4_LIBRARIES} DIRECTORY )
+  set(LZ4_LIB_NAME liblz4)
+  set(LZ4_STATIC_LIB ${LZ4_LIBS}/${LZ4_LIB_NAME}.a)
+  set(LZ4_SHARED_LIB ${LZ4_LIBS}/${LZ4_LIB_NAME}${CMAKE_SHARED_LIBRARY_SUFFIX})
 else ()
   set(LZ4_FOUND FALSE)
 endif ()
 
 if (LZ4_FOUND)
   if (NOT Lz4_FIND_QUIETLY)
-    message(STATUS "Found the Lz4 library: ${LZ4_LIB_PATH}")
+    message(STATUS "Found the Lz4 library: ${LZ4_LIBRARIES}")
   endif ()
 else ()
   if (NOT Lz4_FIND_QUIETLY)
-    set(LZ4_ERR_MSG "Could not find the Lz4 library. Looked for headers")
-    set(LZ4_ERR_MSG "${LZ4_ERR_MSG} in ${LZ4_SEARCH_HEADER_PATHS}, and for libs")
-    set(LZ4_ERR_MSG "${LZ4_ERR_MSG} in ${LZ4_SEARCH_LIB_PATH}")
+    set(LZ4_ERR_MSG "Could not find the Lz4 library. Looked in ")
+    if ( _lz4_roots )
+      set(LZ4_ERR_MSG "${LZ4_ERR_MSG} in ${_lz4_roots}.")
+    else ()
+      set(LZ4_ERR_MSG "${LZ4_ERR_MSG} system search paths.")
+    endif ()
     if (Lz4_FIND_REQUIRED)
       message(FATAL_ERROR "${LZ4_ERR_MSG}")
     else (Lz4_FIND_REQUIRED)
@@ -63,5 +86,7 @@ endif ()
 mark_as_advanced(
   LZ4_INCLUDE_DIR
   LZ4_LIBS
+  LZ4_LIBRARIES
   LZ4_STATIC_LIB
+  LZ4_SHARED_LIB
 )

cmake_modules/FindSnappy.cmake

@@ -1,4 +1,3 @@
-# Copyright 2012 Cloudera Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -11,47 +10,70 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-
-# - Find SNAPPY (snappy.h, libsnappy.a, libsnappy.so, and libsnappy.so.1)
+#
+# Tries to find Snappy headers and libraries.
+#
+# Usage of this module as follows:
+#
+#  find_package(Snappy)
+#
+# Variables used by this module, they can change the default behaviour and need
+# to be set before calling find_package:
+#
+#  Snappy_HOME - When set, this path is inspected instead of standard library
+#                locations as the root of the Snappy installation.
+#                The environment variable SNAPPY_HOME overrides this veriable.
+#
 # This module defines
 #  SNAPPY_INCLUDE_DIR, directory containing headers
 #  SNAPPY_LIBS, directory containing snappy libraries
 #  SNAPPY_STATIC_LIB, path to libsnappy.a
+#  SNAPPY_SHARED_LIB, path to libsnappy's shared library
 #  SNAPPY_FOUND, whether snappy has been found
 
-set(SNAPPY_SEARCH_HEADER_PATHS
-  ${THIRDPARTY_PREFIX}/include
-)
-
-set(SNAPPY_SEARCH_LIB_PATH
-  ${THIRDPARTY_PREFIX}/lib
-)
+if( NOT "$ENV{SNAPPY_HOME}" STREQUAL "")
+    file( TO_CMAKE_PATH "$ENV{SNAPPY_HOME}" _native_path )
+    list( APPEND _snappy_roots ${_native_path} )
+elseif ( Snappy_HOME )
+    list( APPEND _snappy_roots ${Snappy_HOME} )
+endif()
 
-find_path(SNAPPY_INCLUDE_DIR snappy.h PATHS
-  ${SNAPPY_SEARCH_HEADER_PATHS}
-  # make sure we don't accidentally pick up a different version
-  NO_DEFAULT_PATH
-)
+# Try the parameterized roots, if they exist
+if ( _snappy_roots )
+    find_path( SNAPPY_INCLUDE_DIR NAMES snappy.h
+        PATHS ${_snappy_roots} NO_DEFAULT_PATH
+        PATH_SUFFIXES "include" )
+    find_library( SNAPPY_LIBRARIES NAMES snappy
+        PATHS ${_snappy_roots} NO_DEFAULT_PATH
+        PATH_SUFFIXES "lib" )
+else ()
+    find_path( SNAPPY_INCLUDE_DIR NAMES snappy.h )
+    find_library( SNAPPY_LIBRARIES NAMES snappy )
+endif ()
 
-find_library(SNAPPY_LIB_PATH NAMES snappy PATHS ${SNAPPY_SEARCH_LIB_PATH} NO_DEFAULT_PATH)
 
-if (SNAPPY_INCLUDE_DIR AND SNAPPY_LIB_PATH)
+if (SNAPPY_INCLUDE_DIR AND SNAPPY_LIBRARIES)
   set(SNAPPY_FOUND TRUE)
-  set(SNAPPY_LIBS ${SNAPPY_SEARCH_LIB_PATH})
-  set(SNAPPY_STATIC_LIB ${SNAPPY_SEARCH_LIB_PATH}/libsnappy.a)
+  get_filename_component( SNAPPY_LIBS ${SNAPPY_LIBRARIES} DIRECTORY )
+  set(SNAPPY_LIB_NAME libsnappy)
+  set(SNAPPY_STATIC_LIB ${SNAPPY_LIBS}/${SNAPPY_LIB_NAME}.a)
+  set(SNAPPY_SHARED_LIB ${SNAPPY_LIBS}/${SNAPPY_LIB_NAME}${CMAKE_SHARED_LIBRARY_SUFFIX})
 else ()
   set(SNAPPY_FOUND FALSE)
 endif ()
 
 if (SNAPPY_FOUND)
   if (NOT Snappy_FIND_QUIETLY)
-    message(STATUS "Found the Snappy library: ${SNAPPY_LIB_PATH}")
+    message(STATUS "Found the Snappy library: ${SNAPPY_LIBRARIES}")
   endif ()
 else ()
   if (NOT Snappy_FIND_QUIETLY)
-    set(SNAPPY_ERR_MSG "Could not find the Snappy library. Looked for headers")
-    set(SNAPPY_ERR_MSG "${SNAPPY_ERR_MSG} in ${SNAPPY_SEARCH_HEADER_PATHS}, and for libs")
-    set(SNAPPY_ERR_MSG "${SNAPPY_ERR_MSG} in ${SNAPPY_SEARCH_LIB_PATH}")
+    set(SNAPPY_ERR_MSG "Could not find the Snappy library. Looked in ")
+    if ( _snappy_roots )
+      set(SNAPPY_ERR_MSG "${SNAPPY_ERR_MSG} in ${_snappy_roots}.")
+    else ()
+      set(SNAPPY_ERR_MSG "${SNAPPY_ERR_MSG} system search paths.")
+    endif ()
     if (Snappy_FIND_REQUIRED)
       message(FATAL_ERROR "${SNAPPY_ERR_MSG}")
     else (Snappy_FIND_REQUIRED)
@@ -63,5 +85,7 @@ endif ()
 mark_as_advanced(
   SNAPPY_INCLUDE_DIR
   SNAPPY_LIBS
+  SNAPPY_LIBRARIES
   SNAPPY_STATIC_LIB
+  SNAPPY_SHARED_LIB
 )

setup_build_env.sh

@@ -0,0 +1,27 @@
+#!/bin/bash
+
+set -e
+
+SOURCE_DIR=$(cd "$(dirname "$BASH_SOURCE")"; pwd)
+: ${BUILD_DIR:=$SOURCE_DIR/build}
+
+mkdir -p $BUILD_DIR
+cp -r $SOURCE_DIR/thirdparty $BUILD_DIR
+cd $BUILD_DIR
+./thirdparty/download_thirdparty.sh
+./thirdparty/build_thirdparty.sh
+
+export SNAPPY_HOME=$BUILD_DIR/thirdparty/installed
+export LZ4_HOME=$BUILD_DIR/thirdparty/installed
+# build script doesn't support building thrift on OSX
+if [ "$(uname)" != "Darwin" ]; then
+  export THRIFT_HOME=$BUILD_DIR/thirdparty/installed
+fi
+
+cmake $SOURCE_DIR
+
+cd $SOURCE_DIR
+
+echo
+echo "Build env initialized in $BUILD_DIR."
+

src/parquet/compression/lz4-codec.cc

@@ -20,7 +20,7 @@ namespace parquet_cpp {
 
 void Lz4Codec::Decompress(int input_len, const uint8_t* input,
       int output_len, uint8_t* output_buffer) {
-  int n = LZ4_uncompress(reinterpret_cast<const char*>(input),
+  int n = LZ4_decompress_fast(reinterpret_cast<const char*>(input),
       reinterpret_cast<char*>(output_buffer), output_len);
   if (n != input_len) {
     throw parquet_cpp::ParquetException("Corrupt lz4 compressed data.");