PARQUET-267 - Relax Third Party Dependency Restrictions

.gitignore

@@ -9,4 +9,3 @@ CTestTestfile.cmake
 cmake_install.cmake
 Makefile
 
-thirdparty

.travis.yml

@@ -12,7 +12,7 @@ addons:
     apt:
         packages:
             - libboost-dev
-              #- libsnappy-dev currently handled by thirdparty scipts.
+            - libsnappy-dev
             - libboost-program-options-dev #needed for thrift cpp compilation
             - libboost-test-dev            #needed for thrift cpp compilation
             - libssl-dev                   #needed for thrift cpp compilation
@@ -22,8 +22,8 @@ addons:
             - pkg-config                   #needed for thrift cpp compilation
 
 before_install:
-    - pushd thirdparty
-    # thrift cpp
+    - mkdir $HOME/thirdparty && pushd $HOME/thirdparty
+    # linux: thrift cpp
     - >
       if [ $TRAVIS_OS_NAME == linux ]; then
         wget http://www.us.apache.org/dist/thrift/0.9.1/thrift-0.9.1.tar.gz &&
@@ -34,15 +34,22 @@ before_install:
         make install &&
         popd;
       fi
-    - if [ $TRAVIS_OS_NAME == osx ]; then brew install thrift; fi
-    # snappy and lz4
-    - ./download_thirdparty.sh
-    - ./build_thirdparty.sh
+    # linux: lz4
+    - >
+      if [ $TRAVIS_OS_NAME == linux ]; then
+        wget https://github.com/Cyan4973/lz4/archive/r128.tar.gz &&
+        tar xfz r128.tar.gz &&
+        pushd lz4-r128 &&
+        PREFIX=$HOME/local make install &&
+        popd;
+      fi
     - popd
+    # mac dependencies
+    - if [ $TRAVIS_OS_NAME == osx ]; then brew install thrift snappy lz4; fi
 
 before_script:
     - mkdir build
     - cd build
-    - THRIFT_HOME=$HOME/local cmake ..
+    - THRIFT_HOME=$HOME/local LZ4_HOME=$HOME/local cmake ..
 
 script: make

CMakeLists.txt

@@ -21,9 +21,6 @@ enable_testing()
 # where to find cmake modules
 set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_SOURCE_DIR}/cmake_modules")
 
-set(THIRDPARTY_PREFIX ${CMAKE_SOURCE_DIR}/thirdparty/installed)
-set(CMAKE_PREFIX_PATH ${THIRDPARTY_PREFIX})
-
 # find boost headers and libs
 set(Boost_DEBUG TRUE)
 set(Boost_USE_MULTITHREADED ON)
@@ -47,16 +44,18 @@ add_library(thriftstatic STATIC IMPORTED)
 set_target_properties(thriftstatic PROPERTIES IMPORTED_LOCATION ${THRIFT_STATIC_LIB})
 
 ## Snappy
+set( Snappy_USE_STATIC_LIBS ON )
 find_package(Snappy REQUIRED)
 include_directories(SYSTEM ${SNAPPY_INCLUDE_DIR})
 add_library(snappystatic STATIC IMPORTED)
-set_target_properties(snappystatic PROPERTIES IMPORTED_LOCATION ${SNAPPY_STATIC_LIB})
+set_target_properties(snappystatic PROPERTIES IMPORTED_LOCATION ${SNAPPY_LIBRARIES})
 
 ## LZ4
+set( Lz4_USE_STATIC_LIBS ON )
 find_package(Lz4 REQUIRED)
 include_directories(SYSTEM ${LZ4_INCLUDE_DIR})
 add_library(lz4static STATIC IMPORTED)
-set_target_properties(lz4static PROPERTIES IMPORTED_LOCATION ${LZ4_STATIC_LIB})
+set_target_properties(lz4static PROPERTIES IMPORTED_LOCATION ${LZ4_LIBRARIES})
 
 SET(CMAKE_CXX_FLAGS "-msse4.2 -Wall -Wno-unused-value -Wno-unused-variable -Wno-sign-compare -Wno-unknown-pragmas")
 SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -g -ggdb")

README.md

@@ -2,27 +2,28 @@ 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>
+## Third Party Dependencies
+- boost
+- 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 boost snappy lz4 thrift
+```
+You can also 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 ./bin which contains the libraries to link against as
 well as a few example executables.

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.
@@ -11,57 +10,79 @@
 # 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 - The root of the Lz4 installation.  This is considered before
+#             standard locations. The environment variable LZ4_HOME is also
+#             considered if this variable is not set.
+#  Lz4_USE_STATIC_LIBS - Set to ON to force the use of the static
+#                        libraries.  Default is OFF
+#
+# Variables defined by this module:
+#
+#  LZ4_FOUND              System has Lz4 libs/headers
+#  LZ4_LIBRARIES          The Lz4 libraries
+#  LZ4_INCLUDE_DIR        The location of Lz4 headers
 
-# - 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_FOUND, whether lz4 has been found
+# Determine parameterized roots
+if( Lz4_HOME )
+    list( APPEND _lz4_roots ${Lz4_HOME} )
+endif()
+if( NOT "$ENV{LZ4_HOME}" STREQUAL "")
+    file( TO_CMAKE_PATH "$ENV{LZ4_HOME}" _native_path )
+    list( APPEND _lz4_roots ${_native_path} )
+endif()
 
-set(LZ4_SEARCH_HEADER_PATHS
-  ${THIRDPARTY_PREFIX}/include
-)
+# Try the parameterized roots first
+find_path( LZ4_INCLUDE_DIR NAMES lz4.h
+    PATHS ${_lz4_roots} NO_DEFAULT_PATH
+    PATH_SUFFIXES "include" )
+find_path( LZ4_INCLUDE_DIR NAMES lz4.h )
 
-set(LZ4_SEARCH_LIB_PATH
-  ${THIRDPARTY_PREFIX}/lib
-)
+# Support preference of static libs by adjusting CMAKE_FIND_LIBRARY_SUFFIXES.
+# Patterned after FindBoost.cmake
+if( Lz4_USE_STATIC_LIBS )
+    set( _lz4_ORIG_CMAKE_FIND_LIBRARY_SUFFIXES ${CMAKE_FIND_LIBRARY_SUFFIXES} )
+    if(WIN32)
+        set( CMAKE_FIND_LIBRARY_SUFFIXES .lib .a ${CMAKE_FIND_LIBRARY_SUFFIXES} )
+    else()
+        set( CMAKE_FIND_LIBRARY_SUFFIXES .a )
+    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
-)
+    # customize the cached variable depending on the search
+    set( _lz4_LIBRARIES LZ4_LIBRARIES_STATIC )
+else()
+    set( _lz4_LIBRARIES LZ4_LIBRARIES_DEFAULT )
+endif()
 
-find_library(LZ4_LIB_PATH NAMES liblz4.a PATHS ${LZ4_SEARCH_LIB_PATH} NO_DEFAULT_PATH)
+# Try the parameterized roots first
+find_library( ${_lz4_LIBRARIES} NAMES lz4
+    PATHS ${_lz4_roots} NO_DEFAULT_PATH
+    PATH_SUFFIXES "lib" )
+find_library( ${_lz4_LIBRARIES} NAMES lz4 )
 
-if (LZ4_INCLUDE_DIR AND LZ4_LIB_PATH)
-  set(LZ4_FOUND TRUE)
-  set(LZ4_LIBS ${LZ4_SEARCH_LIB_PATH})
-  set(LZ4_STATIC_LIB ${LZ4_SEARCH_LIB_PATH}/liblz4.a)
-else ()
-  set(LZ4_FOUND FALSE)
-endif ()
+set( LZ4_LIBRARIES ${${_lz4_LIBRARIES}} )
 
-if (LZ4_FOUND)
-  if (NOT Lz4_FIND_QUIETLY)
-    message(STATUS "Found the Lz4 library: ${LZ4_LIB_PATH}")
-  endif ()
-else ()
-  if (NOT Lz4_FIND_QUIETLY)
-    set(LZ4_ERR_MSG "Could not find the Lz4 library. Looked for headers")
-    set(LZ4_ERR_MSG "${LZ4_ERR_MSG} in ${LZ4_SEARCH_HEADER_PATHS}, and for libs")
-    set(LZ4_ERR_MSG "${LZ4_ERR_MSG} in ${LZ4_SEARCH_LIB_PATH}")
-    if (Lz4_FIND_REQUIRED)
-      message(FATAL_ERROR "${LZ4_ERR_MSG}")
-    else (Lz4_FIND_REQUIRED)
-      message(STATUS "${LZ4_ERR_MSG}")
-    endif (Lz4_FIND_REQUIRED)
-  endif ()
-endif ()
+# Restore the original find library ordering
+if( Lz4_USE_STATIC_LIBS )
+  set( CMAKE_FIND_LIBRARY_SUFFIXES ${_lz4_ORIG_CMAKE_FIND_LIBRARY_SUFFIXES} )
+endif()
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args( Lz4 DEFAULT_MSG
+    ${_lz4_LIBRARIES}
+    LZ4_INCLUDE_DIR )
 
 mark_as_advanced(
-  LZ4_INCLUDE_DIR
-  LZ4_LIBS
-  LZ4_STATIC_LIB
-)
+    LZ4_LIBRARIES_STATIC
+    LZ4_LIBRARIES_DEFAULT
+    LZ4_INCLUDE_DIR )
+

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.
@@ -11,57 +10,78 @@
 # 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 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 - The root of the Snappy installation.  This is considered before
+#                standard locations. The environment variable SNAPPY_HOME is also
+#                considered if this variable is not set.
+#  Snappy_USE_STATIC_LIBS - Set to ON to force the use of the static
+#                           libraries.  Default is OFF
+#
+# Variables defined by this module:
+#
+#  SNAPPY_FOUND              System has Snappy libs/headers
+#  SNAPPY_LIBRARIES          The Snappy libraries
+#  SNAPPY_INCLUDE_DIR        The location of Snappy headers
+
+# Determine parameterized roots
+if( Snappy_HOME )
+    list( APPEND _snappy_roots ${Snappy_HOME} )
+endif()
+if( NOT "$ENV{SNAPPY_HOME}" STREQUAL "")
+    file( TO_CMAKE_PATH "$ENV{SNAPPY_HOME}" _native_path )
+    list( APPEND _snappy_roots ${_native_path} )
+endif()
 
-# - Find SNAPPY (snappy.h, libsnappy.a, libsnappy.so, and libsnappy.so.1)
-# This module defines
-#  SNAPPY_INCLUDE_DIR, directory containing headers
-#  SNAPPY_LIBS, directory containing snappy libraries
-#  SNAPPY_STATIC_LIB, path to libsnappy.a
-#  SNAPPY_FOUND, whether snappy has been found
+# Try the parameterized roots first
+find_path( SNAPPY_INCLUDE_DIR NAMES snappy.h
+    PATHS ${_snappy_roots} NO_DEFAULT_PATH
+    PATH_SUFFIXES "include" )
+find_path( SNAPPY_INCLUDE_DIR NAMES snappy.h )
 
-set(SNAPPY_SEARCH_HEADER_PATHS
-  ${THIRDPARTY_PREFIX}/include
-)
+# Support preference of static libs by adjusting CMAKE_FIND_LIBRARY_SUFFIXES.
+# Patterned after FindBoost.cmake
+if( Snappy_USE_STATIC_LIBS )
+    set( _snappy_ORIG_CMAKE_FIND_LIBRARY_SUFFIXES ${CMAKE_FIND_LIBRARY_SUFFIXES} )
+    if(WIN32)
+        set( CMAKE_FIND_LIBRARY_SUFFIXES .lib .a ${CMAKE_FIND_LIBRARY_SUFFIXES} )
+    else()
+        set( CMAKE_FIND_LIBRARY_SUFFIXES .a )
+    endif()
 
-set(SNAPPY_SEARCH_LIB_PATH
-  ${THIRDPARTY_PREFIX}/lib
-)
+    # customize the cached variable depending on the search
+    set( _snappy_LIBRARIES SNAPPY_LIBRARIES_STATIC )
+else()
+    set( _snappy_LIBRARIES SNAPPY_LIBRARIES_DEFAULT )
+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 first
+find_library( ${_snappy_LIBRARIES} NAMES snappy
+    PATHS ${_snappy_roots} NO_DEFAULT_PATH
+    PATH_SUFFIXES "lib" )
+find_library( ${_snappy_LIBRARIES} NAMES snappy )
 
-find_library(SNAPPY_LIB_PATH NAMES snappy PATHS ${SNAPPY_SEARCH_LIB_PATH} NO_DEFAULT_PATH)
+set( SNAPPY_LIBRARIES ${${_snappy_LIBRARIES}} )
 
-if (SNAPPY_INCLUDE_DIR AND SNAPPY_LIB_PATH)
-  set(SNAPPY_FOUND TRUE)
-  set(SNAPPY_LIBS ${SNAPPY_SEARCH_LIB_PATH})
-  set(SNAPPY_STATIC_LIB ${SNAPPY_SEARCH_LIB_PATH}/libsnappy.a)
-else ()
-  set(SNAPPY_FOUND FALSE)
-endif ()
+# Restore the original find library ordering
+if( Snappy_USE_STATIC_LIBS )
+  set( CMAKE_FIND_LIBRARY_SUFFIXES ${_snappy_ORIG_CMAKE_FIND_LIBRARY_SUFFIXES} )
+endif()
 
-if (SNAPPY_FOUND)
-  if (NOT Snappy_FIND_QUIETLY)
-    message(STATUS "Found the Snappy library: ${SNAPPY_LIB_PATH}")
-  endif ()
-else ()
-  if (NOT Snappy_FIND_QUIETLY)
-    set(SNAPPY_ERR_MSG "Could not find the Snappy library. Looked for headers")
-    set(SNAPPY_ERR_MSG "${SNAPPY_ERR_MSG} in ${SNAPPY_SEARCH_HEADER_PATHS}, and for libs")
-    set(SNAPPY_ERR_MSG "${SNAPPY_ERR_MSG} in ${SNAPPY_SEARCH_LIB_PATH}")
-    if (Snappy_FIND_REQUIRED)
-      message(FATAL_ERROR "${SNAPPY_ERR_MSG}")
-    else (Snappy_FIND_REQUIRED)
-      message(STATUS "${SNAPPY_ERR_MSG}")
-    endif (Snappy_FIND_REQUIRED)
-  endif ()
-endif ()
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args( Snappy DEFAULT_MSG
+    ${_snappy_LIBRARIES}
+    SNAPPY_INCLUDE_DIR )
 
 mark_as_advanced(
-  SNAPPY_INCLUDE_DIR
-  SNAPPY_LIBS
-  SNAPPY_STATIC_LIB
-)
+    SNAPPY_LIBRARIES_STATIC
+    SNAPPY_LIBRARIES_DEFAULT
+    SNAPPY_INCLUDE_DIR )