CMake/Utilities/Scripts/BoostScanDeps.cmake

208 lines
8.8 KiB
CMake
Raw Normal View History

Simplify CMake per-source license notices Per-source copyright/license notice headers that spell out copyright holder names and years are hard to maintain and often out-of-date or plain wrong. Precise contributor information is already maintained automatically by the version control tool. Ultimately it is the receiver of a file who is responsible for determining its licensing status, and per-source notices are merely a convenience. Therefore it is simpler and more accurate for each source to have a generic notice of the license name and references to more detailed information on copyright holders and full license terms. Our `Copyright.txt` file now contains a list of Contributors whose names appeared source-level copyright notices. It also references version control history for more precise information. Therefore we no longer need to spell out the list of Contributors in each source file notice. Replace CMake per-source copyright/license notice headers with a short description of the license and links to `Copyright.txt` and online information available from "https://cmake.org/licensing". The online URL also handles cases of modules being copied out of our source into other projects, so we can drop our notices about replacing links with full license text. Run the `Utilities/Scripts/filter-notices.bash` script to perform the majority of the replacements mechanically. Manually fix up shebang lines and trailing newlines in a few files. Manually update the notices in a few files that the script does not handle.
2016-09-27 22:01:08 +03:00
# Distributed under the OSI-approved BSD 3-Clause License. See accompanying
# file Copyright.txt or https://cmake.org/licensing for details.
# Scan the Boost headers and determine the library dependencies. Note
# that this script only scans one Boost version at once; invoke once
# for each Boost release. Note that this does require the headers for
# a given component to match the library name, since this computes
# inter-library dependencies. Library components for which this
# assumption does not hold true and which have dependencies on other
# Boost libraries will require special-casing. It also doesn't handle
# private dependencies not described in the headers, for static
# library dependencies--this is also a limitation of auto-linking, and
# I'm unaware of any specific instances where this would be
# problematic.
#
# Invoke in script mode, defining these variables:
# BOOST_DIR - the root of the boost includes
#
# The script will process each directory under the root as a
# "component". For each component, all the headers will be scanned to
# determine the components it depends upon by following all the
# possible includes from this component. This is to match the
# behaviour of autolinking.
# Written by Roger Leigh <rleigh@codelibre.net>
# Determine header dependencies on libraries using the embedded dependency information.
#
# component - the component to check (uses all headers from boost/${component})
# includedir - the path to the Boost headers
# _ret_libs - list of library dependencies
#
function(_Boost_FIND_COMPONENT_DEPENDENCIES component includedir _ret_libs)
# _boost_unprocessed_headers - list of headers requiring parsing
# _boost_processed_headers - headers already parsed (or currently being parsed)
# _boost_new_headers - new headers discovered for future processing
set(library_component FALSE)
# Start by finding all headers for the component; header
# dependencies via #include will be solved by future passes
# Special-case since it is part of mpi; look only in boost/mpi/python*
if(component STREQUAL "mpi_python")
set(_boost_DEPS "python")
set(library_component TRUE)
file(GLOB_RECURSE _boost_unprocessed_headers
RELATIVE "${includedir}"
"${includedir}/boost/mpi/python/*")
list(INSERT _boost_unprocessed_headers 0 "${includedir}/boost/mpi/python.hpp")
# Special-case since it is a serialization variant; look in boost/serialization
elseif(component STREQUAL "wserialization")
set(library_component TRUE)
file(GLOB_RECURSE _boost_unprocessed_headers
RELATIVE "${includedir}"
"${includedir}/boost/serialization/*")
list(INSERT _boost_unprocessed_headers 0 "${includedir}/boost/serialization.hpp")
# Not really a library in its own right, but treat it as one
elseif(component STREQUAL "math")
set(library_component TRUE)
file(GLOB_RECURSE _boost_unprocessed_headers
RELATIVE "${includedir}"
"${includedir}/boost/math/*")
list(INSERT _boost_unprocessed_headers 0 "${includedir}/boost/math.hpp")
# Single test header
elseif(component STREQUAL "unit_test_framework")
set(library_component TRUE)
set(_boost_unprocessed_headers "${BOOST_DIR}/test/unit_test.hpp")
# Single test header
elseif(component STREQUAL "prg_exec_monitor")
set(library_component TRUE)
set(_boost_unprocessed_headers "${BOOST_DIR}/test/prg_exec_monitor.hpp")
# Single test header
elseif(component STREQUAL "test_exec_monitor")
set(library_component TRUE)
set(_boost_unprocessed_headers "${BOOST_DIR}/test/test_exec_monitor.hpp")
else()
# Default behaviour where header directory is the same as the library name.
file(GLOB_RECURSE _boost_unprocessed_headers
RELATIVE "${includedir}"
"${includedir}/boost/${component}/*")
list(INSERT _boost_unprocessed_headers 0 "${includedir}/boost/${component}.hpp")
endif()
while(_boost_unprocessed_headers)
list(APPEND _boost_processed_headers ${_boost_unprocessed_headers})
foreach(header ${_boost_unprocessed_headers})
if(EXISTS "${includedir}/${header}")
file(STRINGS "${includedir}/${header}" _boost_header_includes REGEX "^#[ \t]*include[ \t]*<boost/[^>][^>]*>")
# The optional whitespace before "#" is intentional
# (boost/serialization/config.hpp bug).
file(STRINGS "${includedir}/${header}" _boost_header_deps REGEX "^[ \t]*#[ \t]*define[ \t][ \t]*BOOST_LIB_NAME[ \t][ \t]*boost_")
foreach(line ${_boost_header_includes})
string(REGEX REPLACE "^#[ \t]*include[ \t]*<(boost/[^>][^>]*)>.*" "\\1" _boost_header_match "${line}")
list(FIND _boost_processed_headers "${_boost_header_match}" _boost_header_processed)
list(FIND _boost_new_headers "${_boost_header_match}" _boost_header_new)
if (_boost_header_processed EQUAL -1 AND _boost_header_new EQUAL -1)
list(APPEND _boost_new_headers ${_boost_header_match})
endif()
endforeach()
foreach(line ${_boost_header_deps})
string(REGEX REPLACE "^[ \t]*#[ \t]*define[ \t][ \t]*BOOST_LIB_NAME[ \t][ \t]*boost_([^ \t][^ \t]*).*" "\\1" _boost_component_match "${line}")
list(FIND _boost_DEPS "${_boost_component_match}" _boost_dep_found)
if(_boost_component_match STREQUAL "bzip2" OR
_boost_component_match STREQUAL "zlib")
# These components may or may not be required; not
# possible to tell without knowing where and when
# BOOST_BZIP2_BINARY and BOOST_ZLIB_BINARY are defined.
# If building against an external zlib or bzip2, this is
# undesirable.
continue()
endif()
if(component STREQUAL "mpi" AND
(_boost_component_match STREQUAL "mpi_python" OR
_boost_component_match STREQUAL "python"))
# Optional python dependency; skip to avoid making it a
# hard dependency (handle as special-case for mpi_python).
continue()
endif()
if (_boost_dep_found EQUAL -1 AND
NOT "${_boost_component_match}" STREQUAL "${component}")
list(APPEND _boost_DEPS "${_boost_component_match}")
endif()
if("${_boost_component_match}" STREQUAL "${component}")
set(library_component TRUE)
endif()
endforeach()
endif()
endforeach()
set(_boost_unprocessed_headers ${_boost_new_headers})
unset(_boost_new_headers)
endwhile()
# message(STATUS "Unfiltered dependencies for Boost::${component}: ${_boost_DEPS}")
if(NOT library_component)
unset(_boost_DEPS)
endif()
set(${_ret_libs} ${_boost_DEPS} PARENT_SCOPE)
#string(REGEX REPLACE ";" " " _boost_DEPS_STRING "${_boost_DEPS}")
#if (NOT _boost_DEPS_STRING)
# set(_boost_DEPS_STRING "(none)")
#endif()
#message(STATUS "Dependencies for Boost::${component}: ${_boost_DEPS_STRING}")
endfunction()
message(STATUS "Scanning ${BOOST_DIR}")
# List of all directories and files
file(GLOB boost_contents RELATIVE "${BOOST_DIR}/boost" "${BOOST_DIR}/boost/*")
# Components as directories
foreach(component ${boost_contents})
if(IS_DIRECTORY "${BOOST_DIR}/boost/${component}")
list(APPEND boost_components "${component}")
endif()
endforeach()
# The following components are not top-level directories, so require
# special-casing:
# Special-case mpi_python, since it's a part of mpi
if(IS_DIRECTORY "${BOOST_DIR}/boost/mpi" AND
IS_DIRECTORY "${BOOST_DIR}/boost/python")
list(APPEND boost_components "mpi_python")
endif()
# Special-case wserialization, which is a variant of serialization
if(IS_DIRECTORY "${BOOST_DIR}/boost/serialization")
list(APPEND boost_components "wserialization")
endif()
# Special-case math* since there are six libraries, but no actual math
# library component. Handle specially when scanning above.
#
# Special-case separate test libraries, which are all part of test
if(EXISTS "${BOOST_DIR}/test/unit_test.hpp")
list(APPEND boost_components "unit_test_framework")
endif()
if(EXISTS "${BOOST_DIR}/test/prg_exec_monitor.hpp")
list(APPEND boost_components "prg_exec_monitor")
endif()
if(EXISTS "${BOOST_DIR}/test/test_exec_monitor.hpp")
list(APPEND boost_components "test_exec_monitor")
endif()
if(boost_components)
list(SORT boost_components)
endif()
# Process each component defined above
foreach(component ${boost_components})
string(TOUPPER ${component} UPPERCOMPONENT)
_Boost_FIND_COMPONENT_DEPENDENCIES("${component}" "${BOOST_DIR}"
_Boost_${UPPERCOMPONENT}_LIBRARY_DEPENDENCIES)
endforeach()
# Output results
foreach(component ${boost_components})
string(TOUPPER ${component} UPPERCOMPONENT)
if(_Boost_${UPPERCOMPONENT}_LIBRARY_DEPENDENCIES)
string(REGEX REPLACE ";" " " _boost_DEPS_STRING "${_Boost_${UPPERCOMPONENT}_LIBRARY_DEPENDENCIES}")
message(STATUS "set(_Boost_${UPPERCOMPONENT}_DEPENDENCIES ${_boost_DEPS_STRING})")
endif()
endforeach()