Once CMAKE_SYSTEM_NAME is known, load a platform-specific
Platform/<os>-Determine
module in order to enable custom determination of the other settings
needed for the CMakeSystem module (e.g. CMAKE_SYSTEM_PROCESSOR). Also
add a hook in Modules/CMakeSystem.cmake.in to allow platform-specific
information to be saved.
Since commit v3.6.0-rc1~287^2 (GetPrerequisites: Fix
gp_resolved_file_type on non-canonical paths, 2016-03-08) we
accidentally convert relative paths (e.g. system dll file names) to
absolute paths even when we do not know the base directory. Fix this by
canonicalizing only paths that are already absolute.
Closes: #16240
Commit 7229ae72 (FindCUDA: Refactor Android(Tegra) support, 2016-04-19)
changed the logic to only add librt to the link line for CUDA versions
6.5 and older. However, newer versions of CUDA still require librt.
Fix bug introduced by commit 7229ae72 (FindCUDA: Refactor Android(Tegra)
support, 2016-04-19). `CUDA_USE_STATIC_CUDA_RUNTIME` should be enabled
by default if `cudart_static` is available, and silently disabled if
it is not.
CMake used to put all header search paths into HEADER_SEARCH_PATHS
attribute. Unfortunately this attribute does not support to declare
a search path as a system include.
As a hack one could add a -isystem /path to the cflags but then include
ordering is not deterministic. A better approach was chosen with this
patch by not filling HEADER_SEARCH_PATHS at all and to populate
the C, C++, and Fortran flags directly. The include paths used by
Xcode should be now identical to the ones used by Unix Makefiles and
Ninja generator.
The Numerical Algorithms Group (NAG) Fortran compiler documents -PIC for
position-independent code and does not have a separate option for PIE.
We added `-PIC` for PIC in commit v2.8.11~174^2 (NAG: Use -PIC for
Fortran position-independent code, 2013-02-18). Follow up for PIE.
Closes: #16236
When using `grep` to filter the output, add the `-a` flag to tell
it never to treat the output as binary. Otherwise when LANG != C
in the environment the non-ascii text may break the filter.
Find modules only detect Debug and Release configurations. All other
configurations will fall back to the configuration listed as the first
entry in `IMPORTED_CONFIGURATIONS`. Switch the order so that `Release`
is listed first, as this is a better fallback than `Debug` for the
`RelWithDebInfo` and `MinSizeRel` configurations. See issue #16091.
This approach is recommended by documentation in `cmake-developer(7)`
added by commit v3.2.0-rc1~286^2~1 (Help: Document IMPORTED_CONFIGURATIONS
target property for Find modules, 2014-12-04).
088f14eb Intel-C: standard flags are also supported in 12.0
27a3ca15 Intel-C: support gnu89 and gnu99 extension flags
cc223e1e Intel-C: declare support for gnu11
Without extensions, functions like `strdup` are not available since they
are actually controlled by feature flags such as _SVID_SOURCE and
_BSD_SOURCE. When using `-std=c11` on Intel, none of these flags are
set, so the functions are not declared properly leading to compile
errors.
Reported-by: Adam J. Stewart <ajstewart426@gmail.com>
Closes: #16226
Do not consider a top-level `.DS_Store` file when deciding whether a
tarball contains exactly one directory whose contents should be used as
the resulting top-level of the extraction.
Fixes#16218.
Suggested-by: Patrice Kouame
This flag is needed for the `<LANG>_VISIBILITY_PRESET` target property.
It has been supported since GCC 4.0, not 4.2 as we previously recorded.
Fixes#16222.
Protect our `$0` reference in the shell as `"$0"`. Otherwise it works
with a space in the path only due to an insecure Windows feature.
Prior to this fix, any installer using the option added by commit
v2.8.9~234^2 (Added CPACK_NSIS_ENABLE_UNINSTALL_BEFORE_INSTALL,
2011-06-11) exposes a local privilege escalation vulnerability.
Reported-by: Amir Szekely <kichik@gmail.com>
Reported-by: Ug_0 Security
8a98cf64 Honor CMAKE_*_LINKER_FLAGS[_<CONFIG>]_INIT set in toolchain files
37d15c39 MSVC: Set all CMAKE_*_LINKER_FLAGS_INIT directly
55c884ed Embarcadero: Set all CMAKE_*_LINKER_FLAGS_INIT directly
aec3c79a Strip CMAKE_*_LINKER_FLAGS[_<CONFIG>] initializer whitespace
Document these variables.
Change our convention for setting these variables from:
set(CMAKE_EXE_LINKER_FLAGS_INIT "...")
to
string(APPEND CMAKE_EXE_LINKER_FLAGS_INIT " ...")
so that any value previously set by a toolchain file will be used.
Port changes from VTK commit fda6a31cb9 (Added Improved FindOpenMP
module, 2015-04-23). Improve use of try_compile to avoid needing
to pass OpenMP flags as libraries.
d5e4516e FindHDF5: Cleanup inconsistent use of HDF5_ROOT
4ece8bdd FindHDF5: Properly fail when required components are not found.
87c1cd9c FindHDF5: Fix h5cc arg parsing to work with homebrew on Mac
These libraries are used for Clang runtime analysis support with
flags like `-fsanitize=memory` and are not actually implicitly
linked libraries.
Fixes#16194.
In commit v3.6.0-rc1~66^2 (CPack/RPM different package names,
2016-05-15) we forgot the non-uppercased fallback variable name
`CPACK_RPM_<COMPONENT>_FILE_NAME`. Add it now.
a66004be Honor CMAKE_<LANG>_FLAGS[_<CONFIG>]_INIT set in toolchain files
cdde77e5 OpenWatcom: Partially modernize platform information modules
f9dbe22c Intel: Do not use GNU-like flags on Windows
5a3ed0d7 Intel: Do not use MSVC-like flags for Fortran
Document these variables.
Change our convention for setting these variables from:
set(CMAKE_C_FLAGS_INIT "...")
to
string(APPEND CMAKE_C_FLAGS_INIT " ...")
so that any value previously set by a toolchain file will be used.
Automate the conversion with:
sed -i 's/set *(\(CMAKE_\(C\|CXX\|Fortran\|RC\|ASM\|${[^}]\+}\)_FLAGS\(_[^_]\+\)\?_INIT \+"\)/string(APPEND \1 /' \
Modules/Compiler/*.cmake Modules/Platform/*.cmake
and follow up with some manual fixes (e.g. to cases that already
meant to append). Also revert the automated changes to contexts
that are not protected from running multiple times.
Migrate from the old `<os>-<cc>.cmake` layout to the modern
`<os>-<id>-<lang>.cmake` layout. Keep settings common to C and C++ in a
`Windows-OpenWatcom.cmake` helper module with an include blocker.
For now just add both C and CXX settings in the helper module.
Directives that are legal inside the %files section of an RPM spec may
contain a variety of characters particularly when specifying %caps which
can include +, _, and space. Watch for parenthesis to determine what
forms the prefix vs. path.
Fixes#14362.
While at it, refactor logic to consolidate the filename extraction
and verification into a single match.
Inspired-by: Ruslan Baratov <ruslan_baratov@yahoo.com>
The official Windows builds of Boost have internal implementations for
`zlib` and `bzip2` libraries used by Boost::iostreams library, e.g.
* boost_bzip2-vc140-mt-1_59.lib
* boost_zlib-vc140-mt-1_59.lib
Add check headers for these so that users can specify them as
components without any warnings.
Reviewed-by: Roger Leigh <rleigh@dundee.ac.uk>
Drop Java_INCLUDE_DIRS and Java_LIBRARIES from the documentation since
we do not provide them. Also add a link to FindJNI to help users find
it. Fixes#16180.
Refactor options out of `Modules/Compiler/Intel-{ASM,C,CXX,Fortran}.cmake`
into a common helper in `Modules/Compiler/Intel.cmake`. Condition
them to be used only on non-Windows hosts where the Intel compiler
is GNU-like instead of MSVC-like.
Previously this worked only because the options were later overridden
by `Modules/Platform/Windows-Intel*.cmake`, but it is cleaner to not
set the options in the first place.
Teach `Modules/Platform/Windows-MSVC.cmake` not to use MSVC options
for Fortran. We use the `__windows_compiler_msvc` for the Intel
Fortran compiler on Windows for other settings, but we do not want
the flags.
Previously this worked only because the options were later overridden
by `Modules/Platform/Windows-Intel*.cmake`, but it is cleaner to not
set the options in the first place.
234deec2 FindHDF5: create all the *_LIBRARIES when using hdf5-config.cmake
963b576f FindHDF5: cache the correct path to the high level libraries
9f2728ce FindHDF5: When component targets not found fallback to compiler wrappers
149539d9 FindHDF5: Handle HDF5 builds with non-suffixed components
f30b0fec FindHDF5: correctly add lang to each component target name.
eecd307e FindLua: Iterate through all include subdirs to find a suitable header
3532cbce FindLua: Check header file against suitable versions
281c707b FindLua: Factor out a helper function for version extraction
92c67b9f FindLua: Prefix private APIs with `_lua_`
We no longer define the ZLIB_LIBRARY cache entry, so we should not mark
it as advanced either. We only use this variable when someone else
defines it in order to remain compatible with older versions of this
module, but we never create the entry ourselves anymore.
Extend the optimization from commit v3.4.0-rc1~264^2 (GetPrerequisites:
Optionally filter "objdump" output for speed, 2015-07-29) to work
on Windows by using `findstr` in place of `grep`.
FindBoost only detects Debug and Release configurations. All other
configurations will fall back to the configuration listed as the first
entry in `IMPORTED_CONFIGURATIONS`. Switch the order so that `Release`
is listed first, as this is a better fallback than `Debug` for the
`RelWithDebInfo` and `MinSizeRel` configurations. See issue #16091.
In commit v3.6.0-rc1~279^2~10 (VS: in Clang/C2 toolset, setup correct
compiler settings, 2016-02-18) a flag mapping was added for the clang
`-Os` flag. However, this collides with a mapping we already had for
the MSVC flag of the same name. This is a symptom of a larger problem
in that the VS generators need a per-toolset flag map (issue #16153).
For now, simply drop the new mapping and drop `-Os` from clang compiler
flags in the MinSizeRel configuration.
Reported-by: Felix Bruns <felixbruns@gmail.com>
Search for both release and debug library variants, and use
SelectLibraryConfigurations to choose the appropriate
library. Also add both release and debug libraries to the
imported targets.
When enabling the CUDA static runtime, the current module always uses
the shared runtime in the REQUIRED_VARS check. This change should
select the correct runtime to be checked for as required based on the
CUDA_USE_STATIC_CUDA_RUNTIME option.
Fixes#16096
Add a ``FIND_LIBRARY_USE_LIB32_PATHS`` global property analogous to the
``FIND_LIBRARY_USE_LIB64_PATHS`` property. This helps find commands on
multilib systems that use ``lib32`` directories and either do not have
``lib`` symlinks or point ``lib`` to ``lib64``.
0bd91ad4 UseJava: Fix race condition creating java class list
89df91b9 Help: Add notes for topic 'java-export-targets'
95d84369 Tests: Add test for exported JARs
5341c0d8 UseJava: Add infrastructure to export targets
d91ec044 Tests/Java: Clean up style of Java test code
Modify logic to unconditionally create an empty class list response file
to instead only create the empty file if no classes are being compiled.
Otherwise, the response file is created via a custom command that
generates the correct list.
This fixes a race condition on certain systems (e.g. HFS on OS X) where
the dependency for creating the "correct" list can have the "same" (due
to limited precision of the filesystem) timestamp as the empty file
created at configure time, resulting in an empty class list being
incorrectly used to invoke `jar` (resulting in `jar` being unhappy that
it was not given any input files). Although one would have some hope
that this doesn't affect "real" projects, it definitely does affect the
JavaExportImport unit test, causing the same to fail randomly when run
on an HFS partition.
Add additional functions to UseJava.cmake to create "exported" targets,
similar to those of C/C++ libraries. In support of this, ensure that the
JAR_FILE property of jar targets is always an absolute path.
63e5eb5f Help: Add notes for 'productbuild' topic
2e3c67d1 productbuild: Add new productbuild cpack generator.
50a3d340 PackageMaker: factor out common code for creating pkg files.
c3a281fd FindBoost: Search in official Windows prebuilt binary library directory
3720d5f4 FindBoost: Test `CMAKE_CXX_COMPILER_{ID,VERSION}` instead of `MSVC*`
We use file(GLOB) to load the list of files produced by compiling the
compiler identification source. Encode square brackets in the path to
the directory so that they are not treated as special characters in the
globbing expression. Otherwise we fail to find any files when the path
contains square brackets and the compiler id is unknown.
Reported-by: Esch Nigma <eschnigma@openmailbox.org>
In commit v2.6.0~3438 (ENH: Better support for adding and removing path,
2006-04-27) an include of "WinMessages.NSH" was added without removing
the original include. Remove it now.
Rename Modules/Platform/<os>-<lang>.cmake files to
Modules/Platform/<os>-Determine-<lang>.cmake to clarify their role.
For compatibility with user-provided modules, load the old names
if they exist.
If e.g. libc.a is passed before any objects the linker raises the follwing
warning: cannot resolve archive libc.a to a compatible library, as no
input files have been encountered
In the end the library is skipped and missing symbols occur.
Signed-off-by: Alexander Stein <alexander.stein@systec-electronic.com>
OS X supports using the SDK for any version equal to or newer than
the deployment target. There is no reason to warn if the versions
do not match exactly.
Suggested-by: James Burgess <jamesrburgess@mac.com>
Suggested-by: Clinton Stimpson <clinton@elemtech.com>
Follow the configure_file by a file(GENERATE) so the resulting cmake
scripts used by FindCUDA for wrapping nvcc calls can now support
generator expressions in the CUDA_NVCC_FLAGS variable.
Prevent accidental inter component dependency setting
since this is a breaking feature in situations where
another CPack module is already using the feature.
It should be enabled if desired since it can cause
issues when upgrading only one of the components
for e.g. configuration instead of all the components
at once.
This patch preserves backward compatibility of
deb package names with previous CMake versions
but similarly to CPack/RPM allows to change
package name format and supports DEB-DEFAULT
setting that produces proper Debian package names.
GIT_VERSION_STRING is unset in _ep_write_gitupdate_script(), hence
git stash is not being called with the "--all" option, even if Git
is new enough to support this.
Move the content to a `ExternalProject-download.cmake.in` file and use
`configure_file` to generate the final script.
Retry logic was not working before because similar script trigger
FATAL_ERROR if 'file(DOWNLOAD ...)' exits with nonzero 'status_code'.
FATAL_ERROR makes the whole chain of commands stop and
'_ep_write_verifyfile_script' retry logic was not used in fact.
Default retry number set to 5 with pauses 0, 5, 5, 15, 60 seconds. Some
space left for future improvements if needed (90, 300, 1200=20min). Can
be controlled by user.
The purpose of the `update` step is to run an update on each build
(subject to `UPDATE_DISCONNECTED`). This is done for version-controlled
source directories. We should do it for a custom `UPDATE_COMMAND` too.
In particular, when `UPDATE_DISCONNECTED` is used we expect the
`skip-update` step to exist.
The `CMAKE_PLATFORM_REQUIRED_RUNTIME_PATH` is used by CMake to always
append `-Wl,-blibpath:/usr/lib:/lib` to the link line. This is needed
by default on these platforms but needs to be overridden in some use
cases (e.g. an environment in which one maintains versioned shared
libraries). Change our logic to set this value only if it not already
set by the user, project, or toolchain file.
Run the `Utilities/Scripts/clang-format.bash` script to update
all our C++ code to a new style defined by `.clang-format`.
Use `clang-format` version 3.8.
* If you reached this commit for a line in `git blame`, re-run the blame
operation starting at the parent of this commit to see older history
for the content.
* See the parent commit for instructions to rebase a change across this
style transition commit.
This supplements the existing library checks, to
cater for the possibility that the libraries are
present but the headers are not. This can happen
when the Boost collections is split up into
multiple packages and not all are installed,
and will avoid the checks silently passing when
the build would subsequently fail.
Packagers may now set their own rpm package
file names or request that rpmbuild tool
chooses one for them. It also supports handing
of situations where one spec file may produce
multiple rpm packages.