The compiler documents symbols _DF_VERSION_ and _VF_VERSION_ but they do
not seem to be available to the preprocessor. Instead we add a vendor
query table entry for Compaq. Running "f90 -what" produces
Compaq Visual Fortran Optimizing Compiler Version ...
This clearly identifies the compiler.
At least one Fortran compiler does not provide a preprocessor symbol to
identify itself. Instead we try running unknown compilers with version
query flags known for each vendor and look for known output. Future
commits will add vendor-specific flags/output table entries.
PathScale Fortran mangles module symbols as "MYSUB.in.MYMODULE" and also
requires "mymodule_" when the module is imported. We cannot provide the
symbol with ".in." mangling so we should not provide "mymodule_" because
it would duplicate the one in the Fortran-provided object file.
In commit "use export all symbols on cygwin" (2003-01-21) we started
passing -Wl,--export-all-symbols when linking shared libraries. Now
cygwin exports all symbols automatically if no symbols are explicitly
exported. When symbols are explicitly exported we want to honor that
narrow interface. Therefore this flag should not be passed.
Change based on patch from issue #10122.
The variable should contain the name of a library needed to link the
symbol equivalent to dlopen. On Cygwin no special library is needed,
and certainly not "gdi32".
Change based on patch from issue #10122.
After discussing with Brad and Clinton:
-the namespace for the imported targets is now "Qt4::", tested with Makefiles, Visual Studio and XCode projects
-the imported targets are always created
-if QT_USE_IMPORTED_TARGETS is set to TRUE (it defaults to FALSE), the QT_QTFOO_LIBRARY variables are set to point to these imported
targets, otherwise the old behaviour is used.
-on OSX if Qt has been found as framework, disable QT_USE_IMPORTED_TARGETS, since cmake doesn't handle the framework directory as location of the library correctly
Alex
-set the type of the IMPORTED libraries to UNKNOWN, this way also on Windows
only the "LOCATION" property has to be set
-the if() around the SET(QT_${basename}_FOUND 1) was useless (always true)
-the mapping of the configuration types DEBUG and PROFILE did not belong here
Alex
This commit syncs FindQt4.cmake again with KDEs version.
Now for every Qt library an imported target with the name
Qt4ImportedTarget__<LIBNAME> is created.
This way we can now finally handle the release and debug versions of the Qt
libraries correctly.
Also, if a Qt-using project A installs a file with exported targets, these
targets now depend on the imported Qt targets, e.g.
Qt4ImportedTarget__QtCore. The location of QtCore is then resolved at
buildtime of project B, which uses the exported targets from project A.
Before this patch the full path to the QtCore on the original build machine
of project A was stored, so this had to match the directory layout on the
build machine for project B.
Alex
While Cygwin supports linking directly to .dll files, the behavior is
now discouraged. All Cygwin packages now provide import libraries of
the form lib*.dll.a and CMake has built the import libraries for years.
We believe it is now safe to stop explicitly searching for .dll files
because their import libraries will always be available when the
corresponding header files are available. Users can always set
find_library cache entries to point at a .dll file by hand if they
really must use one.
Change based on patch from issue #10122.
Commit "Find locally installed software first" made /usr/local the first
prefix searched to be consistent with the Filesystem Hierarchy Standard:
http://www.pathname.com/fhs/
The standard also implies that the root prefix "/" should not have any
package or development files. The "/bin" and "/lib" directories should
have only minimal contents to boot the system. No "/include" ever
exists. This commit re-orders the search path prefix list from
/usr/local
/
/usr
to
/usr/local
/usr
/
to prefer package and development files over low-level system files.
See issue #10136.
On Cygwin /usr/lib == /lib and /usr/bin == /bin. This change also makes
search results report locations as "/usr/..." instead of "/lib/...".
See issue #10122.
We add compiler information files
Compiler/PathScale-<lang>.cmake
to specify PathScale compiler information for C, C++, and Fortran
languages. We use a macro in Compiler/PathScale.cmake to consolidate
the information common to all languages.
The commit "Drop -rdynamic from Linux build rules" removed default use
of the flag on Linux. It was expected to be compatible because any
project using plugins should set ENABLE_EXPORTS on its executables to
export their symbols for use by the plugins in a cross-platform way.
However, it is possible to build without ENABLE_EXPORTS and load plugins
that do not link to any symbols from the executable explicitly. These
plugins may need to see RTTI and other executable symbols needed by the
language implementation. Executables using such plugins were broken by
the change.
If we want to remove the -rdynamic flag in the future we should do so in
a compatible way. At that time we should also remove equivalent flags
on other platforms (like -bexpall on AIX). We will either need a policy
or an explicit API to disable symbol exports on executables.
The primary purpose of the above-mentioned commit was to avoid passing
the -rdynamic flag to compilers on Linux that do not support it. In
this commit we restore the flag but only on GNU and Intel compilers
which are known to support it.
See issue #9985.
Implement Fortran 32/64-bit ABI detection on some platforms. We need to
set CMAKE_SIZEOF_VOID_P correctly in Fortran-only projects so that the
find_library() command knows whether to look for 64-bit binaries. We
also detect ELF binaries to enable RPATH replacement. See issue #10119.
Now the case that both the release- and the debug-version of a library is
handled first, because otherwise we always ran into this branch, since the
debug-only and the release-only branch also set both variables.
Alex
We modify the signature of _HDF5_parse_compile_line to pass the command
line variable name rather than the command line itself. Otherwise the
CMake language MACRO implementation tries to parse the command line as
CMake syntax, which does not like backslashes.
We re-implement this module to support architecture-dependent type
sizes. In the mixed-size case we generate C preprocessor code to select
the detected type size for each architecture.
before this patch -F<framework> dir had to be added manually in some way
when using Qt4 installed as framework and when using FindQt4.cmake directly,
i.e. without UseQt4.cmake. With this patch the framework dir is
automatically added to QT_INCLUDE_DIR when Qt is installed as a framework.
Ok by Clinton, tested already in KDE by Mike Arthur.
Alex
The commit "FortranCInterface: Honor language flags in checks" taught
the FortranCInterface module to pass C and Fortran flags into its
detection and verification checks. We improve on the change to allow
the '=' character in the language flags. This requires passing the
cache entry type with the -D options.
CMake 2.8 was released with the FindHDF5 module setting HDF5_INCLUDE_DIR rather
than the correct plural HDF5_INCLUDE_DIRS. Since this went into a release, it is
now necessary to set the singular for backwards compatibility.
Replace them with CPACK_PACKAGE_NAME. The registry keys involved in this commit are used by Windows to track things in the Add/Remove programs portion of the Control Panel. With '\' characters in the keyname, the calls do not do what they are intended to do and the installed program never shows up in the control panel view. (Details noted in the issue itself.) Thanks to 'killerfox' for the patch.
Default to "" for CMAKE_OSX_DEPLOYMENT_TARGET if CMAKE_OSX_SYSROOT is set. Also, add new error message to detect the case where there is a deployment target, but no SDK has been set. Fix args to STRING REGEX call so that it works even if _sdk_path variable is empty inside sanity check function.
When there is no shared object to link to a second call to find library is
necessary to find the static Python library. Fixes an issue raised on the CMake
mailing list, and it should be included in the next CMake patch release.
We add the macro CMAKE_FORCE_Fortran_COMPILER to the cross-compiling
helper module CMakeForceCompiler.cmake so that toolchain files can force
a Fortran compiler as well as C and C++ compilers. See issue #10032.
CMake does not enable Fortran for its own build, but it needs to find a
Fortran compiler to know if it is possible to enable Fortran tests.
Previously we searched for a hard-coded list of Fortran compilers which
was duplicated from the CMakeDetermineFortranCompiler.cmake module. We
now run CMake on a small test project that enables the Fortran language
and reports the compiler it found. This represents a more realistic
check of whether the Fortran tests will be able to find a compiler.
Previously this module gave only very brief documentation. We extend
the module's documentation to describe CTestConfig.cmake, interaction
with dashboard scripts, and the CTEST_USE_LAUNCHERS option.
We remove the shared library compile/link flags "-fPIC" and "-shared"
because they are not provided by all compilers on Linux. This allows us
to drop code from the Linux-XL-*.cmake files that erases the bad flags.
All other supported compilers already provide their correct flags for
Linux in their own platform information files.
We factor flags from Platform/Linux-PGI-Fortran.cmake into language
independent helper modules
Compiler/PGI.cmake
Platform/Linux-PGI.cmake
and invoke the macros from
Compiler/PGI-<lang>.cmake
Platform/Linux-PGI-<lang>.cmake
This enables general support for the PGI compilers.
The commit "Split GNU compiler information files" intended to move GNU
flags from the platform-wide Platform/SunOS.cmake module into
Platform/SunOS-GNU-<lang>.cmake
using a helper module Platform/SunOS-GNU.cmake to consolidate flags.
However, it accidentally put Fortran flags in the C language module and
left out the Fortran module altogether. This fixes those mistakes.
Several platform-wide linker flag variables are defined in
Modules/Platform/<os>.cmake files for C and then copied by the
Modules/CMake<lang>Information.cmake file for each language.
We now use this approach for the variables
CMAKE_EXE_EXPORTS_${lang}_FLAG
CMAKE_SHARED_LIBRARY_SONAME_${lang}_FLAG
CMAKE_SHARED_LIBRARY_CREATE_${lang}_FLAGS
to avoid duplication for multiple languages in each platform file.
The commit "Split GNU compiler information files" broke the settings of
CMAKE_SHARED_LIBRARY_CREATE_${lang}_FLAGS
CMAKE_SHARED_MODULE_CREATE_${lang}_FLAGS
and started using just "-shared" for them. This worked when tested on newer
Mac machines, but older ones really need "-dynamiclib" and "-bundle" (which are
the documented flags anyway).
This moves GNU compiler info on Windows into new-style modules
Platform/Windows-GNU-<lang>.cmake
using language-independent helper module
Platform/Windows-GNU.cmake
to define macros consolidating the information.
This moves GNU compiler flags into new-style modules
Compiler/GNU-<lang>.cmake
Platform/<os>-GNU-<lang>.cmake
We use language-independent helper modules
Compiler/GNU.cmake
Platform/<os>-GNU.cmake
to define macros consolidating the information.
The CMakeBackwardCompatibilityC module provides some try-compile results
that were automatically provided by CMake 1.4. When performing the
checks for OS X universal binaries we just pick one architecture to get
through the checks without error. Since CMake 1.4 did not support any
universal binaries, projects that want them should not depend on this
compatibility module anyway.
This is a GNU-specific option that should not be specified for all
compilers on Linux. It tells the GNU compiler to pass -export-dynamic
to the linker to export symbols from executables for use by plugins.
Since we provide the ENABLE_EXPORTS target property to do the same thing
in a cross-platform way, there is no need to pass -rdynamic always.
Since the option is not useful for GNU tools and breaks other tools on
Linux we simply remove it from CMAKE_SHARED_LIBRARY_LINK_<lang>_FLAGS.
This also allows us to stop setting the variable in other Linux compiler
files just to erase the bad flag.
See issue #9985.
The Watcom tools do their own command-line parsing and do not accept
double-quotes. Instead we single-quote the target output name when
invoking wlink and other Watcom tools. This fixes support for spaces in
the target output directory path when it is not under the build tree.
-use find_package(PkgConfig) instead of include(UsePkgConfig)
-remove the "if already cached make silent" logic, this is already handled by find_package_handle_standard_args()
-remove the if(WIN32) around pkg-config, it shouldn't be necessary
Alex
We pass CMAKE_C_FLAGS, CMAKE_CXX_FLAGS, and CMAKE_Fortran_FLAGS through
try_compile() for the FortranCInterface Detect and Verify projects.
This honors user-specified compiler flags for each language, thus
supporting flags that affect the Fortran mangling.
The FortranCInterface module should execute with CMake 2.8.0 behavior
even if policies are set differently by the including project. In
particular, it makes use of empty list elements and therefore expects
NEW behavior of CMP0007.
Qt4Macros.cmake: all the "public" macros of FindQt4.cmake
Qt4ConfigDependentSettings.cmake: the code for detecting the
Qt-configuration dependent additional libraries, e.g. when linking statically.
There should be no functional changes in this patch.
The patch reduces the length of FindQt4.cmake from 1700 lines to around 1000
lines, which is still long enough, but this should make the file a easier to
handle (and it is similar to what we do in KDE with FindKDE4Internal.cmake
and KDE4Macros.cmake)
Ok by Clinton.
Alex
This should not change the result (since both should be in the same
directory), but seems a bit more logical and is also in sync with what is done in the KDE version.
Alex
-the mark_as_advanced() calls for the variables coming from qmake are now in
the corresponding section, and not in the section where the include dirs are
foudn
Alex
Some compilers use implicit link options of the form
-lcrt*.o
-lgcc*
-lSystem (on Mac)
-lSystemStubs (on Mac)
that provide system-wide symbols not specific to any language.
These need not be listed explicitly for mixed-language linking.
We teach CMake to remove the above items from the implicit library list
of each language. This change makes it possible to mix GNU compiler
versions in some cases.
BUILD_SHARED_LIBS is now only recognized when calling CUDA_ADD_LIBRARY. If you want the CMAKE_SHARED_LIBRARY_C/CXX_FLAGS to be used, pass SHARED as an argument. This prevents -fPIC from being used on objects destined for executables by default.
We replace "/MD" with ifort-specific flags as follows:
/MD -> /threads /libs:dll
/MDd -> /threads /libs:dll /dbglibs
We also enable the "/MD" equivalent for all Fortran configurations.
Previously multithreaded dll runtimes were used for release builds and
threaded static runtimes for debug builds. For mixed Fortran C/C++
projects, this led to link warnings for Debug but not for Release.
See issue #8744.
We add Intel and MinGW Fortran linker options to create the import
library portion of a DLL. This allows other binaries to link to a
Fortran DLL.
We also update the Fortran test to use a .def file to specify exports
since there is no __declspec(dllexport) markup syntax in Fortran.
This commit teaches CMake about the g95 compiler from
http://www.g95.org
We use 'G95' as the compiler id string, and add some basic flags.
See issue #9241.
although #9621 did not happen anymore with cmake 2.8.0, probably because
GET_PROPERTY(... GLOBAL ...) now makes the result variable empty instead of
simply not touching it, using FUNCTION() instead of MACRO() makes sure that
the _EnabledFeatures variables is always empty before the GET_PROPERTY()
call (and does not still have the old value from the previous call)
Alex
1. Add STATUS output "Could NOT find Boost" if boost is not found
which brings FindBoost closer in behavior to most CMake find modules.
2. Add an option: Boost_DETAILED_FAILURE_MSG to output
Boost_ERROR_REASON on a non-REQUIRED find if this
is desired by the developer. This is done because the error messages
are rather long and software with optional Boost dependencies might
not like them showing up by default, especially since this wasn't done
before.
3. Add mention of Boost_ADDITIONAL_VERSIONS close to top
of file since this seems to be the most common problem brought
up on the mailing list (maybe people will notice it there)
4. Added additional check for intel compiler which probably isn't
necessary but ultimately should be cleaner if CMAKE_CXX_COMPILER_ID
sticks around.
5. Added my name to the Copyright list
This commit re-writes Borland compiler build rules. We split the rules
into modern <os>-<id>-<lang> information modules but share a common
macro between languages to avoid duplication.
We also address a bug in the previous rules that would build some target
types against the static Borland runtime and others against the shared
Borland runtime in one build tree. Now we always use the shared runtime
as is the default in the rules for MS tools.
This teaches the FindVTK module to use the Config mode of find_package()
to search for VTKConfig in the common case. The old search method based
on find_path() and UseVTK is now used only to search for VTK 4.0.
This approach avoids the need to update the module for each new VTK
version because find_package(VTK) automatically searches "lib/vtk*". It
also addresses issue #9105 since find_package searches lib64 paths too.
The regex used by CMAKE_PARSE_IMPLICIT_LINK_INFO to detect link lines
should not match lines that happen to have ".../ld.../..." in them. A
linker name should match only as the last component of a path.
See issue #9666.
This commit teaches the CMAKE_PARSE_IMPLICIT_LINK_INFO function to log
the regex it uses to detect the linker invocation line. The regex is
computed from the CMAKE_LINKER if it is found, so it might change. A
strange value might match the wrong line and cause implicit link info
extraction to fail.
See issue #9666.
This commit teaches the CMAKE_PARSE_IMPLICIT_LINK_INFO function to log
its actions. We store the log in CMakeFiles/CMakeOutput.log at the top
of the project build tree. This will make diagnosis of implicit link
information parsing problems easier.
The commit "fix for bug 4772" added a work-around to CMakeCInformation
for platform config files that put compiler information in the system
file (like SunOS flags for GCC). This commit adds the same work-around
for CXX and Fortran. It is necessary to support enabling these
languages separately from C in other subdirectories.
The commit "Avoid (Unix|Windows)Paths.cmake multiple include" added
include blockers that prevent the files from multiple inclusion, so it
is safe to include the system information files from every language.
See issue #4772 and issue #9656.
Block multiple inclusion because "Modules/CMakeCInformation.cmake"
includes "Platform/${CMAKE_SYSTEM_NAME}" even though the generic
module "CMakeSystemSpecificInformation.cmake" already included it.
The extra inclusion is a work-around to address issue #4772 without
intrusive platform file changes. Once those changes are made the
work-around and these include blockers can be removed. See issue #9656.
This commit re-orders the search path prefix list from
/
/usr
/usr/local
to
/usr/local
/
/usr
so that locally-installed software is preferred.
This makes the search consistent with the Filesystem Hierarchy Standard:
http://www.pathname.com/fhs/
See issue #9657.
Most problems are fixed (or rather worked-around) by making long '====='
separators pre-formatted (i.e. prefixed with two spaces). In order to
preserve visual view, the code examples themselves are prefixed with 3
spaces.
This commit fixes the following man warnings:
$ cmake --help-man - | LANG=C MANWIDTH=80 man --warnings -l - > /dev/null
<standard input>:6024: warning [p 105, 1.7i]: can't break line
<standard input>:6027: warning [p 105, 2.7i]: cannot adjust line
<standard input>:6027: warning [p 105, 2.8i]: can't break line
<standard input>:7142: warning [p 117, 7.8i]: can't break line
<standard input>:7171: warning [p 117, 11.8i]: can't break line
<standard input>:8878: warning [p 136, 9.0i]: can't break line
<standard input>:8887: warning [p 136, 11.5i]: cannot adjust line
<standard input>:8887: warning [p 136, 11.7i]: can't break line
<standard input>:8904: warning [p 136, 14.2i]: can't break line
Patch from Modestas Vainius. See issue #9659.
GNU/kFreeBSD = FreeBSD kernel + userspace with glibc. Linux.cmake
doesn't contain anything too OS specific, so we can forward to it.
Here are outputs of /bin/uname on author's machine:
uname -p ==> i386
uname -o ==> GNU/kFreeBSD
uname -s ==> GNU/kFreeBSD
uname -r ==> 5.4-1-686
Patch from Modestas Vainius. See issue #9659.
On Debian moc and uic from Qt3 have '-qt3' suffixes. The latter names
might come from Qt4, so prefer the version-specific names.
Patch from Modestas Vainius. See issue #9659.
This module was contributed under a BSD-like license. We added CMake's
OSI-approved BSD License on top of it. With the author's permission,
this commit removes the old license text and puts the author's copyright
notice in the block referring to the new license.
-remove the parentheses in the if() conditions, they don't change the
result, and without them these lines are identical to the ones in
FindQt4.cmake from KDE
-mention which qmake was used for finding Qt
Alex
This commit creates target and directory properties to enable the Intel
interprocedural optimization support on Linux. Enabling it adds the
compiler option '-ipo' and uses 'xiar' to create archives.
See issue #9615.
Other minor changes:
* Enhanced documentation & added examples
* _INCLUDE_DIRS and _LIBRARIES was being set regardless of _FOUND
* Fixed status message to include library rather than include dir
* Improved detection of MSVC compiled libraries
* Added a variable (GTEST_BOTH_LIBRARIES) for libgtest + libgtest_main
This commit adds some default initial C flags for the XL compiler. The
most important is "-qhalt=e" which causes the compiler to error-out on
non-severe error messages. This is necessary to get try-compiles to
fail when bad arguments are passed to a function.
This moves platform-independent XL compiler flags into separate
"Compiler/XL-<lang>.cmake" modules. Platform-specific flags go in
"Platform/<os>-XL-<lang>.cmake" modules.
Since Haiku does not have /usr (and therefore /usr/local), this commit
changes the default install prefix to the equivalent directory of
/boot/common.
See issue #9607.
Cutil was never intented to be used outside of the SDK. The removal of this
code is in support of this. The CUDA_SDK_ROOT_DIR will continue to be
supported, in case users wish to use this to find files in the SDK. There are
also two examples of how to use CUDA_SDK_ROOT_DIR to find header files and
libraries if users so wish.
The CUDA_PROPAGATE_HOST_FLAGS was incorrect in that it prevented the CUDA_NVCC_FLAGS_CONFIG variable from getting filled.
Also, added a search path for the CUDA SDK install on Macs.
Brad King
David Cole
James Bigler
Alexander Neundorf
Clinton Stimpson
Philip Lowman
Zach Mullen
Bill Hoffman
Will Dicharry
Marcus Hanwell
Mathieu Malaterre
Eric Noulard
Dave Partyka