| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
|
|
|
| |
am deducing that the value of CMAKE_OSX_ARCHITECTURES_DEFAULT is responsible for the failure, although I cannot reproduce it on other builds or even by running the test via ctest interactively *on* the continuous dashboard's build...
|
|
|
|
| |
that runs on 10.3 and earlier does not understand the compiler flag it maps to...
|
|
|
|
| |
specify the target deployment runtime OS version of the built executables on Mac OSX. Thanks to Mike Jackson for the patch.
|
| |
|
|
|
|
| |
with Qt 4.5 on Mac.
|
| |
|
|
|
|
| |
speaking, this changes behavior from cpack 2.6, but now that cpack returns a non-zero exit code when it encounters an error, and it is an error to try to use a generator that is not available... It makes sense to turn these off by default since not everybody has these generators installed. It is easy for a project to turn these options back on if they need to: simply set(CPACK_BINARY_TBZ2 ON) or set(CPACK_BINARY_ZIP ON) before include(CPack) in your CMakeLists.txt...
|
|
|
|
| |
doesn't have to have everything in order for FLTK_FOUND to be true. #7809
|
|
|
|
| |
of CPack. Add a test of it in the CPackComponents test. Thanks to Clinton Stimpson for the patch.
|
| |
|
|
|
|
|
| |
only looks for FL/Fl.h. Verified: all FLTK header files in 1.1.9 are .H ...
how bizarre.
|
|
|
|
| |
installers for OSXX11 applications. Also modify the X11 test to build such an installer on Mac builds that test CPack and have X11 available. Thanks to Wes Turner for the patch.
|
| |
|
|
|
|
|
| |
Gentoo 1.37 system installed boost). Fixes issue #8404 reported on mailing
list.
|
|
|
|
|
|
|
| |
mark_as_advanced variables removed, documentation cleaned up and OSX stuff
isolated to it's own section, support added for DOXYGEN_SKIP_DOT, support
added to call FindPackageHandleStandardArgs to avoid output on every CMake
run.
|
|
|
|
| |
gcc -dumpversion code and regex to a function.
|
| |
|
|
|
|
| |
OPENTHREADS_LIBRARIES.
|
| |
|
| |
|
|
|
|
| |
examples, addressed autolinking issue on MSVC
|
|
|
|
| |
prior to 1.35. Eliminate "lib" prefix except on MSVC.
|
|
|
|
| |
including Issue #6926
|
|
|
|
| |
called more than once (Issue #7460)
|
| |
|
|
|
|
| |
the auspices of forcing the user to use autolinking, but it only did this squashing on the first call to FindBoost. Subsequent calls to FindBoost would not have Boost_LIBRARIES squashed so this code was doing nothing. If you link your target_link_libraries() against dynamic boost libraries it appears from tools like Dependency Walker that the pragma calls to autolink to the static boost libraries are ignored. It's therefore too late to make this squash apply to all calls to FindBoost because that would break users that have not setup autolinking properly. For now this fix is largely cosmetic since the original code never worked anyways (see version 1.5 introduced on 4/22/08).
|
|
|
|
| |
call since it would be useless and confusing after an initial configure.
|
|
|
|
| |
workaround for Issue #8378. Resolves Issue #8099
|
|
|
|
|
| |
regression for bjam users on Win32 introduced in 1.4.2.4 (7/13/08). This
commit partially or completely resolves Issues #8173, #8326, #7943, #7725!
|
|
|
|
| |
reporting version x.y and not x.y.z
|
| |
|
|
|
|
| |
Alex
|
|
|
|
|
|
| |
but just a macro argument -> make it a real variable
Alex
|
|
|
|
| |
Alex
|
|
|
|
|
|
|
|
|
| |
now generates the project so that the environment variable VERBOSE is set to
1 when make is executed by kdevelop (and additionally this didn't work,
since CMAKE_GENERATOR never matches KDevelop3, this is now in
CMAKE_EXTRA_GENERATOR)
Alex
|
|
|
|
|
| |
Also eliminated superfluous CXXTEST_FOUND assignment and cleaned up the code
and added additional documentation. Tagged v1.0.
|
|
|
|
| |
followed by multiple calls to ENABLE_LANGUAGE. Use find_program to set the CMAKE_INSTALL_NAME_TOOL variable so it gets saved in the cache as a full path to the install_name_tool executable rather than a simple set which eventually goes out of scope.
|
| |
|
|
|
|
| |
it properly by using the SYMBOLIC source file property to indicate to WMake when the sentinel file is not actually written by the update step.
|
|
|
|
| |
some cases to breakaway from tests without resetting to FALSE (BUG: 8188).
|
|
|
|
|
|
| |
svn and fix bug #8290
Alex
|
|
|
|
| |
UPDATE_COMMAND handling. Output a -complete sentinel in synch with the -install sentinel, but do not list it as an OUTPUT of the custom command. That breaks the chaining of add_custom_commands between custom targets, but allows for a file-level dependency expression that will cause proper incremental rebuilds. When earlier targets rebuild, subsequent dependent targets will also rebuild. CVS and SVN update commands are always out-of-date so that they always run to get the latest source. To suppress that behavior on a per-external project basis use an explicit empty string for UPDATE_COMMAND. The source will still be checked out from the repository prior to the update step by the download step.
|
| |
|
|
|
|
|
|
| |
On 64-bit Windows there may be two Program Files folders, one for 32-bit
binaries and one for 64-bit binaries. When we compute
CMAKE_SYSTEM_PREFIX_PATH we should put both folders in the path.
|
|
|
|
|
| |
Add QT_TRANSLATIONS_DIR pointing to the Qt translation files, and docs for it.
Also add docs for QT_BINARY_DIR.
|
|
|
|
| |
testing framework within CTest
|
|
|
|
| |
installing. If none specified default to the cmake used to configure the outer/aggregating project.
|
|
|
|
| |
CMake, and running Squish tests from ctest
|