Commit Graph

9148 Commits

Author SHA1 Message Date
Alexander Neundorf e10e3bc86e ENH: add compiler id for IAR compiler (http://www.iar.com/)
ENH: don't run endian test again if the variable is already set

Alex
2007-05-24 08:33:05 -04:00
Alexander Neundorf 1835c327ff STYLE: use "items" instead od "item"
Alex
2007-05-24 08:18:46 -04:00
Andy Cedilnik 4df8826d6c STYLE: Nightly Version update 2007-05-24 05:26:22 -04:00
Brad King 5639c69d39 ENH: Removed unused code now that INSTALL_FILES and INSTALL_PROGRAMS are not targets. 2007-05-23 18:23:17 -04:00
Brad King 9d75da9eaf ENH: Removed unused code now that INSTALL_FILES and INSTALL_PROGRAMS are not targets. 2007-05-23 18:22:52 -04:00
Brad King 6a145fb2be BUG: Report proper error message when project does not build the first time. Also added hack to rebuild subproject several times for Xcode. The generator should be fixed and the hack removed. 2007-05-23 17:58:14 -04:00
Brad King ea44d68cb6 BUG: Need to use GetRealDependency for custom command file-level dependencies. 2007-05-23 17:21:08 -04:00
Brad King c3474ad029 ENH: Executable bar should rebuild when its generated header changes. It does not need to link to the foo library anymore. 2007-05-23 17:01:26 -04:00
Brad King ba7780a3c4 ENH: Fixed INSTALL_FILES and INSTALL_PROGRAMS commands to not create targets. No targets of type cmTarget::INSTALL_FILES or cmTarget::INSTALL_PROGRAMS are created, so we do not need to check for them everywhere anymore. 2007-05-23 15:40:12 -04:00
Brad King 303b4312b2 BUG: Target names in the COMMAND part of a custom command should not create a file-level dependency that forces the command to rerun when the executable target rebuilds, but the target-level dependency should still be created. Target names in a DEPENDS should do both a target-level and file-level dependency. Updated the BuildDepends test to check that this works. 2007-05-23 13:30:23 -04:00
Brad King c288da754a BUG: Target names in the COMMAND part of a custom command should not create a file-level dependency that forces the command to rerun when the executable target rebuilds, but the target-level dependency should still be created. Target names in a DEPENDS should do both a target-level and file-level dependency. Updated the BuildDepends test to check that this works. 2007-05-23 13:27:00 -04:00
Brad King 702d785c9a BUG: Add ./ to custom command executables in the top of the build tree even when the path is generated by target name replacement. 2007-05-23 12:05:05 -04:00
Brad King 2d051a6709 ENH: Unify design of CMakeCCompilerId.c, CMakeCXXCompilerId.cpp, and CMakePlatformId.h. BUG: Do not violate system-reserved symbol namespace _[A-Z]. 2007-05-23 11:00:54 -04:00
Alexander Neundorf 526b72e12c COMP: don't user string::clear(), fix warnings about unused variables
Alex
2007-05-23 08:24:29 -04:00
Andy Cedilnik 1d84b36935 STYLE: Nightly Version update 2007-05-22 17:10:02 -04:00
Alexander Neundorf 6e2fd2c2ca BUG: now the toolchain file is configured into the buildtree, otherwise e.g.
CMAKE_SOURCE_DIR can't be used there
ENH: modify CMakeCCompilerId.c and .h so that sdcc can compile them. As they
were the preprocessor produced:

 9 "test.c"
static char const info_compiler[] = "INFO:compiler["
# 40 "test.c"
""

"]";

and the mixing of the preprocessing directives and the string constants
didn't work.

Alex
2007-05-22 12:48:16 -04:00
Alexander Neundorf bef8d3580b COMP: compile fix
Alex
2007-05-22 11:05:59 -04:00
Alexander Neundorf b37cc6072f COMP: compile fix and remove warning
Alex
2007-05-22 10:42:37 -04:00
Alexander Neundorf 4878c00905 ENH: add the IMPORT keyword to ADD_EXECUTABLE(), which generates an
"imported" executable target. This can then be used e.g. with
ADD_CUSTOM_COMMAND() to generate stuff. It adds a second container for
"imported" targets, and FindTarget() now takes an additional argument bool
useImportedTargets to specify whether you also want to search in the
imported targets or only in the "normal" targets.

Alex
2007-05-22 10:24:59 -04:00
Alexander Neundorf 6be930f0cf STYLE: move the two CMAKE_SHARED_LIBRARYC/CXX_FLAGS for gcc from
CMakeGenericSystem.cmake to gcc.cmake

Alex
2007-05-22 09:15:00 -04:00
Andy Cedilnik b759765134 STYLE: Nightly Version update 2007-05-22 04:14:07 -04:00
Alexander Neundorf 71359f5b9d STYLE: use a separate source file for generating CMakeSystem.cmake if CMAKE_TOOLCHAIN_FILE is used
Alex
2007-05-21 11:26:40 -04:00
Alexander Neundorf c8dd1caff9 BUG: don't fail if a compiler is given in CMAKE_C/CXX_COMPILER but it can't
be found in the path

Alex
2007-05-21 10:58:04 -04:00
Alexander Neundorf 3b4aa3bf35 BUG: always search for ar, ranlib, etc. except under MSVC -> this should fix the mingw fortran test
-also generate the fortran test with the kdevelop generator

Alex
2007-05-21 10:15:42 -04:00
Andy Cedilnik a7a1e40afa STYLE: Nightly Version update 2007-05-21 05:03:43 -04:00
Brad King 7416d0d67d BUG: Fix name of project to build for LoadCommandOneConfig now that it has been renamed for new name of LoadCommand test. 2007-05-20 10:11:48 -04:00
Brad King daca282db4 BUG: Use @ONLY substitution to configure CMakeSystem.cmake. 2007-05-20 10:08:38 -04:00
Andy Cedilnik fe1f147dc0 STYLE: Nightly Version update 2007-05-20 02:39:38 -04:00
Brad King 5e45ca6fbf BUG: Finish fixing test for new name. 2007-05-19 10:15:24 -04:00
Bill Hoffman 27a0c92ffc ENH: try to tone down kwstyle 2007-05-19 10:11:01 -04:00
Bill Hoffman e2ea9391e6 BUG: fix project name for test 2007-05-19 10:10:29 -04:00
Brad King 4259c7c1e4 COMP: Fix for borland now that components list check is const. 2007-05-19 09:55:18 -04:00
Andy Cedilnik 7cd767ca48 STYLE: Nightly Version update 2007-05-18 20:51:40 -04:00
Ken Martin df1a77c225 ENH: some cleanup, condensing some tests, removing arguments that were not needed but rather were cut and paste copies etc 2007-05-18 15:11:20 -04:00
Alexander Neundorf 9670a8c485 ENH: 2nd try to move stripping out of cpack and to install time, now if
CPACK_STRIP_FILES is true (or contains a list of files), everything will be
stripped, if it's empty or false they won't be stripped

Alex
2007-05-18 14:41:09 -04:00
Alexander Neundorf df8f550029 ENH: add install/strip target for makefile generators if strip was found
Alex
2007-05-18 13:43:39 -04:00
Alexander Neundorf a18d286635 ENH: move hack to fix "new cmake on old build tree on OSX doesn't have CMAKE_INSTALL_NAME_TOOL in the cache" from
cmInstallTargetGenerator.cxx to Darwin.cmake

Alex
2007-05-18 11:57:29 -04:00
Alexander Neundorf d64dc9b5bb COMP: force a recompile on VS71
Alex
2007-05-18 11:45:18 -04:00
Brad King 92c0f41251 ENH: Use CMAKE_PLATFORM_IMPLICIT_LINK_DIRECTORIES from platform files to block link directories. 2007-05-18 11:36:23 -04:00
Alexander Neundorf 15a7414a60 COMP: if a new cmake runs on an old build tree, set CMAKE_LINKER to link to make it link
Alex
2007-05-18 10:55:35 -04:00
Alexander Neundorf 891f2130cf COMP: fix link rules with nmake, the linker command has to be converted to shortpath form for nmake
Alex
2007-05-18 10:32:22 -04:00
Brad King 9bfb3cacbb BUG: Replace "with space" in custom command argument tests with "w s" to still have whitespace but be shorter. The test was failing because the custom command line length was simply too long for the VS IDE. 2007-05-18 09:33:59 -04:00
Bill Hoffman fda955e752 STYLE: fix documentation for command 2007-05-18 09:30:21 -04:00
Brad King ea66edaf5e ENH: Added quick means to turn on verbose output for debugging this test. 2007-05-18 09:18:09 -04:00
Brad King 07fe611fc6 BUG: Added carrot (^) to characters that need quoting. The solaris shell needs it. 2007-05-18 09:17:36 -04:00
Alexander Neundorf d9bbc3f45f STYLE: fdcorrect comments about FC/CC
Alex
2007-05-18 09:16:07 -04:00
Brad King 8780da000a BUG: If the Fortran CompilerId source fails to compile it should not be a failure. It is only expected to work for Fortran90 compilers. 2007-05-18 09:08:49 -04:00
Alexander Neundorf aee311a89d STYLE: fix line lengths
Alex
2007-05-18 08:49:06 -04:00
Brad King 42a272428b BUG: Need to use GetSafeDefinition when assigning to a string. 2007-05-17 17:43:32 -04:00
Brad King 9a5b4eba97 BUG: All variable accesses should produce watch callbacks, including IF(DEFINED <var>) ones. Instead we define a new access type for IF(DEFINED) so that the error does not show up for backward compatibility variables. 2007-05-17 17:40:59 -04:00