Commit Graph

473 Commits

Author SHA1 Message Date
Brad King 3a6c480908 Merge topic 'cygwin-module-prefix'
1dcc977 Cygwin: Use 'cyg' prefix for module DLLs (#10122)
2010-12-16 14:00:06 -05:00
Brad King 72163de0a5 Merge topic 'NAG-Fortran'
09d1c10 FortranCInterface: Recognize NAG Fortran module symbols
af2ad90 Add NAG Fortran compiler information files
24cc3d4 Recognize the NAG Fortran compiler
83892c4 Allow Fortran platform files to set empty values
fe3f878 Detect object files in implicit link information
2010-12-16 13:59:52 -05:00
Yaakov Selkowitz 1dcc9777a7 Cygwin: Use 'cyg' prefix for module DLLs (#10122)
Cygwin now uses the prefix 'cyg' for plugin DLLs instead of 'lib'.
2010-12-13 14:21:07 -05:00
Brad King af2ad90991 Add NAG Fortran compiler information files
On Linux the NAG Fortran compiler uses gcc under the hood to link.  Use
"-Wl,-v" to pass "-v" to the underlying gcc compiler to get verbose link
output.  Detect the NAG Fortran directory (using -dryrun) and then honor
object files in the directory referenced in the implicit link line.
Pass real linker options with "-Wl,-Xlinker,".  The -Wl, gets through
the NAG front-end and the -Xlinker gets through the gcc front-end.
2010-12-09 18:12:34 -05:00
Bill Hoffman cddcad5102 Fix incremental linking for VS2010 with nmake or make.
VS2010 deprecated /INCREMENTAL:YES.  This change makes
/INCREMENTAL the flag to use for incremental linking with
VS2010.
2010-12-09 13:32:48 -05:00
Brad King eda7841fd2 Pass Mac linker flag through PGI compiler using "-Wl,"
The Mac linker defines -headerpad_max_install_names and the GCC
front-end passes this flag through.  The PGI compiler does not know
about this flag, so we must use -Wl,-headerpad_max_install_names to pass
it to the linker instead.
2010-11-22 08:43:13 -05:00
Brad King 14d5ea7438 Merge topic 'mingw-long-object-lists'
5f05a3c MinGW: Support long object file lists
2d9bb33 Evaluate <OBJECT_DIR> rule variable for executables
2010-09-22 13:52:24 -04:00
Brad King b7539e3832 Merge topic 'compaq-fortran-moddir-flag'
c04af06 Add module-dir flag for Compaq Visual Fortran (#11248)
2010-09-21 10:41:02 -04:00
Brad King c04af06009 Add module-dir flag for Compaq Visual Fortran (#11248)
This compiler uses "-module:" to specify where modules should be placed.
2010-09-20 09:53:45 -04:00
Brad King 5f05a3c25e MinGW: Support long object file lists
Use a combination of response files and the archiver to support long
object file lists that do not fit in the Windows command-line length
limit.  This can work only with GCC >= 4 because the MinGW GCC 3.x
front-ends do not support response-file syntax.
2010-09-17 09:25:36 -04:00
Brad King 048c905f05 Merge topic 'hpux-required-runtime-path-10571'
11a917d HP-UX: Always add /usr/lib to rpath (#10571)
2010-09-08 11:08:13 -04:00
Brad King 02e3f42a6a Merge topic 'watcom_fixes'
3b7da53 Fix for bug 10388, fix various default flags.
2010-09-08 11:08:01 -04:00
Brad King 11a917d04e HP-UX: Always add /usr/lib to rpath (#10571)
Set CMAKE_PLATFORM_REQUIRED_RUNTIME_PATH in the HP-UX platform file to
tell CMake to pass -Wl,+b,/usr/lib no matter whether RPATH is enabled or
not.  This corrects the behavior of -Wl,+nodefaultrpath to look in this
default library path as documented.
2010-09-07 08:44:19 -04:00
Bill Hoffman 3b7da5396d Fix for bug 10388, fix various default flags. 2010-09-02 14:35:17 -04:00
Brad King eae45a67e7 Search MacPorts /opt/local prefix on Mac
Include this prefix in CMAKE_SYSTEM_PREFIX_PATH so that it will be used
for all find* commands.  Previously only find_library and find_path
would look under /opt/local/lib and /opt/local/include, respectively.
2010-09-01 08:50:51 -04:00
Brad King da6fc97f19 Merge topic 'tg/BlueGeneP'
7083c81 Add platform files for BlueGene/P systems
2010-08-17 15:14:07 -04:00
Brad King 32ebe21ad8 Merge topic 'intel-response-files'
4b1e5f0 Pass objects to Intel linker using a response file
2010-08-17 15:09:12 -04:00
Todd Gamblin 7083c81db7 Add platform files for BlueGene/P systems
BlueGeneP-base:    Internal base shared by static and dynamic files
BlueGeneP-static:  Platform file for all-static builds
BlueGeneP-dynamic: Platform file for "default" dynamic builds
2010-08-13 12:17:57 -04:00
Brad King cc31f89c17 Merge topic 'module-header-spelling'
2cde67a Modules: Fix spelling 'To distributed' -> 'To distribute'
2010-08-10 14:33:47 -04:00
Brad King 4b432328a2 Merge topic 'mingw-response-files'
b03f4ec No response files with GNU ld <= 2.16 (#10913)
2010-08-10 14:33:01 -04:00
Arjen Verweij 4b1e5f0152 Pass objects to Intel linker using a response file
Use response files for C and CXX languages with the Intel compiler on
Windows.  We already used them for Fortran.  This enables creation of
libraries and executables with a very large number of object files.
2010-08-10 09:34:12 -04:00
Todd Gamblin 2cde67a781 Modules: Fix spelling 'To distributed' -> 'To distribute' 2010-08-09 08:48:31 -04:00
Brad King b03f4ec09d No response files with GNU ld <= 2.16 (#10913)
Older GNU ld does not support the @FILE syntax for response files.
Check the ld version on MinGW and MSYS before enabling the syntax.
2010-08-06 09:34:53 -04:00
Pino Toscano 90467d6294 GNU/Hurd platform support fixes (#9873)
Define use of -rpath-link and --export-dynamic flags.  The toolchain on
Hurd is GNU-based so we can use the values from Linux.
2010-07-29 10:43:03 -04:00
Brad King 5444bd6ca3 Merge branch 'tru64-make-includes' 2010-06-15 14:03:26 -04:00
Brad King d9b2da139d Merge branch 'mingw-response-files' 2010-06-15 13:58:58 -04:00
Brad King c592df8377 Tru64: Use full-path include directives in Makefiles (#10569)
Tru64's make(1) resolves relative paths in "include" directives with
respect to the includer.  This is inconsistent with all other known make
tools.  Note that this make tool treats the path literally so we cannot
use our standard FULL path code which escapes spaces.  Instead qualify
the paths with $(CMAKE_BINARY_DIR) to avoid the problem.
2010-06-14 13:06:39 -04:00
Brad King 6f9bb934d4 Merge branch 'sunCC-5.11-rpath-link' 2010-06-08 09:23:11 -04:00
Brad King 6cf1ccd632 Merge branch 'per-config-link-flags' 2010-06-07 14:36:28 -04:00
Brad King ec66c9db8c Merge branch 'cygwin-exe-export-all' 2010-06-07 14:30:21 -04:00
Brad King 693f98c105 Fix rpath-link flag for SunPro C++ 5.11 on Linux
Commit 82c081ba (Fix rpath-link flag for SunPro C++ on Linux,
2009-07-13) taught CMake to pass '-rpath-link' because SunPro C++ 5.9
does not support '-Wl,'.  Now SunPro C++ 5.11 does not recognize the
option without using '-Wl,'.  Detect whether to use '-Wl,' based on the
output of "sunCC -flags".
2010-06-07 10:15:58 -04:00
Brad King 9622fe5e9a Watcom: Use LINK_FLAGS and STATIC_LIBRARY_FLAGS
Add the <LINK_FLAGS> rule variable in Watcom command lines.
2010-06-01 14:40:28 -04:00
Yaakov Selkowitz cd3a4f0025 Cygwin: Export all symbols with ENABLE_EXPORTS
The ENABLE_EXPORTS property exports all symbols from executables on
UNIX-like platforms, typically for use by plugins.  Honor this behavior
on Cygwin.  See issue #10122.
2010-05-27 14:41:38 -04:00
Brad King 571dc74891 Recognize Clang C and C++ compilers (see #10693)
Map to the platform and compiler information for GNU because the
compilers are command-line compatible for common operations.  Later we
can add Clang-specific features as necessary.  We honor the preferred
capitalization is "Clang", not the common mis-spelling "CLang".
2010-05-17 14:11:20 -04:00
David Cole e0acb6ca5f Fix Windows-cl.cmake so that at most one MSVC** variable is defined.
The expectation of users of the MSVC60, MSVC70, MSVC71, MSVC80, MSVC90
and the new MSVC10 variables is that at most one of them will be set
for any given build tree. This change enforces that expectation for
build trees using Makefile generators. It also fixes the one mismatch
in that expectation to be found in the Visual Studio generator world:
previously, the VS 7.1 generator would set *both* MSVC70 and MSVC71;
now, it only sets MSVC71.

With these changes, user expectations are now met, and the recently
introduced CheckCompilerRelatedVariables test should pass everywhere.
2010-05-05 17:08:06 -04:00
David Cole 2938263636 Fix missing set of MSVC10 and add CheckCompilerRelatedVariables test. 2010-04-30 00:46:24 -04:00
Alex Neundorf d827b9f21d -add basic search directories for the "Generic" platform
As reported on the mailing list, find_path/file/library/program() basically don't work
at all if CMAKE_FIND_ROOT_PATH is set and searching in the host system directories
is disabled. This patch adds /include, /lib and /bin to the search directories, so they
will be appended to CMAKE_FIND_ROOT_PATH so this will work for the "Generic" platform (embedded
systems without OS)

Alex
2010-04-14 22:09:19 +02:00
Chuck Atkins e750761857 OpenBSD: Work-around static/runtime linker inconsistency
Detect the runtime linker's search path and add to the compile time
linker's search path.  This is needed because OpenBSD's static linker
does not search for shared library dependencies in the same places as
the runtime linker.
2010-03-26 09:33:28 -04:00
Bill Hoffman 570ceeb4a4 Teach CMake how to work with G95 on mingw. 2010-03-23 11:32:40 -04:00
Brad King 00477de1c9 Use response file for objects on MinGW and MSYS
Windows command lines are limited to about 32K so we need to use
response files for linking very large lists of object files.

See issue #10401.
2010-03-11 09:46:18 -05:00
Brad King 4eba05de42 Suppress GNU flag -fPIC on Windows
Commit "Modernize GNU compiler info on Windows" (2009-12-02) reorganized
GNU flags on Windows but let -fPIC slip through for compilation of
objects in shared libraries.  While this flag is valid on most GNU
compiler platforms we need to suppress it in Windows-GNU.cmake just as
we already do in CYGWIN-GNU.cmake.
2010-02-19 08:23:31 -05:00
David Cole 6ee87b2e5c Fix issue #10155 - default value of CMAKE_OSX_DEPLOYMENT_TARGET should always be the empty string. When the value of CMAKE_OSX_DEPLOYMENT_TARGET is the empty string, the -mmacosx-version-min flag should not show up on the compiler command line. The logic for selecting default value of CMAKE_OSX_SYSROOT is orthogonal to and independent of the value of the deployment target. The default value for CMAKE_OSX_SYSROOT is the SDK that corresponds to the current version of Mac OSX on which cmake is running. 2010-01-29 11:56:35 -05:00
Brad King 3684f62340 Do not export all symbols from DLLs on Cygwin
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.
2010-01-21 15:03:32 -05:00
Brad King 41024b006b Fix CMAKE_DL_LIBS on Cygwin
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.
2010-01-21 15:03:17 -05:00
Brad King 1804c6bfe2 Add PathScale shared library flags on Linux
We add platform-specific compiler information files

  Platform/Linux-PathScale-<lang>.cmake

to enable -fPIC and -shared flags for shared libraries.
2010-01-21 09:09:27 -05:00
Brad King e46e8fb937 Do not find cyg*.dll on Cygwin
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.
2010-01-13 14:12:29 -05:00
Brad King 12dcf9e4fb Search prefix /usr before root prefix /
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.
2010-01-13 13:36:03 -05:00
Brad King 9ef3f8e820 Restore -rdynamic in Linux build rules
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.
2010-01-13 08:13:46 -05:00
Brad King 7b106a6fb3 Create Linux GNU compiler flag consolidation macro
This macro will be used for GNU compiler flags that are specific to
Linux but not to any language.
2010-01-13 08:13:19 -05:00
David Cole 8d88de4b4a Fix issue with SDK not matching initial deployment target chosen by setting the MACOSX_DEPLOYMENT_TARGET environment variable. The problem was that we were setting the initial SDK value based on our own internal default value for deplyment target rather than the user's environment variable choice. The solution is to base the default value for the SDK on the deployment target variable after initially caching the deployment target... Every time I'm in this code I think I leave it cleaner, only to be proven otherwise. Let's give this one a whirl. Bleh. 2009-12-23 12:08:25 -05:00