Escaped the @var@ in the file writes - this was being expanded at file
write and so not causing a reconfigure at the right time. I also took
care of build up lists of lists in the variables, especially important
for things like MPI_EXTRA_LIBRARY. Added some error checking, and use
the tmp_dir for initial cache file.
On Windows the limit for command line arguments is 8192 characters, and
this was limiting longer paths with some of our more nested projects
such as Library. Placing the -D arguments into CMAKE_CACHE_ARGS will
write out an initial cache file, that will be passed to CMake with a -C
argument as the initial cache.
By forcing the cache variables we preserve the existing behavior with
-D, to change the values of cache variables in our inner projects.
If a .sln file refers to a project file with a leading ".\", as in
".\foo.vcxproj" instead of just "foo.vcxproj" or a full path then
msbuild behaves strangely. Whenever target foo is built as a dependency
of another target, msbuild brings multiple configurations up to date
instead of just the requested configuration!
Refer to all project files by full path to avoid this behavior.
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.
The Numerical Algorithms Group (NAG) Fortran compiler does not document
a preprocessor macro to identify it. Check for identifying output using
the -V option.
Teach CMakeFortranInformation to use default flags only for variables
that have not been set at all, rather then not set or empty. This will
allow platform or compiler-specific information files to set empty
values without getting the defaults.
The NAG Fortran compiler implicitly passes object files by full path to
the linker. Teach CMakeParseImplicitLinkInfo to parse object files that
match some tool-specific regular expression.
FindTCL.cmake switched variables in the FIND_LIBRARY invocation. The
FIND_LIBRARY() statement for TCL used the TK variables and vice versa.
This patch reverses that into the right usage.
Closes debian issue 600245.
The soname generation code was compile-time selected instead of runtime
selected. The result is that a Mac-compiled cmake used to cross-compile
Mac -> Unix generates an soname of the form libfoo.x.y.so instead of
libfoo.so.x.y as expected. Instead do a runtime check based on the
target platform.
Inspired-By: George Staikos <staikos@kde.org>
Avoid tracing dependencies of GLOBAL_TARGET targets. The build system
generators are not designed to handle any dependencies that may be
discovered. Global targets are only generated by CMake and never have
commands that reference targets built in the project anyway.
The exception is when building CMake itself there is a special case to
use the just-built "cmake" binary in the "install" target so that CMake
can replace itself on Windows. Even in this special case we do not want
to let the "install" target depend on the "cmake" target. Doing so
breaks cases like "make -j4 install".
Commit "merge in changes for beos support" (2006-12-04) added a realpath
call for every directory parsed out of a PATH-style environment
variable. No reason was given in the commit message or comments.
The call incorrectly resolves symlinks in referenced paths. Remove it.
If BeOS support really needs it then it can be restored for that
platform with a full explanation.
The cmCustomCommandGenerator::GetCommand method completely replaces the
purpose of this method. Re-implement GetRealLocation inline at the only
remaining call site and remove it.
The Makefile, VS, and Xcode generators previously duplicated some custom
command line generation code. Factor this out into a separate class
cmCustomCommandGenerator shared by all generators.
A custom command may name a target created by add_custom_target in its
DEPENDS field. Treat this case as a target-level dependency only since
a custom target provides no standard file on which to add a file-level
dependency.
Imported targets do not themselves build, but we can follow dependencies
through them to find real targets. This allows imported targets to
depend on custom targets that provide the underlying files at build
time.
Commit 60cd72d0 (Cleaned up generation of symbolic rules, 2006-02-15)
incorrectly made these Makefile targets .PHONY even though the build
rule touches an actual file. Correct it so that the copy_f90_mod and
touch steps do not happen on every "make".
Also, improve the documentation of the fixup_bundle and fixup_bundle_item
functions to clarify that plugin type "libs" need to be copied into
the bundle *before* calling fixup_bundle.
Commit e93a4b4d34 changed the way that
the libs parameter to fixup_bundle is interpreted. Before the commit,
the libs were copied into the bundle first and then fixed up. After
the commit, the copy was skipped, assuming the libs were in the bundle
in the first place, and then the fixups occurred as before.
However, before the commit, it was possible to name a lib from outside
the bundle, and have it copied in and then fixed up. Its resolved
embedded name was always inside the bundle before. After, its resolved
embedded name was just the same as its resolved name, which is in its
original location, and not necessarily inside the bundle.
This manifested itself as a problem with the ParaView call to
fixup_bundle and its many plugins. Previously, ParaView had simply
passed in the list of plugin file names as they existed in the build
tree, and left the copying into the bundle up to the fixup_bundle
function. When built with CMake 2.8.3 (the first version to contain
the above named commit) the fixup_bundle call would inadventently
fixup libraries in the build tree, not libraries that were in the
bundle. Furthermore, the plugins would not be in the final bundle.
This points out the fact that the fix for the bugs made by the above
commit was a backwards-incompatible change in behavior.
This commit makes it an error to try to fixup an item that is not
already inside the bundle to make the change in behavior apparent
to folks who were depending on the prior copy-in behavior: now,
they should get an error, and hopefully, reading the new and
improved documentation, should be able to resolve it in their
projects by adding code to install or copy in such libraries prior
to calling fixup_bundle.
Whew.
Commit 34e1ac24 (Create Fortran info variables for .mod behavior,
2010-11-12) incorrectly taught GetFortranModuleDirectory to return a
relative path. We really want to use "." as the module directory only
as a workaround for compilers that do not do so by default. Therefore
we need this default only when generating the compiler command line and
not when scanning dependencies.
Revert the previous change to GetFortranModuleDirectory and apply the
change only at one of its call sites.
Commit 3827991e (...fix...not being able to remove try compile code,
2008-03-26) introduced a loop of RemoveFile attempts to overcome
anti-virus locks on recently created try_compile executables. Fix the
logic in this loop to work when the file is already missing.
KWSys Robot
Marcus D. Hanwell
Brad King
Bill Hoffman
Kai Wasserbäch
Ben Boeckel
Rolf Eike Beer
David Cole
Alex Neundorf