* Add a warning if the user sets Boost_ROOT which is not correct
* Clarify directions to the user for viewing debugging messages
* Move the CMAKE_FIND_LIBRARY_SUFFIXES tweak outside of a for loop
* Fixed issue 11204: FindBoost.cmake had trouble discovering libraries
when both -sgd and -gd libraries were available by adding a new option
Boost_USE_STATIC_RUNTIME.
Backwards compatibility of searching for first -gd and then -sgd on
WIN32 is maintained unless the user sets Boost_COMPAT_STATIC_RUNTIME to
false (or they have set Boost_USE_STATIC_RUNTIME).
* Fixed issue 8529: FindBoost was unable to detect boost libraries compiled against
STLport, by reworking the way the Boost ABI tag is calculated. There are additional
ABI tag options available now as well.
* Boost_DEBUG now reports the full list of filenames being searched for when
find_library is called.
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.
Commit 7a18dd8e (Add searching of variables, 2003-03-07) added method
cmCursesMainForm::JumpToCacheEntry to search for cache entries whose
names match a given search string. The method also had a useless
argument "int idx" probably left from earlier development iterations and
hard-coded in all calls to the value '-1'. The method compared this
argument to the "NumberOfVisibleEntries" member which at the time was of
type "int" also.
Commit ff1f8d0b (Fix or cast more integer conversions in cmake,
2010-06-29) changed the type of "NumberOfVisibleEntries" to size_t to
fix other integer conversion warnings. An unsigned type makes sense
given the purpose of the member. However, this caused the '-1' signed
value to be converted to a large unsigned value in the above-mentioned
comparison, leading to incorrect behavior.
Fix the problem by removing the useless argument and the comparison.
eccc7d5 Merge branch 'allow_upper_case_cpp_and_others' into resolve/fix_target_name_with_dot_vs10/allow_upper_case_cpp_and_others
ed37fc3 VS2010: Set IntDir for utility and global targets.
e79e412 VS2010: Honor PROJECT_LABEL target property (#10611)
530ade6 Fix targets with . in the name for VS 10 IDE.
The new first arg in the test is the critical one
to prove that the new NMake specific code works.
The additional colons in the middle of the arg
stream work fine everywhere else, but not on
dash1.kitware with Visual Studio 7.1. Just avoid
the failure for now by removing the unnecessary
new args from the test.
VS2010 uses IntDir as the location for writing log files for
what happens during custom build steps. With no IntDir settings,
all ExternalProject usage within the same CMakeLists.txt file
would result in multiple utility targets all trying to use the
same custom build log files.
With parallel builds, they would try to use them simultaneously
and result in file access errors, preventing the builds from
completing successfully.
Now each utility target has its own IntDir setting, and so, its
own custom build rule log files.
The nightly dashboard showed that the following
platforms had difficulties dealing with "bin dir"
and/or "check command line" as directory and
file names:
AIX
Borland 5.5, 5.6 and 5.8
IRIX
NMake 6.0
OpenBSD
VS 7.1
Watcom
Re-visit later, after the release, to use spaces
in the bin dir and in the target name where possible.
The solution seems hackish, but it works: for
NMake only, prepend a no-op command before each
real command that begins with ".
This is really a work-around for an NMake problem.
When a command begins with ", nmake truncates the
first argument to the command after the first :
in that arg. It has a parsing problem.
Workaround..., hackish..., but it should solve
the issue for #9963 and its related friends.
Also, modify the CustomCommand test to replicate
the problem reported in issue #9963. Before the
NMake specific code change, the test failed.
Now, it passes. Ahhhhhh.
The SetError method automatically starts the message with the name of
the command. Fix up calls to it so that we get "file DOWNLOAD..."
instead of "file FILE(DOWNLOAD ...". Also reduce length of long lines
containing these calls.
CUDA 3.2 on Windows systems changed the layout of the library paths. This adds the extra
directories needed to locate the new files.
I also explicitly disable emulation mode for CUDA 3.1+. This feature was deprecated in
3.0 and ultimately removed in 3.1. The script errors out if CUDA_BUILD_EMULATION is
turned on. I didn't want to ignore emulation mode (even with a warning - which most
people may not even see) and have users confused as to why it wasn't working.
Dereferencing a 0-pointer is undefined behavior, not a deterministic
crash. Use a 1-pointer instead. This also avoids a warning by Clang
about the undefined behavior.