]>
git.ipfire.org Git - thirdparty/ccache.git/log
Joel Rosdahl [Wed, 9 Apr 2014 20:11:25 +0000 (22:11 +0200)]
Fix bug in common_dir_prefix_length
Based on a patch by Douglas Graham <douglas.graham@ericsson.com>.
Joel Rosdahl [Wed, 8 Jan 2014 20:13:22 +0000 (21:13 +0100)]
Require CCACHE_SLOPPINESS=pch_defines,time_macros opt-in to enable PCH handling
The background is that since ccache runs the preprocessor on the header to
be precompiled, the preprocessor removes #defines and ccache then hashes
the output, which means that changes in #defines are not detected even
though the resulting .gch file will have different behavior.
Joel Rosdahl [Wed, 8 Jan 2014 19:34:07 +0000 (20:34 +0100)]
doc: Mention that --ccache-skip currently does not mean "skip hashing option"
Joel Rosdahl [Wed, 8 Jan 2014 19:29:33 +0000 (20:29 +0100)]
Don't hash compiler option -fdebug-prefix-map=
The -fdebug-prefix-map option may be used in combination with
CCACHE_BASEDIR to reuse results across different directories.
Based on a patch by Pawel Sikora.
Ville Skyttä [Sat, 10 Aug 2013 21:48:54 +0000 (00:48 +0300)]
Spelling fixes
Joel Rosdahl [Fri, 30 Aug 2013 20:29:28 +0000 (22:29 +0200)]
Avoid creating ccache directory when CCACHE_DISABLE is set
Joel Rosdahl [Thu, 8 Aug 2013 20:20:12 +0000 (22:20 +0200)]
Bail out on option -gsplit-dwarf which produces multiple output files
Michael Meeks [Thu, 8 Aug 2013 20:07:27 +0000 (22:07 +0200)]
Treat zero length object files as invalid
Fix for bug 9972.
Joel Rosdahl [Thu, 8 Aug 2013 16:06:16 +0000 (18:06 +0200)]
Fix path canonicalization in make_relative_path when path doesn't exist
Joel Rosdahl [Thu, 8 Aug 2013 16:05:30 +0000 (18:05 +0200)]
Handle some dirname() special cases as expected
Joel Rosdahl [Thu, 8 Aug 2013 16:04:44 +0000 (18:04 +0200)]
Correct make_relative_path comment
Ville Skyttä [Sun, 31 Mar 2013 18:34:29 +0000 (21:34 +0300)]
Fix test suite failure on GCC 4.8, take three
Joel Rosdahl [Sun, 31 Mar 2013 18:22:58 +0000 (20:22 +0200)]
Fix test suite failure on GCC 4.8, take two
Joel Rosdahl [Wed, 20 Mar 2013 21:18:16 +0000 (22:18 +0100)]
Fix test suite failure on GCC 4.8
GCC 4.8 includes /usr/include/stdc-predef.h implicitly, and this shows up
in generated .d files.
Ville Skyttä [Sat, 16 Mar 2013 19:35:47 +0000 (21:35 +0200)]
Comment spelling fix
Joel Rosdahl [Mon, 4 Feb 2013 21:24:08 +0000 (22:24 +0100)]
Improve execute() not to rely on special exit codes internally
Bernhard Bauer [Tue, 16 Aug 2011 08:23:55 +0000 (10:23 +0200)]
Support -Xclang compiler option.
Joel Rosdahl [Sun, 6 Jan 2013 16:56:50 +0000 (17:56 +0100)]
Prepare for v3.1.9
Joel Rosdahl [Sat, 5 Jan 2013 17:10:36 +0000 (18:10 +0100)]
Handle disappearing temporary stdout file gracefully
If the temporary stdout file has disappeared, it's most likely because
another ccache process is doing cleanup on a too small cache.
Joel Rosdahl [Sat, 5 Jan 2013 17:08:36 +0000 (18:08 +0100)]
Improve error reporting when failing to create stdout/stderr file in execute()
Joel Rosdahl [Sat, 5 Jan 2013 15:47:04 +0000 (16:47 +0100)]
Don't leak log file descriptor to executed commands
Joel Rosdahl [Thu, 3 Jan 2013 16:51:10 +0000 (17:51 +0100)]
Improve x_unlink's temporary filename
The XXXXXX part was unused.
Joel Rosdahl [Sun, 4 Nov 2012 10:38:13 +0000 (11:38 +0100)]
Don't create cache subdirectories in read-only mode
Andrew Stubbs [Mon, 10 Sep 2012 13:57:23 +0000 (14:57 +0100)]
Fix bug testing with ecryptfs
The test script assumes that zero-length files take up zero disk space,
that the block size is 4k, and that the disk usage will be the file size
rounded up to the nearest block size. Unfortunately these were not true for
ecryptfs.
Additionally, running the test.sh from another directory did not work due
to a bad location assumption in another test.
Alfred Landrum [Fri, 7 Sep 2012 18:39:26 +0000 (20:39 +0200)]
Handle EAGAIN during copy_fd
In util.c, write may return EAGAIN if the file descriptor is a full pipe.
Currently, if this occurs, the log message "ccache: FATAL: Failed to copy fd"
is printed and ccache exits. We've seen this locally when the recorded stderr
output was exceptionally large due to many compilation warnings.
Joel Rosdahl [Sat, 11 Aug 2012 09:02:05 +0000 (11:02 +0200)]
Prepare for v3.1.8
Joel Rosdahl [Sat, 11 Aug 2012 09:01:34 +0000 (11:01 +0200)]
Use C89-style comments
Joel Rosdahl [Sat, 11 Aug 2012 09:01:16 +0000 (11:01 +0200)]
Add quicktest Makefile target
Joel Rosdahl [Thu, 9 Aug 2012 13:09:16 +0000 (15:09 +0200)]
Rename static_assert macro to avoid clash with any preexisting static_assert
Patrick von Reth [Thu, 2 Aug 2012 10:58:22 +0000 (12:58 +0200)]
make git version macro work if compiled outside of src dir
Lubos Lunak [Thu, 26 Jul 2012 12:45:16 +0000 (14:45 +0200)]
hash also Clang plugins
Joel Rosdahl [Sun, 29 Jul 2012 21:11:10 +0000 (23:11 +0200)]
Simplify generation of relative automatic dependency target
Joel Rosdahl [Sun, 29 Jul 2012 21:01:29 +0000 (23:01 +0200)]
Minor code cleanup
Joel Rosdahl [Sun, 29 Jul 2012 20:53:20 +0000 (22:53 +0200)]
Add test for rewriting of absolute dependency arguments
Lalit Chhabra [Sun, 29 Jul 2012 20:52:51 +0000 (22:52 +0200)]
Make dependency arguments relative to increase cache hits
Joel Rosdahl [Sun, 29 Jul 2012 20:07:44 +0000 (22:07 +0200)]
Correct doc string of make_relative_path
Joel Rosdahl [Sat, 28 Jul 2012 09:50:06 +0000 (11:50 +0200)]
Improve get_relative_path and add unit tests
Joel Rosdahl [Fri, 27 Jul 2012 13:59:13 +0000 (15:59 +0200)]
Canonicalize paths when computing path relative to base directory
This fixes a bug when current working directory contains a "/./" part.
Based on a patch by Eric Blau.
Joel Rosdahl [Thu, 31 May 2012 20:38:49 +0000 (22:38 +0200)]
Correctly handle backslash in strings in hash_source_code_string
Joel Rosdahl [Thu, 31 May 2012 19:38:39 +0000 (21:38 +0200)]
Clean up code formatting
Joel Rosdahl [Wed, 7 Mar 2012 20:10:59 +0000 (21:10 +0100)]
Don't crash when stderr is closed
Joel Rosdahl [Wed, 1 Feb 2012 21:12:29 +0000 (22:12 +0100)]
Add -c if missing when compiling a header file
This is to please g++ 4.1.
Joel Rosdahl [Mon, 16 Jan 2012 21:35:39 +0000 (22:35 +0100)]
Don't crash when current working directory doesn't exist
Joel Rosdahl [Sun, 8 Jan 2012 14:37:50 +0000 (15:37 +0100)]
Prepare for v3.1.7
Joel Rosdahl [Sun, 8 Jan 2012 12:05:28 +0000 (13:05 +0100)]
Update NEWS
Joel Rosdahl [Sun, 8 Jan 2012 11:59:02 +0000 (12:59 +0100)]
Also recognize -specs=file in addition to --specs=file
Joel Rosdahl [Sun, 8 Jan 2012 09:40:17 +0000 (10:40 +0100)]
Revert GCC bug compatibility for -MTarg and -MQarg
The bug addressed by commit
8c3039d676d71564e98d50f1ca8ba029d1efa46c has
been fixed in a recent GCC version. Since ccache currently can't change
behavior depending on compiler version, we have to choose the least
troublesome behavior.
Joel Rosdahl [Sat, 7 Jan 2012 15:01:45 +0000 (16:01 +0100)]
Fix minor memory leaks
Joel Rosdahl [Fri, 6 Jan 2012 15:03:36 +0000 (16:03 +0100)]
Correct log message when unify mode is enabled
Joel Rosdahl [Fri, 6 Jan 2012 15:02:42 +0000 (16:02 +0100)]
Hash environment variables that affect the preprocessor output
Joel Rosdahl [Thu, 5 Jan 2012 14:58:50 +0000 (15:58 +0100)]
Hash mtime or content of GCC plugins specified with -fplugin=
Joel Rosdahl [Thu, 5 Jan 2012 14:49:26 +0000 (15:49 +0100)]
Use hash_compiler for explicit --specs= options as well
Joel Rosdahl [Thu, 5 Jan 2012 14:49:03 +0000 (15:49 +0100)]
Refactor code into a hash_compiler function
Joel Rosdahl [Tue, 20 Dec 2011 21:21:46 +0000 (22:21 +0100)]
Improve description on how to fix bad object files in the cache
Joel Rosdahl [Mon, 19 Dec 2011 19:59:53 +0000 (20:59 +0100)]
Make failure to create files in cache fatal
Previously, such failures were erroneously flagged as "compiler produced
stdout".
Joel Rosdahl [Mon, 19 Dec 2011 19:58:00 +0000 (20:58 +0100)]
Make failure to create cache directories fatal
Joel Rosdahl [Sat, 17 Sep 2011 21:55:22 +0000 (23:55 +0200)]
Remove unused print_executed_command function
Joel Rosdahl [Sun, 11 Sep 2011 19:01:28 +0000 (21:01 +0200)]
Handle non-writable CCACHE_DIR gracefully
Ville Skyttä [Mon, 22 Aug 2011 20:40:03 +0000 (23:40 +0300)]
Sync mdfour.c license header with other files, notably fixing FSF's address.
Joel Rosdahl [Sun, 21 Aug 2011 14:42:11 +0000 (16:42 +0200)]
Tweak news markup
Joel Rosdahl [Sun, 21 Aug 2011 14:36:29 +0000 (16:36 +0200)]
Prepare for v3.1.6
Joel Rosdahl [Wed, 3 Aug 2011 18:10:57 +0000 (20:10 +0200)]
Don't crash when getcwd() fails
Joel Rosdahl [Wed, 27 Oct 2010 19:27:33 +0000 (20:27 +0100)]
Rewrite argument to --sysroot if CCACHE_BASEDIR is used
Based on a patch by Bo <yebo9098@gmail.com>.
Jon Bernard [Mon, 30 May 2011 16:42:04 +0000 (12:42 -0400)]
Fix stats display alignment for preprocessing
Joel Rosdahl [Sun, 29 May 2011 16:14:08 +0000 (18:14 +0200)]
Minor spelling tweak
Joel Rosdahl [Sun, 29 May 2011 15:45:28 +0000 (17:45 +0200)]
Prepare for v3.1.5
Joel Rosdahl [Sat, 21 May 2011 18:09:44 +0000 (20:09 +0200)]
Only link with -lm if needed
Some systems don't have libm.
Joel Rosdahl [Sat, 21 May 2011 17:18:46 +0000 (19:18 +0200)]
Move up "can't use precompiled header" stat so that cache size is at end
Joel Rosdahl [Wed, 4 May 2011 20:34:12 +0000 (22:34 +0200)]
Fix minor memory leak in get_cwd()
Found and suggested by Bo Rydberg.
Grigory Entin [Sun, 17 Apr 2011 20:22:44 +0000 (22:22 +0200)]
Support rewriting paths in -F/-iframework GCC options
Joel Rosdahl [Sun, 17 Apr 2011 15:55:14 +0000 (17:55 +0200)]
Be bug compatible with GCC for -MTarg and -MQarg
It's strange, but apparently, GCC behaves differently for -MT arg and
-MTarg (and similar for -MQ): in the latter case, but not in the former,
an implicit dependency for the object file is added to the dependency file.
Joel Rosdahl [Wed, 13 Apr 2011 20:21:54 +0000 (22:21 +0200)]
Log original command line
Joel Rosdahl [Wed, 13 Apr 2011 19:58:29 +0000 (21:58 +0200)]
Correctly handle -MF/-MT/-MQ with concatenated argument last on commandline
Joel Rosdahl [Sat, 9 Apr 2011 19:44:37 +0000 (21:44 +0200)]
Update copyright for 2011
Joel Rosdahl [Sun, 3 Apr 2011 12:42:33 +0000 (14:42 +0200)]
Add separate counter "called for preprocessing"
Joel Rosdahl [Sun, 3 Apr 2011 11:22:57 +0000 (13:22 +0200)]
Include strerror(errno) in error messages when system calls fail
Joel Rosdahl [Sun, 9 Jan 2011 19:56:49 +0000 (20:56 +0100)]
Prepare for version 3.1.4
Joel Rosdahl [Mon, 3 Jan 2011 20:41:36 +0000 (21:41 +0100)]
Cope with file systems (e.g. FAT) that don't handle symlinks
Joel Rosdahl [Sat, 18 Dec 2010 09:01:15 +0000 (10:01 +0100)]
Close file handle in create_cachedirtag on write error
Joel Rosdahl [Tue, 14 Dec 2010 20:11:36 +0000 (21:11 +0100)]
Remove corrupt manifest files so that they won't block direct mode hits
Joel Rosdahl [Mon, 13 Dec 2010 19:46:06 +0000 (20:46 +0100)]
Avoid creating empty file before creating new manifest file
Joel Rosdahl [Sun, 12 Dec 2010 14:18:06 +0000 (15:18 +0100)]
Compensate for bug in gzputc in zlib 1.2.5
gzputc(f, -1) returns -1 (== EOF) in zlib 1.2.5 but should return 255.
Joel Rosdahl [Sun, 28 Nov 2010 16:43:46 +0000 (17:43 +0100)]
Prepare for version 3.1.3
Joel Rosdahl [Sun, 28 Nov 2010 16:04:38 +0000 (17:04 +0100)]
Handle -MFarg, -MTarg and -MQarg compiler options correctly
GCC apparently (and undocumentedly) accepts these.
Joel Rosdahl [Sun, 28 Nov 2010 14:23:05 +0000 (15:23 +0100)]
Correct location of rpl_* defines to outside "#if !HAVE_VSNPRINTF"
Joel Rosdahl [Mon, 22 Nov 2010 20:09:30 +0000 (21:09 +0100)]
portability: Disable support for long double in snprintf replacement
ccache doesn't link correctly on HP-UX 11.00 when support for long double
is enabled.
Joel Rosdahl [Thu, 18 Nov 2010 17:22:06 +0000 (18:22 +0100)]
portability: Handle system *sprintf functions with bad type signatures
Joel Rosdahl [Wed, 17 Nov 2010 20:20:24 +0000 (21:20 +0100)]
portability: Add Holger Weiß's improved snprintf replacement
Joel Rosdahl [Sun, 21 Nov 2010 19:16:55 +0000 (20:16 +0100)]
Update NEWS for version 3.1.2
Joel Rosdahl [Sun, 21 Nov 2010 16:06:09 +0000 (17:06 +0100)]
test: New attempt at coping with unsynchronized clocks for new include file
Joel Rosdahl [Sun, 21 Nov 2010 12:15:51 +0000 (13:15 +0100)]
Add Björn Jacke as a contributor
Joel Rosdahl [Sun, 21 Nov 2010 10:37:42 +0000 (11:37 +0100)]
Bail out on too hard compiler options -fdump-*
Joel Rosdahl [Sun, 21 Nov 2010 10:36:13 +0000 (11:36 +0100)]
Put -E compiler option check in the compopts table
Joel Rosdahl [Wed, 17 Nov 2010 20:12:22 +0000 (21:12 +0100)]
portability: Support systems without /dev/zero
Joel Rosdahl [Sun, 14 Nov 2010 10:00:04 +0000 (11:00 +0100)]
Correctly handle "#line N" (without path) directive for AIX's preprocessor
Joel Rosdahl [Thu, 11 Nov 2010 22:17:02 +0000 (23:17 +0100)]
Add notes about the AIX compiler's preprocessor output
Joel Rosdahl [Thu, 11 Nov 2010 20:14:40 +0000 (21:14 +0100)]
Improve documentation of hashed information
Joel Rosdahl [Tue, 9 Nov 2010 21:57:28 +0000 (22:57 +0100)]
Fixed typo in x_calloc
Joel Rosdahl [Tue, 9 Nov 2010 21:40:43 +0000 (22:40 +0100)]
Make x_malloc and x_calloc always return NULL for zero byte allocations
This helps making sure that the code handles NULL return values regardless
of platform.
Björn Jacke [Tue, 9 Nov 2010 20:33:46 +0000 (21:33 +0100)]
Fix calloc success check
On Tru64 I had issues with this: If the number of elements or the size is 0
then you may get a NULL pointer back, which is not a sign of an error.