]> git.ipfire.org Git - thirdparty/ccache.git/log
thirdparty/ccache.git
7 years agoRearrange files into suitable subdirectories, e.g. src and doc
Joel Rosdahl [Wed, 7 Feb 2018 11:14:49 +0000 (12:14 +0100)] 
Rearrange files into suitable subdirectories, e.g. src and doc

The top directory has become too crowded.

7 years agoSimplify initialize() slightly
Joel Rosdahl [Tue, 6 Feb 2018 20:58:19 +0000 (21:58 +0100)] 
Simplify initialize() slightly

7 years agoTreat unreadable conf file like missing instead of a fatal error
Joel Rosdahl [Tue, 6 Feb 2018 20:49:55 +0000 (21:49 +0100)] 
Treat unreadable conf file like missing instead of a fatal error

This reverts 0b18af47 and implements a better solution.

The major reason is to keep ccache transparent, i.e. to still have the
following behavior:

* If ccache returns exit code == 0, then any produced stderr comes from
  the compiler and only from the compiler.
* If ccache returns exit code != 0, then ccache may print error messages
  of its own.

The reason is that autoconf configure scripts have been known to fail
tests if the compiler emits anything to stderr even if the produced
result actually works. (And printing to stdout is also a no-no.)

7 years agoDocument caveats related to colored warnings from compilers
Joel Rosdahl [Mon, 5 Feb 2018 20:58:55 +0000 (21:58 +0100)] 
Document caveats related to colored warnings from compilers

7 years agoDocument more pitfalls with hard links
Joel Rosdahl [Mon, 5 Feb 2018 18:11:02 +0000 (19:11 +0100)] 
Document more pitfalls with hard links

7 years agoUse strftime instead of obsolete asctime function
Joel Rosdahl [Sun, 4 Feb 2018 09:47:06 +0000 (10:47 +0100)] 
Use strftime instead of obsolete asctime function

7 years agoFix most warnings from cppcheck 1.72
Joel Rosdahl [Sun, 4 Feb 2018 09:40:04 +0000 (10:40 +0100)] 
Fix most warnings from cppcheck 1.72

7 years agoChange “improvements” to “enhancements” in NEWS.txt
Joel Rosdahl [Sun, 4 Feb 2018 09:33:54 +0000 (10:33 +0100)] 
Change “improvements” to “enhancements” in NEWS.txt

Every change is hopefully an improvement.

7 years agoFix use-after-free error in 316b966b
Joel Rosdahl [Sat, 3 Feb 2018 21:56:11 +0000 (22:56 +0100)] 
Fix use-after-free error in 316b966b

7 years agoHash host compilers invoked by nvcc
Joel Rosdahl [Sat, 3 Feb 2018 20:16:26 +0000 (21:16 +0100)] 
Hash host compilers invoked by nvcc

See #145.

7 years agoIntroduce and use ARRAY_SIZE macro
Joel Rosdahl [Sat, 3 Feb 2018 09:42:22 +0000 (10:42 +0100)] 
Introduce and use ARRAY_SIZE macro

7 years agoHash '-ccbin' argument
Sam Gross [Tue, 1 Nov 2016 17:45:26 +0000 (10:45 -0700)] 
Hash '-ccbin' argument

7 years agounittest: Pass errmsg to CHECKM when available
Joel Rosdahl [Wed, 31 Jan 2018 21:15:45 +0000 (22:15 +0100)] 
unittest: Pass errmsg to CHECKM when available

From PR #169.

7 years agoUpdate config.guess and config.sub to modern versions
Joel Rosdahl [Wed, 31 Jan 2018 21:00:46 +0000 (22:00 +0100)] 
Update config.guess and config.sub to modern versions

As suggested in PR #169.

7 years agoOnly pass implicit -MQ to preprocessor if needed
Joel Rosdahl [Tue, 30 Jan 2018 21:38:29 +0000 (22:38 +0100)] 
Only pass implicit -MQ to preprocessor if needed

Fixes #209.

7 years agoUpdate NEWS
Joel Rosdahl [Tue, 30 Jan 2018 21:13:32 +0000 (22:13 +0100)] 
Update NEWS

7 years agotest: Add and fix cpp1 suite
Joel Rosdahl [Tue, 30 Jan 2018 21:07:52 +0000 (22:07 +0100)] 
test: Add and fix cpp1 suite

I messed up when adapting Anders's work in f58741b5 to the new test
layout.

7 years agoUpdate NEWS
Joel Rosdahl [Tue, 30 Jan 2018 20:55:17 +0000 (21:55 +0100)] 
Update NEWS

7 years agoImprove docs and clean up
Joel Rosdahl [Tue, 30 Jan 2018 20:51:17 +0000 (21:51 +0100)] 
Improve docs and clean up

7 years agoPass the cpp1 flags more selectively
Anders F Björklund [Mon, 26 Jun 2017 20:51:17 +0000 (22:51 +0200)] 
Pass the cpp1 flags more selectively

The -fdirectives-only flag is needed for *both* cpp and cc/c++,
but the -frewrite-includes is only needed for cpp (not cc/c++).

Also make it easier to follow along in the gcc manual by passing
-fpreprocess explicitly, just like we pass the language to clang.

7 years agoAdd test case for directives-only/rewrite-includes
Anders F Björklund [Sun, 25 Jun 2017 19:49:15 +0000 (21:49 +0200)] 
Add test case for directives-only/rewrite-includes

Use run_second_cpp = false (also known sometimes as "cpp1"),
but instruct the preprocessor to only handle include directives.

Make sure that any headers are still processed, and that both
the #define and -D work as usual (with regular preprocessing).

7 years agoHandle semi-preprocessed source code with cpp1
Anders F Björklund [Sun, 25 Jun 2017 18:41:24 +0000 (20:41 +0200)] 
Handle semi-preprocessed source code with cpp1

When using gcc -fdirectives-only or clang -frewrite-includes,
we still need to pass preprocessor arguments to the compiler.

Also make sure to set the language for clang (but not for gcc),
since the file no longer contains regular preprocessed source.

7 years agoClean up
Joel Rosdahl [Tue, 30 Jan 2018 20:36:19 +0000 (21:36 +0100)] 
Clean up

7 years agoFail if boolean env vars are set with common negative strings
Mostyn Bramley-Moore [Fri, 23 Jun 2017 18:20:54 +0000 (20:20 +0200)] 
Fail if boolean env vars are set with common negative strings

Boolean environment variables have strange semantics: if the variable
is set then the configuration setting is considered "true".  This means
users can easily be mistaken that eg CCACHE_DISABLE=false means that
the "disable" setting is set to false.

To avoid too much backwards-incompatibility, we now exit with an error if
some common negative-sounding (case-insensitive) values:
"0", "false", "disable", "no".

All other values (including the empty string) are still considered to mean
"true".

Resolves https://github.com/ccache/ccache/issues/182

7 years agotest: Rearrange files into test and unittest directories
Joel Rosdahl [Mon, 29 Jan 2018 19:57:18 +0000 (20:57 +0100)] 
test: Rearrange files into test and unittest directories

* The unit tests formerly located in the test directory have been moved
  to the unittest directory.
* The integration test suite test.sh has been made a bash script and
  renamed to test/run. Its test suites are located in test/suites.

7 years agoPrefer bash, when running test.sh script
Anders F Björklund [Fri, 26 Jan 2018 18:12:57 +0000 (19:12 +0100)] 
Prefer bash, when running test.sh script

Look for bash, to make it possible to run "make test"
on platforms where it is not the /bin/sh implementation.

It is possible to build ccache without it, when only
running the limited unit tests (with "make quicktest").

7 years agoMerge branch '3.3-maint'
Joel Rosdahl [Sun, 28 Jan 2018 15:24:14 +0000 (16:24 +0100)] 
Merge branch '3.3-maint'

* 3.3-maint:
  Prepare for v3.3.6

7 years agoPrepare for v3.3.6 3.3-maint v3.3.6
Joel Rosdahl [Sun, 28 Jan 2018 15:22:29 +0000 (16:22 +0100)] 
Prepare for v3.3.6

7 years agoTweak code style and comments
Joel Rosdahl [Sun, 28 Jan 2018 15:20:30 +0000 (16:20 +0100)] 
Tweak code style and comments

7 years agoMerge branch '3.3-maint'
Joel Rosdahl [Sun, 28 Jan 2018 15:12:47 +0000 (16:12 +0100)] 
Merge branch '3.3-maint'

* 3.3-maint:
  Update NEWS
  test: Verify that UNCACHED_ERR_FD works
  Only write hash debug output if there is any data
  Fix broken UNCACHED_ERR_FD

7 years agoUpdate NEWS
Joel Rosdahl [Sun, 28 Jan 2018 15:08:31 +0000 (16:08 +0100)] 
Update NEWS

7 years agotest: Verify that UNCACHED_ERR_FD works
Joel Rosdahl [Sun, 28 Jan 2018 15:06:36 +0000 (16:06 +0100)] 
test: Verify that UNCACHED_ERR_FD works

7 years agoOnly write hash debug output if there is any data
Joel Rosdahl [Sun, 28 Jan 2018 15:05:07 +0000 (16:05 +0100)] 
Only write hash debug output if there is any data

7 years agoFix broken UNCACHED_ERR_FD
Luboš Luňák [Fri, 26 Jan 2018 07:47:30 +0000 (08:47 +0100)] 
Fix broken UNCACHED_ERR_FD

It shouldn't be closed on exec since that's exactly the point, to leak
it to whatever ccache calls.

7 years agotest: Fix nvcc tests when using compiler with options
Joel Rosdahl [Sat, 27 Jan 2018 22:01:07 +0000 (23:01 +0100)] 
test: Fix nvcc tests when using compiler with options

7 years agotest: Remove superfluous blank lines
Joel Rosdahl [Sat, 27 Jan 2018 22:00:01 +0000 (23:00 +0100)] 
test: Remove superfluous blank lines

7 years agotest: Split into one file per suite
Joel Rosdahl [Sat, 27 Jan 2018 21:18:37 +0000 (22:18 +0100)] 
test: Split into one file per suite

The test.sh file has simply grown unpleasantly large, so it's time to
split it up. Some suites still refer to stuff defined in earlier suites,
but we'll have to live with that for now.

7 years agoUpdate NEWS
Joel Rosdahl [Sat, 27 Jan 2018 21:02:25 +0000 (22:02 +0100)] 
Update NEWS

7 years agoClean up code related to CUDA/nvcc
Joel Rosdahl [Sat, 27 Jan 2018 20:57:22 +0000 (21:57 +0100)] 
Clean up code related to CUDA/nvcc

Remove tabs from test.sh, use lowercase names for local variables and
other minor fixes.

7 years agotest: Fix nvcc_ldir tests when gcc is a masquerading ccache
Joel Rosdahl [Sat, 27 Jan 2018 20:25:19 +0000 (21:25 +0100)] 
test: Fix nvcc_ldir tests when gcc is a masquerading ccache

7 years agoBug fixes and tests for nvcc
Clemens Rabe [Wed, 23 Aug 2017 07:33:23 +0000 (09:33 +0200)] 
Bug fixes and tests for nvcc

Feature: Added cuda compiler in separate travis job and implemented
         tests for nvcc.

Feature: Added support for nvcc compiler options --compiler-bindir/-ccbin,
         --output-directory/-odir and --libdevice-directory/-ldir.
         Added tests for these options in test.sh.

Bug fix: Original patch had a statement to avoid using the preprocessed
         input files for nvcc when run_second_cpp is false. Otherwise,
         when a build is necessary and the preprocessed output was used,
         nvcc results with a compiler error. The patch simply ensures
         run_second_cpp is always set to true for the cuda compiler.

Bug fix: The -optf and -odir options are only accepted for nvcc. For other
         compilers, they behave like '-o ptf' resp. '-o dir'.
         A test was added to check this behaviour.

7 years agotest: Fix failure in gcc pch suite when using compiler with options
Joel Rosdahl [Sat, 27 Jan 2018 14:26:15 +0000 (15:26 +0100)] 
test: Fix failure in gcc pch suite when using compiler with options

Need to remember any extra compiler options in REAL_COMPILER as well.

7 years agotest: Find and use real compiler
Joel Rosdahl [Mon, 22 Jan 2018 20:59:31 +0000 (21:59 +0100)] 
test: Find and use real compiler

The real compiler is now found in the same way as ccache itself does it,
so we no longer have to rely on the uncached_compile hack.

7 years agotest: Check that dependency file contains the correct object name
Joel Rosdahl [Mon, 22 Jan 2018 09:56:06 +0000 (10:56 +0100)] 
test: Check that dependency file contains the correct object name

7 years agoRefactor and clean up code for copying results to/from the cache
Joel Rosdahl [Sun, 21 Jan 2018 21:17:24 +0000 (22:17 +0100)] 
Refactor and clean up code for copying results to/from the cache

Lots of cruft and copy pasta has collected during the years.

* It's enough to let get_file_from_cache detect missing files in the
  cache; no need to check before.
* Made the order of the handling of the different file types consistent
  in various places.
* Removed the tmp_cov, tmp_su and tmp_dwo variables. I cannot
  see why they just aren't redundant since the corresponding output_*
  variables already exist with the same value. Except for the "GCC has
  some funny rule" part, but that seems bogus anyway since tmp_* and
  output_* must be equal for things to work. Let's just ignore it for
  now.
* Removed the support for generating placeholders for .cov and .su files
  if the compiler doesn't create them. Modern compilers always create
  them, and I want to reduce code complexity.

7 years agoImprove file size/number counters for overwritten files
Joel Rosdahl [Sun, 21 Jan 2018 08:34:51 +0000 (09:34 +0100)] 
Improve file size/number counters for overwritten files

When ccache overwrites a cached file, the counters are still increased
as if a new file was stored in the cache, for instance when recaching.

This is now fixed by checking if a cached file exists before storing the
cached file and update file size/number counters with the correct
size/count delta.

7 years agoMerge branch '3.3-maint'
Joel Rosdahl [Sun, 21 Jan 2018 09:53:10 +0000 (10:53 +0100)] 
Merge branch '3.3-maint'

* 3.3-maint:
  doc: Improve instructions on how to compile in different directories
  Split .incbin directive in ccache source code to enable cache hits

7 years agodoc: Improve instructions on how to compile in different directories
Joel Rosdahl [Sun, 21 Jan 2018 08:13:45 +0000 (09:13 +0100)] 
doc: Improve instructions on how to compile in different directories

7 years agoSplit .incbin directive in ccache source code to enable cache hits
Joel Rosdahl [Sun, 21 Jan 2018 07:54:17 +0000 (08:54 +0100)] 
Split .incbin directive in ccache source code to enable cache hits

7 years agoMerge branch '3.3-maint'
Joel Rosdahl [Sat, 13 Jan 2018 18:38:26 +0000 (19:38 +0100)] 
Merge branch '3.3-maint'

* 3.3-maint:
  Prepare for v3.3.5
  doc: Tweak automatic cleanup description

7 years agoPrepare for v3.3.5 v3.3.5
Joel Rosdahl [Sat, 13 Jan 2018 18:34:47 +0000 (19:34 +0100)] 
Prepare for v3.3.5

7 years agodoc: Tweak automatic cleanup description
Joel Rosdahl [Sat, 13 Jan 2018 16:43:34 +0000 (17:43 +0100)] 
doc: Tweak automatic cleanup description

7 years agoFix code style
Joel Rosdahl [Thu, 11 Jan 2018 21:33:50 +0000 (22:33 +0100)] 
Fix code style

7 years agoFix compilation with MinGW32
Orgad Shaneh [Wed, 24 May 2017 12:31:23 +0000 (15:31 +0300)] 
Fix compilation with MinGW32

7 years agoAdd link to a good log post on how to get help from maintainers
Joel Rosdahl [Thu, 11 Jan 2018 20:44:27 +0000 (21:44 +0100)] 
Add link to a good log post on how to get help from maintainers

7 years agoMerge branch '3.3-maint'
Joel Rosdahl [Thu, 11 Jan 2018 20:08:01 +0000 (21:08 +0100)] 
Merge branch '3.3-maint'

* 3.3-maint:
  Update NEWS
  Don't skip compiler argument following AFFECTS_CPP without TAKES_ARG
  Clean up
  Clean up redundant expression
  -fdebug-prefix-map is supported in clang 3.8 and newer
  Fix leaked file descriptor
  Clang emits warnings for unused linker arguments, respect that!
  doc: Improve markup
  Add some fallthrough comments to work around GCC compiler warnings

7 years agoUpdate NEWS
Joel Rosdahl [Thu, 11 Jan 2018 19:59:57 +0000 (20:59 +0100)] 
Update NEWS

7 years agoDon't skip compiler argument following AFFECTS_CPP without TAKES_ARG
Joel Rosdahl [Thu, 11 Jan 2018 19:51:38 +0000 (20:51 +0100)] 
Don't skip compiler argument following AFFECTS_CPP without TAKES_ARG

Bug found by Mike Gelfand.

Fixes #203.

7 years agoClean up
Joel Rosdahl [Thu, 11 Jan 2018 13:41:51 +0000 (14:41 +0100)] 
Clean up

7 years agoClean up redundant expression
orbitcowboy [Mon, 8 Jan 2018 12:46:40 +0000 (13:46 +0100)] 
Clean up redundant expression

7 years ago-fdebug-prefix-map is supported in clang 3.8 and newer
Tom Hughes [Tue, 12 Dec 2017 16:52:04 +0000 (08:52 -0800)] 
-fdebug-prefix-map is supported in clang 3.8 and newer

It's not documented until newer versions of clang, but support was
added in 3.8: https://reviews.llvm.org/rL250094

7 years agoFix leaked file descriptor
Andreas Huber [Thu, 7 Dec 2017 13:35:43 +0000 (14:35 +0100)] 
Fix leaked file descriptor

7 years agoClang emits warnings for unused linker arguments, respect that!
Edward Z. Yang [Fri, 30 Jun 2017 14:35:29 +0000 (07:35 -0700)] 
Clang emits warnings for unused linker arguments, respect that!

If ccache concludes the invocation with/without linker arguments
is the same as before, then it may show/fail to show a warning
when it should.  I know we're not supposed to rely on the
is clang check, but this solves it in normal cases.

Fixes #189.

Signed-off-by: Edward Z. Yang <ezyang@fb.com>
7 years agodoc: Improve markup
Joel Rosdahl [Thu, 11 Jan 2018 08:01:07 +0000 (09:01 +0100)] 
doc: Improve markup

7 years agoAdd some fallthrough comments to work around GCC compiler warnings
Geert Bosch [Fri, 23 Jun 2017 20:45:33 +0000 (16:45 -0400)] 
Add some fallthrough comments to work around GCC compiler warnings

7 years agoSimplify is_precompiled_header a bit
Orgad Shaneh [Sun, 18 Jun 2017 09:28:04 +0000 (12:28 +0300)] 
Simplify is_precompiled_header a bit

Avoid repeating calls to get_extension

7 years agoMerge branch '3.3-maint'
Joel Rosdahl [Wed, 10 Jan 2018 20:10:45 +0000 (21:10 +0100)] 
Merge branch '3.3-maint'

* 3.3-maint:
  Document how automatic cleanup works
  Make ccache -c/--cleanup not take limit_multiple into account
  test: Fix printing of failure reason with glob characters
  Fix grammar in cleanup_* function names
  test: Unset all CCACHE variables in environment
  cleanup: Improve logging
  test: Check more cases of path normalization
  Don't overwrite cached .d file with a version with absolute paths
  test: Fix verbose printing of test case name with space
  Update NEWS
  Don't use cache when building precompiled header with changed deps
  Always use include DEFAULT_SLOPPINESS when overriding CCACHE_SLOPPINESS

7 years agoRename HACKING.txt to CONTRIBUTING.md and improve text
Joel Rosdahl [Sun, 31 Dec 2017 13:25:40 +0000 (14:25 +0100)] 
Rename HACKING.txt to CONTRIBUTING.md and improve text

7 years agoDocument how automatic cleanup works
Joel Rosdahl [Sat, 6 Jan 2018 15:25:34 +0000 (16:25 +0100)] 
Document how automatic cleanup works

7 years agoMake ccache -c/--cleanup not take limit_multiple into account
Joel Rosdahl [Sat, 6 Jan 2018 14:41:13 +0000 (15:41 +0100)] 
Make ccache -c/--cleanup not take limit_multiple into account

Now "ccache -c/--cleanup" will trim the cache to the configured max
limits instead of 80% (limit_multiple) of the max limits. This feels
like a more intuitive behavior, and it also makes it behave like
documented.

7 years agotest: Fix printing of failure reason with glob characters
Joel Rosdahl [Sat, 6 Jan 2018 14:08:29 +0000 (15:08 +0100)] 
test: Fix printing of failure reason with glob characters

7 years agoFix grammar in cleanup_* function names
Joel Rosdahl [Sat, 6 Jan 2018 13:30:50 +0000 (14:30 +0100)] 
Fix grammar in cleanup_* function names

7 years agotest: Unset all CCACHE variables in environment
Joel Rosdahl [Sat, 6 Jan 2018 13:00:54 +0000 (14:00 +0100)] 
test: Unset all CCACHE variables in environment

Gotta catch 'em all.

As suggested by Radu Rendec.

7 years agocleanup: Improve logging
Joel Rosdahl [Thu, 4 Jan 2018 21:55:58 +0000 (22:55 +0100)] 
cleanup: Improve logging

7 years agotest: Check more cases of path normalization
Joel Rosdahl [Thu, 4 Jan 2018 20:09:10 +0000 (21:09 +0100)] 
test: Check more cases of path normalization

7 years agoDon't overwrite cached .d file with a version with absolute paths
Joel Rosdahl [Sat, 30 Dec 2017 19:49:25 +0000 (20:49 +0100)] 
Don't overwrite cached .d file with a version with absolute paths

The .d file has already been put in the cache by to_cache (and
potentially converted to relative paths using
use_relative_paths_in_depfile).

This fixes what seems to be a mistake in 38301c0d.

7 years agotest: Fix verbose printing of test case name with space
Joel Rosdahl [Thu, 28 Dec 2017 18:16:20 +0000 (19:16 +0100)] 
test: Fix verbose printing of test case name with space

8 years agoUpdate NEWS
Joel Rosdahl [Wed, 19 Jul 2017 20:45:25 +0000 (22:45 +0200)] 
Update NEWS

8 years agoDon't use cache when building precompiled header with changed deps
Tor Arne Vestbø [Thu, 15 Jun 2017 12:40:16 +0000 (14:40 +0200)] 
Don't use cache when building precompiled header with changed deps

If one of the included files in a precompiled header is touched,
changing only its mtime, clang will produce a fatal error when the
precompiled header is then used:

    file 'foo.h' has been modified since the precompiled header 'foo.pch' was built

We need to take this into account when producing precompiled headers,
so that we don't pick out stale objects from the cache. This works
fine for direct mode, but in preprocess mode we don't have enough
information to decide if the object is still valid, so we skip the
cache entirely in that mode.

8 years agoAlways use include DEFAULT_SLOPPINESS when overriding CCACHE_SLOPPINESS
Tor Arne Vestbø [Thu, 15 Jun 2017 09:52:34 +0000 (11:52 +0200)] 
Always use include DEFAULT_SLOPPINESS when overriding CCACHE_SLOPPINESS

Otherwise the tests will have problems with backdated files.

8 years agoUpdate NEWS
Joel Rosdahl [Wed, 19 Jul 2017 20:04:44 +0000 (22:04 +0200)] 
Update NEWS

8 years agoGet rid of global variable and improve name
Joel Rosdahl [Wed, 19 Jul 2017 20:04:34 +0000 (22:04 +0200)] 
Get rid of global variable and improve name

8 years agoEnable the unifier option, also in direct mode
Anders F Björklund [Tue, 28 Mar 2017 05:29:47 +0000 (07:29 +0200)] 
Enable the unifier option, also in direct mode

8 years agoAdd debugging option of printing unify output
Anders F Björklund [Tue, 28 Mar 2017 05:21:34 +0000 (07:21 +0200)] 
Add debugging option of printing unify output

8 years agoMerge branch '3.3-maint'
Joel Rosdahl [Wed, 21 Jun 2017 20:15:09 +0000 (22:15 +0200)] 
Merge branch '3.3-maint'

* 3.3-maint:
  Fix detection of missing argument to -opt/--options-file

8 years agoFix detection of missing argument to -opt/--options-file
Joel Rosdahl [Wed, 21 Jun 2017 20:14:35 +0000 (22:14 +0200)] 
Fix detection of missing argument to -opt/--options-file

Closes #171.

8 years agoMerge branch '3.3-maint'
Joel Rosdahl [Thu, 15 Jun 2017 18:32:18 +0000 (20:32 +0200)] 
Merge branch '3.3-maint'

* 3.3-maint:
  Update log message and docs when using SLOPPY_FILE_STAT_MATCHES
  Tweak comment
  Add comment about why SLOPPY_INCLUDE_FILE_[CM]TIME compares using >=t

8 years agoUpdate log message and docs when using SLOPPY_FILE_STAT_MATCHES
Tor Arne Vestbø [Thu, 15 Jun 2017 12:50:00 +0000 (14:50 +0200)] 
Update log message and docs when using SLOPPY_FILE_STAT_MATCHES

The size is no longer part of this check.

8 years agoTweak comment
Joel Rosdahl [Thu, 15 Jun 2017 18:24:38 +0000 (20:24 +0200)] 
Tweak comment

8 years agoAdd comment about why SLOPPY_INCLUDE_FILE_[CM]TIME compares using >=t
Tor Arne Vestbø [Thu, 15 Jun 2017 12:49:28 +0000 (14:49 +0200)] 
Add comment about why SLOPPY_INCLUDE_FILE_[CM]TIME compares using >=t

8 years agoTweak comment
Joel Rosdahl [Thu, 15 Jun 2017 18:24:38 +0000 (20:24 +0200)] 
Tweak comment

8 years agoAdd comment about why SLOPPY_INCLUDE_FILE_[CM]TIME compares using >=t
Tor Arne Vestbø [Thu, 15 Jun 2017 12:49:28 +0000 (14:49 +0200)] 
Add comment about why SLOPPY_INCLUDE_FILE_[CM]TIME compares using >=t

8 years agoMerge branch '3.3-maint'
Joel Rosdahl [Sat, 25 Mar 2017 20:24:57 +0000 (21:24 +0100)] 
Merge branch '3.3-maint'

* 3.3-maint:
  Update NEWS
  Add test for directory in ignore_headers_in_manifest
  Fix DIR_DELIM_CH for win32 and non-win32 platforms, values were mixed up
  Update NEWS
  Tweak source code format
  Support using multiple -fdebug-prefix-map
  Fixed the check for empty object file vs diagnostics
  Removed time_of_compilation check wrt SLOPPY_FILE_STAT_MATCHES
  Revert "Better parsing of g* Options, and only use last argument"
  Test that the order of debug options is kept

8 years agoUpdate NEWS
Joel Rosdahl [Sat, 25 Mar 2017 20:24:17 +0000 (21:24 +0100)] 
Update NEWS

8 years agoAdd test for directory in ignore_headers_in_manifest
Joel Rosdahl [Sat, 25 Mar 2017 20:23:06 +0000 (21:23 +0100)] 
Add test for directory in ignore_headers_in_manifest

I should have added this test in a61d0f24 – that would have prevented
issue #165.

8 years agoFix DIR_DELIM_CH for win32 and non-win32 platforms, values were mixed up
Oded Shimon [Sun, 19 Mar 2017 08:56:20 +0000 (10:56 +0200)] 
Fix DIR_DELIM_CH for win32 and non-win32 platforms, values were mixed up

8 years agoUpdate NEWS
Joel Rosdahl [Sat, 25 Mar 2017 19:41:05 +0000 (20:41 +0100)] 
Update NEWS

8 years agoTweak source code format
Joel Rosdahl [Sat, 25 Mar 2017 19:39:48 +0000 (20:39 +0100)] 
Tweak source code format

8 years agoAdd docker target for running test with Docker
Anders F Björklund [Fri, 24 Mar 2017 17:44:38 +0000 (18:44 +0100)] 
Add docker target for running test with Docker

Add standard container definition for running the tests,
to make things easier for developers on other platforms.

8 years agoSupport using multiple -fdebug-prefix-map
Anders F Björklund [Fri, 17 Mar 2017 19:21:09 +0000 (20:21 +0100)] 
Support using multiple -fdebug-prefix-map

Previously only the last one was actually being applied to the hash.
Note that the compiler will actually only *use* one of all the paths.

Closes #163.