| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| | | |
|
| | | |
|
| | |
| | |
| | |
| | | |
doesn't match other platforms, but it should be sufficient to verify that build.ninja file still exists
|
| | | |
|
|\ \ \
| |/ / |
|
| |\ \
| | |/
| |/| |
Skip unusable javah tests
|
| | | |
|
| | |
| | |
| | |
| | | |
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The previous sanity check was to see if the less-optimistic
WhereIs could find javah, and skip if not, since java_where_javah
looks many more places than SCons will actually look. This
was insufficient - some openjdk packages don't clean up the
alternatives link to javah even if a java version which doesn't
have it becomes the selected version. Add a check on found
paths to detect a mismatch.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
For the three tests which fail, skip if the command finder (WhereIs), which
is much less aggressive than the java_where_j* routines, doesn't find javah.
Split the JAVAH test file into two files, so that the part which
doesn't call javah can run even if the rest is skipped.
Did the header cleanup and partial reformats on those.
Leave a comment in the Repository/JavaH test, and do some reformatting
there as well. Test needs rework, but for now doesn't fail so leave alone.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Update file headers
Remove references to javah that are not used
Black a couple of SConstcuts
No functional changes - prep work for javah workarounds.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |\ \
| | | |
| | | | |
drop unnecessary soname check on OpenBSD
|
| | |/
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The changelog for the 4.1 release says that
- No longer automatically disable setting SONAME on shared libraries
on OpenBSD.
this bit was left out.
|
| |\ \
| | | |
| | | | |
fix test some test expected stderr regexp
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
On OpenBSD the output is
sh: <filename>: cannot execute - Permission denied
|
| | |/
| | |
| | |
| | |
| | |
| | | |
On OpenBSD the output is
sh: <filename>: cannot execute - Is a directory
|
| |/
| |
| |
| |
| |
| | |
leaky-handles requires a /proc/ partition which isn't present on most
posix-y platforms: OpenBSD and darwin lacks a procfs, while on FreeBSD
must be explicitly mounted by the user (is disabled by default.)
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If the substable element is a callable, only set for_signature
true if the mode is SUBST_SIG. Previously it was set even if
the mode was SUBST_RAW.
Fixes #4037
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |\ |
|
| | |
| | |
| | |
| | | |
output.
|
| |\ \
| | | |
| | | |
| | | | |
Manually resolve conflict in CHANGES.txt
|
| | | | |
|
| | | | |
|
| |_|/
|/| |
| | |
| | | |
rebuild build.ninja preserves the file
|
|\ \ \
| | | |
| | | | |
update ninja tool to support $ in scons command line args
|
| | | |
| | | |
| | | |
| | | | |
from https://github.com/mongodb/mongo/commit/dee0f733cdb3cab7714326574efc8a1ae4ec750f
|
|\ \ \ \ |
|
| | | | | |
|
| | |_|/
| |/| | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Maintenance: SCons has multiple classes named Base, which is a bit
unfortunate. Some already use a context-qualified name, like BuilderBase,
which seems preferable. Do that for ScannerBase, but leave the name Base
in the SCons.Scanner package in case *external* users are depending on
SCons.Scanner.Base working. SCons internally no longer uses that name.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |\ \ \
| | |/ /
| |/| | |
Do not convert throwaway input from gcc tool.
|
| | | |
| | | |
| | | |
| | | | |
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|\ \ \ \
| |/ / / |
|
| |\ \ \
| | | |/
| | |/| |
|
| | | | |
|
| | | |
| | | |
| | | |
| | | | |
MSVC/hierarchical.py to use File('StdAfx.pch') Best to start recommending File() or Node from env.PCH() as value for PCH.
|
| | | | |
|
| | | |
| | | |
| | | |
| | | | |
value will disable PCH being added to compile and link lines
|
| | | | |
|
| | | |
| | | |
| | | |
| | | | |
expected files
|
| | | |
| | | |
| | | |
| | | | |
explain how the dir/file fixture in pch_gen's logic is working
|
| |/ / |
|
| | | |
|
|/ / |
|
| | |
|
|/ |
|
|
|
|
|
|
|
|
| |
Tweaked some tests which directly set values for _exe, _obj,
which are available from the framework. Done for consistency,
none of these were doing these wrong.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|\
| |
| | |
Two small test tweaks
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Modernize the C in test/Libs/LIBS.py to avoid clang errors (gcc
and MSVC only warn here).
Change where test list is sorted in SCons/cppTests.py.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|/ |
|