a3335709 Enable libuv on Solaris 10
e4ac7c2b libuv: Compile as C 90 on Solaris 10
c9b3c562 libuv: Link to 'rt' library on Solaris 10 to get semaphores
f0698692 libuv: Implement mkdtemp on Solaris 10
5651257f libuv: automatically skip ifaddrs on Solaris 10
In commit 0bbd993f (Make CMAKE_HOST_SYSTEM_NAME available in scripting
context, 2016-12-26) we added a call to `uname` that checks for a zero
return value. However, on Solaris the `uname(2)` manual [1] says that
on success a non-negative value is returned. Fix our return code check
so that we detect the `SunOS` name correctly.
[1] https://docs.oracle.com/cd/E53394_01/html/E54765/uname-2.html
This was added upstream but does not make sense in CMake because we
may replace the library with a logical target name that will not
work inside a `try_compile`.
When `COMPILE_PDB_NAME` is used without `COMPILE_PDB_OUTPUT_DIRECTORY`
we cannot fall back on the link `PDB_OUTPUT_DIRECTORY` for an object
library because it has no link step.
Fixes: #16674
This reverts commit v3.7.0-rc1~275^2 (Ninja: Use full path for all
source files, 2016-08-05). Unfortunately using absolute paths can
cause incorrect rebuilds due to ninja limitations. The ninja
manual [1] explains:
> ... using absolute paths, your depfile may result in a mixture of
> relative and absolute paths. Paths used by other build rules need
> to match exactly.
Passing an absolute path to a source file to the compiler while using a
relative path in the ninja build manifest can cause such mixture and
lead to incorrect rebuilds. Simply revert the change for now.
Note that there was a follow-up to the original change in commit
v3.7.0-rc2~10^2 (Ninja: Fix RC language depfile generation with
cmcldeps, 2016-10-13). We don't need to revert that because that
change made the relevant code cleverly adapt to whatever variable
we use to reference the source file.
[1] https://ninja-build.org/manual.html#_depsFixes: #16675
Issue: #13894
624021a0 Add test for object library dependencies
5da9266a Xcode: Always track object library dependencies via hacky Makefile
e80ac953 Xcode: Record dependency information also for object libraries
9293e57d Xcode: Collect dummy rules during iteration and emit afterwards
33a1d727 Makefile: Allow adding post-build rules to object libraries
62a1e292 Autogen: Add hint to SKIP_AUTOMOC in error message
e98465cf Autogen: Add Quote function and use it for logging
96f6f392 Autogen: Add AUTOMOC_DEPEND_FILTERS documentation
94e3f82c Autogen: Add AUTOMOC_DEPEND_FILTERS test
70ebf35c Autogen: Add AUTOMOC_DEPEND_FILTERS support
2a6fd432 Autogen: Single point of return in Run() method
e9a8a207 Autogen: Log simplifications
d49176e9 libarchive: Avoid using isblank
ce8f117f libarchive: Avoid declaration after statement in C code
e0f725f8 libarchive: Fix use of ssize_t in archive_entry.h
ae4861ec libarchive: Define __LA_DEPRECATED consistently
fecb70dd Update CMake pre-cached values for libarchive 3.3.0
a59a7ee9 Merge branch 'upstream-LibArchive' into update-libarchive
c206211a LibArchive 2017-02-19 (100ee75a)
cdce7c61 libarchive: Update script to get 3.3.0
Adjust the image format to create the initial `dmg`, and convert only
when necessary. This eliminates spurious
hdiutil create -format UDRW
failures I'm seeing which may be related to an anti-virus scanner.
Running
hdiutil create -format UDRW -verbose
indicates a "resource is busy" issue when unmounting an image.
In commit v3.8.0-rc1~87^2~1 (VS: added support for C#, 2017-01-09) we
removed what looked like a no-op streaming operation but in fact it
is responsible for applying indentation. Restore the line.
Try `-fopenmp=libiomp5` in addition to `-fopenmp=libomp`. Leave the
latter as the default that we try first when the compiler id is `Clang`.
If that does not work we will still fall back to trying all the
candidates and eventually get to the iomp5 variant.
