[cfe-dev] [3.8 Release] RC1 has been tagged

Hans Wennborg via cfe-dev cfe-dev at lists.llvm.org
Wed Jan 20 09:23:46 PST 2016


On Wed, Jan 20, 2016 at 5:25 AM, Dimitry Andric <dimitry at andric.com> wrote:
> Unfortunately I'm having lots of trouble with rc1 at this point:
> * libcxxabi can't build, because it requires unwind.h, which we do not yet have on FreeBSD 10.x (Ed Maste is working on it for 11.x, but that is not ready for general consumption).
> * The test-release.sh script has no option to disable only libcxxabi, you can only disable libcxx, libcxxabi and libunwind together (maybe this can be improved)

Yes, I'd be happy to take a patch for this, or I suppose you could
just hack it out locally when building.

> * Last time I hand-built libcxx, it still had a lot of test failures in the locale parts, but I haven't had time to investigate.

Did we have that for 3.7 too?

> * OpenMP does not support i386-freebsd, so I have to disable it there

That's perfectly fine. Including OpenMP by default is new for this
release, so if it causes any problems, just exclude it.

> * Last but not least: the host compiler on FreeBSD 10.x is clang 3.4.1 (the last version that can build without C++11 support), and it crashes with a segfault during building of CGBlocks.cpp.  I'll need to find some way to work around this failure, since we cannot upgrade the compiler easily on FreeBSD 10.x.

This sounds like the biggest problem. Is there a PR for the crash? I
suppose the alternatives are either to try not to tickle the crash in
our source, or fixing the 3.4.1 compiler?

> I also had to hack the test-release.sh script to fix a number of problems that I encountered during the 3.7.1 release, but haven't gotten to upstreaming them.  E.g. the way the source code is checked out with symlinks all over the place does not work, and I need to add a custom patch to clang-tools-extra to make the tests succeed, because there is a race condition in the Makefile.

:-( What's not working with the symlinks? Please send patches. I'm
sorry for the trouble here.

Thanks,
Hans



More information about the cfe-dev mailing list