LLVM 9.0.1-rc2 has been tagged

Hi,

I've tagged LLVM 9.0.1-rc2. Testers can begin testing and uploading binaries.
If all goes well, this will be the last -rc.

-Tom

For this rc, I used two patches, from:

* https://bugs.llvm.org/show_bug.cgi?id=42892 - After r356631, Sanitizer-i386-Test faills to link on FreeBSD
* https://bugs.llvm.org/show_bug.cgi?id=42894 - FreeBSD needs -pthread link flag for dynamic ASan tests

Uploaded:

SHA256 (clang+llvm-9.0.1-rc2-amd64-unknown-freebsd11.tar.xz) = 72e3a872d5ff2311ec1d68e40dcc86022655c87e24583339b8331744647d8982
SHA256 (clang+llvm-9.0.1-rc2-i386-unknown-freebsd11.tar.xz) = b0bbcdc7d164502620fd4f11ef7e16e58c18aadf94441aea38d1890b951ac9da

Main test results on amd64-freebsd11:

My initial amd64 multilib Gentoo testing reveals no issues left, except
for LLDB that is buggy as usual.

Hello

Works great on Debian & Ubuntu.

I was just a bit surprise to see the python path changes in a dot releases (site-packages=>dist-packages)

Thanks

Sylvestre

Windows looks good.

$ sha256sum LLVM-9.0.1-rc2*.exe
131aa6b03a0c090ee9a5a9cd5e31fb27241e344c26380c83adb2f53b2e929f5a
LLVM-9.0.1-rc2-win32.exe
4ba66e7ea09d433dc6f41ebdb26510427ddeb3377224498fbdb839472d7505ce
LLVM-9.0.1-rc2-win64.exe

It was built with the attached batch file.

Thanks,
Hans

build_llvm_901-rc2.bat|attachment (4.9 KB)

Hello

Works great on Debian & Ubuntu.

I was just a bit surprise to see the python path changes in a dot releases (site-packages=>dist-packages)

I'm guessing this was caused by 19c735edfb7e937894463d0b315896ad972e44da.
What exactly changed for you?

Michał, was this expected?

-Tom

The install paths changed. Instead of installing some python stuff in site-packages, it installed it in dist-packages.

This wasn't a big deal to be clear!

Sylvestre

Hello,

I finally figured out that I still need Python 2.7 in order to get
tests running correctly on macOS. But this lead to a bunch of test
fails when using the script:

FAIL: LLVM :: Bindings/Go/go.test (23671 of 61786)
FAIL: ThreadSanitizer-x86_64 :: Darwin/norace-objcxx-run-time.mm
(52223 of 61786)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/assign.pass.cpp
(55919 of 61786)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/ctor_copy.pass.cpp
(55920 of 61786)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/ctor_default.pass.cpp
(55921 of 61786)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/rethrow_if_nested.pass.cpp
(55922 of 61786)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/rethrow_nested.pass.cpp
(55924 of 61786)
FAIL: libc++ ::
std/language.support/support.exception/propagation/make_exception_ptr.pass.cpp
(55933 of 61786)
FAIL: libc++ ::
std/language.support/support.exception/propagation/rethrow_exception.pass.cpp
(55934 of 61786)
FAIL: libc++ :: std/thread/futures/futures.promise/dtor.pass.cpp
(57862 of 61786)
FAIL: libc++ ::
std/thread/futures/futures.promise/set_exception.pass.cpp (57866 of
61786)
FAIL: libc++ ::
std/thread/futures/futures.promise/set_exception_at_thread_exit.pass.cpp
(57867 of 61786)
FAIL: libc++ :: std/thread/futures/futures.async/async.pass.cpp (57895 of 61786)
FAIL: libc++ :: std/thread/futures/futures.shared_future/get.pass.cpp
(57899 of 61786)
FAIL: libc++ ::
std/thread/futures/futures.task/futures.task.members/dtor.pass.cpp
(57902 of 61786)
FAIL: libc++ ::
std/thread/futures/futures.task/futures.task.members/make_ready_at_thread_exit.pass.cpp
(57913 of 61786)
FAIL: libc++ ::
std/thread/futures/futures.task/futures.task.members/operator.pass.cpp
(57915 of 61786)
FAIL: libc++ :: std/thread/futures/futures.unique_future/get.pass.cpp
(57922 of 61786)

Are these expected or should they be followed up on somehow? I am
sorry I am not more useful here - I have a fast mac pro that can run
the builds and tests but I am not to plugged into to the rest of the
internals and release process. If it's not useful that I work on this
let me know.

The binaries can still be downloaded from my GDrive:
https://drive.google.com/file/d/19sV_IUYC6WxmgR0SYEL4wEYd90jEPHo0/view?usp=sharing
SHA256: 1a8c62bc37962abc90f4caee9620dc01e4413196c344a9472dfc6f0dc9fd1176

Thanks,
Tobias

Hello,

I finally figured out that I still need Python 2.7 in order to get
tests running correctly on macOS. But this lead to a bunch of test
fails when using the script:

FAIL: LLVM :: Bindings/Go/go.test (23671 of 61786)

I've seen this go failure before on macOS, so it is somewhat expected.
Do the rest of these tests fail with 9.0.0 too ?

-Tom

Good idea. I’ll start a 9.0.0 build to see.

Tom,

Seems like the same test failures happened on 9.0.0:

FAIL: LLVM :: Bindings/Go/go.test (22699 of 61768)
FAIL: ThreadSanitizer-x86_64 :: Darwin/norace-objcxx-run-time.mm
(52206 of 61768)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/assign.pass.cpp
(55902 of 61768)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/ctor_copy.pass.cpp
(55903 of 61768)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/ctor_default.pass.cpp
(55904 of 61768)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/rethrow_if_nested.pass.cpp
(55905 of 61768)
FAIL: libc++ ::
std/language.support/support.exception/except.nested/rethrow_nested.pass.cpp
(55906 of 61768)
FAIL: libc++ ::
std/language.support/support.exception/propagation/make_exception_ptr.pass.cpp
(55916 of 61768)
FAIL: libc++ ::
std/language.support/support.exception/propagation/rethrow_exception.pass.cpp
(55917 of 61768)
FAIL: libc++ :: std/thread/futures/futures.promise/dtor.pass.cpp
(57845 of 61768)
FAIL: libc++ ::
std/thread/futures/futures.promise/set_exception.pass.cpp (57848 of
61768)
FAIL: libc++ ::
std/thread/futures/futures.promise/set_exception_at_thread_exit.pass.cpp
(57851 of 61768)
FAIL: libc++ :: std/thread/futures/futures.async/async.pass.cpp (57880 of 61768)
FAIL: libc++ :: std/thread/futures/futures.shared_future/get.pass.cpp
(57882 of 61768)
FAIL: libc++ ::
std/thread/futures/futures.task/futures.task.members/dtor.pass.cpp
(57884 of 61768)
FAIL: libc++ ::
std/thread/futures/futures.task/futures.task.members/make_ready_at_thread_exit.pass.cpp
(57895 of 61768)
FAIL: libc++ ::
std/thread/futures/futures.task/futures.task.members/operator.pass.cpp
(57897 of 61768)
FAIL: libc++ :: std/thread/futures/futures.unique_future/get.pass.cpp
(57904 of 61768)

I am guessing we should file bugs for these? But not block 9.0.1 on it.

Thanks,
Tobias

Hi,

Uploaded ARM & AArch64:
b0e4160da05a2a734cc2e73b925003a7595798f9544c522fa3e3e5815451b861
clang+llvm-9.0.1-rc2-aarch64-linux-gnu.tar.xz
2c5c08fd4a5fb290d3f148b806bb708832239030fe39529746de3e746bbf5cdd
clang+llvm-9.0.1-rc2-armv7a-linux-gnueabihf.tar.xz

Same as rc1.

Cheers,
Diana

Well, we did notice it in the original patch but I've forgotten about it
when backporting. To be honest, I have mixed feelings about it but I
don't think there's a good way of fixing the original problem without
changing the path.

If this is not a big deal to Sylvestre, I'd like to keep it as-is.
To the best of my knowledge, the path change happens on Debian only,
and specifically because Debian patches Python for the reasons I don't
know. In any case, the dist- path seems more correct for distribution
packages.