On Sat, Mar 28, 2020 at 07:24:08PM +0000, Ken Moffat via blfs-dev wrote:
> On Sat, Mar 28, 2020 at 06:44:50PM +0000, Ken Moffat via blfs-dev wrote:
> >
> > Just a note for everyone here who doesn't follow blfs-book :
> >
> > With llvm added in /opt/llvm and at the front of my PATH (test:
> > clang --version reports 10.0.0) I'm again seeing a failure from rust
> > issue 69225 https://github.com/rust-lang/rust/issues/69225 which is
> > why I was building 1.41.1 with its shipped llvm rather than old
> > system llvm-9.0.1 (as documented in the wiki for firefox and rust).
> >
> > Failing the test added in 1.41.1 for that problem makes me think
> > _my_ build is not safe to use. I know that Pierre only had 4
> > failures, but accidentally deleted the log (yeah, I too have done
> > that often enough) so I don't know for certain if something in my
> > rustc-1.42.0 build was fubar'd, but for the moment I'm not going to
> > procede with measuring 1.42.0 until I can try to get a handle on
> > this.
> >
> > My build used all 8 cores and my own CFLAGS, CXXFLAGS so not
> > relevant to the SBU and space figures for the book.
> >
> I've had an idea. I think that something from llvm in /usr maybe
> got pulled in by the running version from /opt. Will backup, then
> try installing llvm-10.0.0 in /usr and retry rustc.
>
> ĸen
Yes, that was it. It seems that putting a new version of llvm in
/opt while an old version is in /usr is NOT a good idea.
Interestingly, the llvm test output when I did that was somewhat
short:
[1/3] Preparing lit tests
[2/3] cd /scratch/working/llvm-10.0.0.src/tools/clang/bindings/python &&
/usr/bin/cmake -E env
CLANG_LIBRARY_PATH=/scratch/working/llvm-10.0.0.src/build/lib /usr/bin/python
-m unittest discover
........s...s.s.......s...s.............s.....................................................................................
----------------------------------------------------------------------
Ran 126 tests in 0.274s
OK (skipped=6)
[2/3] Running all regression tests
-- Testing: 51963 tests, 8 workers --
Testing: 0.. 10.. 20.. 30.. 40.. 50.. 60.. 70.. 80.. 90..
Testing Time: 412.44s
Expected Passes : 38655
Expected Failures : 73
Unsupported Tests : 13235
And that was all. No idea about the return status, I was expecting
it to report errors so my script has '|| true' after the tests.
When I used a prefix of /usr I got a lot of ninja output as tests
were compiled, then loads of messages and eventual FAIL reports,
followed by
Testing Time: 795.94s
********************
Failing Tests (7):
AddressSanitizer-x86_64-linux :: TestCases/throw_invoke_test.cpp
AddressSanitizer-x86_64-linux-dynamic :: TestCases/throw_invoke_test.cpp
ThreadSanitizer-x86_64 :: Linux/double_race.cpp
ThreadSanitizer-x86_64 :: inlined_memcpy_race.cpp
ThreadSanitizer-x86_64 :: inlined_memcpy_race2.cpp
ThreadSanitizer-x86_64 :: memcmp_race.cpp
ThreadSanitizer-x86_64 :: memcpy_race.cpp
Expected Passes : 41831
Expected Failures : 106
Unsupported Tests : 14339
Unexpected Failures: 7
FAILED: CMakeFiles/check-all
So the version I installed in /opt did not run all its tests. I
will try to remember never to use /opt for llvm, or at least never
if a version is already installed in /usr.
ĸen
--
When alle is ſayed and all is done, ye must chooſe your faces wisely,
for soon enouff ye will be playing with fyre."
The Nice and Accurate Prophecies of Agnes Nutter, Prophecy 5004
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
