llvm 14.0.5 #97618
llvm 14.0.5 #97618
Conversation
BrewTestBot
added
automerge-skip
carlocab
added
long build
SMillerDev
added
the
CI-long-timeout
|
Might be able to rerun now, though Monterey CI are either slow or not running. Will need an |
|
We'll probably need an |
Bo98
removed
the
CI-long-timeout
LLDB is always such a pain on Linux. I'm traveling for work at the moment. Will have a look later today. |
|
Funnily enough, the code change that I think is causing this error seems to have been motivated by a Homebrew-installed Python: llvm/llvm-project@f07ddbc |
carlocab
force-pushed
the
bump-llvm-14.0.0
branch
from
733f0ad to
2decd87
Compare
|
Our Python 3.7 doesn't have the An upstream general fix would probably be something like this (albeit not one necessary to apply ourselves since the var option is easier): diff --git a/lldb/bindings/python/get-python-config.py b/lldb/bindings/python/get-python-config.py
index 32d82a54c1..cc3d9aaceb 100755
--- a/lldb/bindings/python/get-python-config.py
+++ b/lldb/bindings/python/get-python-config.py
@@ -42,24 +42,29 @@ def main():
else:
raise
elif args.variable_name == "LLDB_PYTHON_EXE_RELATIVE_PATH":
- tried = list()
+ tried_exe = set()
+ tried_prefix = set()
exe = sys.executable
- prefix = os.path.realpath(sys.prefix)
+ prefix = sys.prefix
while True:
try:
print(relpath_nodots(exe, prefix))
break
except ValueError:
- tried.append(exe)
+ tried_exe.add(exe)
+ tried_prefix.add(prefix)
if os.path.islink(exe):
exe = os.path.join(os.path.realpath(os.path.dirname(exe)), os.readlink(exe))
continue
+ elif os.path.islink(prefix):
+ prefix = os.path.realpath(prefix)
+ continue
else:
print("Could not find a relative path to sys.executable under sys.prefix", file=sys.stderr)
- for e in tried:
- print("tried:", e, file=sys.stderr)
- print("realpath(sys.prefix):", prefix, file=sys.stderr)
- print("sys.prefix:", sys.prefix, file=sys.stderr)
+ for e in tried_exe:
+ print("tried exe:", e, file=sys.stderr)
+ for e in tried_prefix:
+ print("tried prefix:", e, file=sys.stderr)
sys.exit(1)
elif args.variable_name == "LLDB_PYTHON_EXT_SUFFIX":
print(sysconfig.get_config_var('EXT_SUFFIX')) |
carlocab
force-pushed
the
bump-llvm-14.0.0
branch
from
2decd87 to
5194846
Compare
carlocab
added
CI-long-timeout
carlocab
added
the
CI-long-timeout
|
Linux build failed with Not sure why it can't find that header. Also, why is it trying to build |
|
ARM failed with: Odd. |
cho-m
removed
the
CI-long-timeout
|
Can't reproduce the ARM Monterey build failure locally, with I'm going to guess that the failure is either transient or a regression in Xcode 13.3. |
|
On all macOS except for Only on Intel macOS:
I'll need to investigate the rest for LLVM 14 support. |
macOS: 12.3-arm64 |
|
This pull request has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. To keep this pull request open, add a |
danielnachun
force-pushed
the
bump-llvm-14.0.0
branch
from
18467df to
03a9b56
Compare
|
I rebased this now that Qt is updated to 6.3.1. I think we still need some way to make sure there is no garbage in |
cho-m
added
the
CI-long-timeout
There was a problem hiding this comment.
I think this will be ready to merge once CI finishes (maybe ~1 hour to go). I am guessing the CI audit failures are caused by long runtime leading to GitHub API auth issues, but I don't think this blocks us from merging.
Last chance for anyone to add any review comments before we trigger a merge. Unless critical, most requested changes may be better to carry over to follow up PR(s).
|
@cho-m let's publish the bottles as soon as CI finishes. There has been ample opportunity for review feedback and all the bottles have built successfully. We have multiple things we should follow up on but they should all be handled in separate PRs. The main reason this is so slow is because of the |
cho-m
removed
the
CI-long-timeout
Created with
brew bump-formula-pr.