Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Build 13.0.1-rc1 #160

Merged
merged 93 commits into from
Jan 9, 2022
Merged

Build 13.0.1-rc1 #160

merged 93 commits into from
Jan 9, 2022

Conversation

h-vetinari
Copy link
Member

conda-forge-admin and others added 30 commits August 7, 2020 20:41
Cross compiling for aarch64 and ppc64le
automerged PR by conda-forge/automerge-action
Disable -Werror,-Wundef-prefix=TARGET_OS as they are not yet defined in the 10.9 SDK used for osx-64
automerged PR by conda-forge/automerge-action
@conda-forge-linter
Copy link

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

@h-vetinari
Copy link
Member Author

h-vetinari commented Dec 29, 2021

@wolfv, has mamba run through label/rc PRs successfully already? [snip]

EDIT: mea culpa, the merge had removed those lines from the cbc, and though I checked it before-hand, I must have managed to get it wrong... Sorry.

@h-vetinari
Copy link
Member Author

@conda-forge/clangdev

No idea what's driving up the OSX compile time so badly (perhaps just a very slow agent), but IMO this is ready.

@xhochy
Copy link
Member

xhochy commented Dec 30, 2021

Can you try to use lief_dev as we do in Tensorflow? Could be that the long times stem from there?

@h-vetinari
Copy link
Member Author

Unfortunately no change with lief_dev. I have to say it's a crazy kind of compile time regression, which is especially surprising given that it's a patch release. 13.0.0 compiled in 3h on OSX (incl. all the conda-stuff), now 6h are not enough for 13.0.1...

@h-vetinari
Copy link
Member Author

h-vetinari commented Jan 9, 2022

Could be that the long times stem from there?

In case it helps, I heavily redacted the raw log from the last run to get some idea about where the time is spent:

18:15:44 ##[section]Starting: Run OSX build
18:21:55 BUILD START: ['libclang-cpp13-13.0.1.rc1-default_haeb2ba7_0.tar.bz2', 'clang-13-13.0.1.rc1-default_haeb2ba7_0.tar.bz2', 'libclang-cpp-13.0.1.rc1-default_haeb2ba7_0.tar.bz2', 'clang-13.0.1.rc1-h9d310f0_0.tar.bz2', 'clangxx-13.0.1.rc1-default_haeb2ba7_0.tar.bz2', 'libclang-13.0.1.rc1-default_haeb2ba7_0.tar.bz2', 'clang-format-13-13.0.1.rc1-default_haeb2ba7_0.tar.bz2', 'python-clang-13.0.1.rc1-default_h53e2bf8_0.tar.bz2', 'clang-format-13.0.1.rc1-default_haeb2ba7_0.tar.bz2', 'clang-tools-13.0.1.rc1-default_haeb2ba7_0.tar.bz2', 'clangdev-13.0.1.rc1-default_haeb2ba7_0.tar.bz2']
18:22:17 Downloading source to cache: clang-13.0.1rc1.src_c8359a51f7.tar.xz
         # run global build.sh
18:22:56 -- Build files have been written to: $SRC_DIR/build
18:27:01 [ 20%] Building CXX object lib/Basic/CMakeFiles/obj.clangBasic.dir/Targets/AMDGPU.cpp.o
18:48:57 [ 40%] Building CXX object lib/Tooling/CMakeFiles/obj.clangTooling.dir/ArgumentsAdjusters.cpp.o
19:19:12 [ 60%] Building CXX object tools/extra/clang-tidy/cppcoreguidelines/CMakeFiles/obj.clangTidyCppCoreGuidelinesModule.dir/ProTypeVarargCheck.cpp.o
19:54:53 [ 80%] Building CXX object tools/extra/clangd/CMakeFiles/obj.clangDaemon.dir/index/Background.cpp.o
20:14:12 Resource usage statistics from building clang_packages:
         # output libclang-cpp13
20:14:12 Packaging libclang-cpp13
20:14:25 + make install VERBOSE=1
20:14:31 [ 20%] Built target obj.clangLex
20:14:35 [ 40%] Built target obj.clangDriver
20:14:41 [ 61%] Built target obj.clangStaticAnalyzerCheckers
20:14:49 [ 81%] Built target obj.clangTidyModernizeModule
20:15:07 Resource usage statistics from bundling libclang-cpp13:
         # output clang-13
20:15:27 Packaging clang-13
20:15:43 + make install
20:17:39 [ 11%] Built target clang-resource-headers
20:21:26 [ 20%] Building CXX object lib/Lex/CMakeFiles/obj.clangLex.dir/PPCallbacks.cpp.o
20:50:10 [ 30%] Building CXX object lib/Sema/CMakeFiles/obj.clangSema.dir/SemaStmtAttr.cpp.o
21:04:36 [ 40%] Building CXX object lib/Driver/CMakeFiles/obj.clangDriver.dir/XRayArgs.cpp.o
21:34:11 [ 50%] Building CXX object lib/Tooling/CMakeFiles/obj.clangTooling.dir/ArgumentsAdjusters.cpp.o
22:04:12 [ 60%] Building CXX object lib/StaticAnalyzer/Checkers/CMakeFiles/obj.clangStaticAnalyzerCheckers.dir/StreamChecker.cpp.o
22:19:33 [ 70%] Building CXX object tools/extra/clang-tidy/utils/CMakeFiles/obj.clangTidyUtils.dir/ExceptionAnalyzer.cpp.o
22:49:45 [ 80%] Building CXX object tools/extra/clang-tidy/modernize/CMakeFiles/obj.clangTidyModernizeModule.dir/DeprecatedHeadersCheck.cpp.o
23:26:08 [ 90%] Building CXX object tools/extra/clang-doc/CMakeFiles/obj.clangDoc.dir/HTMLGenerator.cpp.o
23:48:48 Resource usage statistics from bundling clang-13:
         # output libclang-cpp
23:49:00 Packaging libclang-cpp
23:49:12 + make install VERBOSE=1
23:54:47 [ 20%] Building CXX object lib/Lex/CMakeFiles/obj.clangLex.dir/PPCallbacks.cpp.o
00:15:38 [ 28%] Building CXX object lib/Sema/CMakeFiles/obj.clangSema.dir/SemaDecl.cpp.o
00:15:47 ##[error]The operation was canceled.
00:15:47 ##[section]Finishing: Run OSX build

It's worth noting that at this point, we still have many outputs left (see second line above):

  • build.sh
  • 'libclang-cpp13-13.0.1.rc1-default_haeb2ba7_0.tar.bz2'
  • 'clang-13-13.0.1.rc1-default_haeb2ba7_0.tar.bz2'
  • 'libclang-cpp-13.0.1.rc1-default_haeb2ba7_0.tar.bz2'
  • 'clang-13.0.1.rc1-h9d310f0_0.tar.bz2'
  • 'clangxx-13.0.1.rc1-default_haeb2ba7_0.tar.bz2'
  • 'libclang-13.0.1.rc1-default_haeb2ba7_0.tar.bz2'
  • 'clang-format-13-13.0.1.rc1-default_haeb2ba7_0.tar.bz2'
  • 'python-clang-13.0.1.rc1-default_h53e2bf8_0.tar.bz2'
  • 'clang-format-13.0.1.rc1-default_haeb2ba7_0.tar.bz2'
  • 'clang-tools-13.0.1.rc1-default_haeb2ba7_0.tar.bz2'
  • 'clangdev-13.0.1.rc1-default_haeb2ba7_0.tar.bz2'

@h-vetinari
Copy link
Member Author

So on OSX, packaging clang-13 now needs 3.5h, whereas on linux, it takes 40 seconds!

20:08:04 Packaging clang-13
20:08:44 Packaging libclang-cpp

Could it be that something from build.sh is not picked up correctly when processing the outputs?

@isuruf
Copy link
Member

isuruf commented Jan 9, 2022

Fixed now. I marked as clang_bootstrap 11.rc builds as broken to avoid different compilers in top level build vs output level build envs which caused make to recompile.

@h-vetinari
Copy link
Member Author

Awesome, thanks! Glad this could be resolved. So then there should be no impediment to merging anymore?

@isuruf isuruf merged commit 1cae51f into conda-forge:rc Jan 9, 2022
@h-vetinari h-vetinari deleted the rc branch January 9, 2022 11:55
@h-vetinari
Copy link
Member Author

h-vetinari commented Jan 9, 2022

I don't see any packages uploaded for this... I cannot really think of a reason why the CI wouldn't trigger?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

10 participants