infra: avoid excessive objdump calls and adjust all.sh - #16083
Conversation
We're currently doing excessive objdump calls in the bad build check. We only need it once to calculate the sanitizer keywords. Also adjust all.sh so it's no longer needed to set up specific Python packages to build the base images. This also makes it very easy to quickly adjust which base images to build by modifying the file. This is how it was earlier which IMO works better when adjusting the base images. Signed-off-by: David Korczynski <david@adalogics.com>
|
/gcbrun |
|
/gcbrun trial_build.py htslib json-c javacpp jansson janet --fuzzing-engines libfuzzer --sanitizers address |
Updates LLVM usage to match [6d1ace547c9f](llvm/llvm-project@6d1ace547c9f) Indexer-PiperOrigin-RevId: 956634975
|
The trial run failed because the indexer is pushed out of sync with our llvm version. See 0f4d4d3#commitcomment-198554014 Consequently, no projects were build.
|
Signed-off-by: David Korczynski <david@adalogics.com>
|
/gcbrun trial_build.py htslib json-c javacpp jansson janet --fuzzing-engines libfuzzer --sanitizers address |
|
The above succeeded: |
|
/gcbrun trial_build.py binutils boringssl bzip2 c-blosc ghostscript openssl giflib haproxy gstreamer libpng ffmpeg libjxl libavc muduo muparser --fuzzing-engines libfuzzer --sanitizers address |
Updates LLVM usage to match [cbc5a226cbf8](llvm/llvm-project@cbc5a226cbf8) Indexer-PiperOrigin-RevId: 974715562
We're currently doing excessive objdump calls in the bad build check. We only need it once to calculate the sanitizer keywords.
Also adjust all.sh so it's no longer needed to set up specific Python packages to build the base images. This also makes it very easy to quickly adjust which base images to build by modifying the file. This is how it was earlier which IMO works better when adjusting the base images.