When I removed deterministic_build_blacklist, I broke the remove_build_metadata step:
https://logs.chromium.org/logs/chromium/buildbucket/cr-buildbucket.appspot.com/8935870544761739680/+/steps/remove_build_metadata/0/stdout
Traceback (most recent call last):
File "/b/swarming/w/ir/kitchen-workdir/src/tools/determinism/remove_build_metadata.py", line 158, in <module>
sys.exit(main())
File "/b/swarming/w/ir/kitchen-workdir/src/tools/determinism/remove_build_metadata.py", line 154, in main
return remove_metadata(options.build_dir, options.recursive, bitness)
File "/b/swarming/w/ir/kitchen-workdir/src/tools/determinism/remove_build_metadata.py", line 112, in remove_metadata
with open(os.path.join(BASE_DIR, 'deterministic_build_blacklist.json')) as f:
IOError: [Errno 2] No such file or directory: '/b/swarming/w/ir/kitchen-workdir/src/tools/determinism/deterministic_build_blacklist.json'
step returned non-zero exit code: 1
It apparently fails silently, so I didn't notice until just now.
But remove_build_metadata.py is the wrong approach anyhow: Instead of writing the wrong bytes in the build and then fixing that up in a post build step, we should just write the right bytes in the first place.
Now that we're on lld on windows, we're also almost there already. remove_build_metadata.py tries to zap timestamps in zip files, but tikuta already made us write zip files without timestamps for everything but remoting already (issue 875279), so it's not really a problem that the step fails and does nothing.
We should just remove the step instead.
Comment 1 by thakis@chromium.org
, Sep 9