Skip to content

AtlasBuildScripts: Update release notes script

Edward Moyse requested to merge emoyse/athena:update-release-notes into master

I'm opening this now to get some comments. The idea is to use gitlab labels to make the release notes a bit more descriptive. This requires the use of a gitlab token, and then if you run the command as: ./Build/AtlasBuildScripts/prepare_release_notes.py release/23.0.4 nightly/master/2022-07-12T2101 -t INSERT_TOKEN_HERE then you get the following output:

Release notes for release/23.0.4

The release release/23.0.4 was built from the tag nightly/master/2022-07-12T2101

This is the list of merge requests that were included since the previous release release/23.0.3:

  • 471824 Initial CI artifacts copying prototype Test, Tools
  • 472043 GeoExporter: Add missing QDebug include Geometry
  • 472054 AODMerge_Skeleton.py: Set AthenaEventLoopMgr.EventPrintoutInterval = 100 Reconstruction
  • 471981 Updating bs-streamerinfos in master (ATR-25602) Trigger
  • 471672 CI: Sync master and 22.0 Run 3 tests Test, Tools

If you use the --group-merge-requests option, then it will try to group MRs with the same main labels. Example output is:

Release notes for release/23.0.4

The release release/23.0.4 was built from the tag nightly/master/2022-07-12T2101

This is the list of merge requests that were included since the previous release release/23.0.3:

  • Test, Tools
    • 471824 : Initial CI artifacts copying prototype
    • 471672 : CI: Sync master and 22.0 Run 3 tests
    • 471662 : CI: Update Run 4 geometry tag, add digest check and enable PHYS test
  • Geometry
    • 472043 : GeoExporter: Add missing QDebug include
    • 471474 : Sweeping !54951 (merged) from master to 22.0. cppcheck uninitialised member
  • Reconstruction
    • 472054 : AODMerge_Skeleton.py: Set AthenaEventLoopMgr.EventPrintoutInterval = 100
  • Trigger
    • 471981 : Updating bs-streamerinfos in master (ATR-25602)

There's quite a lot in here that isn't perfect - I don't like how I've hard-coded the approved labels, for example, and the code could really do with a general clean-up, but hopefully this is useful as a proof-of-principle.

Edit: I forgot to mention that I took the opportunity to move from argparse to optparse, since argparse is deprecated since 2.7 https://docs.python.org/2/library/optparse.html

This was discussed in SW coordination and it was agreed to accept this, and use the default option: https://indico.cern.ch/event/1183702/

cc @dario @chlee and @fwinkl @aundrus @elmsheus @nstyles

Edited by Edward Moyse

Merge request reports