Skip to content
GitLab
Projects Groups Snippets
  • /
  • Help
    • Help
    • Support
    • Community forum
    • Submit feedback
  • Sign in
  • athena athena
  • Project information
    • Project information
    • Activity
    • Labels
    • Members
  • Repository
    • Repository
    • Files
    • Commits
    • Branches
    • Tags
    • Contributors
    • Graph
    • Compare
    • Locked Files
  • Jira
    • Jira
  • Merge requests 149
    • Merge requests 149
  • CI/CD
    • CI/CD
    • Pipelines
    • Jobs
    • Schedules
  • Deployments
    • Deployments
    • Releases
  • Packages and registries
    • Packages and registries
    • Container Registry
  • Analytics
    • Analytics
    • Value stream
    • CI/CD
    • Code review
    • Repository
  • Activity
  • Graph
  • Jobs
  • Commits
Collapse sidebar
  • atlasatlas
  • athenaathena
  • Merge requests
  • !33564

Debugging changes to (MuonSegment)PseudoJetAlgorithm

  • Review changes

  • Download
  • Email patches
  • Plain diff
Merged Teng Jian Khoo requested to merge khoo/athena:master-DEBUG-MSPJG into master Jun 09, 2020
  • Overview 8
  • Commits 5
  • Pipelines 3
  • Changes 10

Addresses issue noted on ATLJETMET-1248, where the debug message attempts to access a pointer that has been moved.

@rbielski I tested with q431 on DEBUG -- this hit an error, no segfault

Expanded: To fix the failing test I found myself needing to make a bunch more changes to fix some dodgy configuration in JetRecAlgTestCfg.py and took advantage of this to clean up the ghost configuration, which had some redundancies introduced. Notably:

  • GhostScale was hardcoded in at least one place, so I just eliminated this property entirely. I generally dislike string parsing, but in this case I thought the best thing was to just check for "Ghost" in the label, rather than have people write a non-ghost label and then have the C++ prepend "Ghost" (which would also cause problems if we want e.g. HLT to have distinct container names).
  • I added a check in PseudoJetMerger so that it should not accept a list of PJCs if there is not exactly one non-ghost container (though I don't require this to be the first one, which is something we could consider enforcing). I'd actually even prefer that we explicitly configure the constituents separately from the ghosts, but we can discuss this.

I also switched JetRecAlgTestCfg to use the new JetTrimming class and made sure this was actually doing something useful (the previous setup with JetTrimmer just did nothing every event because it had no JetPseudojetRetriever). This seems to work...

FYI @delsart, @wbalunas, @chala

Edited Jun 10, 2020 by Teng Jian Khoo
Assignee
Assign to
Reviewers
Request review from
Time tracking
Source branch: master-DEBUG-MSPJG