Fix memory leak in MuPatTrackBuilder

Follow up of !37460 (merged), apparently I forgot one call-chain leading MuPatHitTool.

This merge request affects 2 files:

• MuonSpectrometer/MuonReconstruction/MuonRecTools/MuonRecToolInterfaces/MuonRecToolInterfaces/IMuonTrackFinder.h
• MuonSpectrometer/MuonReconstruction/MuonTrackMakers/MuonTrackMakerAlgs/MuonSegmentTrackMaker/src/MuPatTrackBuilder.cxx

Adding @goetz ,@rosati ,@wleight ,@nkoehler as watchers

added MuonSpectrometer master review-pending-level-1 labels

Valgrind suggests that this change doesn't fix the problem - not sure why.

CI Result SUCCESS (hash 8f09d95b)

	Athena	AthSimulation	AthGeneration	AnalysisBase
externals
cmake
make
required tests
optional tests

Full details available on this CI monitor view
Athena: number of compilation errors 0, warnings 0
AthSimulation: number of compilation errors 0, warnings 0
AthGeneration: number of compilation errors 0, warnings 0
AnalysisBase: number of compilation errors 0, warnings 0
For experts only: Jenkins output [CI-MERGE-REQUEST-CC7 23674]

marked as a Work In Progress

Hi Walter, if this is still a work in progress change, could you edit the title to start with "WIP:" please?

I did that already. I just noticed that I forgot to recompile on package for the valgrind-test I ran yesterday. The test is re-running now, more news in a few hours.

• Walter

Ah, ok, no problem!

I found one more leaking code path.

The setup looks somewhat suspicious to me, let me tag @nkoehler, @emoyse and @rosati for insight.

It turns out the AlgTool MuonTrackSteering owns three private instances of the AlgTool MooTrackBuilder

1. ToolHandle<IMuonTrackBuilder> m_trackBTool{this, "TrackBuilderTool", "Muon::MooTrackBuilder/MooMuonTrackBuilder"};
2. ToolHandle<MooTrackBuilder> m_mooBTool{this, "MooBuilderTool", "Muon::MooTrackBuilder/MooMuonTrackBuilder"};
3. ToolHandle<IMuonTrackRefiner> m_trackRefineTool{this, "TrackRefinementTool", "Muon::MooTrackBuilder/MooMuonTrackBuilder"};

The latter wasn't clean up.

Since all three have the same name, assume in rel 21 the same public instance of Muon::MooTrackBuilder was used for all purposes. Not we have three separate (isolated) copies owned by the same parent. I really wonder if this is intentional and make sense.

• Walter

added 1 commit

• 24ee6704 - MuonTrackSteering: Clean up also IMuonTrackRefiner-instance of MooTrackBuilder

Compare with previous version

unmarked as a Work In Progress

This particular leak is now fixed (just confirmed by valgrind). There is another code-path from MuonInsideOutRecoTool to an instance of MuPatHitTool that is leaking. But that's for another MR.

This merge request affects 3 packages:

• MuonSpectrometer/MuonReconstruction/MuonRecTools/MuonRecToolInterfaces
• MuonSpectrometer/MuonReconstruction/MuonTrackMakers/MuonTrackMakerAlgs/MuonSegmentTrackMaker
• MuonSpectrometer/MuonReconstruction/MuonTrackMakers/MuonTrackMakerTools/MuonTrackSteeringTools

Affected files list will not be printed in this case

Adding @goetz ,@rosati ,@wleight ,@nkoehler as watchers

CI Result SUCCESS (hash 24ee6704)

	Athena	AthSimulation	AthGeneration	AnalysisBase
externals
cmake
make
required tests
optional tests

Full details available on this CI monitor view
Athena: number of compilation errors 0, warnings 0
AthSimulation: number of compilation errors 0, warnings 0
AthGeneration: number of compilation errors 0, warnings 0
AnalysisBase: number of compilation errors 0, warnings 0
For experts only: Jenkins output [CI-MERGE-REQUEST-CC7 23770]

resolved all threads

added review-approved label and removed review-pending-level-1 label

merged

mentioned in commit a72e7482

added sweep:ignore label