fix crash in pre4 Assertion `seedCollection==nullptr' failed.

[slava77]

seedCollection needs to be reset to zero on a legal exit from the method, otherwise next call will give an assert

should fix crashes reported in https://hypernews.cern.ch/HyperNews/CMS/get/relval/3121.html
Tested with http://cms-project-relval.web.cern.ch/cms-project-relval/failures/CMSSW_7_2_0_pre4/standard/WElSkim2012D__RelVal_wEl2012D_Job1.tar.gz

[cmsbuild]

A new Pull Request was created by @slava77 (Slava Krutelyov) for CMSSW_7_2_X.

fix crash in pre4 Assertion `seedCollection==nullptr' failed.

It involves the following packages:

RecoTracker/ConversionSeedGenerators

@nclopezo, @cmsbuild, @Degano, @StoyanStoynev, @slava77 can you please review it and eventually sign? Thanks.
@ghellwig, @GiacomoSguazzoni, @rovere, @VinInn, @gpetruc, @cerati, @venturia this is something you requested to watch as well.
You can sign-off by replying to this message having '+1' in the first line of your reply.
You can reject by replying to this message having '-1' in the first line of your reply.

[slava77]

+1

for #5022 fe57c40

checked in pre4 that the assert is gone

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next CMSSW_7_2_X IBs unless changes or unless it breaks tests.

[VinInn]

Why the assert did not trigger on the usual Matrix?
(I introduced the assert just to make sure such a bug did not sneak in...)

[VinInn]

Thanks Slava for the fix

[slava77]

we don't have complex enough events in the matrix tests

the crash I was looking at was following the warning message from the previous event

%MSG-e TooManyClusters:   PhotonConversionTrajectorySeedProducerFromSingleLeg:photonConvTrajSeedFromSingleLeg  21-Aug-2014 17:38:32 CEST Run: 207454 Event: 1868
625186
Found too many clusters (28767), bailing out.

[VinInn]

The old problem of testing all possible branch combinations in integration tests: notoriously impossible...
One needs synthetic, ad-hoc, unit test for that.

[venturia]

Hi,

for this specific kind of problem a possible way to "enhance" the rate of pathological events is to run in the test a special configuration where these thresholds are artificially low so that their are triggered more often. Or you could create a skim of odd events. I used to run it privately for the tracker with the purpose of monitoring their rate.

On Aug 21, 2014, at 22:01 , David Lange <[email protected]mailto:[email protected]> wrote:

Merged #5022#5022.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/5022#event-155753647.