-
Notifications
You must be signed in to change notification settings - Fork 4.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
AK5 to AK4 for JEC's, Redux #5350
AK5 to AK4 for JEC's, Redux #5350
Conversation
…ng to fix various runtests by setting the GT to run2_mc
…ies runtest. Also fixing a couple of stray AK5s from python files more than one layer deep in the package
Helgrind identified a race condition in MultiTrackSelector. The problem was StringCutObjectSelector is not const thread safe (i.e. it has a const function which modifies its internal state) and therefore only one thread at a time can call it. Therefore MultiTrackSelector can not simultaneously talk to multiple events. However, converting the module to a stream module is thread safe.
Errors reported by ubsan (undefined behavior) with GCC 4.9.1: PixelChannelIdentifier.h:40:22: runtime error: left shift of negative value -1 PixelChannelIdentifier.h:41:26: runtime error: left shift of negative value -1 PixelChannelIdentifier.h:42:23: runtime error: left shift of negative value -1 PixelChannelIdentifier.h:43:22: runtime error: left shift of negative value -1 PixelChannelIdentifier.h:44:25: runtime error: left shift of negative value -1 (1) Change ctor arguments to be unsigned int. Left shifting by a negative number is udefined behavior. (2) ~0 == -1, left shifting such number is undefined behavior. Signed-off-by: David Abdurachmanov <[email protected]>
RecoLocalTracker/SubCollectionProducers -- Fix bug on demand offline
The following adds support for Address Sanitizer (asan) IBs using mainstream branch. Because a full stack of CMSSW is not compiled with ASan we need some extra steps. ASan run-time must be linked into main executables statically to force early ASan initialization. Signed-off-by: David Abdurachmanov <[email protected]>
…treamModule Converted MultiTrackSelector from global to stream
…metry Provide only const access to ME0EtaPartitions
…nEventGenerator-4tau-CMSSW_7_2_X_2014-09-15-1400 restore assert ValidationEventGenerator 4tau decays
DDValue Cleanup
… variables rather than code copy and pastes as is currently
@rappoccio (considering that github occasionally shows more than there really is, I tried to merge locally and see diffs
I'm guessing the last one is actually the needed stuff, not the 60K lines Could you please rebase |
@slava77 yes, sorry, I have a rebase. Should I close this PR and open a new one or use this one? |
you can try here, you will see if the history come out cleaner (I have my doubts ;) ) .. if not, make a new one |
…mssw into rappoccio_AK5toAK4_72x_noHLT
OK merged here. It doesn't seem to have gotten cleaner hehehehe. I will make a new PR then. |
For the record : new PR is #5588. |
Update of PR #5299, including suggestions from @Martin-Grunewald and
@mmarioncern.
Here is the last expected change for the AK5 -> AK4 transition. The
AK4 and AK8 global tags are now present at RECO and HLT level. I have
moved the defaults from AK5/AK7 to AK4/AK8 now. I also added services
for ALL available payloads (AK1,2,3,...10). The "default" is AK4.
I have coordinated with JME Trigger contacts to include the requisite
changes to CondDB for the HLT-related software changes. However, this
will still break the Run 1 matrix tests (5.1 and 25.0) with this
current PR.
Additional changes were made to fix unit tests by changing the global
tag from "startup" to "run2_mc", which fixed the broken unit tests.
Tested on the following workflows :