LHCb merge requestshttps://gitlab.cern.ch/lhcb/LHCb/-/merge_requests2020-02-18T17:10:16+01:00https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/1184work towards make test_public_headers_build pass2020-02-18T17:10:16+01:00Paul Seyfertwork towards make test_public_headers_build pass * add missing include rules needed by headers that aren't compiled regularly within a package (including dependencies on public_headers that aren't tracked by cmake)
* Associators/AssociatorsBase
* Event/DigiEvent
* Event/Tr... * add missing include rules needed by headers that aren't compiled regularly within a package (including dependencies on public_headers that aren't tracked by cmake)
* Associators/AssociatorsBase
* Event/DigiEvent
* Event/TrackEvent
* Kernel/LHCbKernel
* add missing include statements to headers that were usually fulfiled by implementation files
* Day/MDF
* Det/DetDesc
* Kernel/SOAExtensions
* Kernel/LHCbKernel
https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2158Resolve LHCBPS-1845 "Decommission GaudiObjDesc"2021-01-23T02:08:51+01:00Marco Clemencicmarco.clemencic@cern.chResolve LHCBPS-1845 "Decommission GaudiObjDesc"See LHCBPS-1845 (to close once both master and run2-patches versions are merged)See LHCBPS-1845 (to close once both master and run2-patches versions are merged)https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2188Fix dependency order between TrackEvent and SOAExtensions2019-10-24T00:07:25+02:00Gerhard RavenFix dependency order between TrackEvent and SOAExtensionshttps://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2219Fix creation of TString from property2019-12-09T17:02:45+01:00Ben CouturierFix creation of TString from propertyFix creation of TString from property. I also removed the use of the obsolete USE_DDUPGRADE from Det/LbDD4hepFix creation of TString from property. I also removed the use of the obsolete USE_DDUPGRADE from Det/LbDD4hepBen CouturierBen Couturierhttps://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2220RichUtils - Remove relic dependency on VectorClass2020-02-04T17:50:18+01:00Christopher Rob Jonesjonesc@hep.phy.cam.ac.ukRichUtils - Remove relic dependency on VectorClasshttps://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2228Remove dependencies on vectorclass in Kernel packages for non intel architect...2019-12-09T11:15:23+01:00Stefan RoiserRemove dependencies on vectorclass in Kernel packages for non intel architecturesNote: In LHCbMath the Similarity tests are excluded on non intel platforms as they test vector codes. They also test the scalar execution which is now skipped.Note: In LHCbMath the Similarity tests are excluded on non intel platforms as they test vector codes. They also test the scalar execution which is now skipped.https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2231OTDet for non intel architectures2021-01-29T22:43:13+01:00Stefan RoiserOTDet for non intel architecturesfallback onto scalar version in Det/OTDet when compiling on non intel architectures.fallback onto scalar version in Det/OTDet when compiling on non intel architectures.https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2244Fix check for non-x86 architectures2019-12-05T08:58:52+01:00Rosen MatevFix check for non-x86 architecturesFixes issue introduced by !2228Fixes issue introduced by !2228https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2260Minimize new refs produced by validateWithRef2021-01-29T22:43:13+01:00Rosen MatevMinimize new refs produced by validateWithRef- For tests using `validateWithRef`, only the preprocessed stdout
is written out on failure. Standard preprocessors would filter out
the summary tables, which is why the counter tables, TTree and histogram
summaries are extracte...- For tests using `validateWithRef`, only the preprocessed stdout
is written out on failure. Standard preprocessors would filter out
the summary tables, which is why the counter tables, TTree and histogram
summaries are extracted and written out at the end.
In this way irrelevant lines would not be part of the ref files,
thereby reducing diffs and conflicts.
- Fix a bug in the counter extraction where any component (tool) with
a dot in the name would not be considered in the comparison.
- Catch exceptions in the validators and put the traceback in a `validator_exception` field.
- Expose the counter parsing with a standalone `extract_counters`.
- Add a preprocessor `GroupMessages` that groups multi-line messages.
- Add a preprocessor `BlockSkipper` that improves upon the `BlockSkipper`
from Gaudi (can be moved there).
goes together with Moore!339https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2263Add 'expected' OMP stderr messages to test exclusions2020-02-04T17:48:52+01:00Christopher Rob Jonesjonesc@hep.phy.cam.ac.ukAdd 'expected' OMP stderr messages to test exclusionsRunning clang builds generates stderr messages like
```
OMP: Warning #96: Cannot form a team with 64 threads, using 1 instead.
OMP: Hint Consider unsetting KMP_DEVICE_THREAD_LIMIT (KMP_ALL_THREADS), KMP_TEAMS_THREAD_LIMIT, and OMP_T...Running clang builds generates stderr messages like
```
OMP: Warning #96: Cannot form a team with 64 threads, using 1 instead.
OMP: Hint Consider unsetting KMP_DEVICE_THREAD_LIMIT (KMP_ALL_THREADS), KMP_TEAMS_THREAD_LIMIT, and OMP_THREAD_LIMIT (if any are set).
```
These are due to the fact we use `OMP_THREAD_LIMIT=1` to limit OMP to one thread, and are unavoidable (no way to suppress them).
This MR adds them to the exclusion line skipper in LHCbExclusions.pyhttps://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2266Switch from deprecated GaudiKernel/Counters.h to Gaudi/Accumulators.h2020-02-04T22:49:04+01:00Christopher Rob Jonesjonesc@hep.phy.cam.ac.ukSwitch from deprecated GaudiKernel/Counters.h to Gaudi/Accumulators.hFixes warning
```
/usera/jonesc/LHCbCMake/Feature/Gaudi/InstallArea/x86_64-centos7-clang8-opt/include/GaudiKernel/Counters.h:13:11: warning: warning: deprecated header, use <Gaudi/Accumulators.h> instead [-W#pragma-messages]
# prag...Fixes warning
```
/usera/jonesc/LHCbCMake/Feature/Gaudi/InstallArea/x86_64-centos7-clang8-opt/include/GaudiKernel/Counters.h:13:11: warning: warning: deprecated header, use <Gaudi/Accumulators.h> instead [-W#pragma-messages]
# pragma message "warning: deprecated header, use <Gaudi/Accumulators.h> instead"
^
1 warning generated.
```Rosen MatevRosen Matevhttps://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2268Fix range v3 deprecation warnings2020-02-04T18:08:31+01:00Marco CattaneoFix range v3 deprecation warningsThis fixes deprecation warnings in the clang builds when building against range::v3 version 0.9, which will become the default in the next LCG release.
This should have no effect at all on gcc builds and the ci-test
I am taking the...This fixes deprecation warnings in the clang builds when building against range::v3 version 0.9, which will become the default in the next LCG release.
This should have no effect at all on gcc builds and the ci-test
I am taking the liberty of adding lhcb-lcg-dev4 label, if successful the LHCb x86_64-centos7-clang8-* and Rec x86_64-centos7-clang8-dbg builds in that slot should have many less compilation warnings
This MR is independent of Rec!1843 but can be tested in the same ci-testhttps://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2292Exclude histogram headers in LHCb ref preprocessor2021-01-12T17:40:11+01:00Rosen MatevExclude histogram headers in LHCb ref preprocessorWith LHCb!2260 counter and histogram summary tables are appended
explicitly to preprocessed references to create new refs.
Histogram header lines were not filtered by the preprocessor,
which leads to them appearing twice in the new refs,...With LHCb!2260 counter and histogram summary tables are appended
explicitly to preprocessed references to create new refs.
Histogram header lines were not filtered by the preprocessor,
which leads to them appearing twice in the new refs, which in
turn makes the new refs invalid.
This MR fixes that by extending the list of exclusions.https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2307Replace run time with compile time dispatch LHCbMath similarity2020-05-15T18:01:11+02:00Christopher Rob Jonesjonesc@hep.phy.cam.ac.ukReplace run time with compile time dispatch LHCbMath similarityCloses lhcb/LHCb#71. As detailed there, there is no change in application behaviour (the capabilities found at run time did not matter anyhow, thanks to lhcb/LHCb!2265).
Closes lhcb/LHCb#71. As detailed there, there is no change in application behaviour (the capabilities found at run time did not matter anyhow, thanks to lhcb/LHCb!2265).
https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2308Always write out new ref if different from old2020-02-04T17:48:03+01:00Rosen MatevAlways write out new ref if different from oldPreviously a new ref was only written if the validation failed. Now it is always written if different.
In addition, add a check whether the preprocessor is idempotent (might lead to failures in downstream projects if dedicated problemat...Previously a new ref was only written if the validation failed. Now it is always written if different.
In addition, add a check whether the preprocessor is idempotent (might lead to failures in downstream projects if dedicated problematic preprocessors are used).
Also, add workaround for gaudi/Gaudi#108 which is a follow up from !2260 (the bug got manifested in the `brunel-upgrade-paramKalman-auxiliaries` test)https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2314Fix GroupMessages preprocessor2020-02-04T22:30:51+01:00Rosen MatevFix GroupMessages preprocessorFix idempotency of the GroupMessagess preprocessor (introduced in
!2260). Now it correctly matches lines containing `\n`
(such as after running on its own output).
Also, avoid adding extra new lines characters in grouped messages.Fix idempotency of the GroupMessagess preprocessor (introduced in
!2260). Now it correctly matches lines containing `\n`
(such as after running on its own output).
Also, avoid adding extra new lines characters in grouped messages.https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2324Fixed retrieval of sensibility in counter checking2020-02-04T22:30:38+01:00Sebastien PonceFixed retrieval of sensibility in counter checkingIt was not dealing well with cases where 2 algorithms has long names, cut to the same final string
In such case the string is appended with number of distinguish the different algos, which was breaking the algo name matchingIt was not dealing well with cases where 2 algorithms has long names, cut to the same final string
In such case the string is appended with number of distinguish the different algos, which was breaking the algo name matchinghttps://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2326Add skylake_avx512+vecwid256-centos7-gcc9-* platforms to the list of known ones2020-02-04T22:43:21+01:00Olli LuptonAdd skylake_avx512+vecwid256-centos7-gcc9-* platforms to the list of known onesShould fix [this](https://lhcb-nightlies.web.cern.ch/logs/tests/nightly/lhcb-gaudi-head/2508/skylake_avx512+vecwid256-centos7-gcc9-opt/LHCb/#failures_89) test failure in the nightlies.Should fix [this](https://lhcb-nightlies.web.cern.ch/logs/tests/nightly/lhcb-gaudi-head/2508/skylake_avx512+vecwid256-centos7-gcc9-opt/LHCb/#failures_89) test failure in the nightlies.https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2330Get libgit2 sources from LHCb repository2020-11-13T12:44:24+01:00Marco Clemencicmarco.clemencic@cern.chGet libgit2 sources from LHCb repositoryTo build the Git CondDB access library (GitEntityResolver) in old branches we download the sources of `libgit2` from the official repository.
Unfortunately one of the build machines (lblhcbpr13) does not have external connectivity, an...To build the Git CondDB access library (GitEntityResolver) in old branches we download the sources of `libgit2` from the official repository.
Unfortunately one of the build machines (lblhcbpr13) does not have external connectivity, and cannot access the official `libgit2` repository, so I copied the tar file to the LHCb repository created for LBCORE-1808, so that it can be downloaded from any build machine.
This change should be propagated to the other branches that require downloading of `libgit2`.https://gitlab.cern.ch/lhcb/LHCb/-/merge_requests/2334Fixed counter checking for counters with '|' in the name2020-02-04T22:30:26+01:00Sebastien PonceFixed counter checking for counters with '|' in the nameThis should fix tests in Gauss (lhcb-gauss-dev slot) where the counter parsing fails. E.g. gauss-config-2015-pythia8
This is an extra fix on top of !2324. So should only be merged after itThis should fix tests in Gauss (lhcb-gauss-dev slot) where the counter parsing fails. E.g. gauss-config-2015-pythia8
This is an extra fix on top of !2324. So should only be merged after itAlex PearceAlex Pearce