atlasexternals merge requestshttps://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests2024-03-25T20:02:41+01:00https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1114Add prmon to AthGenerationExternals2024-03-25T20:02:41+01:00Alaettin Serhan MeteAdd prmon to AthGenerationExternalsAdd `prmon` to the list of external packages for `AthGeneration` so that transform jobs can find it out of the box.
cc: @elmsheus @walkerr @tsulaiaAdd `prmon` to the list of external packages for `AthGeneration` so that transform jobs can find it out of the box.
cc: @elmsheus @walkerr @tsulaiahttps://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1113Draft: Updates for python 3.11 with AnalysisBase2024-03-28T15:16:30+01:00Johannes ElmsheuserDraft: Updates for python 3.11 with AnalysisBaseThis MR updates the AnalysisBase externals to python 3.11 (ATLINFR-5217) and packages version to the new dev3python311 LCG nightly versions (SPI-2493) except:
* setuptools
* Lhapdf 6.5.4 and download it from the Lhapdf webpage for now.
...This MR updates the AnalysisBase externals to python 3.11 (ATLINFR-5217) and packages version to the new dev3python311 LCG nightly versions (SPI-2493) except:
* setuptools
* Lhapdf 6.5.4 and download it from the Lhapdf webpage for now.
When updating to setuptools 69.1.0 there are issues with compiling numpy since the distutils python package has changed and will be removed in python 3.12 btw., so we will have to update the cmake configuration for several packages at some point.
Updated the TBB build configuration from make to cmake.
Add in addition AnalysisBaseExternals/x86_64-el9-gcc13 to the CI pipeline.
Let's keep MR in draft until we decide to move to python 3.11 at some point. Tagging @akraszna, @rhauser, @nstyles, @jchapman, @jcatmore for info.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1112Always set CMP0069 to new to enforce LTO use if requested2024-03-19T15:58:50+01:00Benjamin Morganben.morgan@warwick.ac.ukAlways set CMP0069 to new to enforce LTO use if requestedCMake 3.9 introduced support for link time optimization, but the `AtlasFunctions` module sets `cmake_minimum_required` to 3.6. This disables LTO even in CMake versions above 3.9 and means the `CMAKE_INTERPROCEDURAL_OPTIMIZATION` flag and...CMake 3.9 introduced support for link time optimization, but the `AtlasFunctions` module sets `cmake_minimum_required` to 3.6. This disables LTO even in CMake versions above 3.9 and means the `CMAKE_INTERPROCEDURAL_OPTIMIZATION` flag and associated target property cannot be used to enable LTO either locally or globally.
Retain CMake minimum version to avoid surprises, set the LTO related policy CMP0069 available from CMake 3.9 to enforce use of LTO if requested.
There should be no side effects to setting this policy. Its primary benefit is to allow projects using `AtlasFunctions` like Athena/AthSimulation to configure with `cmake -DCMAKE_INTERPROCEDURAL_OPTIMIZATION=ON ...`, or to set the `INTERPROCEDURAL_OPTIMIZATION` property on individual targets. CMake will then ensure the flags appropriate for the compiler are added to the build/link commands as needed, if CMake knows them.
The need for this feature comes from ATLASSIM-6058 on using LTO to further optimization the `AtlasGeant4` "big" library.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1111Update CI testing to LCG_104d and LCG_104d_ATLAS_92024-03-13T11:48:51+01:00Johannes ElmsheuserUpdate CI testing to LCG_104d and LCG_104d_ATLAS_9Update the CI testing to LCG_104d and LCG_104d_ATLAS_9 for ROOT 6.28/12.Update the CI testing to LCG_104d and LCG_104d_ATLAS_9 for ROOT 6.28/12.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1109GeoModel version bumped to 5.3.02024-03-12T19:47:27+01:00Vakhtang TsulaiaGeoModel version bumped to 5.3.0Athena has already been switched to GeoModel 5.3.0 using override variables in build scripts. This MR catches up with the latest GeoModel version. Once this change makes it into Athena main, the overriders in the build scripts can be cle...Athena has already been switched to GeoModel 5.3.0 using override variables in build scripts. This MR catches up with the latest GeoModel version. Once this change makes it into Athena main, the overriders in the build scripts can be cleaned up.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1105Trion Client as External2024-03-26T23:59:16+01:00Xiangyang JuTrion Client as ExternalThe MR is to install TritonClient to the ATLASExternals.
The TritonClient depends on the following packages, which are installed in the LCG:
* [gRPC](https://grpc.io)
* [Protobuf](https://developers.google.com/protocol-buffers)
* [ABSL...The MR is to install TritonClient to the ATLASExternals.
The TritonClient depends on the following packages, which are installed in the LCG:
* [gRPC](https://grpc.io)
* [Protobuf](https://developers.google.com/protocol-buffers)
* [ABSL](https://abseil.io)
* [C-ares](https://c-ares.haxx.se)
TritonClient is wrapped in a tar file located at: `https://portal.nersc.gov/cfs/m3443/xju/tritonclient_v24.02.tar.bz2`, for developping and testing. It is taken from the [client](https://github.com/triton-inference-server/client/tree/r24.02/src/c%2B%2B) repository from the branch `r24.02`.
The current implementation builds well **only if** the `LD_LIBRARY_PATH` is pre-configured: `export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:/cvmfs/sft.cern.ch/lcg/releases/LCG_104d_ATLAS_5/grpc/1.48.0/x86_64-el9-gcc13-opt/lib`. Because when building TritonClient, the `grpc_cpp_plugin` executable is called. It raises an issue of not finding its dependent, shared library. The `LD_LIBRARY_PATH` is then needed to solve that issue.Xiangyang JuXiangyang Juhttps://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1098Draft: Compress debug sections2024-01-22T16:33:38+01:00Reiner HauserDraft: Compress debug sectionsWhen separating the debug sections from libraries and executables, compress them if possible. This works with the GNU/system objcopy but not with llvm-objcopy, so it depends on the CMAKE_COMPILER_ID.
The default is zlib compression, but...When separating the debug sections from libraries and executables, compress them if possible. This works with the GNU/system objcopy but not with llvm-objcopy, so it depends on the CMAKE_COMPILER_ID.
The default is zlib compression, but once CentOS 7 is gone this can be changed to --compress-debug-sections=zstd and one gains a bit more.
The gains seen in TDAQ were about a factor 2-3 reduction in size of the debug information. In total the binary part of a full TDAQ build went from 4.3 GByte to 1.6 GByte (plus a few hundred MByte of arch independent stuff).
So far there were no drawbacks seen, gdb etc. works, although it obviously has to decompress the information during the debug session.
Tagging @akraszna , do you think there are any other drawbacks ?
I will put a few numbers for athena libraries here once I figure out how to use my private version without building all externals.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1082Draft: FindTBB.cmake Update, main branch (2023.11.07.)2023-11-13T11:49:49+01:00Attila KrasznahorkayDraft: FindTBB.cmake Update, main branch (2023.11.07.)Made `FindTBB.cmake` wrap TBB's own CMake configuration. Thereby not requiring our code to parse the TBB headers for the version information of the library.
This is to fix ATLINFR-5140.
Note that `FindTBB.cmake` is such a fundamental p...Made `FindTBB.cmake` wrap TBB's own CMake configuration. Thereby not requiring our code to parse the TBB headers for the version information of the library.
This is to fix ATLINFR-5140.
Note that `FindTBB.cmake` is such a fundamental part of our build setup that I'll need to run a bunch more tests on this update before we could merge this in.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1076Add Acts to the AthSimulation externals2023-10-23T13:47:11+02:00Johannes Junggeburthjohannes.josef.junggeburth@cern.chAdd Acts to the AthSimulation externalsHi everybody,
if we want to use the new Muon sensitive detectors in AthSimulation for testing purposes, we need the new ReadoutGeometry which is in turn relying on the Acts surfaces et al.
Tagging: @jchapman, @jcatmore, @schaarsc, @mb...Hi everybody,
if we want to use the new Muon sensitive detectors in AthSimulation for testing purposes, we need the new ReadoutGeometry which is in turn relying on the Acts surfaces et al.
Tagging: @jchapman, @jcatmore, @schaarsc, @mbandierhttps://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1018Add CLHEP to AnalysisBaseExternals2023-08-15T08:58:22+02:00Baptiste Ravinabaptiste.ravina@cern.chAdd CLHEP to AnalysisBaseExternalsIn !998 we tried to introduce the BootstrapGenerator package recommended by the ATLAS Statistics Forum into AnalysisBaseExternals, in order to write a CP algorithm for it and introduce it into standard analysis code in release 22. The MR...In !998 we tried to introduce the BootstrapGenerator package recommended by the ATLAS Statistics Forum into AnalysisBaseExternals, in order to write a CP algorithm for it and introduce it into standard analysis code in release 22. The MR was closed due to concerns over licensing.
Instead, we now would like to make our own bootstrapping algorithm, based this time on CLHEP.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1015CPack Cleanup, master branch (2023.06.02.)2023-06-02T15:18:21+02:00Attila KrasznahorkayCPack Cleanup, master branch (2023.06.02.)Simplified the [CPack](https://cmake.org/cmake/help/latest/module/CPack.html) configuration in [AtlasCMake](Build/AtlasCMake) a bit.
The goal is to try to ensure that RPM and TGZ packages would be created correctly (i.e. with the same l...Simplified the [CPack](https://cmake.org/cmake/help/latest/module/CPack.html) configuration in [AtlasCMake](Build/AtlasCMake) a bit.
The goal is to try to ensure that RPM and TGZ packages would be created correctly (i.e. with the same layout as before) with either setting of `ATLAS_USE_CUSTOM_CPACK_INSTALL_SCRIPT`. Since as @miholzbo found a few weeks ago, RPMs are not correctly set up currently with `ATLAS_USE_CUSTOM_CPACK_INSTALL_SCRIPT=FALSE`.
:fingers_crossed: that this will work correctly in all circumstances. I did a fair amount of testing, but we'll see... :worried:https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/1001Celeritas, master branch (2023.03.01.)2023-11-28T05:36:42+01:00Attila KrasznahorkayCeleritas, master branch (2023.03.01.)This is a first attempt at building [Celeritas](https://github.com/celeritas-project/celeritas) as an external project for [Athena](https://gitlab.cern.ch/atlas/athena/-/tree/master/Projects/Athena) (and eventually [AthSimulation](https:...This is a first attempt at building [Celeritas](https://github.com/celeritas-project/celeritas) as an external project for [Athena](https://gitlab.cern.ch/atlas/athena/-/tree/master/Projects/Athena) (and eventually [AthSimulation](https://gitlab.cern.ch/atlas/athena/-/tree/master/Projects/AthSimulation).
I had to adjust the [VecGeom](External/VecGeom) and [Geant4](External/Geant4) builds to make it possible to build Celeritas. And even with those modifications, the way that [HepMC3](https://gitlab.cern.ch/hepmc/HepMC3) is picked up by the build is still pretty fragile. But I guess as a first exercise this could be good enough.
Pinging @pcanal, @leggett, @calaf, @bmorgan.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/993Draft: Make Python build a bit more flexible2023-02-07T11:29:09+01:00Tadej Novaktadej.novak@cern.chDraft: Make Python build a bit more flexibleMake custom Python builds a bit more flexible so one can actually override the version.
- Define `ATLAS_BUILD_PYTHON_VERSION_MAJOR` and `ATLAS_BUILD_PYTHON_VERSION_MINOR` in projects that may build python themselves.
- Remove all hardcod...Make custom Python builds a bit more flexible so one can actually override the version.
- Define `ATLAS_BUILD_PYTHON_VERSION_MAJOR` and `ATLAS_BUILD_PYTHON_VERSION_MINOR` in projects that may build python themselves.
- Remove all hardcodes of `python3` and `python3.9` (the former is probably overkill, but at least we're ready for `python4` now 😊).
This allows me to build with Python 3.10 (note that LHAPDF does not build with it, but this is a completely separate issue).
Keeping draft to allow more discussion first.
/cc @akrasznahttps://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/992Add clang test to the CI2023-03-23T15:13:31+01:00Tadej Novaktadej.novak@cern.chAdd clang test to the CIAdd clang test to the CI. Depends on fixes needed for AnalysisBase.Add clang test to the CI. Depends on fixes needed for AnalysisBase.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/929ARM CI Build(s), master branch (2022.02.21.)2023-10-04T13:29:22+02:00Attila KrasznahorkayARM CI Build(s), master branch (2022.02.21.)Enabled the build of a couple of the projects on ARM64 as part of the CI.
The `k8s-arm` tag was advertised on: https://mattermost.web.cern.ch/it-dep/channels/aarch64-arm64 Apparently it's not quite "official" yet, but I wanted to jump o...Enabled the build of a couple of the projects on ARM64 as part of the CI.
The `k8s-arm` tag was advertised on: https://mattermost.web.cern.ch/it-dep/channels/aarch64-arm64 Apparently it's not quite "official" yet, but I wanted to jump on it already. :wink:
Note that I encountered some issues with the LCG GCC installation and Xerces-C package while playing with this MR. See: https://sft.its.cern.ch/jira/browse/SPI-2432
Pinging @elmsheus and @fwinkl.https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/845Use a better way to find PYTHONHOME from the found python, to avoid guessing...2021-09-27T13:31:41+02:00Will ButtingerUse a better way to find PYTHONHOME from the found python, to avoid guessing...Hi @akraszna
First off ... I didn't mean to push to a branch in this repo, I should have sent this to my fork and didnt even expect to have permission to push here, but it's happened now, so hope we can continue here anyway ....
I wa...Hi @akraszna
First off ... I didn't mean to push to a branch in this repo, I should have sent this to my fork and didnt even expect to have permission to push here, but it's happened now, so hope we can continue here anyway ....
I was resurrecting the demonstrator project you created a while back for a stat software release and when running it on my local machine I noticed that if I told it not to build python as part of the release build then my local python installation was picked up but the PYTHONPATH it inferred from it's path (`/usr/local/bin/python`) would then cause problems because that was just a symlink to the real location of python. The fix was using a snippet of cmake from the original FindPython module which infers the path by calling python.
Do you think this is a sensible improvement to the code?Attila KrasznahorkayAttila Krasznahorkayhttps://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/719WIP:Add clang-tidy to clang build2020-09-03T11:17:53+02:00Stewart Martin-HaughWIP:Add clang-tidy to clang buildAdd clang-tidy to clang build with `bugprone` and `performance` checks all enabled.
Sample output:
```
/scratch/smh/athena/Calorimeter/CaloCnv/CaloDetMgrDetDescrCnv/src/CaloMgrDetDescrCnv.cxx:77:8: warning: the variable 'ctx' is copy-co...Add clang-tidy to clang build with `bugprone` and `performance` checks all enabled.
Sample output:
```
/scratch/smh/athena/Calorimeter/CaloCnv/CaloDetMgrDetDescrCnv/src/CaloMgrDetDescrCnv.cxx:77:8: warning: the variable 'ctx' is copy-constructed from a const reference but is only used as const reference; consider making it a const reference [performance-unnecessary-copy-initialization]
auto ctx = Gaudi::Hive::currentContext();
^
const &
```
Overall we end up with a lot of warnings - maybe more than we really want to fix? e.g.
```
/scratch/smh/athena/Calorimeter/CaloTrackingGeometry/src/CaloSurfaceBuilder.cxx:1124:17: warning: narrowing conversion from 'double' to 'float' [bugprone-narrowing-conversions]
float eta = etaMin+(ie+0.5)*deta;
```
@christos has a set of checks:
https://gitlab.cern.ch/ATLAS-EGamma/Software/Reconstruction/clang-tidy-checks/-/blob/master/.clang-tidy
which might be a good starting point.
It might also be useful to be able to steer this from within athena rather than atlasexternals if we end up making lots of changes to the list of warnings.
Cheers,
Stewarthttps://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/642WIP: LCG Hashed RPM Usage, master branch (2020.03.23.)2020-04-02T08:34:44+02:00Attila KrasznahorkayWIP: LCG Hashed RPM Usage, master branch (2020.03.23.)This is first stab at using the "hashed RPMs" provided by the SFT group.
There are 2 main parts to it:
- The `lcg_need_rpm(...)` function was updated to set up a dependency on the "hashed RPM" instead of the "link RPM";
* For ...This is first stab at using the "hashed RPMs" provided by the SFT group.
There are 2 main parts to it:
- The `lcg_need_rpm(...)` function was updated to set up a dependency on the "hashed RPM" instead of the "link RPM";
* For this to work, the hash ID of the packages had to be tweaked a bit;
- A post-installation script had to be introduced for setting up the symbolic links after the "hashed RPMs" are installed.
It's the second part that is quite tricky. I took @rhauser's script, and tried to tweak it according to my own taste. :stuck_out_tongue: A couple of points about this:
- For now I didn't actually look at `LCG_97rc4`, as the installed files for that on http://cern.ch/lcgpackages is a bit of a mess. Instead I've set up the script (`rpmPostInstall.sh.in`) to work with existing releases. In practice I was experimenting with [LCG_96b](http://lcginfo.cern.ch/release/96b/).
- The script itself is quite self-reliant at this point, apart from not knowing automatically where to look for the LCG installation. :frowning: I was hoping that I could rely on `$RPM_INSTALL_PREFIX` for this, but of course the installation prefix can be (and usually is) different for the LCG and ATLAS packages. So instead we would have to run (a)yum with the `LCG_INSTALL_PREFIX` environment variable set to have this script behave correctly.
- I didn't bother trying to set up a post-uninstallation script. Unfortunately because of how this script works, uninstallation is just not a well-defined operation for these symbolic links. (When uninstalling a given ATLAS RPM, it's nearly impossible to tell which symbolic links may need to be deleted...)
To give you a bit of a taste, the code produces this type of output when installing an RPM into a Docker image:
```
Step 9/11 : RUN yum install -y RPMTestProject_1.0.0_x86_64-centos7-gcc8-opt
---> Running in 1e3fdb07698b
Loaded plugins: changelog, fastestmirror, kernel-module, ovl, protectbase,
: tsflags, versionlock
Loading mirror speeds from cached hostfile
252 packages excluded due to repository protections
Resolving Dependencies
--> Running transaction check
---> Package RPMTestProject_1.0.0_x86_64-centos7-gcc8-opt.noarch 0:1.0.0-1 will be installed
--> Processing Dependency: mcutils-f0b2f_1.3.3_x86_64_centos7_gcc8_opt for package: RPMTestProject_1.0.0_x86_64-centos7-gcc8-opt-1.0.0-1.noarch
--> Processing Dependency: Boost-eebf1_1.70.0_x86_64_centos7_gcc8_opt for package: RPMTestProject_1.0.0_x86_64-centos7-gcc8-opt-1.0.0-1.noarch
--> Running transaction check
---> Package Boost-eebf1_1.70.0_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-96 will be installed
--> Processing Dependency: Python-e553a_2.7.16_x86_64_centos7_gcc8_opt for package: Boost-eebf1_1.70.0_x86_64_centos7_gcc8_opt-1.0.0-96.noarch
---> Package mcutils-f0b2f_1.3.3_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-96 will be installed
--> Processing Dependency: heputils-8ac2c_1.3.2_x86_64_centos7_gcc8_opt for package: mcutils-f0b2f_1.3.3_x86_64_centos7_gcc8_opt-1.0.0-96.noarch
--> Running transaction check
---> Package Python-e553a_2.7.16_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-96 will be installed
--> Processing Dependency: sqlite-472f2_3280000_x86_64_centos7_gcc8_opt for package: Python-e553a_2.7.16_x86_64_centos7_gcc8_opt-1.0.0-96.noarch
---> Package heputils-8ac2c_1.3.2_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-95 will be installed
--> Processing Dependency: gcc_8.2.0_x86_64_centos7 for package: heputils-8ac2c_1.3.2_x86_64_centos7_gcc8_opt-1.0.0-95.noarch
--> Running transaction check
---> Package gcc_8.2.0_x86_64_centos7.noarch 0:3.0.0-1 will be installed
---> Package sqlite-472f2_3280000_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-96 will be installed
--> Finished Dependency Resolution
Beginning Kernel Module Plugin
Finished Kernel Module Plugin
Dependencies Resolved
================================================================================
Package Arch Version Repository Size
================================================================================
Installing:
RPMTestProject_1.0.0_x86_64-centos7-gcc8-opt noarch 1.0.0-1 local 150 k
Installing for dependencies:
Boost-eebf1_1.70.0_x86_64_centos7_gcc8_opt noarch 1.0.0-96 lcg-testing 17 M
Python-e553a_2.7.16_x86_64_centos7_gcc8_opt noarch 1.0.0-96 lcg-testing 20 M
gcc_8.2.0_x86_64_centos7 noarch 3.0.0-1 lcg_contrib 121 M
heputils-8ac2c_1.3.2_x86_64_centos7_gcc8_opt noarch 1.0.0-95 lcg 20 k
mcutils-f0b2f_1.3.3_x86_64_centos7_gcc8_opt noarch 1.0.0-96 lcg-testing 28 k
sqlite-472f2_3280000_x86_64_centos7_gcc8_opt noarch 1.0.0-96 lcg-testing 7.1 M
Transaction Summary
================================================================================
Install 1 Package (+6 Dependent packages)
Total download size: 165 M
Installed size: 813 M
Downloading packages:
--------------------------------------------------------------------------------
Total 18 MB/s | 165 MB 00:09
Running transaction check
Running transaction test
Transaction test succeeded
Running transaction
Installing : gcc_8.2.0_x86_64_centos7-3.0.0-1.noarch 1/7
Replacing /afs/cern.ch/cms/CAF/CMSCOMM/COMM_ECAL/dkonst/GCC/lcgcmake-install -> /opt/lcg ...
Installing : sqlite-472f2_3280000_x86_64_centos7_gcc8_opt-1.0.0-96.noar 2/7
Checking for post install in /opt/lcg/sqlite/3280000-472f2/x86_64-centos7-gcc8-opt
Invoking post install /opt/lcg/sqlite/3280000-472f2/x86_64-centos7-gcc8-opt/.post-install.sh
Replacing /build/jenkins/workspace/lcg_release_tar/BUILDTYPE/Release/COMPILER/gcc8binutils/LABEL/centos7/install -> /opt/lcg [or ] ...
Installing : Python-e553a_2.7.16_x86_64_centos7_gcc8_opt-1.0.0-96.noarc 3/7
Checking for post install in /opt/lcg/Python/2.7.16-e553a/x86_64-centos7-gcc8-opt
Invoking post install /opt/lcg/Python/2.7.16-e553a/x86_64-centos7-gcc8-opt/.post-install.sh
Replacing /build/jenkins/workspace/lcg_release_tar/BUILDTYPE/Release/COMPILER/gcc8binutils/LABEL/centos7/install -> /opt/lcg [or ] ...
Installing : Boost-eebf1_1.70.0_x86_64_centos7_gcc8_opt-1.0.0-96.noarch 4/7
Checking for post install in /opt/lcg/Boost/1.70.0-eebf1/x86_64-centos7-gcc8-opt
Invoking post install /opt/lcg/Boost/1.70.0-eebf1/x86_64-centos7-gcc8-opt/.post-install.sh
Replacing /build/jenkins/workspace/lcg_release_tar/BUILDTYPE/Release/COMPILER/gcc8binutils/LABEL/centos7/install -> /opt/lcg [or ] ...
Installing : heputils-8ac2c_1.3.2_x86_64_centos7_gcc8_opt-1.0.0-95.noar 5/7
Checking for post install in /opt/lcg/MCGenerators/heputils/1.3.2-8ac2c/x86_64-centos7-gcc8-opt
Invoking post install /opt/lcg/MCGenerators/heputils/1.3.2-8ac2c/x86_64-centos7-gcc8-opt/.post-install.sh
Replacing /build/jenkins/workspace/lcg_release_tar/BUILDTYPE/Release/COMPILER/gcc8binutils/LABEL/centos7/install -> /opt/lcg ...
Installing : mcutils-f0b2f_1.3.3_x86_64_centos7_gcc8_opt-1.0.0-96.noarc 6/7
Checking for post install in /opt/lcg/MCGenerators/mcutils/1.3.3-f0b2f/x86_64-centos7-gcc8-opt
Invoking post install /opt/lcg/MCGenerators/mcutils/1.3.3-f0b2f/x86_64-centos7-gcc8-opt/.post-install.sh
Replacing /build/jenkins/workspace/lcg_release_tar/BUILDTYPE/Release/COMPILER/gcc8binutils/LABEL/centos7/install -> /opt/lcg [or ] ...
Installing : RPMTestProject_1.0.0_x86_64-centos7-gcc8-opt-1.0.0-1.noarc 7/7
Setting up links under "/opt/lcg" using "http://cern.ch/lcgpackages/tarFiles/releases/LCG_96b_x86_64-centos7-gcc8-opt.txt"
Creating link: /opt/lcg/LCG_96b/sqlite/3280000/x86_64-centos7-gcc8-opt -> ../../../sqlite/3280000-472f2/x86_64-centos7-gcc8-opt
Creating link: /opt/lcg/LCG_96b/MCGenerators/mcutils/1.3.3/x86_64-centos7-gcc8-opt -> ../../../../MCGenerators/mcutils/1.3.3-f0b2f/x86_64-centos7-gcc8-opt
Creating link: /opt/lcg/LCG_96b/MCGenerators/heputils/1.3.2/x86_64-centos7-gcc8-opt -> ../../../../MCGenerators/heputils/1.3.2-8ac2c/x86_64-centos7-gcc8-opt
Creating link: /opt/lcg/LCG_96b/Boost/1.70.0/x86_64-centos7-gcc8-opt -> ../../../Boost/1.70.0-eebf1/x86_64-centos7-gcc8-opt
Creating link: /opt/lcg/LCG_96b/Python/2.7.16/x86_64-centos7-gcc8-opt -> ../../../Python/2.7.16-e553a/x86_64-centos7-gcc8-opt
Verifying : gcc_8.2.0_x86_64_centos7-3.0.0-1.noarch 1/7
Verifying : Boost-eebf1_1.70.0_x86_64_centos7_gcc8_opt-1.0.0-96.noarch 2/7
Verifying : sqlite-472f2_3280000_x86_64_centos7_gcc8_opt-1.0.0-96.noar 3/7
Verifying : heputils-8ac2c_1.3.2_x86_64_centos7_gcc8_opt-1.0.0-95.noar 4/7
Verifying : mcutils-f0b2f_1.3.3_x86_64_centos7_gcc8_opt-1.0.0-96.noarc 5/7
Verifying : Python-e553a_2.7.16_x86_64_centos7_gcc8_opt-1.0.0-96.noarc 6/7
Verifying : RPMTestProject_1.0.0_x86_64-centos7-gcc8-opt-1.0.0-1.noarc 7/7
Installed:
RPMTestProject_1.0.0_x86_64-centos7-gcc8-opt.noarch 0:1.0.0-1
Dependency Installed:
Boost-eebf1_1.70.0_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-96
Python-e553a_2.7.16_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-96
gcc_8.2.0_x86_64_centos7.noarch 0:3.0.0-1
heputils-8ac2c_1.3.2_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-95
mcutils-f0b2f_1.3.3_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-96
sqlite-472f2_3280000_x86_64_centos7_gcc8_opt.noarch 0:1.0.0-96
Complete!
```
And then I see the following in the image:
```
[bash][atlas]:workdir > ls -l /opt/lcg/
total 28
drwxr-xr-x 3 root root 4096 Mar 23 18:32 Boost
drwxr-xr-x 6 root root 4096 Mar 23 18:32 LCG_96b
drwxr-xr-x 4 root root 4096 Mar 23 18:32 MCGenerators
drwxr-xr-x 3 root root 4096 Mar 23 18:32 Python
drwxr-xr-x 4 root root 4096 Dec 13 14:20 binutils
drwxr-xr-x 1 root root 4096 Mar 23 18:32 gcc
drwxr-xr-x 3 root root 4096 Mar 23 18:32 sqlite
[bash][atlas]:workdir > ls -l /opt/lcg/LCG_96b/
total 24
drwxr-xr-x 3 root root 4096 Mar 23 18:32 Boost
-rw-r--r-- 1 root root 301 Mar 23 18:32 LCG_externals_x86_64-centos7-gcc8-opt.txt
-rw-r--r-- 1 root root 245 Mar 23 18:32 LCG_generators_x86_64-centos7-gcc8-opt.txt
drwxr-xr-x 4 root root 4096 Mar 23 18:32 MCGenerators
drwxr-xr-x 3 root root 4096 Mar 23 18:32 Python
drwxr-xr-x 3 root root 4096 Mar 23 18:32 sqlite
[bash][atlas]:workdir > ls -l /opt/lcg/LCG_96b/Boost/1.70.0/
total 0
lrwxrwxrwx 1 root root 51 Mar 23 18:32 x86_64-centos7-gcc8-opt -> ../../../Boost/1.70.0-eebf1/x86_64-centos7-gcc8-opt
[bash][atlas]:workdir > more /opt/lcg/LCG_96b/LCG_externals_x86_64-centos7-gcc8-opt.txt
PLATFORM: x86_64-centos7-gcc8-opt
VERSION: 96b
COMPILER: GNU 8.3.0
sqlite; 472f2; 3280000; ./sqlite/3280000/x86_64-centos7-gcc8-opt;
Boost; eebf1; 1.70.0; ./Boost/1.70.0/x86_64-centos7-gcc8-opt; Python-2.7.16-e553a
Python; e553a; 2.7.16; ./Python/2.7.16/x86_64-centos7-gcc8-opt; sqlite-3280000-472f2
[bash][atlas]:workdir > more /opt/lcg/LCG_96b/LCG_generators_x86_64-centos7-gcc8-opt.txt
PLATFORM: x86_64-centos7-gcc8-opt
VERSION: 96b
COMPILER: GNU 8.3.0
mcutils; f0b2f; 1.3.3; ./MCGenerators/mcutils/1.3.3/x86_64-centos7-gcc8-opt; heputils-1.3.2-8ac2c
heputils; 8ac2c; 1.3.2; ./MCGenerators/heputils/1.3.2/x86_64-centos7-gcc8-opt;
[bash][atlas]:workdir >
```
I.e. It "sort of" works. :wink:
As long as we synchronise how TDAQ sets up its post-install script with what we do, this could work...
Also pinging @akazarov, @fwinkl, @wlampl and @emoyse. For this to work, we'll need to make sure that text files like this will continue to be published by SFT on their "repository webpage".https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/378WIP: Adding CommonStatTools as external package2018-11-20T17:29:31+01:00Iacopo VivarelliWIP: Adding CommonStatTools as external packageAdded CommonStatTools as per request at https://its.cern.ch/jira/browse/ATLINFR-2751Added CommonStatTools as per request at https://its.cern.ch/jira/browse/ATLINFR-2751https://gitlab.cern.ch:8443/atlas/atlasexternals/-/merge_requests/223WIP:Remove GAUDI_V20_COMPAT, HAVE_GAUDI_PLUGINSVC and ATLAS_GAUDI_V21 flags, ...2022-02-17T09:59:05+01:00Stewart Martin-HaughWIP:Remove GAUDI_V20_COMPAT, HAVE_GAUDI_PLUGINSVC and ATLAS_GAUDI_V21 flags, fix for ATLINFR-1086Fairly trivial, but GAUDI_V20_COMPAT is no longer used in master so it can be removed from the AthenaExternals build flags.
The HAVE_GAUDI_PLUGINSVC and ATLAS_GAUDI_V21 flags are still used but I am tackling this in https://gitlab.cer...Fairly trivial, but GAUDI_V20_COMPAT is no longer used in master so it can be removed from the AthenaExternals build flags.
The HAVE_GAUDI_PLUGINSVC and ATLAS_GAUDI_V21 flags are still used but I am tackling this in https://gitlab.cern.ch/atlas/athena/merge_requests/8424
Cheers,
Stewart