ROOT Team Meeting

Europe/Zurich
32/1-A24 (CERN)

32/1-A24

CERN

40
Show room on map
Axel Naumann (CERN)
Description
CERN number: 71400 Extension: 109284483#

Shift/Danilo: - PRs: we are approaching 100 PRs. Perhaps we can go through them and
close/merge the ones we really (don't) care about?

- Forum: all fine, except this one
https://root-forum.cern.ch/t/std-complex-and-clone-function/31743 which
has the features of being severe

Philippe is taking the shift this week.

I/O meeting: no meeting last week nor this week.

PPP meeting: roofit and rdataframe presentation.

No TMVA meeting, at least not officially.

Planning meeting: nope?

Infrastructure: Guilherme installed a few ROOT only node but not all quite ready.
gcc 8 warnings still left.

need to disable cuda on Ubuntu because not compatible with gcc 8.

Vassil: we don’t have enough nodes and missing coverage.

Axel: The new setup has found several issues that went by unnoticed, several related to the future binaries. Examples include a bug in python 3 that prevents us from having a working PyROOT on most current platforms, or runtime cxx modules not working with cfitsio. Other problems are clang-HEAD and tbb (missing symbols).

Despite trying a vast amount of permutations I cannot come up with a combination that covers what we need AND works. I have vetoed cfitsio for runtime cxx modules; I did not manage to get a python 3 build to work on an old enough python 3 that does not yet have the python bug (Ubuntu 14).

I see only one viable way forward:
- disable failing builds at the cost of coverage for everything that is not required by the release binaries (e.g. python 3 - we never published python 3 binaries)
- but keep those failures that would prevent us from building our binaries.

This will force us to focus on the failures immediately relevant for the release; we can take care of the remaining failures after the release.

Let me know if people come up with a better solution or agree with the plan.

Enric: The python3 issue should be now fixed with Python 3.2.4

Releases: mostly all in Danilo’s hand.  

Oksana: Last part of compression algorithm ready.  Left in all test the all uses to test the user code as-is.

ACAT:  Enric will upload one for PyROOT.   Olivier will write an abstract on New Graphics.  Enrico maybe something on RDataFrame.

Javier: fix a few things related to RDataFrame reading too much, now as efficient as file based splitting.  This is for Totem.  Next: mainly FCC things.

Enric: working with Javier/Totem.  when reading remote files, multi-thread RDataFrame was taking longer.  Will present Totem result to PPP meeting.   Also working on PyROOT, discover a bug in one of the macro on Python 3 related to garbage collection.  3.7.2 has the fix

Enrico: looking at conda build of ROOT, including some of its dependencies (xrootd, davix).  Working nicely.  He is building nightly master and 6.14.  Wanted to add a job running roottest against it and discover deficiency in roottest CMakeFiles.  Also help with RDataFrame over reads.  

Stefan W: working on roofit feature (see PPP meeting).  having RDataFrame to Python array would be very useful and is also completed.   For Wednesday for SHIFT, added/updated tutorials to use miniAOD with more statistics and is now closer to real analysis.

Stephan H: replacing linked list with a vector to improve the performance/memory space.   Because TIter is used a lot, I wrote a wrapper to adapt std::vector to useable there.

Bertrand: worked with Axel on infrastructure.  I added cfitsio to Windows.  I also PyROOT experimental is building but I am not sure what is supposed to work or not.   I also need to start working on being able to open the ‘new’ canvas on Windows.

Jakob: Working on new RawFile for v7.

Olivier: worked on forum/mail backlog.  Working on student PR.   Parallel coordinate fix.  I needed to revert partially to allow picking it.  I will carefully look at the consequence.

Vassil: working on release notes.  I am working on removing rdict and rootmap.  rdict needs performance measurement.  For rootmap, it seems to be triggering a bug in PyROOT, I need Philippe to take a look.

Yuka: recently on exporting c++ module to CMSSW, SetROOTSYS is completely broken …. dladdr is broken … I tried to replace it with readlink of the binary and it worked. but of course it only work with root.exe itself not cmssw.  I also try to just respect the environment variable but it does fail during the build.  See ROOT-3358

Oksana: abstract for ACAT. CHEP proceedings.  Benchmarking infrastructure tweaks.    Removal of hard coded number for Compression setting.   Done in two steps.  For now just adding constants, later will introduce a setting class (with possibly more flexibility than just algorithm and level).

Guilherme: Infrastructure and JIRA issues for releases.  In particular fail if C++14 requested but not available.  Would like to raise to require CMake 3.9 (current release is 3.13).

Philippe’s plate: cms schema evolution issue, roofit workspace issue when reading old files, valgrind reports
TTReader performance, end of process/cxxmodule/SetClassInfo

There are minutes attached to this event. Show them.
    • 16:00 16:05
      News 5m
    • 16:05 16:15
      Shift handover 10m
    • 16:15 16:25
      Meeting Summaries and Plans 10m
      • I/O
      • PPP
      • TMVA
      • Planning / Godparents /...
    • 16:25 16:40
      Topics 15m
      • Release status
      • CI infra changes
      • ACAT Abstract
    • 16:40 16:55
      Round table 15m
    • 16:55 17:00
      A.O.B. 5m