ROOT Team Meeting

Europe/Zurich
32/1-A24 (CERN)

32/1-A24

CERN

40
Show room on map
Danilo Piparo (CERN)
Description

Zoom link in announcement email; please contact rootdev@cern.ch if you did not receive it!

Zoom Meeting ID
97374667082
Host
Axel Naumann
Alternative hosts
Bertrand Bellenot, Lorenzo Moneta, Enrico Guiraud, Jakob Blomer, Danilo Piparo, Vincenzo Eduardo Padulano
Useful links
Join via phone
Zoom URL

# ROOT Meeting 2023-12-18

https://indico.cern.ch/event/1359002/

Notetaker: Jonas Rembser

## Project lead handover

Axel to Danilo

## Shift handover

Olivier to Marta

Best-effort for shift work to all of us during the break.

January: revising the shifter pool and responsabilities

## Meeting summaries

### IO meeting

  * No guests, quick roundtable
  * Set the stage for RNtuple limit testing
  * Next meeting on first working Friday in January
  * First PR for RNTupleMerger was merged, work continuing

### PPP

  * Marco Meyer from Japan was talking about ROOT in gravitational wave
    analysis with Apacke Kafka.
  * They use time series analysis.
  * Wants to potentially use RDataFrame in the future.
  * Now in touch with him via emails to help with adaptation of RDataFrame
  * RDataSource for gravitational wave analysis format

### RooFit

  * Small RooFit meeting last week with Jonas Lorenzo and Zeff only
  * Discussed some ATLAS related issues and forum posts, since not many people
    were connected
  * Talked about avoiding interpreter memory limit by avoiding the use of the
    RooWorkspace factory language in HistFactory
      * acted on right after
  * Next meeting in January


## Topics

### CMS Issue with very large number of libraries being loaded with ROOT 6.28 (LLVM 13)

https://github.com/cms-sw/cmssw/issues/43077

Problem is that a particular symbol is not emitted, and then the linker tries
to find this in all the possible libraries in the interpreter.

Needs to be fixed by January 23 for CMS to adopt 6.30.

A hotfix was prepared by emitting that missing symbol (static TString member).

However, the underlying issue is still a ticking timebomb in the interpreter.

Vassil offered to work on this in the second week of January.

Danilo: is it possible to do this earlier?

Jonas H.: What Vincenzo has as a fix is more a compliance fix for C++.
          The problem is maybe not the interpreter, but could be ROOT Meta or PyROOT.
          The "fix" is correct, but we need to understand what the actual problem is.

Vassil: This fix is good enough, the interpreter has sailed quite well without
        this problem fixed.

Jonas H.: But if the behavior changed with the LLVM upgrade, this could also
          have other consequences.

Danilo: Shazad is testing the hotfix now, let's see what the result is and then
        decide tomorrow.

### Issues in JIRA and GH

  * Hundreds of open JIRA tickets and GitHub issues and PRs.
  * Please take a look at your areas.
  * We will come back to this in January.

### NGT involvement

Possible synergy topics:

  * reflection
  * data layouts

### POW presentation

  * presentation on the 15th of January
  * Danilo will send around draft at January 10

## Going throught the issues

What about the flatpack issue?

  * Not to commit to anything, it should just be a user effort

Discussion about https://github.com/root-project/root/issues/14230

  * Assigned to Danilo, to be followed up

## ACAT abstracts

Danilo: Can we have more than one abstract? Last year it was eight.

Jonas. H: We could do an abstact on Cling, explaining what patches we still
          have to do, and how we will benefit from Clang REPL.

RooFit: New R & D already exhausted last year at CHEP and PyHEP.

RDataFrame:
  * The bulk processing will probably not be ready by then.
  * Possible RDataFrame presentation to show RDF with RNTuple as a backend.

Deadline is the 5th of January.

## Round Table

### Jonas

  * Addressed several forum posts, work on RooFit bugs that were discovered by
    forum users (e.g. asymmetry plots with RooSimultaneous don't work)
  * Several technical improvements in RooFit, and deprecation of a few older
    things, most notably the old test statistics classes
  * Continuing the support of RooFit AD classes, also together with Abhi, the
    IRIS-HEP student
  * Avoiding the RooWorkspace factory language in HistFactory to avoid JIT
    memory limit

### A.O.B.

Jonas: when will we have a ROOT 6.30.04 patch release?

Danilo: not this year.

There are minutes attached to this event. Show them.
    • 16:00 16:01
      Find notetaker 1m
    • 16:01 16:05
      News 4m
    • 16:05 16:10
      Shift handover 5m
    • 16:10 16:20
      Meeting Summaries and Plans 10m
      • I/O
      • TMVA
      • RooFit
      • PPP
      • Planning / Godparents /...
      • LIM
    • 16:20 17:00
      Topics 40m
      • CMS Issue with very large number of libraries being loaded with ROOT 6.28 (LLVM 13) https://github.com/cms-sw/cmssw/issues/43077
      • Issues in JIRA and GH
      • NGT involvement
      • POW presentation
    • 17:00 17:25
      Round Table 25m
    • 17:25 17:30
      A.O.B. 5m
      • Release lunch