Choose timezone
Your profile timezone:
ROOT Team Meeting
ROOT Team Meeting
Zoom Meeting ID
63815034196
Host
Danilo Piparo
Alternative host
Philippe Canal
Useful links
Join via phone
Zoom URL
Shift handover: Jack -> Olivier
Proposals
- Stephan: XRootD workshop in Lyon, invitation for a team member to join, Jacob?
- Jacob: I am interested, I will take a look
- Jonas H.: generally PRs are opened, reviewed and eventually approved by other members of the team, github forbids you self approving PRs you are an author of, however if somebody open a PR, you can push to it to make changes, you can then approve it, effectively circumventing the self approval policy, i want to confirm we agree on what to do in this situation
- Jonas R. external contributors just do what you tell them, i don't see a difference between asking them to change and approve, and doing it yourself then approving, what would you do in that case where the guy opening the PR is an external contributor? would you loop in someone elese from the team?
- Jonas H. I think so yes, if we don't trust the external contributor to be the second pair of eyes in this code review
- Danilo: there is no technical enforcement through github, so we can just agree on it
- Jonas H. it is not written down
- Danilo: then we should add it to the internal docs
- Jonas R. i would like it if we can put in the guidelines for the external contributors too: external contributions that are not trivial, reviewed by one root team members asking changes, require 2 different pairs of eyes
- Danilo: agreement: we always try to have at least an additional pair of eyes for the approval of non trivial external contributions requiring changes
Liaisons
- Everyone off
- LHCb invited us to give a talk on ROOT and RNTuple
Round table
- Aaron: working on the migration
- Danilo: question for Jonas R., when do you forsee the ROOT full migration?
- Jonas R. cppjit releases in around 2 weeks, then it depends on how "clean" we want to get, rebasing is mechanical, but getting rid of the patches would take some time
- Aaron: some patches aren't trivial
- Danilo: we should aim towards the end of september to match with the release timeline
- Jonas R.: That's the timeline I would like to have, mid to end of september, if by then we're not in the state we'd like to be in then we'll just accept the divergence
- Stephan: do we know the list of env variables ROOT reacts to?
- ...
- Stephan volunteers to compile this list
- Jonas R.: we lost one of our benchmarking nodes, upgraded ntpl-perf01 and now network adapter is not showing up after reboot
- Danilo: maybe we reach out to someone we know to turn it on and off
- Jonas H. I tried switching it on and off again ten times and it didn't work
- my takeaway message: we should not take the open source tools for granted too much
- Koen: developped a script for
- Jonas R. does it bother us that during the cmake configuration step we spawn one python interpreter per module to import, lasting around 10s (30% of our cmake configure time) , instead of doing it the way Koen does with one interpreter that checks all?
- Koen: I could upstream it to ROOT
- Jonas H. do we really care about those 10 seconds?
- Stephan: it won't cost us anything
- Jonas H. except more code that could break
- Stephan: let's see it then decide
- Philippe: reviewing PRs touching core, if they are not urgent, I am being very conservative and won't approve untill ALL tests pass including the ones that look seemingly unrelated
There are minutes attached to this event.
Show them.