In the room: Stephan, Giacomo, Florine, Jonas Rembser, Jonas Hahnfeld, Aaron, Vipul, Martin, Marco Clemencic, Silia, Stephan
Zoom: Olivier, Danilo, Serhan, Devajith, Lukas, Marta, Oliver Lantwin, David Lange, Juraj Smiesko, Liz Sexton-Kennedy, Philippe, Giulio Eulisse, Florian Uhlig, Phat Srimanobhas, Peter Van Gemmeren, John Chapman
Giulio: would be good to indicate when you plan to change LLVM version of ROOT on the plot, or when it was changed in the past
Serhan: How will ROOT7 work out? New histogramming will live alongside, or old one will stop getting supported?
Stephan: alongside, we’re even planning to have converter functions. ROOT7 should not be completely breaking your code, we want to provide a transition path with new classes that you can adopt at your own pace
Marco Clemencic: happy to mention interoperability. Wondering about RFile though. Can I opena. file written with TFile with RFile?
Stephan: Yes, the on-disk file format is not going to change. We change only the interface and object ownership model
Oliver: in addition to R classes there’s several different ways of doing things in ROOT. Might be useful to have migration guide/ suggestions to move old code to new code.
Stephan: good idea! We try whenever possible to update tutorials to show modern ways of doing things. Of course for what you suggest is to have some general document to provide a transition path.
Juraj: How frozen the new R classes things will be?
Stephan: classes or functions in Experimental namespace can be changed at any point. Once they come out of Experimental, then you can consider the API to be stable (backwards-compatible).
Juraj: Does this mean you will come up with a new letter?
Stephan: Will be R for all new modern classes
Vincenzo: Slide 22, Backported to 6.36, will be there in the next patch release
Oliver: Can this be backported to 6.30? Would allow us to compare the new things in ROOT 6.36 separately from this blocker issue.
Giulio: For the file reading, it was always clear it was a nice to have. For the RDataFrame bit, I thought this was already in the PoW. What would you need the collaboration of ALICE for?
John: For the RooFit docs, we’ll go back and ask which topics to prioritize. We have one ATLAS person coming to the workshop, it would be good to discuss it there.
Oliver: Where are these quarterly reports announced? I don’t get emails.
Stephan: You can contact Danilo, maybe you are not in the mailing list.
Florian: As you probably know MacOS is quite picky about the version of the SDK. One of the experiment of ours are using still ROOT 6.30 and we finally found a way to pass the right SDK to all cling features. Would ROOT be interested in getting a merge request?
Danilo: This would be quite interesting, let’s discuss this.
Florian: Compile ROOT 6.30 with clang 16 on MacOS 15 you get error messages. You have the possibility to pass a different SDK to CMake. This is taken all over the place correctly except rootcling. rootcling at that point always takes the latest version of the SDK. In principle there are some minor changes to pass the right SDK also to rootcling and enables to at least compile older ROOT versions with newer clang versions on MacOS.