Notes from Cauldron 2025 GCC GSoC BoF

[Martin Jambor <mjambor@suse.cz>]

Hello,

I have uploaded my slides from GSoC Bof at the Cauldron to the
conference pretalx:
https://conf.gnu-tools-cauldron.org/media/opo25/submissions/FHSXQA/resources/mjambor-gsoc_bof-cau_PcBrscg.pdf

The discussion has mostly followed the topics outlined there.  We have
agreed to the following changes in the program for the next year
(assuming we get selected) in the following three areas:

1. Better selection of candidates
  - When they write proposals for projects, we will ask mentors to
    include  pointers into what applicants should look into to learn
    about the topic...
  - ...and to think of a way to verify they can do basic stuff (testsuite
    results, get info obtainable from gdb or something similar)

2. Engagement with the community
  - We should require subscribing to the gcc@gcc.gnu.org mailing list
  - We should require public status reports
    - I will prepare a template (what they have studied, what they have
      done, what they have struggled with).
    - I will reach out to gcc-rust people to make sure this requirement
      make sense to their projects too.
  - We did not discuss having an introductory call but I think it is a
    good idea and will try to hold one.

3. Evaluations
  - There should be reasonable patches at least posted for review before final
    evaluation.  Only truly research projects, or projects which morphed
    into research ones for good reason should be an exception.
  - We should strongly advise to have some kind of RFC or minor patches posted
    even before the mid-term evaluation

Please feel free to continue the discussion here, it is understandable a
lot of people could not attend.

(Also, I hope I did not forget anything but please do remind me if I have :-)

I would like to thank a lot everyone participating in the BoF but
particularly my co-org-admins: Thomas Schwinge who has been helping me a
lot and has stepped in many times when I was not available over the last
two years and to David Edelsohn who has helped with quite a few
difficult selection decisions.

Martin