Hello Paul, Sorry for my late reply, I took me a while until I bumped into your email today. I'll try to keep my reply short, then we can expand given your interest.
2018-02-26 14:28 GMT+01:00 Paul Gevers <elb...@debian.org>: > On 22-02-18 10:12, Paul Gevers wrote: >> Technically my half year sabbatical leave has started last week. One of >> the projects I put my mind to is to work on materializing bikesheds. Yay! The problem is not only technical, but it carries with it some initial FUD or design proposal. Some of the questions to answer with such FUD are: - Who has access to bikesheds? - What do we want bikesheds for? - Given the above replies, do we want as a project support that idea? >> If I haven't missed progress, we are currently basically in the state >> where Joerg brought us in 2015. I was wondering what work needs to be >> done (and which first) to progress. >> >> I found the dak bikeshed branch of Joerg¹. I guess what is needed now is >> to rebase that branch on the current state of DAK and then start >> experimenting with wanna-build to see how this could/should interface >> with DAK to actually do the intended builds? Is that correct? (Does it >> make sense for me to try and setup my own DAK and wanna-build to play >> with this?) >> >> Is there any documentation at all (or even code) about ideas on the >> wanna-build side apart from the stuff that was discussed in the BOF at >> Debconf15 in Heidelberg? I vaguely remember discussing some design with few involved people at DC15 to overhaul the whole Debian Build infrastructure. However nothing happened and we are still maintaining last century Perl spaguetti we inherited from those days, but it had nothing to do with bikesheds. If you are interested, I could share those ideas with you but I leave that for another reply or thread. >> The DSA was also mentioned as stakeholder in this, should they be added >> to this thread and if so, what is their role in this? No need to get DSA involved at this stage, only after design proposal has been done, you can share with DSA for their input on how to best bring up an official Debian service and providing associated hardware resources. >> I hope you can help me to get started, such that I can help you to get >> this long standing idea into a working state. Which benefits bikesheds bring to you as individual? >> Paul >> >> ¹ https://ftp-master.debian.org/users/joerg/dak.git/ > > @wanna-build team: Neil suggested that instead of hacking on > wanna-build, you may be interested to switch to e.g. OBS²? Not sure how > I can help with such a task, but I am willing to dive into it. I am among official OBS package maintainers and obs.debian.net DNS owner, so I would not mind to push forward such idea with your support, that'd be probably awesome. However I do not yet see official Debian adopting OBS as their main wanna-build service, which does not mean it cannot be done, but it is a lot of work and it also impacts on several teams workflows. > Paul > > ² http://openbuildservice.org/ > In summary, I see three different projects you could focus on: 0. Implementing bikesheds in current infrastructure 1. Design bikesheds and discuss it's requirements 2. Overhaul current infrastructure (leaving bikesheds out of scope) 3. Setup and maintain OBS as debian.net service 4. Study and propose a design on how to replace wanna-build with OBS 5. Find something else and enjoy sabbatical If any of the above tickle you and would like me to expand on them, giving more details, please do not hesitate to let me know Regards, -- Héctor Orón -.. . -... .. .- -. -.. . ...- . .-.. --- .--. . .-. <free spam> -- Would you like to make a donation towards the upcoming Debian conference? Brochure: https://media.debconf.org/dc18/fundraising/debconf18_sponsorship_brochure_en.pdf ** https://debconf18.debconf.org/sponsors/become-a-sponsor/ ** </free spam>