Piotr Król wrote:
I don't understand argument about running significant ("enough") of the infrastructure. Why maintainers of platforms do not run their part of infrastructure which support those platforms?
I think this is a key point. It's a lot easier to develop centralized solutions, so that's a common mistake. And even if not strictly or intentionally centralized then there's often at least a knowledge gap, case in point your request for more integration information.
If you want to maintain any particular release as a long term branch, announce your intent and we'll set up a branch!
I'm not sure what benefit it would give to community or to us, but we have to maintain v4.0.x probably until PC Engines hardware EOL.
I think it would be fantastic if that happened on coreboot.org!
Maybe it would not benefit you in any way immediately, but it would probably not be a disadvantage for you either, and it would for sure look great for the project as a whole.
And maybe, just maybe, some colleagues will pitch in because they see value in what is then in practice a stable branch.
What I really try is to highlight various problems 3mdeb see over 6 years of coreboot development.
That's super valuable and I'm much appreciate your input!
Toolchain stability and reproducibility is something I discussed. I even started to write something here: https://docs.dasharo.com/osf-trolling-list/build_process/
Yes. Docker is a dark pattern, yet like `curl | bash` it prevails.
Yeah, our coreboot-sdk has problem having only python3, SeaBIOS do not understand that: https://github.com/coreboot/seabios/blob/master/Makefile#L25
At least this was problem on 4.13 tag.
I appreciate that SeaBIOS likes to still support python2, but when all distributions choose to end support for python2 it'll take more effort in SeaBIOS. It could be simple though, maybe the attached patch is already enough? It looks like all python scripts are called through $(PYTHON) rather than executed directly, so the #! command doesn't actually matter so much.
//Peter