W dniu czw, 11.01.2018 o godzinie 22∶30 +0200, użytkownik Mart Raudsepp napisał: > On Thu, 2018-01-11 at 21:27 +0100, Michał Górny wrote: > > W dniu czw, 11.01.2018 o godzinie 22∶17 +0200, użytkownik Mart > > Raudsepp > > napisał: > > > On Thu, 2018-01-11 at 20:45 +0100, Michał Górny wrote: > > > > # ARM64 Profiles > > > > -arm64 default/linux/arm64/13.0 > > > > dev > > > > +arm64 default/linux/arm64/13.0 > > > > exp > > > > arm64 default/linux/arm64/13.0/desktop > > > > exp > > > > -arm64 default/linux/arm64/13.0/desktop/systemd > > > > dev > > > > +arm64 default/linux/arm64/13.0/desktop/systemd > > > > exp > > > > arm64 default/linux/arm64/13.0/developer > > > > exp > > > > arm64 default/linux/arm64/13.0/systemd > > > > exp > > > > -arm64 default/linux/arm64/17.0 > > > > dev > > > > +arm64 default/linux/arm64/17.0 > > > > exp > > > > arm64 default/linux/arm64/17.0/desktop > > > > exp > > > > -arm64 default/linux/arm64/17.0/desktop/systemd > > > > dev > > > > +arm64 default/linux/arm64/17.0/desktop/systemd > > > > exp > > > > arm64 default/linux/arm64/17.0/developer > > > > exp > > > > arm64 default/linux/arm64/17.0/systemd > > > > exp > > > > > > With this I'll need to run through all of these profiles with > > > repoman -e=y and wait a long time, instead of just the two (well, > > > with > > > 17.0 now 4) profiles that I actually care about and checks enough. > > > I > > > also will see a TON of issues from the musl profiles down below > > > that > > > main block (it doesn't inherit base or something). > > > > > > This would make arm64 work completely impossible, so NAK from me. > > > > repoman has --include-arches option for a reason. Profile statuses > > are > > not your private convenience. > > That doesn't help whatsoever due to musl. Also not for running things > on a slower arch (mips in this case). > > > > Additionally if depgraph wouldn't be broken anymore, we would be > > > moving > > > them to stable, not some separate "dev" step. > > > > > > Same goes for the mips main block changes. > > > > But it is broken, and it won't become less broken if we keep ignoring > > the fact and not reporting new breakages just because one developer > > can't fix his workflow. > > Your patch simply removes dev completely, with no reason for it to > exist anymore then. If depgraph isn't broken, it'd be stable. There'd > be no reason for dev. Dev is dev BECAUSE it has a broken depgraph, but > is aspiring towards not. > My workflow is just fine. >
The point of dev is to bring staging warnings so that we can improve the depgraph. If you really insist, we can start with dev status for arm64 & mips. But it'd be more readable if we worked on only one arch simultaneously. -- Best regards, Michał Górny
