On Fri, 2019-10-18 at 17:11 +1100, Daniel Axtens wrote: > Stephen Finucane <step...@that.guru> writes: > > > On Fri, 2019-10-18 at 13:41 +1100, Daniel Axtens wrote: > > > It's been deprecated since 2.0 with the new REST API. That API is > > > now pretty solid, and git-pw is good. Drop the old API. > > > > > > Provide a page letting people know that the API is gone if they > > > access any of the old pages. > > > > > > This breaks pwclient, which only supports the old API. So we delete > > > a few things that referred to it or used it, including some old tools. > > > > I'm still not convinced that dropping this is something we _have_ to do > > right now. I know we said we'd remove it in 3.0 but it's not hurting > > anyone, as both the code and the docs are pretty well isolated, and > > doesn't seem to impact our plans to drop Python 2.7 (which will start > > hurting us as it ties us to Django 1.11). It should go at some point, > > but, only once we're sure there's nothing significant talking to it. > > With that said, if you really, really want to do this, I won't block > > but we *have* to get pwclient migrated first. We simply can't drop this > > until that happens because we're breaking workflows without replacement > > (see comments below on removal of scripts from 'tools' for a taster). > > I've said it below, but I'm confident I'll be able to get this done > > well before we release 3.0 so this isn't a serious concern. > > > > Lots of comments, mostly repeating what I've said here, below. > > > > Stephen
Per the above, I haven't applied this to master when applying the rest of these PW3 patches. I have, however, pushed the rebased version to master. If/when pwclient is migrated, we can look at merging this (though even then, I'd really rather _not_ break things for people as long as we don't have to - kernel maintainers can be a grumpy bunch :)). Stephen _______________________________________________ Patchwork mailing list Patchwork@lists.ozlabs.org https://lists.ozlabs.org/listinfo/patchwork