Public SourceMage Discussion List

[Jeremy Blosser <jblosser-smgl AT firinn.org>]

On Apr 11, Jaka Kranjc [lynx AT mages.ath.cx] wrote:
> On Wednesday 11 of April 2007 22:07:59 Jeremy Blosser wrote:
> > Thanks.  Please add that bugs requesting verification can be CLOSED after
> > waiting a week without response.  Please also clean up the query links;
> > right now we have 2 for the old way and 2 for the new way.  I think we
> > need:
> >
> > - list of requested (not yet approved) -rc integrations (FIXED + -rc
> >   integration requested)
> > - list of approved pending -rc integrations (-rc integration granted, not
> >   CLOSED)
> > - list of integrated -rc bugs not yet CLOSED (subject matches 
> > [integrated],
> >   not CLOSED)
> >
> > ...and the same 3 for stable.  The last one is one we don't have now but
> > allows tracking what we think is in the tarball as well as things that are
> > waiting to be verified.  It will also help transition from the old 
> > workflow
> > to the new.
> I added the two new ones to ease the transition, but they are incomplete 
> and 
> temporary.
> Why have separate queries for stable and stable-rc? Most of the time stable 
> integration requests are coupled with stable-rc ones, so the lists would be 
> needlessly duplicated.

Because when they aren't duplicated they get mixed together in the list,
and it makes it more difficult to mark them all CLOSED after I make the
tarball (the -rc ones that aren't getting integrated to stable shouldn't
get CLOSED until they eventually get to stable).  Or even not mark them
CLOSED, just to check they all got caught correctly.  Right now I have to
eyeball each in the list to see which is -rc only, then go back and only
mark/deal with those.

I think as time goes on that "most of the time they're for both" will
become less and less true; it's already going that way from what I can see,
closer to 50%.

> We can make the new queries without the integrated flag being there, so I 
> don't see a reason to complicate the matter with searching the titles. One 
> can use the exsisting queries and then easily spot the [integrated] bugs.

Well, the goal would be to get an integrated state in place to query
against, so this was just a placeholder for that.  Again, it's me speaking
as a person making tarballs -- which ones are and aren't integrated
obviously are a big deal there.

> as for the doc:
> "When a bug is fixed in test, " ... should also mention that we need the 
> commit id of the fix and preferably also the version of the grimoire it was 
> fixed in, eg. test 2.3 (this helps when cleaning old bugs that didn't get 
> integrated)
> 
> It also doesn't define when people should or shouldn't ask for integration 
> to 
> what at all, the discussed general guidelines would do.

Yeah.

Attachment: pgprA_dBcWkkk.pgp
Description: PGP signature