Re: Patches

[Alfred M. Szmidt]

   > Please don't post just references to mailing list archives and
   > expect me to take patches from there.  Those are munged and I'm
   > not going to deal with that.  Just repost them.

   How do you feel about patches in the savannah patch manager?  In
   the past, a fair amount of time has gone by before some patches
   have been looked at and applied.  So by keeping them in savannah,
   we have less of a chance of losing track of them.  On the other
   hand, savannah does not send patches to the list inline.

Don't know about Roland, but I know how I feel about the SV Patch
Manager.  I think it pure and simply sucks.  Yes, it keeps track of
patches, so they don't get lost; but that is it.

The patch manager doesn't support sending patches via email, replying
to notifications (they don't get stored as follow ups), poking at the
database via email etc.  And as Neal noted, they don't get in lined.


As for time between a patch has been looked at and applied.  That is
exactly the same as before and this isn't something a patch manager
will ever solve.

Only way to solve that I guess might be to have some "vague" policy
about what kind of patches may get committed without a OK from the
gods (Roland, Thomas, Marcus) and let people like say Marco commit
them.  Now if Marco--or who ever--doesn't have time to commit a patch,
then giving commit access to more people who will promise to follow
these rules (and have some kind of sense what is a OK patch and what
is a broken patch) might work.  Then for patches that don't follow
this vauge policy, if the gods OK it, then said people can commit the
patch.  can commit the patch.

Obviously this will only work for patches that are OK for committing,
so all broken patches must be fixed in either case by someone.  And
one shouldn't blindly follow this "policy", a patch might still be
broken for whatever reason so the people committing them must have
some sense of what is right and wrong in either case.



But what do I know, this is the third time I pitch in at trying to
solve the issue of getting patches committed in a "timely" manner...