PQM needs to take over the world

[Matthew D. Fuller]

On Wed, Jul 18, 2007 at 05:52:04PM -0700 I heard the voice of
Wichmann, Mats D, and lo! it spake thus:
> 
> Maybe it's just the way we use it (incorrectly?), but for someone
> used to a centralized model, the disconnected-mode behavior of PQM
> will rapidly drive you to feel like PQM is your personal enemy #1.

That's the way I've always felt about it.  Which is to say, PQM is an
interesting and apparently pretty solid solution to a particular
problem, but it's not _my_ problem.  PQM is essentially an
asynchronous process, which lets you do a lot of things (like run a
half hour test suite) that you can't reasonably do in a synchronous
process.  It's very nice to be able to decouple that sort of thing,
fire it off, and move on.

But I don't _want_ an asynchronous process in the step to "commit to
mainline".  Committing to a common "mainline" should be the same
process and the same time consumption as commiting to my branch.
Using PQM in that situation (as opposed to one of the ones mentioned
above) doesn't feel like "using a handy tool", it feels like "using a
hacked up process to work around the inadequate tool".  No matter how
powerful or flexible async may be, it's an alternative to sync; not a
replacement.


Now, mind you, in my environments, handing out ssh access is usually a
perfectly reasonable step, so I don't so much suffer right now.
Still, a way to build in auth, and be able to do more fine-grained
permissions (like "xyz can read, but not write", and "abc can only
write to this branch", etc) would be an awful nice tool to have in the
box.



-- 
Matthew Fuller     (MF4839)   |  fullermd at over-yonder.net
Systems/Network Administrator |  http://www.over-yonder.net/~fullermd/
           On the Internet, nobody can hear you scream.