Re: [Fab-user] Fabric 0.9 alpha 1

[Jeff Forcier]

On Fri, May 1, 2009 at 9:57 AM, s s <address@hidden> wrote:

> Something pythonic based on django would be better just due to the nature of
> the project, in my opinion but whatever...

Trac's git support is quite a bit worse than Redmine's, or I would
have used that. I was actually involved in a Django-based Trac/Redmine
type app (pair of apps), but it's a lot of work and I decided it
wasn't worth it to keep going.

There is Roundup which I didn't like too much, but it wasn't that
horrid either (outside of not actually integrating with Git, that is).

> They can't submit tickets at all now so any improvement is infinite.

They sure can, it's called email! I realize that "email and then wait
for the dev to put it into a TODO list" isn't as tactile as "submit
Web form and have it show up on the Web right away" but the difference
isn't so large IMO. However, the ">1 person working on fixing the same
thing because we can't see who's assigned to what task" bit is
definitely problematic and probably shoots my earlier thinking (the
previous sentence) out of the water.

I think my problem was that I was looking at Fabric's earlier
development; while lots of folks had forked and watched it, there
wasn't a whole lot of actual patching or pull requests going on that I
saw, thus not as much need for divvying up work. It looks like a lot
more people are interested now that I've stirred things up a bit.

> Unfuddle?  A local trac installation (I have plenty of hosting space, I run
> a hosting company...)?

As mentioned, Trac's Git support is overall weaker than Redmine's, and
as I'm a fan of "best tool for the job" I'd rather install a Redmine
with the branch problem fixed than a Trac with a still-crappy Git
plugin :(

Didn't like Unfuddle when I looked at it, and again, I prefer to
self-host whenever possible.

> Anything would be better than waiting for perfect.

I know, I know :( It's probably naive of me to expect the "just email"
approach to really hold any water, but I'd rather *not* have to move
to not-perfect plan A and then have to move a 2nd time to more-perfect
plan B. However, I probably will, because, yes, perfect is the enemy
of good. I just wish there were even some (IMO) good options instead
of mediocre ones.

Right now I'm focusing on getting the documentation and code in better
shape so I can get to the beta, but making some sort of decision on
the ticketing problem is basically the next thing on my list. So
despite all my hand-wringing here, expect that I will take some kind
of action as soon as I can.

Best,
Jeff