[Adium-devl] Version Control

[Colin Barrett]

On Feb 6, 2008, at 3:58 AM, Evan Schoenberg wrote:

> Peter pointed out one use case - folks who might have commit access
> but have primary association with another organization, a difference
> we choose to make clear: a pidgin developer (with access, probably
> enforced only by verbal agreement, to the Purple plugin), an MIT
> person whose involvement is limited to Zephyr, etc...

What about, for example, you? You're both a pidgin developer and an  
adium developer. What would your affiliation be?

> Another is for patches. I don't know about this system, but mtn allows
> the committee to specify a searate author for the change (the patch
> writer).  Making our affiliations clear keeps -that- clear.

Mercurial allows you to maintain the patch author's name in the  
"author" field if they send it in the appropriate form (as a bundle by  
using hg bundle, or by serving their own repository and having us pull  
from it) automatically, or you can use the --user flag on the hg  
commit command to include it yourself manually.

> Another is for distributed vcs. This is what I was talking about
> before. A repo (say, by Apple, where they maintain a branch of Adium
> which uses iChat's backend for AIM) may have a branch be pushed to our
> rep with our permission. It's useful to tell at a glance that -these-
> changes came from outside while -those- were made by an Adiym  
> developer.

If we really have so many people working on the project that we don't  
know who each of them are, that's a wonderful problem to have, but  
it's not something we are dealing with right now, let's be honest.

> On the other hand, I see no disadvantages. What do you see?

Primarily, it is annoying to set up. I must either 1) have colin at adiumx.com 
  be in my global .hgrc or 2) modify the .hgrc of the new repository  
every time I make anew clone to include "username = Colin Barrett <colin at adiumx.com 
 >", and if I forget, then there are commits made by "Colin Barrett <timber at lava.net 
 >" instead.

I see minimal gain for quite a bit of hassle, at least on my end.

-Colin