[pacman-dev] [11] library symbols naming

Aurelien Foret aurelien at archlinux.org
Fri Feb 17 03:10:28 EST 2006

VMiklos wrote:
>> http://frugalware.org/~vmiklos/patches/libpacman-proposed/_alpm.diff
> is there any reason why this was not commented?
> 1) as reviewing the patches takes sometimes weeks, we had to create our
> own tree
> 2) so every patch has to be backported manually
> 3) this patch is rather trivial
> 4) it implements Aurel's idea (not a Frugalware one)
> 5) you know that the patch was affected almost everything, and yes, it
> no longer applies

> sorry, but this is a bit annoying. i really don't want to port it again
> to your cvs

6) if you had read my TODO completely, you would have notice that I 
suggested this change to be the last one to be done before a release.
But you chose to do it and send it anyway...

Should I feel sorry because you're working on a modified version of 
pacman sources, and that you have to backport your changes to the 
official sources in order send us patches, or that you have to merge our 
cvs code with yours?
To be honest, I don't think so.
You're maintaining a parallel version of pacman, and I'm not responsible 
for this choice.

By the way, the fact that a patch does not apply cleanly is absolutely 
not an issue.
For your information, I already applied it to my local code sources, 
although I didn't committed it yet. I'm still reviewing it (especially 
the need to rename symbols for db_XXX functions).

> possible sollutions i see:
> - speeding up reviewing the patches a bit (apply/reject them in a week)
> - giving cvs write access

I don't see the need for it. If it takes more than a week to review and 
commit, what's the deal?
I can't even call the current pacman cvs code a beta version, it is 
still code in development.
So, there's no reason to put us under pressure.

More information about the pacman-dev mailing list