Getting xserver patches reviewed

[Dodji Seketeli]

On Mon, 26 Nov 2007 07:48:24 -0500,
Bernardo Innocenti <bernie at codewiz.org> wrote :

[...]

> 
> Please let's not use Bugzilla as a patch queuing system.
> It's a recipe for 
> 
>  * Lack of public reviews
> 
>  * Small patches never being submitted due to laziness
>    of filing a bug in bugzilla
> 
>  * Patches bitrotting for months in bugzilla before anyone
>    notices

Well, when a patch submited to the mailing list touches a subsystem I am
interested in, I usually create a bug myself in bugzilla and attach the
patch there so that I don't forget, and I reply on the list saying that
I have attached the patch to a bug. Maybe I should also systematically
add the patch sender to the CC list of the bug ...


> The LKML can effortlessly handle *thousands* of patches per
> release.  Surely there is not yet a scalability problem for
> Xorg.

Well, the kernel people also thought at the beginning they could live
without bugzilla ... and experience has shown that it could be
usefull even for them.

Using Bugzilla has some advantages :

 + It allows easy reporting of the open bugs. One can and drill from
there to see if patches were submitted to these bugs etc...
Even if you don't want to attach the patches to bugzilla, there could
be a small note pointing the patch on the mailing list.

 + it allows patch comments threads not being buried in the middle of
other discussions ...

In any case, I agree that we cannot force people to submit patches to
bugzilla, and that is not a problem because nothing prevents
maintainers to reference the submitted patches from bugzilla bugs later.


-- 
Dodji Seketeli
http://www.seketeli.org/dodji