svn commit: r214817 - head/sys/teken

Alexander Best arundel at freebsd.org
Fri Nov 5 09:38:44 UTC 2010


On Fri Nov  5 10, Mark Linimon wrote:
> On Fri, Nov 05, 2010 at 01:21:03AM +0000, Alexander Best wrote:
> > maybe, because nearly all PRs (most of them including patches) they
> > have submitted via GNATS in the past remain unnoticed and thus they've
> > gotten tired of reporting issues and submitting patches if nobody
> > seems to care? ;)
> 
> We actually do better these days w/rt userland bugs (and some src
> bugs) than we were a few years ago.  I'm always looking for suggestions
> on how we can get more committers interested in PRs (especially those
> with patches).

unfortunately this might come too late. in 2006 and 2007 apple e.g. submitted
quite some patches and tried contributing some work back to the freebsd
project. since most of their PR didn't trigger any attention at all, they
seem to have completely stopped sending in patches.
even if a higher response rate in connection with GNATS can be achieved, these
parties (GNU/kFreeBSD, apple, ...) and their "manpower" are probably lost
forever for the freebsd project. since their patches didn't trigger any
interest, it is very likely that they have no intentions anymore to contribute
any work back: reasonably.

expecially in the apple case that's a huge loss. personally i think any
contributions by such companies like apple need to have high priority. maybe
having a single committer who will interact with apple or other major
companies might even be a good idea. this doesn't mean that user contributions
aren't as valuable as contributions by companies, but the fact that certain
companies have quite a massive "manpower", the impact of ignoring their patches
is far more severe.

if a company wished to have their work pushed back into the freebsd tree this
seems to be quite a complex process. maybe 2 or 3 years ago nvidia requested
some VM changes for their closed source driver to work on freebsd amd64. they
decided to skip the PR databse and directly interacted with developers (i think
mostly alc@). so that kinda prooves that the PR databse is not suited very well
to get patches committed to freebsd; instead contacting with developers
directly seems to be far more successful.

but again (as mark pointed out many times) freebsd doesn't have a culture of
bug busting.

cheers.
alex

> 
> mcl

-- 
a13x


More information about the svn-src-head mailing list