httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Behlendorf <>
Subject Re: cvs commit: apache/src mod_rewrite.h Makefile.tmpl (fwd)
Date Mon, 18 Nov 1996 00:30:49 GMT
On Sun, 17 Nov 1996, Jim Jagielski wrote:
> Rob Hartill wrote:
> > 
> > I hope that after 1.2 we get some peer-review back into the development
> > process.
> I've said that before... We really need to get back to this thing being
> a group, with group decisions and group vote, and just a bunch of
> independent coders. Maybe moving to the "release leader" only
> commiting things might make that better. I don't really think that
> is a good idea, since it puts even more pressure on the rel-leader,
> but those of us with commit priv's really should:
>    o use it sparingly
>    o and after group consensus

I understand Rob and Jim's concerns about this.  I also definitely appreciate
Rob's concerns about commits so close to release.  But I also do believe that
those with commit privileges should be able to use it to fix typos and smaller
bugs where they have a good deal of confidence in it being the right fix.  Peer
review, in my opinion, does not end at the commit stage - I specifically watch
every patch which crosses though apache-cvs.  

I'd prefer to pose the question, where is the acceptible line which must be
crossed before requiring three +1 votes and no vetos before committing a patch? 
Was my mod_rewrite across that line?  Was rob's adding of comments to warn
about USR1 across that line?  I'll agree that line gets more conservative the
closer to release we get.  Things like the "satisfy"  functionality definitely
cross it, at this stage. 

Finally, group consensus can only work if folks start speaking up about
patches.  Many times, patches get committed lacking three +1 votes not because
they haven't been presented to new-httpd or discussed, but because they've only
gotten one or two +1 votes over the course of a couple weeks, and no vetos, yet
they are necessary to fix a bug or implement a feature we've all expressed a
desire for.  Should that situation be interpreted as a failure or a success for
the patch?  The only thing that can rectify that situation is for more
diligence on the part of the members of this list.  Do we need more blood?
More folks interested in actively testing scripts and such?  I know a couple of
those who were very active at one point have pulled back a bit (myself

Okay, so, let's move on.



View raw message