httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rodent of Unusual Size <Ken.C...@Golux.Com>
Subject Re: cvs commit: apachen/src/main http_main.c
Date Sat, 24 Jan 1998 19:18:05 GMT
Marc Slemko wrote:
> On Sat, 24 Jan 1998, Rodent of Unusual Size wrote:
> >
> > It *was* discussed, starting last Sunday.  You made no comments.  No
> > conclusions were drawn, except that leading underbars were out due to
> > reservation restrictions, and my objection to the overloading of the
> Exactly.  Discussing does not mean bring it up then if no decision is
> reached, pick your own thing.

If it's something that needs to be done, and the discussion isn't going
anywhere or has died completely, and two of the developers simultaneously
volunteer to put *some* sort of stake in the ground, then I think it
is entirely reasonable for someone to pick something so that at least
the issue is addressed.  It can be retrofitted later, but IMNSHO endless
discussion about how to address an open issue is part of what a) keeps
the beta cycle so long and b) commit-then-review is supposedly going to
help address.

> Whatever you call it, this smerits discussion before committing because it
> implacts things forever since you were not just trying to change one
> function, but to establish a standard.

Yes, it merits discussion, and no you didn't participate, and no I don't
think we should hold things up because people decide - later - that they
have an opinion.  And no, I wasn't (and am not) trying to establish "ap__"
as a standard, but to avoid the inappropriate overloading of one that's
sort-of evolving ("ap_").  I had hopes, which is why the commit log
is phrased the way it is, but it's not a big deal for me.

>              I am seeing an awful lot of clutter around lately with
> commits and uncommits and recommits, etc. that could have been easily
> avoided.

You must be looking at something other than what I'm seeing; the only
such case I can think of is the _APACHE_mumble wrapper issue.  That
consumed maybe ten minutes, cumulative total, of everyone's time.  If
that had gone through a pre-discussion, it would have been a lot more.
Everything else in the last week-plus has been moving forward.  This
is how commit-then-review is supposed to work, and why we're experimenting
with it: to see if it really does.  So far it looks like it *is* working -
at least to my eyes.  The confusion between Jim and Dean had nothing
to do with pre- or post-discussion; Dean simply wasn't aware that Jim
was doing a phased commit.

> If Rasmus needs it right now then he can take 2 seconds and change his
> tree right now.

While we discuss and debate and argue for days over a trivial change.
Ludicrous.  Right now there's something that works.  Rasmus doesn't
care what it gets called as long as it isn't new_connection().  We can
rename it to anything else at any time and the only person that will
be affected is Doug.  If we left it alone and debated forever, then
*two* people would be affected by the ultimate change, Doug *and* Rasmus.
And the rest of us indirectly through a tiresome point-discussion rather
than a strategic one.

>                  I'm not sure I see the problem...

Well, it was there notwithstanding.

#ekn	P-)}

View raw message