httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@lnd.com>
Subject RE: 1.3.13 - Complete some Win32 PRs???
Date Thu, 13 Apr 2000 16:19:32 GMT
> From: rbb@covalent.net [mailto:rbb@covalent.net]
> Sent: Thursday, April 13, 2000 7:33 AM
> 
> > First off, the list has grown fairly long in changes 1.3.13/dev
> > Many of these changes are Win32, so I don't know that there is
> > sufficient momentum to begin targeting a date for 1.3.13.
> 
> Personally, I would like to move away from fixing 1.3, but 
> that is a bit
> of selfishness on my part.

Haleluja!  But, since we don't have a 2.0 product, we have to
be real.  Once 2.0-beta is out the door, I'm in your boat :-)
 
> Bill, I hope that you are not considering stopping posting patches.

Enough of that.  -I'm- going to keep posting patches.  I'm talking
about Win32 fixes for 1.3.x and reworking -other- people's patches
(which may not have been particularly clean) and need massaging into
the 1.3.13/dev tree.

> I am very hopeful that with a release of 2.0 we can disple the 
> myth that Apache is not suitable for a Windows machine.  I think 
> 1.3 on Windows may be something of a lost cause.  It is hard to 
> maintain, and there are serious limitations just because we are 
> forced to use POSIX calls.  

Actually, it works quite nicely as a testbed machine.  Before I
kill the AT&T commercial servers, I regularly try things out here
in NT.  95?  Well many seem to enjoy Apache given the amount of
discussion on www-servers.windows.

> Having said that, here are the answers as I see them, others 
> may see things differently.
> 
> > Q.  [all] Are we moving to set a date to 1.3.13?
> 
> I hope not.  If we release 1.3.13, I want it to be limited in 
> scope.  I know the NetWare guys have some changes that they 
> want to get into the 1.3 tree, I would like to see 1.3.13 wait
> until they are ready so that we don't have to have a 1.3.14.  

That's what I wanted to know, and I think 1.3.13 is a great number
to end that cycle on :-)

Well NetWare gurus, what timetable might we be looking at?

> > Q.  [all] If not, is a 1.3.13/int-Win32 release appropriate?
> 
> I don't think so.  Only because the NetWare guys have patches too.

I guess that's my question.  1.3.13/int-NetWare etc... is this
something to consider when we have 'completed' a major set of 
platform-specific fixes.  If the answer is no, then it's no, 
but I was just curious.

> > Q.  If I send a set of -clean-, -tested- diff -u 's of the 
> >     incomplete Win32 work, that is;  mod_proxy cleanup, taskbar
> >     icon upon minimize, JJ Keijser's Win95/98 run-as-service,
> >     and the most compact mod_info changes possible, 
> 
> I don't know how I feel about the taskbar icon, there are 
> people who don't think it belongs there.  I have no strong 
> feelings.  If these are all posted individually, I think you 
> will find people willing to look at them and commit as 
> appropriate.  (Especially if they are ported to 2.0 at the
> same time.)

That's what I REFUSE TO DO.  They were posted here by diligent
coders.  No feedback.  Enough of that nonsense.

Let's hash out the feature set, I'll propose the fix to those
and those alone.  If I have a committer, I will list the patch
sets to the tree, and we can argue it's merits.  Then (and only
then) will I massage them into the 1.3.13/dev tree.

I'm asking for a commit on a feature set before I waste the time
to bring the old patches up to date and 'apachify' them as needed.

If there are x-plat diffs, I will post those well ahead to give
everyone a chance to assure those didn't break any other platform.

I will bundle each issue so debugging is easier.  I'm moving to
fixing the problem with a single patch, but only a single problem
per patch.  I'm still waiting on you to review the md5 patch
(which was, unfortunately, several issues since the src/support 
code still referred to the ap tree, and not apr move.  You
did create the issue when you apr'ed md5.  You've also left a
dangling declaration of ap_slack() around (plus SLACK_HIGH etc.)

> >     and most importantly, a patch to hold the console window 
> >     open (until it is dismissed) if the server exits with an
> >     error (so the GUI fools can see what they broke),
> 
> I am definately doubleplus 1 for this.

So you are reading www-server.windows too, on occasion?  :-P
 
> >     Will an individual commit to committing these to 1.3.13,
> >     and the alpha-4?  You've seen the quality of my patches,
> >     and these changes only touch the Win32 code path.
> 
> I think you'll have a hard time getting somebody to commit to 
> committing these until we have seen the code.  I will commit 
> to looking at the patches though.  :-)

That's the friggin point.  You (personally) have seen them.  Esp.
the JJ Keijser Win95/service patch.  (You personally posted that 
very day, 10.6.99).  So has anyone that was watching the feed.
No feedback to the coder that month, at least to the list.

> > Many patches have taken months to hit the tree, and other
> > appropriate patches have been ignored -without comment-.  
> > This is really not acceptable.  Respond with -something-.  
> > Everyone gets a chuckle out of the kudo messages to the 
> > group - I don't believe we take the time to do so for any
> > or most of these contributors.  That is fine if you commit
> > their work and credit them.  But if you don't ack them at 
> > all, you've lost a contributor.
> 
> Unfortunately, this is a recurring problem in the group.  We 
> all work on the sections that are important to us.  I haven't
> even looked at the 1.3 tree for almost a year, because I don't
> care about it anymore.  I see the future as 2.0.  We all TRY 
> to ack patches as they come in, but it isn't
> always easy to get to them all.

And amen to that.  We have our nitches.  I'm offering to be a
reviewer in this capacity, and do the testing to assure the old
patches don't break anything.  I will even post a pre-release
dev binary and solicit users of www-server.windows to break it,
if that might be appropriate.  I want to leave other coders 
with time to do what -they- want to work on.

> > I'm really not supprized we don't see more patches from these
> > individuals.  I thank Ryan and FirstBill, and several others,
> > for wasting the time to explain how my patches are good, bad,
> > appropriate or not.  I don't see such comments in the archived
> > newsgroups to many of these contributors.
> 
> I don't think of it as wasted time.  I hope you continue to submit
> patches, most of them are very good, if a bit long sometimes.  :-) 

Did we miss the point here?  We aren't talking about me.  I used
that example because there are plenty of would-be-contributors
who didn't get the benefit of such mentoring.  Not even a flame
to push them in the right direction.

Thanks for the feedback

Bill

Mime
View raw message