Mailing-List: contact new-httpd-help@apache.org; run by ezmlm
Precedence: bulk
Reply-To: new-httpd@apache.org
From: "William A. Rowe, Jr." <wrowe@lnd.com>
To: <new-httpd@apache.org>
Subject: RE: alpha 2 going out tomorrow.
Date: Fri, 31 Mar 2000 13:09:32 -0600
Message-ID: <000801bf9b44$d91b4e30$345985d0@corecomm.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Importance: Normal
In-Reply-To: <Pine.LNX.4.10.10003311345210.9678-100000@shell.ntrnet.net>

> I'll look at it after the alpha.  I am grabbing the most 
> recent tree and
> running with it after testing.  All the other patches you 
> have submitted
> in the last few days will get reviewed and committed right after the
> alpha.  I don't want to take any chances that we miss the 
> alpha because of
> a new bug.

Thanks for the pre-alpha quick looks, and TIA for committing them.

As long as declaration of ap_get_file_mtime shall change, I would
like to see alpha 2 released with those sorts of fundimentals.  
Specific bug fixes in apr don't worry me 1/10th as much.  If we 
are asking folks to beat on/experiment with/correct the alpha, 
let's provide the cleanest declarations.  If apr will not return 
time_t's - I would advise against advertising that we do.

That's why you haven't seen my Win95 GetFileAttributesEx workaround,
waiting to post it till post-alpha 2.  (Besides, I don't like the
workaround myself).

Please state that alpha 2 does _not_ work with Win95, that it is
written for NT threading and IO completion ports; but that we 
are interested in patches that work around the limitations of Win95.

As for a workaround to Win95 CreateIOCompletionPort, I'm really
starting to think, we don't need 4 ports (and 98 offers no better
reliability than 95) - do we want 2 mpm's (or a late binding mpm 
code path resolution in pre_config) for 95/98 and NT/2000?  I'm
thinking these paths are so horridly different, and the workaround 
run-time testing will bog the NT performance.  If we setup the 
callbacks at runtime to hook to the NT or 9x code, we can optimize 
the NT path without worring about the 9x path.

That's not a thought for alpha 2, but at least consider my first 
point anyway,

Bill