httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@covalent.net>
Subject Re: Apache 2_0_31 is now rolled
Date Sat, 02 Feb 2002 18:42:59 GMT
RM postscript: the tarball is also missing docs/manual/faq/support.html
for whatever reason.


From: "Brad Nicholes" <BNICHOLES@novell.com>
Sent: Friday, February 01, 2002 5:54 PM


>   So what is the verdict on the messed up #ifdef in scoreboard.c if .31
> goes beta?  Are we going to include the fixed version or patch it in the
> release notes?

Changing this file when we reroll anyway to clean up the tarball seems like
no issue to me;

  1. The #ifdef v.s. #if resolves identically on OS2/Unix/Win32.  Nearly all 
     other platforms supported it (get the code from an #if or #ifdef.)  

  2. It fixes Netware.  If they download an older -alpha, they get nothing.
     If they download -beta it works.  The code didn't change, the conditions
     to exclude Netware-broken code changed.

Isn't this [part] of the point to -alpha, -beta, -gold succession?  Trivial
mistakes might be corrected, while code flaws (as rbb pointed out) can command
a new dot release.  Yea - it's a slippery slope, but this bug was pretty firmly
at the top of the hill.

If we want to wait for -beta, I'd suggest tagging the original -alpha files as
APACHE_2_0_31-alpha, then move this patched file's tag APACHE_2_0_31.  If we
did something similar to -beta, do the same, add a -beta tag, then push the 
APACHE_2_0_31 tag for the file or two that misconfigured _this_code_ on a
given platform.

Fixing ./configure built with the wrong libtool, correcting an #if FEATURE
for some specific platform(s), or correcting missing/too many files seems like
trivial changes.  They don't change the CODE BASE that we released.

Bill





Mime
View raw message