apache-0.2...

[Brian Behlendorf <br...@wired.com>]

compiles fine on SGI's and BSDI, and nothing appears broken (waiting for 
rst's test suite :)  Had to apply B27 to get .cgi scripts to run though, 
so I vote ++++1 to get that into 0.3.

I presume someone's keeping score and will be building 0.3 this weekend, 
so here's my votes on the patches in patches/for_Apache_0.2/:

B18_redirect.txt: +1
Clean compile, and it's logical.  No loss in functionality.  I also like 
the way the lines that are added have a "Patch B18" next to it. 

B39_CRLF: +1
We should try and conform wherever possible to the HTTP specs.  Correcting
the output of scripts is a different issue.  
HOWEVER
This script apparently obliterates the 401 response messages.  a closer 
look at the patch reveals that while "LF"'s were being replaced with 
"\015\012" everywhere, the 401 error response had a "LFLF" replaced with 
"\015\012" rather than "\015\012\015\012".  I put the fixed one at 
in /export/pub/httpd/patches/for_Apache-0.2/B39_CRLF_needs_b18-fixed.txt

B40_trailing_slash: +1
Bug fix? looks fine to me.

E25_custom_error_responses: 0
It's *almost* there.  In some situations I want to be able to control 
exactly what gets displayed - for example, hit 
http://www.hotwired.com/Login/ and fail to log in, and you'll see the 
message we give for 401 messages.  The directory name for these 
custom responses should be a variable in httpd.conf or srm.conf.  
Finally, the patch should include a change to the Makefile to inform the 
user that the -DCUSTOM_ERROR_RESPONSES option must be set in order for 
this to be turned on.  I'm not giving this a -1 because I feel this is 
better than nothing, and nothing is harmed by putting this in - the 
changes I want made could be done later.

E38-alias_malloc-2.txt: +1
As shown before, this definitely improves the memory usage.
I'll submit a patch making the CERN patch a compile-time option.

E37.load_cutoff.txt: +1
This is only an option for those on HPUX or SUNOS, so I can't test it
(so maybe I shouldn't vote) but if it works for others then I think it 
should be part of the distribution.

Hope this helps.

	Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@hotwired.com  brian@hyperreal.com  http://www.hotwired.com/Staff/brian/