0.7.2b

[Rob Hartill <ha...@ooo.lanl.gov>]

the reason my 0.7.2 crashed overnight was that the signal handler for
the parent detecting the child had died only caught the first
dead child...  hmmmmm

0.7.2b *should* fix that.


Randy, I haven't noticed any core dumps, but I'm not using VIRTUAL HOST
at all. Are you ?  It wouldn't surprise me if that's hosed right now.


-=-=-=

On another topic, an idea for "last-modified"

How about if when apache is reading cgi script headers, and it sees a
last modified header, it runs its "if-modified-since" check. If the
thing hasn't been modified, Apache can tell the client, then zap the
CGI script, or soak up its output.

Checking for if-modified-since in CGI is a real drag, beacause of the
different formats the date can take (e.g. Netscape sends the if-mod-since
date in the old HTTP date format)

-=-=-=-

Anything happening on the logging front ?

some thoughts on that... If we do go for the separate process idea, would
it be reasonable for Apache to not do name lookups, and leave that to the
log process which could have a cache of its own. The only problem I see
is that some authorization is done on client name pattern matching. Yup
that's a big problem. Is there any way around it ?


--
Rob Hartill                           
http://nqcd.lanl.gov/~hartill/

Re: 0.7.2b

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

On Wed, 7 Jun 1995, Rob Hartill wrote:
> Randy, I haven't noticed any core dumps, but I'm not using VIRTUAL HOST
> at all. Are you ?  It wouldn't surprise me if that's hosed right now.

I'll try 0.7.2b with VH later today.

> -=-=-=
> 
> On another topic, an idea for "last-modified"
> 
> How about if when apache is reading cgi script headers, and it sees a
> last modified header, it runs its "if-modified-since" check. If the
> thing hasn't been modified, Apache can tell the client, then zap the
> CGI script, or soak up its output.
>
> Checking for if-modified-since in CGI is a real drag, beacause of the
> different formats the date can take (e.g. Netscape sends the if-mod-since
> date in the old HTTP date format)

I think convergence is being reached on that though - didn't Netscape fix 
that in 1.1?

Also, where is the if-modified-since date passed along to the CGI 
script?  I can't seem to find it in any environment variable.... 
 
> -=-=-=-
> 
> Anything happening on the logging front ?
> 
> some thoughts on that... If we do go for the separate process idea, would
> it be reasonable for Apache to not do name lookups, and leave that to the
> log process which could have a cache of its own. The only problem I see
> is that some authorization is done on client name pattern matching. Yup
> that's a big problem. Is there any way around it ?

Could you turn on reverse-DNS lookups only when it's needed for protected 
areas?

	Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com  brian@hyperreal.com  http://www.[hyperreal,organic].com/