httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rodent of Unusual Size <Ken.C...@Golux.Com>
Subject Re: cvs commit: apache-1.3/src/support checkgid.c Makefile.tmpl
Date Mon, 29 Oct 2001 12:31:47 GMT
Marc Slemko wrote:
> 
> It is completely bogus to start adding support
> programs to check every possible error condition

Hardly what is happening here.

> It results in all the error checking code being
> duplicated in two places

Since there is NO 'error checking code' for this
case in the server proper, there is no duplication.

> The proper solution is to either have Apache do
> more checking itself before it daemonizes or change
> apachectl to handle errors that happen after it
> daemonizes.

Since AFAICS this condition can only reliably be
checked by actually calling setgid(), the first would
seem to potentially require some major work.  And
the latter introduces timing issues that are worse
than the proposed solution.

I disagree with your remarks in any event, considering
all the noise on the list in the past about 'the way
for non-httpd apps to do things with/to the config is
through preprocessing' -- this is just preprocessing
that doesn't store the config in LDAP, or XML, or
whatever, but does a band-aid check the server doesn't.

> It is not to duplicate all the error checking code
> in external programs.

Again, that ain't what's happening here, so a little less
hyperbole, eh?

I'll look into what's involved in doing the setgid()
before daemonising, but if it's a hassle on 1.3 I'm
going to stick with this band-aid solution because it
*does* address an existing problem.  2.0 can be done better.
-- 
#ken	P-)}

Ken Coar, Sanagendamgagwedweinini  http://Golux.Com/coar/
Author, developer, opinionist      http://Apache-Server.Com/

"All right everyone!  Step away from the glowing hamburger!"

Mime
View raw message