httpd-apreq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Joe Schaefer <>
Subject Re: cvs commit: httpd-apreq-2/env/cgi_test/t/conf .cvsignore
Date Sun, 12 Oct 2003 15:45:48 GMT
Randy Kobes <> writes:


> I'm not sure if this is a Win32 thing or not, but when
> I tried running the cgi tests from within env/t/, it
> crashed due to the
>    LoadModule apreq_request_test_module ...
>    LoadModule apreq_access_test_module ...
>    LoadModule apreq_redirect_test_module ...
> directives - if I commented those out (as well as the
> directives in the associated <Location></Location> blocks),
> then things were OK. So I couldn't run the two tests
> concurrently. I thought this was due to conflicts with
> mod_apreq - there was a problem linking the c-modules on
> Win32 that required inclusion of -lmod_apreq (just as what
> happens with the perl glue). This may be peculiar to Win32
> - I was going to try to put together the Makefiles on
> Linux to see what happens there, and also try, on
> Win32, to see how one can build the c-modules and perl glue
> without having to explicitly add -lmod_apreq.

The cgi "script" compiles and links against libapreq_cgi before
the tests actually run, so I don't think this is a linking issue
on Win32 (btw- Unix libapreq_cgi is only built statically, maybe
the Win32 build should do this also).

The question I'm wondering about is how httpd sets up the subprocess
environment for the cgi-script prior to execution.  Maybe httpd's
mod_cgi has a bug on Win32?  Or maybe we need to use mod_cgid

Oh, well. I just would have liked to see the cgi tests incorporated
into env/t (as you originally planned) instead of their own location.
Maybe we can move them back in the future, after we've sorted all
this oddness out.

Joe Schaefer

View raw message