Mapserv.exe under Apache doesn't finish requests

[Frank Warmerdam]

On Fri, 18 Feb 2005 14:53:15 +0100, Dijken, M. van (Martin)
<Martin.van.Dijken at rijnmond.politie.nl> wrote:
> Hey Frank,
>
> > Is the WFS you are operating against public?  If so, I would
> > encourage you to prepare a minimal map that demonstrates your
> > problem so that others can reproduce it and possibly fix it.
>
> Afraid not. It's a very closed source thing that runs on a closed network.

Martin,

Well, closed source wouldn't be a problem, but given that it is
on an internal network it may be very difficult for anyone outside
to help much.  You could try and reproduce the problem against
a public WFS.

> > If you boil it down, you may need to submit it via bugzilla.
>
> Hehe, the boiling down part is where I've got the problem. I don't know any
> more how to debug this. Got any hints?

By boiling down, I mean removing everything unnecessary to
reproduce the problem.  For instance, get rid of all other layers,
and reduce things to a direct mapserv invocation that demonstrates
the problem.  As a developer, I like to get a mapfile, a dataset (if
needed) and the exact arguments passed to mapserv via the url
so I can just run things in the debugger and see the problem.

In this case that may not be possible unless you can reproduce
the problem against a public WFS.

> > I will note that there are limitations on MapServers (aka
> > OGR's) GML reader.  Perhaps the WFS server is returning a GML
> > result that is unreadable to OGR?
>
> Hmm, that sounds rather unlikely.

You are correct.  I failed to take into account that it is working fine
on WinXP.

I did notice the following message in your log:

  HTTP request: id=12, found in cache, skipping.

I wonder if there could be any issue with the caching logic.  If there
is a way of disabling this it might be worth a try.  I'm not a libcurl
guru so I don't know what is involved.  Daniel might be the best person
to ask, but I he may not be reading this far in this message.

I don't know what the right way of disabling caching is, but you could
try hacking maphttp.c.  In the function msHTTPExecuteRequests()
add in the line "bCheckLocalCache = MS_FALSE;" just after the
"char debug = MS_FALSE;" declaration.  That should disable use of
caching.

If disabling the caching solves the problem then I would encourage
you to repost your analysis with the title about "HTTP Caching Bug"
or similar.  That should get Daniel's attention.

Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent