Re: [mapserver-users] Mapserver6 calls exit() in FastCGI-mode

[Frank Warmerdam]

On 11-04-20 03:59 AM, Erik Gustafson wrote:

The problem is that the error from Mapserver 6, apart from its user
unfriendlyness for ows-requests, is reported by writeError() in
mapserv.c which calls exit() even if we are running under FastCGI. As
this error is non fatal I consider this a bug.


Erik,

For reference, I do *not* consider this a bug.  In my opinion it is
perfectly acceptable for MapServer processes to terminated sometimes on
an error even when running in FastCGI mode.  Reasonable FastCGI
implementations will just start a new process for the next request so
the only harm is a slight delay as libraries are reinitialized and
DB connections are re-established.

The error you are encountering is not something that will happen often
in practice.

In fact, I believe it is a good practice to have MapServer FastCGI
instances recycled occasionally as it is hard to ensure there are
no memory leaks in the more esoteric parts of the code.

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

___
mapserver-users mailing list
mapserver-users@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/mapserver-users

[mapserver-users] Mapserver6 calls exit() in FastCGI-mode

[Erik Gustafson]

Hello list,

when Mapserver 5 got a WMS request like this:
mapserv?request=getcapabilities&map=mymap.map

it responded with an wms_exception saying that the SERVICE parameter
was missing.

Mapserver 6 can not decide if the request if for a wms service or
another ows service and responds with:
mapserv(): Web application error. Traditional BROWSE mode requires a
TEMPLATE in the WEB section, but none was provided.

The problem is that the error from Mapserver 6, apart from its user
unfriendlyness for ows-requests, is reported by writeError() in
mapserv.c which calls exit() even if we are running under FastCGI. As
this error is non fatal I consider this a bug.


Less important:
I also think mapserver could be a bit more helpful. An error message like:
Traditional BROWSE mode requires a TEMPLATE in the WEB section, but
none was provided. For OWS-requests the SERVICE-parameter is needed.

Or maybe msOWSDispatch could check for all known requests (getmap,
getcapabilities, etc) and return an exception instead of returning
MS_DONE to mapserv.c

--
Erik Gustafson
___
mapserver-users mailing list
mapserver-users@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/mapserver-users