[mapserver-dev] Proposal for OWS Dispatch improvement

Daniel Morissette dmorissette at mapgears.com
Wed Jun 8 09:51:59 EDT 2011

+1 here too, as long as we make sure that the new code deals properly 
with the incompatibilities between various OGC service specs and 
versions. There is stuff like SERVICE being non-existent (or optional?) 
in some older specs, vs newer versions of some specs requiring an 
exception if SERVICE is not set... this exception is sometimes required 
to pass OGC compliance tests.

If I remember correctly, in order to pass OGC tests we added a 
"mode=ows" option that can be used to enforce strict OGC service mode 
and produce an exception if the request does not match any supported 
service type, instead of falling back on the traditional CGI mode which 
is the default behavior when mode=ows is not explicitly set.


On 11-06-08 09:39 AM, Kralidis,Tom [Ontario] wrote:
> +1 on the improvement and no RFC needed.  I echo Assefa's comments.
> ..Tom
> -----Original Message-----
> From: mapserver-dev-bounces at lists.osgeo.org on behalf of Yewondwossen Assefa
> Sent: Wed 08-Jun-11 09:32
> To: Fabian Schindler
> Cc: mapserver-dev at lists.osgeo.org
> Subject: Re: [mapserver-dev] Proposal for OWS Dispatch improvement
> Hi Fabian,
> I agree with the assessment that It is a good improvement. Checking
> quickly, I do not see any issues of doing what you propose if the
> current quirks and functionality described in #2531 are kept.  I also
> agree that there is no need for an RFC.
> best regards,
> On 07/06/2011 5:18 PM, Stephan Meißl wrote:
>> On Tue, 2011-06-07 at 15:13 -0400, Frank Warmerdam wrote:
>>> On 11-06-07 11:56 AM, Fabian Schindler wrote:
>>>>      1. In msOWSDispatch, a preliminary parsing function determines the service,
>>>>         version and the request from the KVP or XML request.
>>> Fabian,
>>> Your idea sounds good to me.  The only caution I will provide is I think
>>> there were some special cases for some services treating the SERVICE
>>> parameter as optional (perhaps WMS?) in some versions and you would need
>>> to take care to replicate this.
>>> We have lots of time to shake out any quirks with regard to this in trunk
>>> before 6.2 is released.  Assuming you get positive feedback from some of
>>> Assefa, Daniel, Tom and Steve who I think cover much of the OGC service
>>> responsibilities then I would suggest you go ahead and prepare the patch.
>>> Stephan can commit the changes.  I don't think an RFC is needed though
>>> perhaps the other guys will see a need for one.
>>> Best regards,
>> Frank, all,
>> you're right, e.g. in WMS 1.0.0 the SERVICE parameter is non-existent.
>> Tom provided a good starting point [1] some time ago.
>> I guess Fabian would be happy to write an RFC although he wouldn't
>> object if he doesn't need to ;). To start with, maybe a more verbose
>> ticket could be used? Anyway, lets see what others think.
>> cu
>> Stephan
>> [1] http://trac.osgeo.org/mapserver/ticket/2531#comment:6
>> _______________________________________________
>> mapserver-dev mailing list
>> mapserver-dev at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/mapserver-dev

Daniel Morissette
Provider of Professional MapServer Support since 2000

More information about the mapserver-dev mailing list