[gdal-dev] WMS Driver defaulting to version 1.1.1

[Timothy Astle]

To be clear, I personally don't have a problem with 1.1.1.  My concern 
is about version negotiation as a WMS client.  I've seen users omit the 
version and then when something doesn't behave as expected, they 
manually (in a web browser) make a request as follows:

http://www.example.com?request=GetCapabilities&service=WMS

But as well all know, that'd return a WMS 1.3.0 service capabilities 
document (assuming the server supports 1.3.0).  So they look at that 
document and cannot identify the problem.  Then they fuddle around and 
eventually realize that GDAL overrides the version and it makes a 
request as follows:

http://www.example.com?request=GetCapabilities&service=WMS&version=1.1.1

I think WMS 1.3.0 has been annoying since day one, especially with 
attention to axis ordering (ISO-ifying). *cough, cough*  However, I 
can't recall seeing a server that didn't support 1.3.0.  I have no 
objection about a user choosing a default version to use by explicitly 
specifying it.  My objection is that if a version is omitted, the 
software chooses an old default (which doesn't fit into typical version 
negotiation with a server) which can cause confusion.

I'm wondering if we're past the 1.3.0 "hate-on" yet? :)  It does seem a 
bit "big brother" to peg OWS client negotiation to a old version in a 
library.  I assume there are reasons, and I've been the first in this 
thread to cite possibly the biggest one.

Thanks for your feedback so far.  Everyone loves a good WMS 1.3.0 
discussion :)

Tim

On 02/03/2016 9:58 AM, Timothy Astle wrote:
> Does anyone know why the WMS Driver defaults to version 1.1.1 instead 
> of 1.3.0 by default?  Does anyone have any objections with this 
> changing to 1.3.0?  It seems like 1.3.0 would fall more into line with 
> the specifications content negotiation.
>
> https://github.com/OSGeo/gdal/blob/trunk/gdal/frmts/wms/wmsdriver.cpp#L96
> http://www.gdal.org/frmt_wms.html
> "WMS version. (optional, defaults to 1.1.1)"
>