4.6.0 final today

[Paul Ramsey]

Fair enough, I can back down, I only really had the opportunity to 
heavily test in the last two days, two months lead time notwithstanding. 
Having run across some of the other quirks recently as a result of that 
(Intersect vs Intersects) I am somewhat sensitive: broken standards 
support is almost worse than none at all (as it raises expectations 
before they are dashed).

<PAUSE>

So, I downloaded CadCorp MapBrowser, and it worked with mapserver 4.4 
and 4.6. Only one problem, and that was when it was given a LineString 
with only one coordinate (not legal GML). Which is a data or a server 
problem, depending on where one wants to point fingers.

So I feel better that our 4.6 problem is very likely self-inflicted and 
not something I should worry about in a larger context. Fire away, sorry 
for the tempest.

Paul

FYI, fun with mapserver versions:

http://mapserver.refractions.net/cgi-bin/mapserv42?map=/home/www/mapserv/maps/victoria-wms.map
http://mapserver.refractions.net/cgi-bin/mapserv44?map=/home/www/mapserv/maps/victoria-wms.map
http://mapserver.refractions.net/cgi-bin/mapserv46?map=/home/www/mapserv/maps/victoria-wms.map

http://mapserver.refractions.net/cgi-bin/mapserv42?map=/home/www/mapserv/maps/elections-wms.map
http://mapserver.refractions.net/cgi-bin/mapserv44?map=/home/www/mapserv/maps/elections-wms.map
http://mapserver.refractions.net/cgi-bin/mapserv46?map=/home/www/mapserv/maps/elections-wms.map 


Both are dual-mode WFS/WMS files, it was fun to see how things changed 
(or not) across versions.

Daniel Morissette wrote:

> Paul Ramsey wrote:
> 
>> Gah. :) When it comes to standards, releasing broken things is really 
>> catastrophic to the ecology of technology. (Speaking as someone who 
>> has a client.) It makes things look randomly broken, even when the 
>> client software is good. I am in a huge MUM rush, it is going to be 
>> hard to make time to DL mapbrowser and try this out, but will try. 
> 
> 
> We've been walking towards this release for almost two months now, it's 
> a bit late to bring this up now at a few hours from the final release. 
> Are there bugs documenting those issues?
> 
> 
>> If there *is* a problem, I would suggest that delay is better than 
>> broken standards support.
>>
> 
> I guess I'd submit this to a vote of the other developers (I have CC'd 
> mapserver-dev to get other developer's opinions). My personal vote goes 
> to NOT delaying the release and including the fix in 4.6.1 (assuming 
> there is an issue).
> 
> Daniel
> 
> 
>> Daniel Morissette wrote:
>>
>>> Paul Ramsey wrote:
>>>
>>>> Daniel,
>>>> Has the 4.6 WFS support been tested with a WFS client like Cadcorp 
>>>> SIS or something else with reasonable GML support? We have thus far 
>>>> failed to be able to consume Mapserver 4.6. While we have no reason 
>>>> to blame mapserver (it could easily be our fault) in general 
>>>> mapserver WFS does not get tested as heavily as the WMS, so it would 
>>>> be nice to have another data point that indicates that our problem 
>>>> is in fact our fault, and not Mapservers.
>>>
>>>
>>>
>>>
>>> You're right that MapServer's WFS doesn't get tested as much as the 
>>> WMS with 3rd-party clients, simply because WFS is not as widely used 
>>> as WMS. We know that Tom and Bart (CC'd on this reply) use it a lot, 
>>> so they may be able to comment on the testing they've done. 
>>> Unfortunately I don't think there is much we'll be able to do on WFS 
>>> before the 4.6 release if we want it out today, so if any fixes are 
>>> required they would have to wait for 4.6.1 (and after MUM3).
>>>
>>> BTW, have you tried testing it with Cadcorp Mapbrowser? How did it 
>>> go? I think we've used MapServer with it succesfully in the past, but 
>>> maybe some recent changes broke something?
>>>
>>> Daniel
>>
>>
>>
>>
> 
>