The strategy pattern worked in regards to server behaviour. It did not 
work well with parsing mis-behaving XML documents. I was just always 
very permissive when it came to elements missing or empty, or in the 
wrong order. I think in order to separate a "pure" and a "permissive" 
XML parsing mechanism, I would need to have two definitions of the spec: 
one that is rigid, and the other that is as flexible as possible (what 
my current WMS one is). That seems like a lot of code/work duplication 
to me.

Richard

Jody Garnett wrote:
> Adrian Custer wrote:
>> 4) WFS will be made permissive to incorrect requrests (but with some
>> mumbling)
>>   
> Um I had this discussion with Richard (for WMS) and also with David for 
> WFS ...
> For WMS we have strategy objects that we delegate to that implement the 
> different policies
> defined by each specification. This practice allows us to define 
> additional strategy object for
> specific service implementations that are have issues....
> 
> I would like to maintain this idea, if I can ask the code to be "pure" 
> spec, all hacks/allowances
> isolated in a manner similar to WMS (or a subclass if that is too hard).
> 
> So Richard mostly raised this issue, I just ask that we provide a 
> technical solution to the problem
> to prevent permissive code from making the code base a mess.
> 
> So time to contact the module maintainer; and I would love to see the 
> design before giving this idea
> a thumbs up.
>> http://news.independent.co.uk/world/fisk/article1204432.ece
>> * FrankW is reading about Lebanon...
>>   
> That is crazy stuff Frank...
>> <acuster> Would it be possible, in future releases, to somehow trim down
>> the geoapi released with Geotools
>> <acuster> to contain only the relevant modules?
>>   
> That may be possible; but we can also pull an interesting trick I have 
> seen in other projects,
> in our reference section only provide javadoc links to the packages that 
> matter to us...
>> <acuster> I don't expect an answer, but am raising the possibility
>> <acuster> Done
>>   
> Darn - caught answering yet again :-)
>> <rgould> so due to a recent bug issue in uDig, I discovered that the WFS
>> plugin in geotools explodes if it encounters a Capabilities document
>> that does not conform to the WFS specification
>> <rgould> the problem is that the part that was missing from the document
>> was not essential to use the service
>> <rgould> now
>> <rgould> a couple years ago, when I was implementing the WMS plugin,
>> James gasped in dismay as he realized I was permitting servers that did
>> not conform to the WMS specification to be accessed normally
>> <rgould> i would like to do the same to WFS
>> <rgould> (so we can access these broken servers)
>> <FrankW> rgould: ++ on permissiveness.
>>   
> Um, only if the code can still be maintained ;-)
>> <rgould> I thought it might be a good idea to come up with a general
>> concensus regarding what to do when something breaks specification
>>   
> :-)
> 1) isolate specification
> 2) create dialects (derrived from the origional specification) that 
> allow for specific breaches of contract in a controlled manner
> 
> This will be come more obvious when we have WFS1.0 and WFS 1.1 support.
> 
> Cheers,
> Jody
> 
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys -- and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Geotools-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/geotools-devel


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to