[Biogeosdi] the EDIT connection

Tue Jun 26 19:04:16 CEST 2007

If you move away the layers from the svn then you have to change back  
what I did...

And still there is the problem that omws is creating the XML  
structure for the getLayers operation based on the file structure  
where the real layers are stored. For example I saw that you had two  
folders there and our code is only getting the layers from first  
"group"/folder. Somehow I dont like recursive XML structures like  
this one, or HTML, because it makes parsing more difficult, but in  
the other hand I understand that you probably want this behavior.

If someone wanna try then using a SAX parser style, like Dave is  
doing with occurrences, will solve the problem and not take into  
consideration the XML structure... At the end it is a kind of  
theoretic problem, our code is trying to flat a non-flat XML structure.

Uff I really need holidays or I will become a ubersupermastergeek  
that noone understand, sorry.

Javi.

On 26/06/2007, at 18:54, Tim Sutton wrote:

> Hi Javi the master php uber hacker geek..
>
> Well done! Yes at the workshop Aimee and I removed the .svn directory
> from teh layers dir but I guess whenever someone updates the problem
> returns. I will move the layers to an out of svn location. Well done!
> Bonus coconut for you!
>
> Regards
>
> Tim
>
> 2007/6/25, Javier de la Torre <jatorre at gmail.com>:
>> Aghhhhhhhhh finally I got it!!!!!
>>
>> It took me 6 hours and at the end I realize I was stupid, like very
>> often happens when you are debugging...
>>
>> I thought nusoap was truncating the messages from the server and that
>> was the reason the layers where not getting display... I though this
>> because when doing error_logs I was seeing truncated messages...
>> After hours looking at the nusoap library and getting used to manage
>> how PHP read sockets I realize that the problem was that you can not
>> trust on the error_log function because it truncates the messages!!!
>> aghh
>>
>> At the end the problem is quite simple, the method is just parsing
>> incorrectly the result from the omws service.
>> And the resoan for that is this line on the OmService class:
>>
>> foreach ($xml->AvailableLayers->LayersGroup->LayersGroup->Layer as
>> $layerelement) {
>>
>> The simplexml parser we are using there is expecting the layers
>> returned by OM exactly in this structure. Now that, I suppose, Tim
>> has added the layers of the om repository into subversion the
>> structure of the result xml is different and this really hacky and
>> simple way of parsing the getLayers operation does not work. I could
>> easily solve it by changing this line to
>>
>> foreach ($xml->AvailableLayers->LayersGroup->LayersGroup[1]->Layer as
>> $layerelement) {
>>
>> But I am not sure what the best option here is... I think Open
>> Modeller should, as I exposed in an earlier message, return only real
>> layers on the server and not just look at the existing files on the
>> server, or at least should not include on the list hidden unix files
>> (like the .svn folders).
>>
>> If you prefer to keep the layers under the repository the we have to
>> tweak more the code...
>>
>> In any case I have commited the code with the fix for the time  
>> being...
>>
>> Cheers.
>>
>> On 25/06/2007, at 18:00, Tim Sutton wrote:
>>
>> > Hi Javi
>> >
>> > If you watch the logs (/tmp/phperrors.log) while clicking onto the
>> > getlayers link you will see only this:
>> >
>> > [25-Jun-2001 09:50:42] nlayers: 1
>> > [25-Jun-2001 09:50:46]
>> > http://omtest.cria.org.br:80/cgi-bin/om_soap_server.cgi<-- ctor  
>> om end
>> > point
>> > [25-Jun-2001 09:50:46]
>> > http://omtest.cria.org.br:80/cgi-bin/om_soap_server.cgi <--
>> > getAvailable Layers endpoint
>> >
>> > When I installed the omtest code on my mac here the mac console  
>> gives
>> > me a bit more detail:
>> >
>> > [Mon Jun 25 12:59:57 2007] [error] nlayers: 1
>> > [Mon Jun 25 13:00:02 2007] [error] [client 200.144.120.115] File  
>> does
>> > not exist: /Users/timlinux/Sites/biogeosdi/frontend/openlayers/
>> > theme/default/style.css
>> > [Mon Jun 25 13:02:06 2007] [error]
>> > http://omtest.cria.org.br:80/cgi-bin/om_soap_server.cgi<-- ctor  
>> om end
>> > point
>> > [Mon Jun 25 13:02:06 2007] [error]
>> > http://omtest.cria.org.br:80/cgi-bin/om_soap_server.cgi <--
>> > getAvailable Layers endpoint
>> > [Mon Jun 25 13:02:06 2007] [error] PHP Notice:  Undefined property:
>> > nusoapclient::$operation in
>> > /Users/timlinux/dev/php/biogeosdi/backend/libs/nusoap/nusoap.php on
>> > line 6837
>> >
>> > Does that mean anything to you?
>> >
>> > Regards
>> >
>> > Tim
>> >
>> > 2007/6/7, Javier de la Torre <jatorre at gmail.com>:
>> >> Give it to me!! I am the master of strange nusoap errors and  
>> you know
>> >> it :D
>> >>
>> >>
>> >> On 06/06/2007, at 20:49, Tim Sutton wrote:
>> >>
>> >> > Hi Guys
>> >> >
>> >> > Yes absolutely its all available as opensource already - they
>> >> can get
>> >> > it using anonymouse svn account. I've been working on getting  
>> the
>> >> > prototype completed yesterday and on and off this week.  
>> Something
>> >> > seems to have been broken with the layer selection screen and Im
>> >> > gettin g a strange error from nusoap at this point.  Ill post  
>> the
>> >> > exact message in a follow up email.
>> >> >
>> >> > Regards
>> >> >
>> >> > Tim
>> >> >
>> >> > 2007/5/30, Javier de la Torre <jatorre at gmail.com>:
>> >> >> Hi all,
>> >> >>
>> >> >> I am trying to finish what we did in the biodgeosdi meeting   
>> and
>> >> >> going through the emails.
>> >> >>
>> >> >> > Some people from partner institutes are visiting us now  
>> for the
>> >> >> EDIT
>> >> >> > project. I have shown them what we have done in Campinas, and
>> >> they
>> >> >> > were
>> >> >> > really interested.
>> >> >> >
>> >> >> > We got some nice interaction from them concerning  
>> Mapbuilder and
>> >> >> other
>> >> >> > softwares, and they were asking specifically for how we went
>> >> >> about to
>> >> >> > get the XMl from the OGC services, and pass it on to PHP.
>> >> >> >
>> >> >> > So, a quick check with all of the team : am I right that our
>> >> code,
>> >> >> > once
>> >> >> > finished, will be available on the web as open source ?  
>> For the
>> >> >> moment
>> >> >> > you have to log into the CVS server.
>> >> >>
>> >> >>
>> >> >>  From my side this is clearly the idea. And expect it really  
>> soon.
>> >> >>
>> >> >> Cheers.
>> >> >> _______________________________________________
>> >> >> biogeosdi mailing list
>> >> >> biogeosdi at lists.tdwg.org
>> >> >> http://lists.tdwg.org/mailman/listinfo/biogeosdi
>> >> >>
>> >> >
>> >> >
>> >> > --
>> >> > --
>> >> > Tim Sutton
>> >> >
>> >> > Visit http://qgis.org for a great Open Source GIS
>> >> > Home Page: http://linfiniti.com
>> >> > Skype: timlinux
>> >> > MSN: tim_bdworld at msn.com
>> >> > Yahoo: tim_bdworld at yahoo.com
>> >> > Jabber: timlinux
>> >> > Irc: timlinux on #qgis at freenode.net
>> >>
>> >>
>> >
>> >
>> > --
>> > Tim Sutton
>> > QGIS Project Steering Committee Member - Release  Manager
>> > Visit http://qgis.org for a great open source GIS
>> > openModeller Desktop Developer
>> > Visit http://openModeller.sf.net for a great open source ecological
>> > niche modelling tool
>> > Home Page: http://tim.linfiniti.com
>> > Skype: timlinux
>> > Irc: timlinux on #qgis at freenode.net
>>
>>
>
>
> -- 
> Tim Sutton
> QGIS Project Steering Committee Member - Release  Manager
> Visit http://qgis.org for a great open source GIS
> openModeller Desktop Developer
> Visit http://openModeller.sf.net for a great open source ecological
> niche modelling tool
> Home Page: http://tim.linfiniti.com
> Skype: timlinux
> Irc: timlinux on #qgis at freenode.net