[mapserver-users] WFS layer - XML/JSON safe chars

[Even Rouault]

Le vendredi 15 avril 2016 16:21:31, Rahkonen Jukka (MML) a écrit :
> Hi,
> 
> I was thinking that the output of GetCapabilities is XML and & must be
> escaped in XML. So if I write & into the abstract and Mapserver works
> correctly it should write my test abstract “test ä &”  escaped perhaps as
> test ä &
> That is the output from the online tool at
> http://www.freeformatter.com/xml-escape.html#ad-output
> 
> So my conclusion was that because Mapserver does not do it right here it
> won’t do it right elsewhere.
> 
> But now I made another test by editing the Geoserver test file “states.shp”
> in Spatialite as UPDATE states set STATE_NAME=’Colo&rado’ WHERE
> STATE_NAME=’Colorado’;
> 
> Output of GetFeature from WFS 1.1.0 from Mapserver 7.0.0  includes
>      <ms:STATE_NAME>Colo&rado</ms:STATE_NAME>
> 
> So my current conclusion is that my Mapserver does not escape “&” correctly
> when it is used inside a mapfile and when I do GetCapabilities request.
> However, when data come from Spatialite database and they are consumed by
> GetFeature then escaping happens correctly.
> 
> We are still comparing different things because you have PostGIS and I have
> Spatialite.
> 
> -Jukka Rahkonen-
> 
> 
> Ian Walberg wrote:
> 
> Jukka,
> 
> I am not sure I understand the test you did.
> 
> The data we are referring to is from the database field for the individual
> features name.
> 
> Regards
> 
> Ian
> 
> From: Rahkonen Jukka (MML) [mailto:jukka.rahkonen at maanmittauslaitos.fi]
> Sent: Thursday, April 14, 2016 10:19 PM
> To: Ian Walberg
> <ian.walberg at airborne.aero<mailto:ian.walberg at airborne.aero>>;
> MapserverList OSGEO
> <mapserver-users at lists.osgeo.org<mailto:mapserver-users at lists.osgeo.org>>
> Subject: Re: WFS layer - XML/JSON safe chars
> 
> Hi Ian,
> 
> Thanks for the clarification.
> I made a simple test by adding an abstract into layer metadata. This shows
> “test ä” as an abstract "wfs_abstract" "test ä”
> 
> This one shows an empty abstract
> "wfs_abstract" "test ä &"
> 
> So it looks like Mapserver doesn’t know to convert “&” into “&” for
> XML.

Actually it works for WFS 1.0, but not for 1.1 & 2.0. The reason is that the 
later use xmlNewChild() that expect values to be already XML-encoded, instead 
of xmlNextTextChild().

Working on a fix

> 
> -Jukka Rahkonen-
> 
> Ian Walberg wrote:
> 
> Jukka,
> 
> Our question was for the data/text read from the database and then
> published as XML by the WFS service.
> 
> So a pub with the name ‘Dog & bone’ created illegal xml as the & is not
> converted to &
> 
> Regards
> 
> Ian
> 
> From: Rahkonen Jukka (MML) [mailto:jukka.rahkonen at maanmittauslaitos.fi]
> Sent: Thursday, April 14, 2016 3:10 PM
> To: Ian Walberg
> <ian.walberg at airborne.aero<mailto:ian.walberg at airborne.aero>>;
> MapserverList OSGEO
> <mapserver-users at lists.osgeo.org<mailto:mapserver-users at lists.osgeo.org>>
> Subject: Re: WFS layer - XML/JSON safe chars
> 
> Hi,
> 
> You can’t have & in feature type names or attribute names in WFS because
> they are defined in the XML schema to be “QName”. The prefix and LocalPart
> are of type NCName https://en.wikipedia.org/wiki/QName and from
> http://stackoverflow.com/questions/1631396/what-is-an-xsncname-type-and-wh
> en-should-it-be-used
> 
> The practical restrictions of NCName are that it cannot contain several
> symbol characters like :, @, $, %, &, /, +, ,, ;, whitespace characters or
> different parenthesis. Furthermore an NCName cannot begin with a number,
> dot or minus character although they can appear later in an NCName.
> 
> -Jukka Rahkonen-
> 
> Lähettäjä: mapserver-users [mailto:mapserver-users-bounces at lists.osgeo.org]
> Puolesta Ian Walberg Lähetetty: 14. huhtikuuta 2016 19:23
> Vastaanottaja: MapserverList OSGEO
> Aihe: [mapserver-users] WFS layer - XML/JSON safe chars
> 
> Folks,
> 
> Can mapserver produce XML or JSON safe characters for WFS served data or do
> we need to ensure the source data has the escaped values.
> 
> We got tripped up today as there were some ‘&’ in our names which we had
> not considered.
> 
> Thanks
> 
> Ian

-- 
Spatialys - Geospatial professional services
http://www.spatialys.com