Hi Goyo, Okay, thanks - this is a bit odd. I only get this error *sometimes*, and other times it works fine! It also worked ok on another machine. I think perhaps it is the server.
It seems to be working ok now. Correction! it *was* working ok - and now it isn't again! Plus, if I go to the website I am getting funny black diamonds - which I wasn't getting 2 minutes ago!! It's obviously some (intermittent) encoding problem. I will send you screenshots off list so you know what I mean So I think we can probably put it down to server error... I will check in the code later today (unless you do so first). Cheers, Dom On Thursday 25 June 2009 00:51:52 Goyo wrote: > Hi Dom, > > I can't reproduce the error, it works well for me. I'm using python2.6 > and ElementTree from the standard library. ¿Maybe a character encoding > issue? > > Note that xml encoding is utf-8: > >>> from urllib2 import urlopen > >>> from owslib.etree import etree > >>> url = > > 'http://www.idee.es/wcs/IDEE-WCS-UTM30N/wcsServlet?service=WCS&request=GetC >apabilities' > > >>> u=urlopen(url) > >>> s = u.read() > >>> s[20:36] > > 'encoding="UTF-8"' > > and that's also my filesystem encoding: > >>> import sys > >>> sys.getfilesystemencoding() > > 'UTF-8' > > At the moment I can think of three possible causes: > > a) The server response is indeed ill-formed, then we don't have to care > (just don't test this). > > b) This is a bug in urllib2 (u.read() returns an invalid string from a > good server) or ElementTree (fails to parse a valid string). > > c) We are supposed to do something with the xml string before passing it > to fromstring. > > I'll try to investigate. Hints welcome. > > Dom, does this has to do with the optional version stuff? does it work > if you add the version parameter to the url? > > Goyo > > El mié, 24-06-2009 a las 13:04 +0100, Dominic Lowe escribió: > > Hi Goyo, > > > > I'm having trouble getting the wcs_idee test to pass. > > > > I can demonstrate the problem with urllib2: > > >>> from urllib2 import urlopen > > >>> from owslib.etree import etree > > > > u=urlopen('http://www.idee.es/wcs/IDEE-WCS-UTM30N/wcsServlet?service=WCS&; > >request=GetCapabilities') > > > > >>> e=etree.fromstring(u.read()) > > > > <snip long traceback> > > <class 'elementtree.ElementTree.ParseError'>: not well-formed (invalid > > token): line 17, column 25 > > > > Basically it's complaining about some characters in the XML here: > > http://www.idee.es/wcs/IDEE-WCS-UTM30N/wcsServlet?service=WCS&request=Get > >Capabilities > > > > Do you have the same problem?? > > > > Cheers, > > Dom > > > > On Monday 22 June 2009 00:29:49 Goyo wrote: > > > Hi all, > > > > > > I didn't feel comfortable with version being mandatory in > > > WebCoverageService calls since it's provided in GetCapabilities > > > response. Or maybe I'm missing something? > > > > > > Find attached patches for wcsBase.py and wcs.py which make this > > > parameter optional in WCSCapabilitiesReader.__init__ and > > > WebCoverageService. The first one is necessary for the second one to > > > work. I also modified the doctest test_idee.txt in order to test > > > WebCoverageService (and WCSCapabilitiesReader.__init__). > > > > > > Cheers > > > > > > Goyo _______________________________________________ Community mailing list [email protected] http://lists.gispython.org/mailman/listinfo/community
