Daniel,
I tried to remove the [ <!ELEMENT > ] bit and it still doesn't work :(
Once I remove the entire <!DOCTYPE> tag it does work, but I guess just
because it doesn't know where to validate it against.
I'll try to see if I can find what bit of the XML isn't deemed valid by
MSXML, but no guarantees that I'll find it out.
Regards,
Steven
Daniel Morissette wrote:
Steven,
The EMPTY element declaration in the DOCTYPE was simply inspired from
the example found in the spec itself (in Annex A.2 of WMS 1.1.1 and
also present in most/all previous revisions).
I'm no XML guru so I'm not sure what the best fix is, but once you
find out I'd invite you to file a ticket documenting the original
problem and the new format that makes MSXML and other parsers happy.
Daniel
Steven M. Ottens wrote:
Hi all,
My mapbuilder/openlayers application wasn't working on IE7, it
complained that the WMS getCapabilities doc generated by Mapserver
wasn't valid XML. I've finally tracked it down to the doctype
declaration:
Mapserver produces
<?xml version='1.0' encoding="ISO-8859-1" standalone="no" ?>
<!DOCTYPE WMT_MS_Capabilities SYSTEM
"http://schemas.opengeospatial.net/wms/1.1.1/capabilities_1_1_1.dtd";
[
<!ELEMENT VendorSpecificCapabilities EMPTY>
]> <!-- end of DOCTYPE declaration -->
The parser removes both the <?xml ..> tag and the <!DOCTYPE ..> tag
from the XML before parsing. So normally the to-be-parsed XML text
starts with:
<!-- end of DOCTYPE declaration -->
Which is a valid piece of XML
However MSXML apparently thinks that the closing > after EMPTY
belongs to the opening < before !DOCTYPE. So instead now the XML text
starts with:
]>
Which obviously isn't valid XML and MSXML fails.
So my questions:
-Anyone else noticed this problem?
-Why is there an (empty!) ELEMENT declaration within the DOCTYPE?
-How can I remove this ELEMENT?
Regards,
Steven
