[
Lists Home |
Date Index |
Thread Index
]
At 10:11 AM 10/26/2002 -0600, Uche Ogbuji wrote:
>Maybe your idea will work for the XInclude part, though I think it's
>likely to
>be messy either way. However wording that says "xpath() processors must
>report an error if they encounter either an unresolved entity" is not really
>meaningful. You rightly couch your draft mostly by reference to XPath, and
>XPath has no concept of an XML entity (besides the sliver in
>unparsed-entity-uri(). I think it's not really meaningful, then to say that
>an XPath processor could encounter an unresolved entity.
It's not meaningful to the XPath processor itself, but it could plausibly
be triggered by an unparsed entity report from the parser feeding the XPath
processor.
Plausible doesn't seem good enough in this case, so I'm leaning toward a
simple warning in the document that XPath resolution may return varying
results depending on the parser used to feed the XPath processor. I could
alternately specify that entity resolution must be performed before the
XPath processor, but that feels like my higher-level spec is reaching down
a few levels into the murk. Hmmm...
Simon St.Laurent
"Every day in every way I'm getting better and better." - Emile Coue
|
