Re: [xslt] DocLoader API Feedback

Ok, the API works like a charm. I made sure to use libxml's URL routines so that I could guarantee that I was comparing two libxml-created absolute URIs, and everything works fine. I have not yet done anything with the dicts. Is sharing a dict only a performance optimization, or will things actually malfunction? In my simple tests (so far), everything seems to work fine without my worrying about the dicts.


On Aug 19, 2004, at 1:08 AM, Daniel Veillard wrote:

On Thu, Aug 19, 2004 at 12:38:51AM -0700, Dave Hyatt wrote:
How do I know which xmlNodePtr is being processed at the time? It seems
like the URI is not enough to sufficiently disambiguate between the
same stylesheet that might be both imported and included by a parent

That information is not provided. A document cache might actually be used for stylesheet since there is xsltDocumentPtr docList; at least for includes.

  <import href="othersheet.xsl"/>
  <import href="othersheet.xsl"/>

Not forbidden, note that import and include are processed very differently.

Is the above legal in XSLT? If it is, should I be giving you the exact
same xmlDocPtr both times?  If I am supposed to give you two different

You should *never* give back twice the same xmlDocPtr at that API
level. This should be very clear from my earlier mails, it replaces an
xmlReadxxx and the documents are subsequently, modified, sometime subtrees
are cut off and merged in the main stylesheet. Do not use it as a cache,
it cannot play that role. It will leak memory, crash and burn !

xmlDocPtrs, then how can I distinguish between the two loader requests,

You always return a different one !


Daniel Veillard | Red Hat Desktop team
veillard redhat com | libxml GNOME XML XSLT toolkit | Rpmfind RPM search engine

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]