RE: XSLT processor performance

Kay Michael writes:
 > Saxon is wrong to reject the ora:output, it should ignore it. It's doing the
 > test "this is an element in an extension element namespace that I don't
 > recognise" before it does the test "this is a top-level element, so I should
 > ignore it if I don't recognise its namespace".

ah, thanks very much. so I'll note that as a bug in Saxon 5.4

 > I don't know how ora:output is designed to work, by the way, but
 > saxon:output is an instruction, not a top-level elements, so it only makes
 > sense to use it within a template body. 

the Oracle thing works in two stages, you declare a named output
specification, and then refer to it later in a template. Unless I
misunderstand, the top-level element sets up a file handle you can
write to at any time, which has its uses. the template-level output
does not open and close a file in its scope, as most (all?) of the
other implementations do

as I say, I may misunderstand it, since it does not actually work for
the way I wanted to test it.

Sebastian


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list