Re: possible to mimic while-like behavior?

[David Carlisle <davidc at nag dot co dot uk>]

David Carlisle
   > It is best to think of xsl:for-each evaluating the template on all the
   > Some xslt engines may in fact evaluate them one at a time, in document
   > order, but they are not obliged to do that.

Mike Brown
 >   Really? Can you give some examples where processing order is ambiguous and
 >   may vary from processor to processor?

Every example of xsl:for-each.

There is no requirement on the processor to evaluate the templates on
each node in any particular order, only that the resulting nodes are
appended to the result tree in document order. One of the main reasons
for using a side effect free language is to allow templates to be
evaluated in any order, your stylesheet may have xsl:for-each
select="chapter" and if you jump straight to chapter 100 then a system
could start rendering chapter 100 and go back and fill in the rest
later, you don't have to wait until it processes the first 99 chapters
before it can display the page you asked for. (Not that any system
currently works this way).

My reason for saying that it helps to think of parallel computation here
was not because any system is implemented that way, but it helps to
answer the faq questions of trying to terminate a loop based on results
of evaluating the template on some of the nodes. If you don't assume the
nodes are processed in document order, it is clear why this can not
work.

David


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