Re: [xslt] [xslt]: Who has a real world use for a high --maxdepth?

[Graham Seaman <graham seul org>]

On Fri, 25 Oct 2002, Jean T. Anderson wrote:

> I'm new to xslt but notice that, so far, only error conditions seem to 
> hit the default xsltMaxDepth of 5000.
> 
I'm running an application that does automated conversion of heavily
tabular html to a text-only css-based layout. As part of it there is a set
of fall-back routines which identify the nested tables with the largest
number of text words (trying to work out which parts of the page are real
content and not just layout or links). That fell over regularly with a
maxdepth of 1000; I bumped it up to 2000 (IIRC) and it fell over on a
small percentage of pages; now on 4000 it's never fallen over (except when
I've (ahem) deliberately introduced a bug).  The need for the high
maxdepth is probably inefficiently written code; but the code works, and
is reasonably fast (ok for realtime). Anyway, at least part of the point
of this ramble is that for many applications it's going to be impractical
to calculate a fixed maxdepth need for a stylesheet, the depth of
recursion is a function of the data, too.

I'm assuming no-ones going to say 'pah > 2000 is nothing, I need
500,000....', but they probably will ;-)
   
Graham