Re: [xsl] 10,000 document()'s

[S Woodside <sbwoodside@xxxxxxxxx>]

I'm assuming you'd rather use XSLT so "just use perl" isn't so useful 
an answer.

On Tuesday, April 8, 2003, at 12:05  PM, Peter Binkley wrote:

> but I gather from a posting by Michael Kay
> (http://www.biglist.com/lists/xsl-list/archives/200212/msg00507.html) 
> that
> all of those document() source trees are remaining in memory 
> throughout the
> transformation, adding up to megabytes of data.

Since document() trees are source trees it would seem as though the 
easy answer is that they must remain in memory. However I wonder what 
would happen if you only call the document() inside of a variable? 
Perhaps when the variable passes out of scope the document() source 
tree will be dumped. Maybe

> Can anyone suggest a strategy? The process doesn't have to be fast, it 
> just
> has to finish.

Using XSLT? You might try another parser just to see what happens. But 
given the limitations of the language, it may not be possible. It all 
depends on whether or not there is a /correct/ way of loading a 
document() source tree only inside a certain scope, and if there is, 
asking one of the parser authors to implement that.

Otherwise you could write a simple shell in another language that loops 
over the file and calls the XSLT with a parameter on the command-line 
that contains the name of a single file.

If you're on unix though you really should look at
  man wc
That will give you the counts you need ... again these could be passed 
in as parameters on the command line and picked up in the XSLT to 
simply add them into the output tree without needing to count the words 
in XSLT.

simon


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