Hi guys,
Thanks for your ideas. 
Geert, creating a list of all image URLs in a an XML file may not work for
this project because our list of images (content) is dynamic and keeps on
changing. Someone may remove/add/update the content files daily.

Conal, I will try out your idea. I hope performance is not an issue as
Cocoon will have to check for physical file every time it encounters an
image URL. But I will test it out and let you guys know.

Also, the other idea I had was to use an XSP page using Java and then use
standard Java File I/O. But looks like XSP can only be used as a 'generator'
which is difficult in my case. The transformation starts off with a source
XML file. And I can get the image URLS only after a 2nd pass of
transformations using XSL.

Anyways, I will send you a update soon!

Thanks again!
Hiral





-----Original Message-----
From: Geert Josten [mailto:[EMAIL PROTECTED] 
Sent: Thursday, February 17, 2005 4:16 AM
To: [email protected]
Subject: Re: XSL File I/O

> See this faq about using XSLT document() in Cocoon, which includes a 
> mention of this bug:
> http://cocoon.apache.org/2.0/faq/faq-xslt.html#faq-6
> 
> This is the "cacheability" problem referred to:
> 
> when you execute an XSLT transform that calls the "document('foo.xml')"
> method, the date stamp of the file foo.xml is not included in the 
> CacheValidity of the transform. So even after you modify foo.xml, the 
> pipeline will continue to serve the cached results.
> 
> Cheers
> Con

Ah! That explains some strange behaviour I experienced.

Does it help to put the XML broking pipe in a non-caching pipeline?

Cheers,
Geert

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]




---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to