cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Robin Green" <>
Subject Re: [C1] Memory problem
Date Thu, 08 Mar 2001 11:06:44 GMT
Ulrich Mayring <> wrote:
>Robin Green wrote:
> >
> > Yes - unfortunately that change has been reversed now (although it is 
> > like that in the latest release) because it caused a serious bug when 
> > parameters were actually being used by the stylesheet.
>Could you elaborate on this serious bug?

Argggh!! How embarassing! I looked back at the "bug report"

and it was not a bug report at all, it was just a question "will this change 
affect me?". I realise I have low credibility now, because after saying 
"Yes, I'll accept this change", and then "I don't know why I accepted that", 
and then "this caused a serious bug", I am now saying - it doesn't cause any 
bugs. Really.

The reason is simple. Caching of each page is always done based on all the 
request parameters (and the user agent, as well). Therefore when the 
XSLTProcessor.hasChanged method is called, the XSLTProcessor only needs to 
return whether the stylesheet has been modified on disk, not whether any of 
the parameters has changed (since if they had, that would not even hit the 
same page in cache!) I was actually right the first time, I just forgot my 

I'll now change it back AGAIN, so that XSLTProcessor ignores the request 
parameters in Monitoring!! Thanks very much Uli for harassing me for an 
explanation - it paid off! ;-)

I also had to fix the same bug in XIncludeProcessor and ProducerFromFile. I 
have to say, the Monitor system is really confusing (I hope we don't make 
the same mistakes in C2) - my redesigned system makes alot more sense 
conceptually, IMHO - even if _some_ of the bugs that it was intended to fix 
have now been fixed already!

Robin Green
Official Cocoon 1.x Dunce

Get Your Private, Free E-mail from MSN Hotmail at

To unsubscribe, e-mail:
For additional commands, email:

View raw message