cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Bart Molenkamp" <b.molenk...@bizzdesign.nl>
Subject RE: Identifier used in FragmentExtractorTransformer
Date Fri, 21 Oct 2005 06:54:55 GMT
Hi,

How is this problem related to caching? It is not that hard for me to provide a patch which
just generates a new identifier, but I'm wondering if it breaks caching...

Bart.

> -----Oorspronkelijk bericht-----
> Van: Joerg Heinicke [mailto:joerg.heinicke@gmx.de]
> Verzonden: donderdag 20 oktober 2005 20:35
> Aan: dev@cocoon.apache.org
> Onderwerp: Re: Identifier used in FragmentExtractorTransformer
> 
> On 20.10.2005 15:15, Bart Molenkamp wrote:
> 
> > My problem is that the id is based on the request uri (and the number of
> > fragments that it extracts during a single transformation). The problem
> > is that the contents of an XML document, and the content of the SVG
> > image in that document (some chart generated from a database query) is
> > different for each user calling that page. But the request uri is always
> > the same, thus the ID for each extracted fragment is also always the
> > same, resulting in each extracted fragment being overwritten in the
> > transient store.
> >
> > This has some nasty side-effects, e.g. when using the browser's back
> > button, or even the possibility to have charts displayed from other
> > users.
> >
> > Therefore, I think it is good to change the way that id's are generated.
> > I was thinking to move the current ID generating code into a protected
> > method, and those who want other ID generation can extend the
> > transformer and overwrite the method.
> >
> > Is this a good idea, and if so, shall I provide a patch? Can it be
> > applied before the 2.1.8 release?
> 
> http://issues.apache.org/bugzilla/show_bug.cgi?id=28724
> 
> Jörg


Mime
View raw message