cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Geoff Howard" <coc...@leverageweb.com>
Subject RE: upload problem (WAS: RE: How to use the file uploading mechanism in cocoon)
Date Sun, 18 May 2003 02:30:55 GMT
The problem with cleanup() was inside the method itself, and not related
to the instanceof - it's now fixed.

Exposing more information from the request may not be necessary.  One reason
I
wanted this was to be able to determine at run time whether uploads were
allowed
from within components.  I've come up with a work-around in the upload.xsp
sample
(which is where I really wanted this the most).

Given that after 2.1 we'll be shooting for a whole new way of handling
uploads,
I'll be content to leave this alone now that the cleanup and the warning are
both
functional.

Geoff

> -----Original Message-----
> From: Geoff Howard [mailto:cocoon@leverageweb.com]
> > -----Original Message-----
> > From: Leo Sutic [mailto:leo.sutic@inspireinfrastructure.com]
> > > -----Original Message-----
> > > From: Geoff Howard [mailto:cocoon@leverageweb.com]
> > >
> > > This makes the cleanup fail on CocoonServlet:1209.
> >
> > Does it? Since the instanceof test is done against the object returned
> > from the requestFactory:

> > 1209:           if (request instanceof MultipartHttpServletRequest) {
> >                     ((MultipartHttpServletRequest) request).cleanup();
> >                 }
> >
> > Should be invoked properly.
>
> Yes, I think you're right - I noticed this last night when I went back to
> look at CocoonServlet.  (The first call I mentioned still won't work
> though).
> What's weird is that I thought I have confirmed that the cleanup is not
> happening.  I'll look at that more carefully.

...

> > > What about exposing something like boolean isMultiPart()?
> >
> > What about exposing a dispose() method instead? In a request-response
> > environment it makes sense to have an explicit disposal point.
>
> We have to clarify where.  I assume you mean (as I did) to expose that on
> the
> environment's wrapper around the request.  I'd be perfectly happy with
> that -
> The environment has the real request, and so can do the
> instanceof test, and
> can determine if MultipartHttpServletRequest.cleanup() can be called.  So,
> +1 from me?



Mime
View raw message