On 11/27/05, Niall Pemberton <[EMAIL PROTECTED]> wrote:
>
> On 11/28/05, Martin Cooper <[EMAIL PROTECTED]> wrote:
> > On 11/27/05, Niall Pemberton <[EMAIL PROTECTED]> wrote:
> > >
> > > Sorry Martin, I'm probably driving you mad. I just attached the 3rd
> patch
> > > before I saw you'd accepted the second :-(
> > >
> > > I'll go away now.
> >
> >
> > ;-)
> >
> > Actually, the second is more robust than the third in any case, since
> the
> > default location for these files is the container's temp directory,
> which is
> > very likely to be cleaned out on a server restart. And if the default
> isn't
> > used, all bets are off anyway.
>
> If the default is used and the  container's temp directory is cleared
> out - then the 2nd patch is stuffed as well - it'll try and copy a
> file that doesn't exist.


Ah, right. Oh well.

> I'm still apprehensive about the performance implications, if people don't
> > think about what they're doing, but at least now DiskFileItem adheres to
> its
> > contract.
>
> I agree, and I think they'll be alot not so large files that'll now
> work well. Although the best solution was your original comment -
> don't stick the FileItem in the Session :-)
>
> You look like you have FileUpload heading towards a release?


That's the general idea, yes. I've temporarily abandoned the enhancements I
had in the wings, on the basis that what's in SVN now has been there so
long, it's stable, and people need a released version, especially the folks
who need portlet support. The bugs are fixed, so it's pretty close, I think.

--
Martin Cooper


> Martin Cooper
> >
> >
> > Niall
> > >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>

Reply via email to