[prev in list] [next in list] [prev in thread] [next in thread]
List: sylpheed
Subject: [sylpheed:28366] Re: Lost mail if no space left on device
From: Seth Kurtzberg <seth () cql ! com>
Date: 2006-08-24 1:49:27
Message-ID: 20060823214927.0a5c7cfc.seth () cql ! com
[Download RAW message or body]
On Wed, 23 Aug 2006 21:20:28 -0400
"Steven M. Bellovin" <smb@cs.columbia.edu> wrote:
> On Thu, 24 Aug 2006 09:39:39 +0900, Hiroyuki Yamamoto <hiro-y@kcn.ne.jp>
> wrote:
>
> > Hello,
> >
> > On Wed, 23 Aug 2006 11:33:06 -0400
> > "Steven M. Bellovin" <smb@cs.columbia.edu> wrote:
> >
> > > > It might be smarter if the program checked for space before deleting
> > > > the message from the server. However, letting a hard disk fill up
> > > > is a bad idea and can lead to other problems. It is also smarter
> > > > to provide adequate storage for your needs. If Hiroyuki can add
> > > > that test without to much degradation in performance then it would
> > > > be an improvement but I wouldn't label it as a bug and wouldn't
> > > > want to take a day in and day out performance hit to protect me
> > > > from something I should avoid anyway. It's not like disk storage
> > > > is a major cost these days.
> > > >
> > > It is most definitely a bug. The check is cheap enough -- look for
> > > ENOSPC after a write has failed. However, even if the check were
> > > expensive it still has to be done.
> >
> > I've added such error checking long time ago. POP3 session will stop
> > immediately if any I/O error (including ENOSPC) occurred (see
> > libsylph/{session.c, pop.c, recv.c}, src/inc.c).
>
> Great.
> >
> > But it could be possible that some bugs creeped in after that. It will
> > be appreciated if someone take a look at what is going on when no disk
> > space left. I'll also look into it.
> >
> Bugs can always happen, even in the best code.
Indeed. This isn't a creeping bug, I believe. It just requires an additional patch \
for the IMAP client case.
Seth Kurtzberg
>
>
> --Steven M. Bellovin, http://www.cs.columbia.edu/~smb
>
>
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic