On Mon, Feb 23, 2004 at 02:46:14PM -0600, Tim wrote:
> On Mon, Feb 23, 2004 at 04:31:00PM +0000, Brian Candler wrote:
> > vpopmail is a fragile and buggy application. Until recently, if you tried to
> > perform two successive authentications, the second authentication would fail
> > if the username was shorter than the first one, because a buffer was not
> > being cleared.
> 
>   Please do not spread such misinformation.  I fixed that bug on
> 03/25/02 and it was released as part of 5.3.4 and all subsequent
> versions.

OK, then it was not fixed in any *stable/production* version. If I
understand rightly, 5.3.x is tagged as a "development" branch.

There have been many people saying "I'm running vpopmail 5.2.1 and I'm
seeing this bug...", and I'm pretty sure 5.2.2 was released only recently.

>   I agree that vpopmail needs work, but there has been tremendous
> improvements in both features and stability especially since the
> project moved to SourceForge.

I'm glad to hear it, as hopefully it will reduce noise here.

Just to recap the start of this thread:
- someone reported a bug (the file 'sqwebmail-ip' becoming owned by root
  rather than the mailbox owner)
- Sam said that only vpopmail people reported this bug, so it's almost
  certainly a vpopmail problem
- someone rushed to the defence of vpopmail, saying that it's fine and
  sqwebmail-ip is nothing to do with vpopmail anyway
- Sam and I pointed out a simple mechanism by which vpopmail may be
  broken and cause the problem seen
- further defence of vpopmail, saying that it works OK in qmail-pop3d
  and bincimap

Ill-considered argument just annoys me. My annoyance was meant to be
directed at those people making those ill-considered arguments, not at the
authors of vpopmail, and I'm sorry if I caused any offence.

Regards,

Brian.

Reply via email to