[RFC PATCH] imapd.c: imapoptions: implement idle timeout

[Philipp Gesang]

-<| Quoting ellie timoney <ellie at fastmail.com>, on Wednesday, 2016-09-07 17:15:52 |>-
> On Wed, Sep 7, 2016, at 05:00 PM, Philipp Gesang wrote:
> > -<| Quoting Philipp Gesang <philipp.gesang at intra2net.com>, on Wednesday,
> > 2016-09-07 08:58:15 |>-
> > > -<| Quoting ellie timoney <ellie at fastmail.com>, on Wednesday, 2016-09-07 11:00:17 |>-
> > > > Hi Philipp,
> > > > 
> > > > On Wed, Sep 7, 2016, at 12:06 AM, Philipp Gesang via Cyrus-devel wrote:
> > > > > [...]
> > > > > -- 
> > > > > 2.4.11
> > > > Is the above indicating that the patch is against Cyrus 2.4.11? Or is
> > > > that just an unrelated signature?
> > > 
> > > AFAIK that signature is inserted by git-format-patch(1). Does it
> > > prevent the patch from applying?
> > > 
> 
> Ohh, that's git telling us what git version you used to make the patch.
> Hah. Okay.
> 
> I didn't try the patch yet cause I wasn't sure what to try applying it
> to.
> 
> > > > If the patch is not against 2.4.11, which version is it against?
> > > 
> > > Repo: https://github.com/cyrusimap/cyrus-imapd.git
> > > On top of: 546e8890256db8f14df544c70c9f6f0a79720fa
> > 
> > Correction, it’s: a546e8890256db8f14df544c70c9f6f0a79720fa
> > Copy-paste failure on my end, sorry.
> > 
> 
> Cool, thanks.
> 
> The principle seems sound, one preliminary comment though:
> 
> > +{ "imapidletimeout", 0, INT }
> > +/* Timeout for idling clients (RFC 2177) in minutes. A value of zero (the
> > +   default) means no timeout is enforced. */
> 
> If this setting were in seconds rather than minutes, then we could
> potentially add tests for it that don't need to hang for an entire
> minute.

 Actually I had seconds in the very first version of the patch
 but then adjusted it according to how that other timeout:

    { "timeout", 32, INT }
    /* The length of the IMAP server's inactivity autologout timer,
       in minutes.  The minimum value is 30, the default.  The default
    …

If I find the time today I will resubmit a seconds-based version.

>         Though looking through our other various timeout settings, it
> seems to be a mixed bag between seconds and minutes.

Indeed.

> I haven't actually tried the patch yet, I'm about to knock off for the
> day. I'll give it a go tomorrow.

Thanks,
Philipp

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.andrew.cmu.edu/pipermail/cyrus-devel/attachments/20160907/dcd6e3c1/attachment.sig>