[merge][ping] Wait on locks

[Robert Collins]

On Mon, 2006-09-25 at 16:33 -0500, John Arbash Meinel wrote:
> Robert Collins wrote:
> > On Mon, 2006-09-25 at 15:52 -0500, John Arbash Meinel wrote:
> >> The attached patch changes bzr so that when attempting to get a
> >> write-lock it will wait up to 1 hour, and give better information to the
> >> user about what is going on.
> >>
> >> It seems to have gotten favorable reviews from users, but I haven't
> >> heard anything from Robert, Martin or Aaron.
> >>
> >> This is a fix for both bugs #43521 and #49556. The branch is available from:
> >> http://bzr.arbash-meinel.com/branches/bzr/0.11-working/wait-on-lock
> >>
> >> The initial work has been sitting around since 10/14 (almost 2 weeks
> >> now), so I'd like to put it to rest.
> > 
> > I think having a longer default is good but maybe 1 hour is excessive? 
> 
> I think 1 hour is excessive as well, just that seemed to be a moderate
> tradeoff between people that wanted infinity.
> 
> I had put it down to 5 minutes. The only reason I did it was the case of
> pushing a lot of changes over a slow link, which may take >5min for the
> push to finish. It probably isn't a common case, though.
> 
> If you prefer 5 min, I'm fine with changing it (or 10min, or whatever,
> it really is fairly arbitrary).
> 
> > 
> > I really dont like the idiom that seems to be turning up of:
> > 
> > +    def test_format_delta(self):
> > +        t = time.time()
> > +
> > +        def check(expected, delta):
> > +            actual = osutils.format_delta(delta)
> > +            self.assertEqual(expected, actual)
> > +
> > +        check('0 seconds ago', 0)
> > +        check('1 second ago', 1)
> > +        check('10 seconds ago', 10)
> > +        check('59 seconds ago', 59)
> > 
> > 'check' in the body is really uninformative. Something like 
> > self.assertIntervalFormatting('0 seconds ago', 0)
> > 
> > would be much clearer.
> 
> Well, a lot of these changes were created at the same time, and this is
> the first that you have reviewed them since. I'm avoiding this idiom
> since you commented that you don't like it. I had just already done it.
> 
> I'll clean up the tests, but you didn't really give much feedback about
> the rest. Do you like it, not like it? What length of time are you
> comfortable with, etc.

I think a 5 minute wait is ok, and certainly better than
'fail-immediately'. An hour long wait seems likely to cause headaches in
shell scripts and the like.

Yes, I like the rest of it.

-Rob
-- 
GPG key available at: <http://www.robertcollins.net/keys.txt>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 191 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060926/f528d2a6/attachment.pgp