Re: 1.7.3. crashing for time to time (SVN)

[Kai Willadsen <kai willadsen gmail com>]

On 30 June 2013 21:57, Gilboa Davara <gilboad gmail com> wrote:
> On Thu, Jun 27, 2013 at 11:53 PM, Kai Willadsen <kai willadsen gmail com> wrote:
> > On 28 June 2013 00:44, Gilboa Davara <gilboad gmail com> wrote:
> > > On Thu, Jun 27, 2013 at 5:21 PM, Chuck Tuffli <chuck tuffli net> wrote:
> > > > I suspect this is the same issue Kai fixed recently for me. Give a
> > > > directory structure like:
> > > >
> > > > a/
> > > >   b/foo.c
> > > >   c/bar.c
> > > >
> > > > meld would generate the error you mention if both foo.c and bar.c had
> > > > modifications. The fix (e79f5611799d0cd88e8ec630b699d9027503a45c I
> > > > think) in git fixed the problem I was seeing.
> > > >
> > > > --chuck
> > >
> > > I rather not pull git right now.
> > >
> > > Kai,
> > >
> > > Just to be certain: Should I simply comment out
> > > self.vc.uncache_inventory() in _search_recursively_iter?
> >
> > Chuck is right, this looks like exactly the same problem. So yeah,
> > that's what I'd try.
> >
> > The behaviour this change fixes is that running a comparison in a
> > different subdirectory to the first comparison that's run, breaks.
> >
> > cheers,
> > Kai
>
> Seems to work just fine.
> Thanks.
>
> Should I contact the Fedora meld maintainer and get the fix applied to
> the official package, or should I simply wait for 1.7.4?

I think that that's a fairly low-risk patch for a real issue... so
yes, trying to get it into the 1.7.3 fedora package would be a good
idea. I think 1.7.4 has most of the remaining breaking changes for the
1.7.x cycle, so after that things should calm down.

cheers,
Kai