Re: svn commit: r924201 - in /subversion/trunk/subversion: libsvn_client/mergeinfo.c libsvn_subr/mergeinfo.c tests/libsvn_subr/mergeinfo-test.c

Paul Burba wrote:
> On Thu, Mar 18, 2010 at 8:44 PM, Julian Foad <julian.foad_at_wandisco.com> wrote:
> > On Thu, 2010-03-18 at 17:34 -0400, Greg Stein wrote:
> >> If somebody manually tweaks an svn:mergeinfo property to include 0,
> >> then is our library going to dump core?
> >>
> >> (I imagine props are checked much earlier, but a new core dump is
> >> always something to question :-P )
> >
> > Good question. Let's see where the mergeinfo prop values are parsed
> > into svn_merge_range_t's.
> >
> > Paul, do you think I should add a check like the following to
> > svn_mergeinfo_parse()?
>
> Hi Julian,
>
> Yes, I think we should do this. However, your patch won't work if
> there is only a single revision or single range, because your two
> checks are inside a "if (rangelist->nelts > 1)" block:

Oops, I saw that at first and then forgot about it.

[...]
> Instead I would put a single check inside parse_rangelist:

OK, done (but explicitly testing for zero, similar to the other explicit
tests in this function) in r925290.

Thanks.
- Julian
Received on 2010-03-19 17:00:05 CET