On Wed, 2016-08-10 at 21:56 +0200, Peter Meerwald-Stadler wrote:
> Hello,
>
> > > starring at the same issue ATM :)
> > > the proposed fix below gets rid of the potential leak, butÂ
> Â
> > > (2) in case value == 0, we'd still overwrite old_value, makes no
> > > sense
> >Â
> > Why not? value is the new value, old_value is the old value. The new
> > value doesn't affect what we should print in the log message for the
> > old value.
>
> in case old_value != 0 and value == 0, we'd log
> key: (data) -> (unset)
> but the correct logging IMHO would be
> key: old_value -> (unset)
>
> new suggestion:
>
> Â Â Â if (pa_proplist_contains(o->proplist, key)) {
> Â Â Â Â Â Â Â old_value = pa_xstrdup(pa_proplist_gets(o->proplist, key));
> Â Â Â Â Â Â Â if (value && old_value && pa_streq(value, old_value))
> Â Â Â Â Â Â Â Â Â Â Â goto finish;
> Â Â Â Â Â Â Â if (!old_value)
> Â Â Â Â Â Â Â Â Â Â Â old_value = pa_xstrdup("(data)");
> Â Â Â } else {
Oh I see. Do you want to send a proper patch?
--Â
Tanu
