Quoting Luca Barbato (2014-10-15 02:26:00) > On 14/10/14 21:32, Anton Khirnov wrote: > > I really do not see how the new code is any better than the old one. > > It is not, as stated. > > > If I'm reading it correctly, it just does the exact same thing in a more > > roundabout way. > > It does the way it is expected, since that function really expects an > array of pointers and coverity figures that part out. > > Not sure if using an explicit array would make the (unlikely) problem of > doing (&pointer)[n] for n > 0 more apparent with a segfault instead of > writing over random memory. >
An array with the size 1 is exactly the same thing as a scalar, so you're not changing anything. And since accessing beyond the first element is just as undefined after the patch as before, I really do not think it can fix any segfaults. Not to mention that any such segfaults would mean a bug in lavr, to be fixed in lavr. -- Anton Khirnov _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
