Bruno Haible <br...@clisp.org> writes: > It was introduced in commit 7d8705539af0c8555badacbdbe9c92ae0abca6f5. > Probably I thought that the uninitialized value doesn't matter, since in this > line it gets multiplied by 0, and 0 * anything is 0. But valgrind reports > an "uninitialised value" in such a case. Therefore you are right: better fix > it.
Ah, okay that makes sense. I didn't focus much on the other side of the equation. > The patch is correct. Any value would be a correct initializer; 0 is perfect > since it is available with the minimum of CPU instructions. > > Thanks! Thanks, I pushed the change. Collin
