On Wed, Oct 05, 2016 at 16:07:09 +0200, Martin Kletzander wrote: > On Wed, Oct 05, 2016 at 09:26:43AM +0200, Jiri Denemark wrote: > >@@ -2648,7 +2647,7 @@ vshReadlineParse(const char *text, int state) > > static vshCommandParser parser, sanitizer; > > vshCommandToken tk; > > static const vshCmdDef *cmd; > >- const vshCmdOptDef *opt; > >+ const vshCmdOptDef *opt = NULL; > > Always a good thing to do, you can't broke something that worked by > initialization. And it's nicer. Even though Peter^Wsome people would > argue that it's pointless, especially for ancient and weird compilers > (i.e. even slightly different compiler version that they're running or > with one different setting). But looking at the function it's easy to > dereference that opt lines below in that function. Same way cmd_exists > can be used without initialization and I believe half of those hundred > variables too. Well, about half of the variables in there are *static*, which means they are initialized. Sure, there are legit uses of static variables in functions and it may even be this case, I don't know, but I know having so many of them in a single function is just not OK. > Basically, I'm trying to say "ACK, you were right, I'm running away too > now!", which translates to "ACK" in tl;dr language. Thanks, I pushed the patch. Jirka -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list