30 Dec
2007
30 Dec
'07
9:43 p.m.
Jeremy wrote:
I guess it's one of the things gcc 4.x is more picky about.
Yeah, I had to go to 4.x after Jorge reported a bunch of sign warnings in my decoding stuff...
Perhaps the problem is that the variables sel_res and size remain in scope in subsequent cases of the switch, yet if those cases are executed then the variables won't be initialised.
If it can tell in the new version that the variables aren't being used in subsequent cases, then I'd expect it to be able to able to tell in the old version.
What happens if you confine the scope of those variables by putting the body of the case clause inside {...} ? Does the compiler still complain?
That works.