Hi, Wolfgang Lux wrote:
Aha, I suspected this. It was me that introduced the "test" in the macro because if (_x) wa snil I would get an exception. I created two versions: one to be used with constants (known to be non-nil) and a generic one.Riccardo Mottola wrote:do you see anything wrong with this macro? #define S2AS(_x) ( (_x) ? [[[NSAttributedString alloc] initWithString: (_x)] autorelease] : nil ) It is then called by passing a string like @"hi". I added the ternary operator to return nil if nil is passed. When used in TalkSoup, I get these two errors: commands.m:368: warning: the address of '_OBJC_INSTANCE_30' will always evaluate as 'true' commands.m: In function '-[TalkSoup(Commands) commandClientinfo:connection:]':If you look carefully you'll notice that this is a warning and not an error. And the compiler is of course right. The condition (_x) expands to (@"hi") and that condition always evaluates to true. So the compiler warns you that in this particular case the else branch of the conditional can never be reached and thus is redundant (and might never be reached by coverage tests).
Perhaps. It was easy to cast it in the macro and silence it, however I have seen popup elswehere in the code and I'll leave it there.main.m: In function '-[NetclassesConnection changeNick:onConnection:withNickname:sender:]': main.m:418: warning: pointer type mismatch in conditional expressionAnd this is also a warning and not an error. It's due to a dumb compiler that expects you to use then and else branches in a conditional expression that have the same type. This is of course the compiler's fault. The compiler should not complain if the expression in one branch is a subtype of the expression in the other branch. I think this was fixed in either gcc version 4.6 or 4.7.
Thank you, Roccardo
<<attachment: riccardo_mottola.vcf>>
_______________________________________________ Discuss-gnustep mailing list [email protected] https://lists.gnu.org/mailman/listinfo/discuss-gnustep
