On Tue, Oct 24, 2006 at 05:51:09PM +0200, Claus Fischer wrote: > Dan, thanks for replying so quickly. > > The bug also happens in 'real' code. > > This is only a very reduced showcase code for the bug, > the bug also occurs in some (pretty big) code of mine > where > if (0) > is actually > if (!some_function(....)) > i.e. if the function failed, jump to the failure label. > > > I believe the -O0 -g does not let gcc optimize the > if (0) away; however, if it does, you'd have to place > some dummy function in the if.
I can only examine testcases that are submitted, not vaguely described real code. Your belief in the last paragraph is obviously incorrect; please take a moment to compile your posted testcase and see what you get. I can reproduce the problem with a more complex condition too. In that case, it's at best a debugging information problem. The compiler has eliminated the goto by transforming the if statement into a conditional jump to the goto target. When there is no code on a line, GDB uses the next line with code. Try "info line 6" for the line with the goto, to see what has happened. -- Daniel Jacobowitz CodeSourcery -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
