* Cristian Rodríguez [2008-11-17 14:00]:
Richard Bos escribió:
The error results in a failed build and no packages :(
Sure it is better to produce no packages than packages that may be broken, in this case function may return random data..
I guess fatal() just calls exit(). The checks are not 100 % true and *do* produce false positives. That's one reason why that's a compiler *warning* and not a compiler *error*. Sure, there are gcc specific macros to declare the function to not return, but that does not work if you have error(FATAL, "") when the error() function returns only when FATAL is specified. "Fixing" errors that are no errors to build packages is not fun, and that's why I also vote to make that a rpmlint check that can be ignored. Yes, I had the case I'm now talking about (error(FATAL)) recently in the crash utility, it's not "theoretic". Regards, Bernhard -- To unsubscribe, e-mail: opensuse-buildservice+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-buildservice+help@opensuse.org