[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
I think you put the wrong priority on the wrong thing.
-Werror should not be an issue. The only reason it is is because gcc
warnings are broken. There are so many of them, there's no sensible
way to choose them in a way that makes sense. They're not upward
If you really want to do something useful concerning -Werror, I say you
start putting pressure on the FSF to release gcc versions with sensible
warnings. Maybe actually number them so that you can create a list of
warnings you actually want, instead of relying on -Wall -W to do something
I mean, for such an old program as gcc, the error subsystem is *still*
experimental: Warning messages mix up false positives, true deadly errors,
and surprising changes from release to release to boot.
I'm not kidding. I'm definitely not a fan of -Werror. It usually leads to
people silencing the warnings by putting deadly casts in the code, and adding
errors, instead of fixing the problems.
The type error warnings are atrocious, the portability warnings are fairly
bad, and the comparison stuff is broken for the most part.
... and if you rely on debian fixes, sooner or later, you're going to run
into the openssl random generation advisory, or worse.
Main Index |
Thread Index |