> Date: Thu, 19 Jan 2006 00:01:13 +0000 > From: "J. Grant" <[EMAIL PROTECTED]> > CC: [email protected] > > > I've looked at the warnings, and while a few of them need to be fixed, > > the absolute majority is just compiler whining about perfectly valid > > code. How many of these warnings go away if you reset the warning > > level back to /W3, the way it was before it was bumped up to /W4 in > > beta4 (for the reasons I cannot understand)? > > It's been set on /W4 for quite a while (since 2005-07-03), I think we > should keep it on this.
But if all it does is spill useless warnings on us, why keep it? > > For that matter, why not disable some of the more annoying false > > warnings in config.h.W32? For example, "FOO uses old-style > > declarator" or "logical operation on address of string constant". > > You're right, old-style declarator should be off, not sure about the > latter one though, it is because of the streq macro which expands to > code which compares the address, does it need to do this as well as the > strcmp ? (I hope we can change all of these macro's to functions as part of > the ISO C clean-up work in 3.82...) Actually, I think streq and strneq should simply go away: they try to optimize where no optimization is needed. But that doesn't change the fact that this warning is useless, as comparing string addresses is perfectly valid _______________________________________________ Make-w32 mailing list [email protected] http://lists.gnu.org/mailman/listinfo/make-w32
