On Mon, Jun 17, 2013 at 1:58 PM, Stephen Kelly <[email protected]> wrote: > Jordan Rose wrote: > >> >> On Jun 16, 2013, at 17:01 , Richard Smith >> <[email protected]> wrote: >> >>> +// FIXME: Why are DeprecatedImplementations and DeprecatedWritableStr >>> +// not in this group? >>> +def Deprecated : DiagGroup<"deprecated", [DeprecatedDeclarations, >>> + DeprecatedRegister]>, >> >> As I understand it, because -Wdeprecated is for user deprecations (i.e. >> __attribute__((deprecated))), not language deprecations. Unless it's for >> GCC compatibility, I actually think it should stay that way, and that you >> should create a new flag for all these new things we're adding (starting >> with warn_increment_bool). > > I made a similar point in the bug report, but it was never addressed: > > http://llvm.org/bugs/show_bug.cgi?id=16336#c1 > > What does GCC do here?
GCC uses -Wdeprecated-declarations for __attribute__((deprecated)), and uses -Wdeprecated for deprecated language features (or at least for access declarations). It does not treat -Wdeprecated-declarations as a subgroup of -Wdeprecated. How about: we move the three separate -Wdeprecated-* warnings into -Wdeprecated, move the two ??? warnings into -Wdeprecated, rename -Wdeprecated-declarations to -Wdeprecated-attribute (with a compatibility synonym to -Wdeprecated-declarations), and remove it from the -Wdeprecated group. _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
