That's good to know! I checked that the problem reproduces on debian wheezy (debian's current stable, a bit less than a year old), but it is fixed on debian sid.
So applications could work around this by doing "mkdir -p hack/linux; touch hack/linux/stddef.h" and then adding "-Ihack" to all their commandlines until this is fixed. Since this is only needed for a limited time (2 years-ish), that seems preferable to me over adding hacks to clang's headers. Any other opinions? If not, I'll drop this patch. In any case, it's probably a good idea to change stddef.h so that the already-present __need_wint_t behaves like in gcc's version: if that's defined, stddef.h only provides win_t and does nothing else. Should I send a patch for just that? On Tue, Apr 29, 2014 at 7:35 AM, Lubos Lunak <l.lu...@centrum.cz> wrote: > On Monday 28 of April 2014, Reid Kleckner wrote: > > Even if we commit this workaround, can we report this as a bug to > upstream > > Linux? > > Already done (http://comments.gmane.org/gmane.linux.kernel/1281756). And > there's actually no public #define NULL in recent Linux headers. > > > The test case you have shows how this definition of NULL is broken > > on x86_64, even in C with gcc: > > > > $ cat t.c > > #define NULL 0 > > int printf(const char *, ...); > > int main() { > > printf("%d %d %d %d %d %d %p\n", 1, 2, 3, 4, 5, 6, > > 0xdeadbeefdeadbeefULL); printf("%d %d %d %d %d %d %p\n", 1, 2, 3, 4, 5, > 6, > > NULL); > > } > > > > $ gcc -w t.c -o t && ./t > > 1 2 3 4 5 6 0xdeadbeefdeadbeef > > 1 2 3 4 5 6 0xdeadbeef00000000 > > -- > Lubos Lunak >
_______________________________________________ cfe-commits mailing list cfe-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
