xingxue marked an inline comment as done. xingxue added inline comments.
================ Comment at: libcxx/include/stdint.h:16 +#endif // _STD_TYPES_T /* ---------------- hubert.reinterpretcast wrote: > mclow.lists wrote: > > I don't think that this will do what you want it to. > > Is this a supported use case? > > > > ``` > > #include <stdint.h> > > #define _STD_TYPES_T > > #include <stdint.h> > > ``` > > > > > The comment is perhaps not clear. We need `<stdint.h>` to be re-enterable > only as long as all previous inclusions have occurred while `_STD_TYPES_T` is > defined. The effect of including `<stdint.h>` without defining `_STD_TYPES_T` > is a strict superset of the effects of including `<stdint.h>` in other > circumstances. Should we adjust the comment? Updated comments as suggested. This patch helps scenarios such as the following on `AIX`. #include <sys/types.h> #include <stdint.h> where `<sys/types.h>` included in the first line defines macro `_STD_TYPES_T`, includes `<stdint.h>`, and then undefines `_STD_TYPES_T`. `<stdint.h>` included in the second line can be entered to get to macros such as `UINT32_MAX`. Since `_STD_TYPES_T` is undefined with the second inclusion, the header guard macro is defined and `<stdint.h>` is no longer re-enterable. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59253/new/ https://reviews.llvm.org/D59253 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
