[PATCH] D159138: [clang][Sema] Fix format size estimator's handling of %o, %x, %X with alternative form

[Nick Desaulniers via Phabricator via cfe-commits]

nickdesaulniers added inline comments.


================
Comment at: clang/test/Sema/warn-fortify-source.c:100-102
+  __builtin_snprintf(buf, 2, "%#x", n);
+  __builtin_snprintf(buf, 2, "%#X", n);
+  __builtin_snprintf(buf, 2, "%#o", n);
----------------
Note that GCC -Wformat-truncation can warn on some of these.

https://godbolt.org/z/jE3axWe1W

Looks like the diagnostic keeps an up and lower bounds on the estimated format string expansion.

Trunk for Clang also warns for these, so is this change a regression? Or are both GCC and Clang (trunk) incorrect?


================
Comment at: clang/test/Sema/warn-fortify-source.c:121-123
+  __builtin_vsnprintf(buf, 2, "%#x", list);
+  __builtin_vsnprintf(buf, 2, "%#X", list);
+  __builtin_vsnprintf(buf, 2, "%#o", list);
----------------
Clang @ trunk and GCC both warn for these 3. Regression?
https://godbolt.org/z/j551hTE5E


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D159138/new/

https://reviews.llvm.org/D159138