https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92061
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jakub at gcc dot gnu.org
--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
We can certainly workaround it, like:
--- gcc/genconditions.c 2019-01-01 12:37:19.064943662 +0100
+++ gcc/genconditions.c 2019-10-11 10:57:11.464595789 +0200
@@ -57,8 +57,9 @@ write_header (void)
\n\
/* It is necessary, but not entirely safe, to include the headers below\n\
in a generator program. As a defensive measure, don't do so when the\n\
- table isn't going to have anything in it. */\n\
-#if GCC_VERSION >= 3001\n\
+ table isn't going to have anything in it.\n\
+ Clang 9 is buggy and doesn't handle __builtin_constant_p correctly. */\n\
+#if GCC_VERSION >= 3001 && __clang_major__ < 9\n\
\n\
/* Do not allow checking to confuse the issue. */\n\
#undef CHECKING_P\n\
@@ -170,7 +171,7 @@ struct c_test\n\
vary at run time. It works in 3.0.1 and later; 3.0 only when not\n\
optimizing. */\n\
\n\
-#if GCC_VERSION >= 3001\n\
+#if GCC_VERSION >= 3001 && __clang_major__ < 9\n\
static const struct c_test insn_conditions[] = {\n");
traverse_c_tests (write_one_condition, 0);
@@ -191,7 +192,7 @@ write_writer (void)
" unsigned int i;\n"
" const char *p;\n"
" puts (\"(define_conditions [\");\n"
- "#if GCC_VERSION >= 3001\n"
+ "#if GCC_VERSION >= 3001 && __clang_major__ < 9\n"
" for (i = 0; i < ARRAY_SIZE (insn_conditions); i++)\n"
" {\n"
" printf (\" (%d \\\"\", insn_conditions[i].value);\n"
but guess it would be better if somebody reports that to LLVM bugzilla first
and see if they are willing to fix.
