diff options
| author | Takashi Kokubun <takashikkbn@gmail.com> | 2026-05-26 14:50:45 -0700 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2026-05-26 14:50:45 -0700 |
| commit | dcfba5725794815fa36e129457f9d748f24371f9 (patch) | |
| tree | 5d0c3f0eebef7e34775e56ebf27bc2890c675740 /internal | |
| parent | ce7f3ed832a50e493e3b6cb592791aa1fe784d85 (diff) | |
After 4bb1977b78 switched RCLASS_PRIME_CLASSEXT_WRITABLE_P to FL_TEST_RAW,
GCC lost the RB_FL_ABLE-derived hint that klass is not a special const,
and started warning that RCLASS_EXT_PRIME(klass) (used by the inlined
RCLASS_WRITE_SUPER in class_associate_super) could yield a near-zero
pointer. Make the existing VM_ASSERT(klass != 0) precondition visible
to the optimizer via RBIMPL_ASSUME so the warning goes away without
re-introducing the FL_TEST overhead.
Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
Diffstat (limited to 'internal')
| -rw-r--r-- | internal/class.h | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/internal/class.h b/internal/class.h index 4366223f84..0773dbad95 100644 --- a/internal/class.h +++ b/internal/class.h @@ -297,6 +297,7 @@ RCLASS_PRIME_CLASSEXT_WRITABLE_P(VALUE klass) { VM_ASSERT(klass != 0, "klass should be a valid object"); VM_ASSERT_BOXABLE_TYPE(klass); + RBIMPL_ASSUME(klass != 0); return FL_TEST_RAW(klass, RCLASS_PRIME_CLASSEXT_WRITABLE); } |