You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Regression: AOT DateTime::add/sub with DateInterval is a silent no-op for format/getTimestamp (re-#33781/#30760, ext/date/php_date.c) #33911
After #33906 / #32691 stamp publishing, AOT DateTime::add/sub with a literal DateInterval updates the LLVM __dt_timestamp slot (or should) but leaves compileTimeDateTimeTimestamp / dateTimeLocalInstants on the receiver stale. DateTime::getTimestamp() and format() prefer the compile-time stamp, so the mutation is a silent no-op vs Zend.
ext/standard/JitDateMutation.php — after mutable add/sub compile-time or LLVM apply, refresh or invalidatecompileTimeDateTimeTimestamp + dateTimeLocalInstants so subsequent format/getTimestamp see the new instant (peer #33906 stamp publish)
Activity
PurHur commentedon Aug 22, 2026
claim: implementer-worker-lane-c — starting this run (lane C: 33911%3==2; north-star #32122/#31968 closed; AOT DateTime::add/sub silent no-op vs Zend after stamp publish).
AOT: refresh DateTime stamps after add/sub so format matches Zend (#3…
AOT: fluent DateTime::modify chain must not drop later mutations (#33935