Repository navigation
Claude / Claude Code Review
completed
Sep 29, 2026 in 24m 52s
Code review found 1 new issue (1 earlier finding still open)
Found 2 candidates, confirmed 2: 1 posted in this review, 1 from earlier reviews still open. See review comments for details.
Details
| Severity | Count |
|---|---|
| 🔴 Important | 2 |
| 🟡 Nit | 0 |
| 🟣 Pre-existing | 0 |
| Severity | File:Line | Issue |
|---|---|---|
| 🔴 Important | Source/WTF/wtf/threads/Signals.cpp:529 |
On Linux the fix in this push works or not per how the thread registered its sigaltstack, not per where the fault happen |
Annotations
Check failure on line 529 in Source/WTF/wtf/threads/Signals.cpp
claude / Claude Code Review
On Linux the fix in this push works or not per how the thread registered its sigaltstack, not per where the fault happened. Since kernel 4.7 uc_stack.ss_flags holds the flags passed to sigaltstack, so the test at Signals.cpp:529 is a per-thread constant.
On Linux the fix in this push works or not per how the thread registered its sigaltstack, not per where the fault happened. Since kernel 4.7 uc_stack.ss_flags holds the flags passed to sigaltstack, so the test at Signals.cpp:529 is a per-thread constant. Registered with SS_ONSTACK: registers are never published and the profiler hang remains. Registered with 0 or SS_AUTODISARM: a handler nested in another alternate-stack handler publishes registers whose sp is on that stack, which the base backed
Loading