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
Commit a0af3bf
Browse filesBrowse the repository at this point in the historyBrowse files
Issue #358 — out-of-order get_output entries after restart_debugging:
- Each launch's output handlers are closure-bound to their own
OutputRingBuffer instance, so an in-flight event from the previous run
(microtask/IPC queue) can never land in the new launch's buffer and take
an early seq. noteUnboundFunctionBreakpoints writes to its launch's
buffer too. No sort added to read(): seq is monotonic per buffer by
construction.
Issue #361 — CodeLLDB DWARF-parser stderr noise in get_output:
- New optional AdapterPolicy.shouldSuppressOutputEvent(category, text)
hook, consulted by the session manager's output capture after the
telemetry drop; suppressed lines stay in debug logs, and policy failures
never drop output.
- lldbShouldSuppressOutputEvent in lldb-policy-shared matches LLDB's DWARF
diagnostics (DIE/DW_AT_ranges, DW_TAG_member bounds, wrapped
continuation fragments) anchored on DWARF-specific tokens; multi-line
events are dropped only when EVERY non-empty line matches. Wired into
the rust and cpp policies.
Issue #366 — Docker JS logpoint/program output lost at run-to-completion:
- js-debug delivers debuggee output as CHILD session events serialized
through the CDP-bridge chain; terminated could outrun them and the
SessionManager teardown dropped whatever was still queued (reliably all
output under Docker's slower scheduling).
- New ChildSessionManager.flushEvents() / MinimalDapClient.flushChildEvents();
the proxy worker's terminal slots (exited/terminated/connection-closed)
drain the chain (two rounds with a 150ms settle for bytes still unread
in the child socket, 2s backstop) before forwarding — output events hit
the FIFO IPC ahead of the terminal signal. Only paid by child-session
adapters.
- Docker e2e now runs the exact issue repro: logpoint + run-to-completion
must capture program stdout and the interpolated LOGPOINT line.
Fixes#358Fixes#361Fixes#366
Co-authored-by: CI Bot <ci@example.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: JF <john.franklin@gmail.com>
Copy file name to clipboardExpand all lines: docs/tool-reference.md
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -960,7 +960,7 @@ Gets the debuggee's output (stdout/stderr/console) captured for a session. Outpu
960
960
961
961
**Notes:**
962
962
- The buffer holds the last 1000 entries per launch; older entries are evicted and counted in `dropped`. Individual entries longer than 8192 characters are cut and flagged `"truncated": true`.
963
-
- Adapter-internal `telemetry` events are filtered out at capture time; all other categories (`stdout`, `stderr`, `console`, `important`, ...) are kept. Adapters that omit a category default to `console`.
963
+
- Adapter-internal `telemetry` events are filtered out at capture time, as are output events an adapter policy declares to be pure adapter noise — e.g. LLDB's harmless DWARF-parser error spew on MinGW-built rust/cpp binaries (issue #361); suppressed lines remain visible in debug logs. All other categories (`stdout`, `stderr`, `console`, `important`, ...) are kept. Adapters that omit a category default to `console`.
964
964
- Works while the program is running and after it finishes — output stays readable until `close_debug_session`. Re-launching a session starts a fresh buffer (seq restarts at 1).
965
965
-`hasMore: true` means more entries matched than `limit` allowed; call again with `since: nextSince`.
966
966
- Incremental polling recipe: call once, remember `nextSince`, and pass it as `since` on the next call — you'll only ever see new output.
0 commit comments