fix: spawn-test stderr persistence and drain-cap deadlock (PR #11 review)

Two bugs caught in supervisor review:

1. Stderr was silently dropped when the Details panel was closed during a
   crash. _on_spawn_done appended to diag_text only when the panel was
   already open, and refresh_dependency() clobbered that text on the next
   field change anyway.
   Fix: stash the result in self._last_spawn; _full_diag_text() appends
   the stderr section whenever diag text is generated; _on_spawn_done
   auto-opens the panel on non-ok outcomes (same pattern as the existing
   auto_open for missing commands).

2. _drain stopped reading once _STDERR_CAP (4 KB) was reached. A process
   that writes more than 4 KB then blocked on a full pipe buffer, never
   exited, and was misclassified as "ok" instead of "crashed".
   Fix: drain to EOF unconditionally; keep only the first _STDERR_CAP bytes.
   Regression test: 64 KB stderr + exit(3) → outcome "crashed".

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

bcc.py

@@ -321,6 +321,7 @@ class ServerEditor(QFrame):
         self._before_change = before_change  # called before any row add/remove in tables
         self._extra = {}  # preserve unknown server fields verbatim
         self._loading = False
+        self._last_spawn: dict | None = None  # most recent spawn_test result
 
         outer = QVBoxLayout(self)
         outer.setContentsMargins(16, 16, 16, 16)
@@ -488,6 +489,7 @@ class ServerEditor(QFrame):
             self.headers.load({})
             self.details_btn.setChecked(False)
             self.diag_text.clear()
+            self._last_spawn = None
             self._set_dep({"status": "unknown", "label": "—"})
             self.args_warn.hide()
             self.args_fix_btn.hide()
@@ -610,7 +612,7 @@ class ServerEditor(QFrame):
         if auto_open and problem and not self.details_btn.isChecked():
             self.details_btn.setChecked(True)  # opens panel (fills text via _toggle_diag)
         if self.diag_card.isVisible():
-            self.diag_text.setPlainText(core.diagnostics_text(self.current_name(), data))
+            self.diag_text.setPlainText(self._full_diag_text())
 
     def _set_dep(self, res: dict):
         status = res.get("status", "unknown")
@@ -620,13 +622,17 @@ class ServerEditor(QFrame):
         color = STATUS_COLORS.get(status, MUTED) if status in ("missing", "warn") else MUTED
         self.dep_label.setStyleSheet(f"color: {color};")
 
+    def _full_diag_text(self) -> str:
+        text = core.diagnostics_text(self.current_name(), self.dump_data())
+        if self._last_spawn and self._last_spawn.get("stderr", "").strip():
+            text += "\n\n── stderr from last Test launch ──\n" + self._last_spawn["stderr"].strip()
+        return text
+
     def _toggle_diag(self, on):
         self.diag_card.setVisible(on)
         self.details_btn.setText("Details ▾" if on else "Details ▸")
         if on and self.isEnabled():
-            self.diag_text.setPlainText(
-                core.diagnostics_text(self.current_name(), self.dump_data())
-            )
+            self.diag_text.setPlainText(self._full_diag_text())
 
     def _recheck(self):
         core.refresh_path_cache()  # re-scan PATH, e.g. after installing a runtime
@@ -684,9 +690,9 @@ class ServerEditor(QFrame):
     def _on_spawn_done(self, result: dict):
         self.spawn_btn.setEnabled(True)
         self.spawn_btn.setText("Test launch")
+        self._last_spawn = result
         outcome = result.get("outcome", "")
         detail = result.get("detail", "")
-        stderr = result.get("stderr", "").strip()
         if outcome == "ok":
             color = GOOD
             label = f"started · {detail}"
@@ -700,9 +706,11 @@ class ServerEditor(QFrame):
         self.dep_dot.setStyleSheet(f"color: {color}; font-size: 14px;")
         self.dep_label.setText(label)
         self.dep_label.setStyleSheet(f"color: {color};")
-        if stderr and self.diag_card.isVisible():
-            current = self.diag_text.toPlainText()
-            self.diag_text.setPlainText(current + f"\n\n── stderr from Test launch ──\n{stderr}")
+        # Auto-open the details panel on non-ok outcomes so stderr is visible.
+        if outcome != "ok" and not self.details_btn.isChecked():
+            self.details_btn.setChecked(True)  # triggers _toggle_diag → fills diag_text
+        elif self.diag_card.isVisible():
+            self.diag_text.setPlainText(self._full_diag_text())
 
 
 # --------------------------------------------------------------------------- #

bcc_core.py

@@ -1114,12 +1114,18 @@ def spawn_test(data: dict, timeout: float = 3.0) -> dict:
     stderr_chunks: list[bytes] = []
 
     def _drain(pipe) -> None:
+        # Read until EOF so the pipe buffer never fills and blocks the subprocess.
+        # Only keep the first _STDERR_CAP bytes; the rest is discarded.
+        captured = 0
         try:
-            while sum(len(c) for c in stderr_chunks) < _STDERR_CAP:
+            while True:
                 chunk = pipe.read(1024)
                 if not chunk:
                     break
-                stderr_chunks.append(chunk)
+                if captured < _STDERR_CAP:
+                    keep = min(len(chunk), _STDERR_CAP - captured)
+                    stderr_chunks.append(chunk[:keep])
+                    captured += keep
         except Exception:
             pass
 

tests/test_core.py

@@ -303,3 +303,18 @@ def test_spawn_test_stdin_not_closed():
         timeout=0.3,
     )
     assert r["outcome"] == "ok", f"expected ok (blocking on stdin), got {r['outcome']}"
+
+
+def test_spawn_test_large_stderr_does_not_deadlock():
+    # A process that writes >4 KB (pipe buffer territory) to stderr then exits non-zero.
+    # Without EOF-draining, the subprocess blocks on a full pipe and is wrongly
+    # classified as "ok" (still running) instead of "crashed".
+    script = "import sys; sys.stderr.write('x' * 65536); sys.stderr.flush(); sys.exit(3)"
+    r = c.spawn_test(
+        {"command": sys.executable, "args": ["-c", script]},
+        timeout=2.0,
+    )
+    assert r["outcome"] == "crashed", f"expected crashed, got {r['outcome']}"
+    assert r["returncode"] == 3
+    # stderr is capped, not the full 64 KB
+    assert len(r["stderr"]) <= c._STDERR_CAP