gh-138004: Encode thread names on Solaris-based platforms as ASCII and fix main/worker threading test

Lib/test/test_threading.py

@@ -2241,6 +2241,7 @@ def __init__(self, a, *, b) -> None:
 
         with warnings.catch_warnings(record=True) as warnings_log:
             CustomRLock(1, b=2)
+
         self.assertEqual(warnings_log, [])
 
 class EventTests(lock_tests.EventTests):
@@ -2360,6 +2361,8 @@ def work():
                 thread = threading.Thread(target=work, name=name)
                 thread.start()
                 thread.join()
+                if not name.isascii() and not work_name:
+                    self.skipTest(f"Platform does not support non-ASCII thread names: got empty name for {name!r}")
                 self.assertEqual(work_name, expected,
                                  f"{len(work_name)=} and {len(expected)=}")
 

Misc/ACKS

@@ -483,6 +483,7 @@ Weilin Du
 John DuBois
 Paul Dubois
 Jacques Ducasse
+Jadon Duff
 Andrei Dorian Duma
 Graham Dumpleton
 Quinn Dunkan

Misc/NEWS.d/next/Core_and_Builtins/2025-08-21-06-31-42.gh-issue-138004.FH2Hre.rst

@@ -0,0 +1 @@
+_thread.set_name() now retries with an ASCII fallback if pthread_setname_np() rejects UTF-8 names on some POSIX-compliant platforms.

Modules/_threadmodule.c

@@ -17,6 +17,11 @@
 #ifdef HAVE_SIGNAL_H
 #  include <signal.h>             // SIGINT
 #endif
+#ifdef HAVE_PTHREAD_H
+#  include <pthread.h>
+#endif
+#include <errno.h>
+#include <string.h>
 
 // ThreadError is just an alias to PyExc_RuntimeError
 #define ThreadError PyExc_RuntimeError
@@ -2517,6 +2522,58 @@ of the main interpreter.");
 #   include <pthread_np.h>
 #endif
 
+/* Helper: encode/truncate and call native API to set thread name.
+ * Return:
+ *   0  : success
+ *   >0 : errno-style native error code (e.g. EINVAL)
+ *  -1  : Python-level error (an exception has been set)
+ */
+static int
+set_encoded_thread_name(PyObject *name_obj, const char *encoding)
+{
+    PyObject *name_encoded = PyUnicode_AsEncodedString(name_obj, encoding, "replace");
+    if (name_encoded == NULL) {
+        /* PyUnicode_AsEncodedString set an exception */
+        return -1;
+    }
+
+#ifdef _PYTHREAD_NAME_MAXLEN
+    if (PyBytes_GET_SIZE(name_encoded) > _PYTHREAD_NAME_MAXLEN) {
+        PyObject *truncated = PyBytes_FromStringAndSize(
+            PyBytes_AS_STRING(name_encoded),
+            _PYTHREAD_NAME_MAXLEN);
+        Py_DECREF(name_encoded);
+        if (truncated == NULL) {
+            /* PyBytes_FromStringAndSize set an exception */
+            return -1;
+        }
+        name_encoded = truncated;
+    }
+#endif
+
+    const char *name = PyBytes_AS_STRING(name_encoded);
+    int rc = 0;
+
+#ifdef __APPLE__
+    rc = pthread_setname_np(name);
+#elif defined(__NetBSD__)
+    pthread_t thread = pthread_self();
+    rc = pthread_setname_np(thread, "%s", (void *)name);
+#elif defined(HAVE_PTHREAD_SETNAME_NP)
+    pthread_t thread = pthread_self();
+    rc = pthread_setname_np(thread, name);
+#elif defined(HAVE_PTHREAD_SET_NAME_NP)
+    pthread_t thread = pthread_self();
+    pthread_set_name_np(thread, name);
+    rc = 0; /* that API returns void */
+#else
+    rc = 0; /* no-op if platform unsupported */
+#endif
+
+    Py_DECREF(name_encoded);
+    return rc;
+}
+
 #if defined(HAVE_PTHREAD_GETNAME_NP) || defined(HAVE_PTHREAD_GET_NAME_NP) || defined(MS_WINDOWS)
 /*[clinic input]
 _thread._get_name
@@ -2590,42 +2647,26 @@ _thread_set_name_impl(PyObject *module, PyObject *name_obj)
     PyInterpreterState *interp = _PyInterpreterState_GET();
     const char *encoding = interp->unicode.fs_codec.encoding;
 #endif
-    PyObject *name_encoded;
-    name_encoded = PyUnicode_AsEncodedString(name_obj, encoding, "replace");
-    if (name_encoded == NULL) {
-        return NULL;
-    }
 
-#ifdef _PYTHREAD_NAME_MAXLEN
-    // Truncate to _PYTHREAD_NAME_MAXLEN bytes + the NUL byte if needed
-    if (PyBytes_GET_SIZE(name_encoded) > _PYTHREAD_NAME_MAXLEN) {
-        PyObject *truncated;
-        truncated = PyBytes_FromStringAndSize(PyBytes_AS_STRING(name_encoded),
-                                              _PYTHREAD_NAME_MAXLEN);
-        if (truncated == NULL) {
-            Py_DECREF(name_encoded);
-            return NULL;
-        }
-        Py_SETREF(name_encoded, truncated);
+    int rc = set_encoded_thread_name(name_obj, encoding);
+    /* Confirm a Python exception was set by the helper.
+    If not, convert to a runtime error (defensive). */
+    if (rc == -1 && PyErr_Occurred()) {
+        return NULL;
     }
-#endif
 
-    const char *name = PyBytes_AS_STRING(name_encoded);
-#ifdef __APPLE__
-    int rc = pthread_setname_np(name);
-#elif defined(__NetBSD__)
-    pthread_t thread = pthread_self();
-    int rc = pthread_setname_np(thread, "%s", (void *)name);
-#elif defined(HAVE_PTHREAD_SETNAME_NP)
-    pthread_t thread = pthread_self();
-    int rc = pthread_setname_np(thread, name);
-#else /* defined(HAVE_PTHREAD_SET_NAME_NP) */
-    pthread_t thread = pthread_self();
-    int rc = 0; /* pthread_set_name_np() returns void */
-    pthread_set_name_np(thread, name);
-#endif
-    Py_DECREF(name_encoded);
     if (rc) {
+        /* If native API refused (EINVAL) and we didn't try ASCII, retry with ASCII. */
+        if (rc == EINVAL && strcmp(encoding, "ascii") != 0) {
+            rc = set_encoded_thread_name(name_obj, "ascii");
+            if (rc == -1 && PyErr_Occurred()) {
+                return NULL;
+            }
+            if (rc == 0) {
+                Py_RETURN_NONE;
+            }
+            /* fall through to raise errno below */
+        }
         errno = rc;
         return PyErr_SetFromErrno(PyExc_OSError);
     }