)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"58628b294d5732e4a3ed9ee0c7178332a475f5ea","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"6875c93b_bd41d5dd","updated":"2026-10-29 15:43:16.000000000","message":"Adding -1 to highlight my comments.","commit_id":"9533671e6a804660acd194bd65ffc0bba7bd322b"},{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"427d0905b0498cf76427b7106bc3bf51582cae4b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"6ab6dca4_f9cff95b","updated":"2026-10-30 05:39:16.000000000","message":"Thanks. The change looks almost good to me. Let me as one question about the test code.","commit_id":"96f4a254592617ab71bba27273d109f7c799ee12"},{"author":{"_account_id":28522,"name":"Hervé Beraud","email":"herveberaud.pro@gmail.com","username":"hberaud"},"change_message_id":"6d35c19b05172eb08274324c830d199984523a5f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":4,"id":"8fbf3e9a_439dd38e","updated":"2026-11-04 10:18:53.000000000","message":"LGTM. Would like to have Daniel\u0027s review too.","commit_id":"3d6719508b3fbaa20fa3d832e34cbbccbcc3ca2d"}],"oslo_service/backend/_threading/service.py":[{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"6874cc87437976bd38e30a422f63f3522aafbaff","unresolved":true,"context_lines":[{"line_number":482,"context_line":"        # that call returns."},{"line_number":483,"context_line":"        self._reload_pending \u003d True"},{"line_number":484,"context_line":"        if self._waiting:"},{"line_number":485,"context_line":"            self._waiting \u003d False"},{"line_number":486,"context_line":"            raise common_service.SignalExit(signal.SIGHUP)"},{"line_number":487,"context_line":""},{"line_number":488,"context_line":"    def _wait_no_fork(self):"}],"source_content_type":"text/x-python","patch_set":1,"id":"e11c425a_b2f37c55","line":485,"range":{"start_line":485,"start_character":12,"end_line":485,"end_character":33},"updated":"2026-10-29 15:26:26.000000000","message":"I think this isn\u0027t needed here ? The flag is updated by _wait_no_fork and _reload_service only reads it.","commit_id":"9533671e6a804660acd194bd65ffc0bba7bd322b"},{"author":{"_account_id":10273,"name":"Adam Harwell","email":"flux.adam@gmail.com","username":"rm_you"},"change_message_id":"8d08440737cf554d406e72bf2b97437c4dcca854","unresolved":false,"context_lines":[{"line_number":482,"context_line":"        # that call returns."},{"line_number":483,"context_line":"        self._reload_pending \u003d True"},{"line_number":484,"context_line":"        if self._waiting:"},{"line_number":485,"context_line":"            self._waiting \u003d False"},{"line_number":486,"context_line":"            raise common_service.SignalExit(signal.SIGHUP)"},{"line_number":487,"context_line":""},{"line_number":488,"context_line":"    def _wait_no_fork(self):"}],"source_content_type":"text/x-python","patch_set":1,"id":"538d042a_6cc3d4e2","line":485,"range":{"start_line":485,"start_character":12,"end_line":485,"end_character":33},"in_reply_to":"43c013b2_e9d20d45","updated":"2026-10-30 03:41:33.000000000","message":"That was bugging me a bit, so I looked again, and I think you were right. A SIGHUP could also land in the finally, before it cleared the flag, on the path where a reload was already pending. Clearing the flag here did not cover that case.\n\nI believe nesting the try/finally inside the try/except will fix it, so a SIGHUP anywhere in that block is caught. This line is then not needed.","commit_id":"9533671e6a804660acd194bd65ffc0bba7bd322b"},{"author":{"_account_id":10273,"name":"Adam Harwell","email":"flux.adam@gmail.com","username":"rm_you"},"change_message_id":"2ab1d74c607f39b50370232ae74c9cd6bcec37bf","unresolved":false,"context_lines":[{"line_number":482,"context_line":"        # that call returns."},{"line_number":483,"context_line":"        self._reload_pending \u003d True"},{"line_number":484,"context_line":"        if self._waiting:"},{"line_number":485,"context_line":"            self._waiting \u003d False"},{"line_number":486,"context_line":"            raise common_service.SignalExit(signal.SIGHUP)"},{"line_number":487,"context_line":""},{"line_number":488,"context_line":"    def _wait_no_fork(self):"}],"source_content_type":"text/x-python","patch_set":1,"id":"43c013b2_e9d20d45","line":485,"range":{"start_line":485,"start_character":12,"end_line":485,"end_character":33},"in_reply_to":"e11c425a_b2f37c55","updated":"2026-10-30 03:04:32.000000000","message":"It is needed for a second SIGHUP that comes while the first one\u0027s SignalExit is still unwinding, before the finally in _wait_no_fork() resets the flag.\n\n- Without this line, the second handler still sees _waiting as True and raises again. That happens outside the try block, so the SignalExit leaves _wait_no_fork(), and the process exits without stopping the service.\n- With it, the second SIGHUP only sets _reload_pending, and the loop runs one reload for both.\n\nI checked this with a scratch test that sends the second SIGHUP from the except clause. Added a comment here to explain it.","commit_id":"9533671e6a804660acd194bd65ffc0bba7bd322b"}],"oslo_service/tests/backend/threading/test_signal_handler.py":[{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"427d0905b0498cf76427b7106bc3bf51582cae4b","unresolved":true,"context_lines":[{"line_number":304,"context_line":"        def reset():"},{"line_number":305,"context_line":"            resets.append(True)"},{"line_number":306,"context_line":"            if len(resets) \u003d\u003d 1:"},{"line_number":307,"context_line":"                for _ in range(100):"},{"line_number":308,"context_line":"                    launcher.signal_handler._handle_signal_cb("},{"line_number":309,"context_line":"                        signal.SIGHUP, None"},{"line_number":310,"context_line":"                    )"},{"line_number":311,"context_line":""},{"line_number":312,"context_line":"        with ("},{"line_number":313,"context_line":"            mock.patch.object(self.conf, \u0027mutate_config_files\u0027),"}],"source_content_type":"text/x-python","patch_set":3,"id":"c12c43e5_2402fc7a","line":310,"range":{"start_line":307,"start_character":0,"end_line":310,"end_character":21},"updated":"2026-10-30 05:39:16.000000000","message":"Could you help me understand why we need this retry here ?","commit_id":"96f4a254592617ab71bba27273d109f7c799ee12"},{"author":{"_account_id":10273,"name":"Adam Harwell","email":"flux.adam@gmail.com","username":"rm_you"},"change_message_id":"3b3962a8839633e8a8bd139627f409b92d4e4afb","unresolved":false,"context_lines":[{"line_number":307,"context_line":"                for _ in range(100):"},{"line_number":308,"context_line":"                    launcher.signal_handler._handle_signal_cb("},{"line_number":309,"context_line":"                        signal.SIGHUP, None"},{"line_number":310,"context_line":"                    )"},{"line_number":311,"context_line":""},{"line_number":312,"context_line":"        with ("},{"line_number":313,"context_line":"            mock.patch.object(self.conf, \u0027mutate_config_files\u0027),"}],"source_content_type":"text/x-python","patch_set":3,"id":"d639e40e_b2de73cc","line":310,"in_reply_to":"c12c43e5_2402fc7a","updated":"2026-10-30 07:28:26.000000000","message":"It isn\u0027t a retry. The loop sends several SIGHUPs during the first restart. The test checks that they give one more reload, not one reload each.\n\nIt needs at least two. With one SIGHUP, the test also passes on code that reloads once per SIGHUP. 100 was more than it needs.\n\nPS4 sends 3 and adds a comment that says this.","commit_id":"96f4a254592617ab71bba27273d109f7c799ee12"}],"releasenotes/notes/threading-no-fork-stop-and-sighup-9431712ba137a07f.yaml":[{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"ae5f902ffad172456780ea565d2ae8e834cff3f7","unresolved":true,"context_lines":[{"line_number":5,"context_line":"    longer hangs at exit in these cases:"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"    - The service\u0027s ``stop()`` raises during a shutdown. The process now"},{"line_number":8,"context_line":"      exits with status 1."},{"line_number":9,"context_line":"    - The service fails to restart on SIGHUP. It is now stopped, and the"},{"line_number":10,"context_line":"      process exits with status 1, as when ``start()`` fails."},{"line_number":11,"context_line":"    - A SIGHUP arrives during ``start()`` or during a restart. It is now"},{"line_number":12,"context_line":"      handled once that call returns, with one reload however many SIGHUPs"},{"line_number":13,"context_line":"      arrived, instead of interrupting the call."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"    A SIGHUP that arrives during a shutdown is now ignored, as a second"},{"line_number":16,"context_line":"    SIGTERM is. Before, it stopped the shutdown and restarted the service."}],"source_content_type":"text/x-yaml","patch_set":1,"id":"8fc8d67c_2121baeb","line":14,"range":{"start_line":8,"start_character":26,"end_line":14,"end_character":1},"updated":"2026-10-29 15:10:33.000000000","message":"These are the new behaviors which haven\u0027t been included by any of the existing releases, right ? Then drop these from release note because no user may face this issue really.","commit_id":"9533671e6a804660acd194bd65ffc0bba7bd322b"},{"author":{"_account_id":10273,"name":"Adam Harwell","email":"flux.adam@gmail.com","username":"rm_you"},"change_message_id":"2ab1d74c607f39b50370232ae74c9cd6bcec37bf","unresolved":false,"context_lines":[{"line_number":5,"context_line":"    longer hangs at exit in these cases:"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"    - The service\u0027s ``stop()`` raises during a shutdown. The process now"},{"line_number":8,"context_line":"      exits with status 1."},{"line_number":9,"context_line":"    - The service fails to restart on SIGHUP. It is now stopped, and the"},{"line_number":10,"context_line":"      process exits with status 1, as when ``start()`` fails."},{"line_number":11,"context_line":"    - A SIGHUP arrives during ``start()`` or during a restart. It is now"},{"line_number":12,"context_line":"      handled once that call returns, with one reload however many SIGHUPs"},{"line_number":13,"context_line":"      arrived, instead of interrupting the call."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"    A SIGHUP that arrives during a shutdown is now ignored, as a second"},{"line_number":16,"context_line":"    SIGTERM is. Before, it stopped the shutdown and restarted the service."}],"source_content_type":"text/x-yaml","patch_set":1,"id":"8c61c2c6_69131418","line":14,"range":{"start_line":8,"start_character":26,"end_line":14,"end_character":1},"in_reply_to":"8fc8d67c_2121baeb","updated":"2026-10-30 03:04:32.000000000","message":"Ah, I guess that\u0027s true. Let me see...\n\nSIGHUP restart with no_fork comes from 1009388, which no release has yet. Released versions end the process on any SIGHUP, and 1009388\u0027s release note already covers that.\n\nI will just keep the stop() item, since that hang is in every release with no_fork (4.2.0 to 4.8.0 according to Claude\u0027s research).","commit_id":"9533671e6a804660acd194bd65ffc0bba7bd322b"},{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"ae5f902ffad172456780ea565d2ae8e834cff3f7","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"342bbfd0_5231e298","line":17,"range":{"start_line":15,"start_character":4,"end_line":17,"end_character":1},"updated":"2026-10-29 15:10:33.000000000","message":"The same here.","commit_id":"9533671e6a804660acd194bd65ffc0bba7bd322b"},{"author":{"_account_id":10273,"name":"Adam Harwell","email":"flux.adam@gmail.com","username":"rm_you"},"change_message_id":"2ab1d74c607f39b50370232ae74c9cd6bcec37bf","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"38cd0bc3_f0b1d3f3","line":17,"range":{"start_line":15,"start_character":4,"end_line":17,"end_character":1},"in_reply_to":"342bbfd0_5231e298","updated":"2026-10-30 03:04:32.000000000","message":"Done","commit_id":"9533671e6a804660acd194bd65ffc0bba7bd322b"}]}
