6.6.0 or 6.6.1 has introduced a segfault

[bdraco]

  File "/Users/bdraco/home-assistant/venv/lib/python3.13/site-packages/aiohttp/http_writer.py", line 217 in write_headers
  File "/Users/bdraco/home-assistant/venv/lib/python3.13/site-packages/aiohttp/client_reqrep.py", line 1467 in send
  File "/Users/bdraco/home-assistant/venv/lib/python3.13/site-packages/aiohttp/client.py", line 746 in _connect_and_send_request
  File "/Users/bdraco/home-assistant/venv/lib/python3.13/site-packages/aiohttp/client.py", line 770 in _request
  File "/Users/bdraco/home-assistant/venv/lib/python3.13/site-packages/aiohttp/client.py", line 1061 in _ws_connect
  File "/Users/bdraco/home-assistant/venv/lib/python3.13/site-packages/uiprotect/websocket.py", line 128 in _websocket_inner_loop
  File "/Users/bdraco/home-assistant/venv/lib/python3.13/site-packages/uiprotect/websocket.py", line 88 in _websocket_loop
  File "/opt/homebrew/Cellar/python@3.13/3.13.3/Frameworks/Python.framework/Versions/3.13/lib/python3.13/asyncio/events.py", line 89 in _run
  File "/opt/homebrew/Cellar/python@3.13/3.13.3/Frameworks/Python.framework/Versions/3.13/lib/python3.13/asyncio/base_events.py", line 2034 in _run_once
  File "/opt/homebrew/Cellar/python@3.13/3.13.3/Frameworks/Python.framework/Versions/3.13/lib/python3.13/asyncio/base_events.py", line 677 in run_forever
  File "/opt/homebrew/Cellar/python@3.13/3.13.3/Frameworks/Python.framework/Versions/3.13/lib/python3.13/asyncio/base_events.py", line 706 in run_until_complete
  File "/Users/bdraco/home-assistant/homeassistant/runner.py", line 154 in run
  File "/Users/bdraco/home-assistant/homeassistant/__main__.py", line 209 in main
  File "/Users/bdraco/home-assistant/venv/bin/hass", line 10 in <module>

I'm still investigating. More details as I find them

[bdraco]

I think its here https://github.com/aio-libs/multidict/pull/1200/files#r2173216892

[asvetlov]

Hmm. Does 6.5.1 work well?

[asvetlov]

Sorry, I'm out of home now, I'll return back to my laptop in the evening.

[bdraco]

6.5.1 doesn't fault

[bdraco]

I think I've found it. PR coming

[bdraco]

diff --git a/multidict/_multilib/hashtable.h b/multidict/_multilib/hashtable.h
index 8d1332a..438acbc 100644
--- a/multidict/_multilib/hashtable.h
+++ b/multidict/_multilib/hashtable.h
@@ -251,8 +251,9 @@ _md_shrink(MultiDictObject *md, bool update)
     for (Py_ssize_t i = 0; i < nentries; ++i, ++old_ep) {
         if (old_ep->identity != NULL) {
             if (new_ep != old_ep) {
-                *new_ep++ = *old_ep;
+                *new_ep = *old_ep;
             }
+            new_ep++;
         } else {
             newnentries -= 1;
         }

I think this will fix it

[bdraco]

It looks like something changed will rules so I can't push a fix

[bdraco]

pretty sure #1222 will fix it