Segfault protocol.cpython

Nobody has claimed this yet.

* **asyncpg version**: 0.27.0
* **PostgreSQL version**: PostgreSQL 15.2 (Debian 15.2-1.pgdg110+1) on x86_64-pc-linux-gnu, compiled by gcc (Debian 10.2.1-6) 10.2.1 20210110, 64-bit
* **Do you use a PostgreSQL SaaS? If so, which? Can you reproduce
the issue with a local PostgreSQL install?**:
* **Python version**: Python 3.11.2 (main, Mar 17 2023, 02:19:18) [GCC 10.2.1 20210110] on linux
* **Platform**: Linux 0aa90c609aa2 4.15.0-210-generic # 221-Ubuntu SMP Tue Apr 18 08:32:52 UTC 2023 x86_64 GNU/Linux
* **Do you use pgbouncer?**: No
* **Did you install asyncpg with pip?**: Yes
* **Can the issue be reproduced under both asyncio and
[uvloop](https://github.com/magicstack/uvloop)?**:
## Error message
In logs on my server I got only

```
May 9 09:50:56 vdska kernel: [586745.282012] python3[2695]: segfault at 0 ip 00007ff211e8be35 sp 00007ffdb02aa3e0 error 6 in protocol.cpython-311-x86_64-linux-gnu.so[7ff211e65000+9f000]
May 9 09:51:16 vdska kernel: [586764.689741] python3[20627]: segfault at 0 ip 00007f1d3a3b4e35 sp 00007f1d3889c1c0 error 6 in protocol.cpython-311-x86_64-linux-gnu.so[7f1d3a38e000+9f000]
```

Start with the reported protocol.cpython-311-x86_64-linux-gnu.so crash logs and the listed Python 3.11.2, PostgreSQL 15.2, and Linux environment. No reproducer, source file, or test is provided, so first establish reproducible steps and identify the failing protocol component; done means the segfault is reproduced, diagnosed, and covered by a regression test.

Written by the indexing model from the issue text.