Auto-Reconnect is broken when reconnect is triggered by _listen

[Sandr0x00]

Bug report

• I confirm this is a bug with Supabase, not with my own application.
• I confirm I have searched the Docs, GitHub Discussions, and Discord.

Describe the bug

When using the example code (test_postgres_changes), the DELETE change triggers our function. As expected.
If we have an automatic reconnect, then we can get lucky and the heartbeat might reconnect (https://github.com/supabase/realtime-py/blob/main/realtime/_async/client.py#L202-L214). Then things keep working.
But if we are unlucky, then the listen function reconnects us (https://github.com/supabase/realtime-py/blob/main/realtime/_async/client.py#L102-L105). In this case we will NEVER receive any realtime-update again.

To Reproduce

Steps to reproduce the behavior, please provide code snippets or a repository:

1. Go to example
2. npx supabase start
3. Setup a table todos (use /tests/supabase/migrations/20240810125543_init.sql)
4. Comment out broken filter for INSERT change (see Example is broken #257) in app.py (https://github.com/supabase/realtime-py/blob/main/example/app.py#L135)
5. Comment out https://github.com/supabase/realtime-py/blob/main/example/app.py#L135 as we don't care about broadcast events for this bug.
6. Insert, update and delete something => see that it works
7. Force a disconnect using docker container restart supabase_realtime_realtime-py-example
8. Check output if INFO - Rejoining to: realtime:test-postgres-changes is in the output. if yes, then things continue to work as _heartbeat did the reconnect
9. If there is no "rejoining"-output, then _listen did the reconnect and then the realtime connection is broken forever.

Expected behavior

Auto-Reconnect to work.

System information

Latest supabase 2.6.8

npx supabase services

        SERVICE IMAGE      │      LOCAL       │ LINKED
  ─────────────────────────┼──────────────────┼─────────
    supabase/postgres      │ 15.8.1.020       │ -
    supabase/gotrue        │ v2.167.0         │ -
    postgrest/postgrest    │ v12.2.0          │ -
    supabase/realtime      │ v2.33.70         │ -
    supabase/storage-api   │ v1.14.5          │ -
    supabase/edge-runtime  │ v1.66.4          │ -
    supabase/studio        │ 20250113-83c9420 │ -
    supabase/postgres-meta │ v0.84.2          │ -
    supabase/logflare      │ 1.4.0            │ -
    supabase/supavisor     │ 1.1.56           │ -

[PatidarManas]

Facing same issue, need this fix to get merged to production
__main__ ERROR Unhandled exception in main: 'AsyncRealtimeClient' object has no attribute '_connect'