Vsevolod Stakhov

Test failure is irrelevant.

> * AUTH and SELECT results should probably be checked in `redis_connect_sync` I totally agree with this point, thank you! For async connections this issue is usually handled by `redis_pool`,...

> The quick workaround is to add `conn:exec()` after all `conn:add_cmd(...)` lines in `redis_connect_sync`, but I think the API is designed rather unfortunately in the first place Yes, that's also...

It's likely fixed by https://github.com/rspamd/rspamd/commit/cdbacf2b135fc49a84df1d0790c2ba77d6d763b7

The first line of the message is clearly malformed.

Hum, cannot it be a source of another problem, as I can still see a line ``` >From - Tue, 30 Aug 2022 15:37:28 GMT ``` in the `message/rfc822` attachment...

I don't understand the issue. It seems that your resolver returns NOREC error that is correctly handled by Rspamd. This error means that MX record is not presented but some...

Yes, your arguments sound correct, thank you.

There is no problem with the koi8-r per se, it is decoded just fine. However, there is no special treatment for the text parts in the `message_to_ucl` and thus the...

We can discuss the changes direction if you want that by the way. I mean a special treating of the text parts.