Taavi Eomäe

@vstakhov I've tested https://github.com/rspamd/rspamd/pull/4449 out. Now validating longer chains works properly, though the (previous) seals are still modified for some reason. Which seems somewhat wrong.

@vstakhov > Or wait, both read and write servers should all be able to accept scripts Maybe in theory, but in practice it keeps on looping and uploading the script...

@vstakhov Yes, it might be intended and desired but it does not work appropriately. It consumes the entirety of `/tmp`, causes logspam and the `neural` module does not work.

> How can it consume any space in /tmp? What do you see in the redis logs? Because each upload attempt creates a temporary file containing that uploaded script. >...

``` 2023-03-15T09:09:27.161653+00:00 RSPAMD-1.LOC-1.NAIDE.EE [mail.err] rspamd[43290]: ; lua; mx_check.lua:88: redis_cache_cb received error: READONLY You can't write against a read only replica. 2023-03-15T09:09:27.163027+00:00 RSPAMD-1.LOC-1.NAIDE.EE [mail.err] rspamd[43290]: ; lua; history_redis.lua:132: got error READONLY...

> Maybe you can read in a loop? Yes, this works to a large extent. Though I kind-of think that it would be much nicer if one could use the...

That's great to hear, will subscribe to that issue.

> > Could in theory provide smooth connection migration when an IRC server goes down > > IRC connections are somewhat stateful, so I'm not sure how this could help,...

> > Well, in theory that state could be migrated and then clients don't have to massively reconnect. > > The same can be said of TCP. Maybe with just...

> So it already supports QUIC in theory. It does not forbid QUIC, supporting it is a whole another topic. > Do any of them care about non-HTTP applications? There's...