[Minor] Remove redundant sub-property

Implement DNSBL checks for Message-ID RHS

Add Content-Description header rules

Add detection for freemail and disposable e-mail usage for message delivery notification

Issue: #4947
Closes: #4947

Apply detection phase if fasttext could not detect language

Issue: #4929

[Minor] Add bad file types commonly found in archives attached to malspam

This header frequently surfaces in spam, mostly advance fee fraud.

See https://github.com/rspamd/rspamd/pull/4888 for the related
discussion.

This reverts commit c78c70c76b.

This reverts commit 5e84a82f87.

Rework composites for spam injected into compromised accounts

Do not read out-of-boundary when doing base64 encoding

Implement backpressure for fuzzy check

Serialize control commands

In fact, we cannot send multiple commands and read them through the pipe. It has
caused multiple weird issues in the past but I can now see clearly how it should
be done. We should send commands and serialize all requests pending to let them
being sent one by one, after reply for the previous command has been received.

[Minor] Update Bootstrap

Fix error in headers_checks.lua

Support UTF8 flag for the received headers

Fix statfile symbols disabling

Issue: #4919

[Fix] Relearn messages in bayes filter

[Minor] Add selector to get rspamd_hostname

When relearning a message, this script is called twice:

1. With `is_unlearn` true and the old `is_spam` value
2. With `is_unlearn` false and the new `is_spam` value

If `is_unlearn` is true, the code should not increase the value of the hash_key. Decreasing it ensures that it behaves like a fresh learn with the new is_spam value.

Treat relearning from ham to spam the same as relearning from spam to ham. `res` is only set to true if the mail is already learned.

Instead of calling get_hostname() for every processed message, we instead reuse the variable value from rspamd boot time