Austin Bonander

I tried to reproduce the `CorruptMessage` error with MySQL 5.6.45 which only supports TLS 1.0 but got a normal `HandshakeFailure` error indicating that it just doesn't support any of the...

On that note though, the `HandshakeFailure` message doesn't really explain anything either which could be confusing to users who don't have intimate knowledge of TLS. I realize that's really the...

This is the handshake error another commenter saw: https://github.com/launchbadge/sqlx/issues/1567#issuecomment-1000254162 `HandshakeFailure` _is_ representative of what the server actually replied, but maybe the `Display` impl for `rustls::Error` could go into more detail...

I will also be adding an entry to SQLx's FAQ to cover this because people often try to use it and RusTLS to connect to older database versions that only...

@ctz @briansmith I finally managed to stumble across a potential repro of the original issue: https://github.com/launchbadge/sqlx/runs/7623647065?check_suite_focus=true Here is a Wireshark capture of the exchange that produced the `corrupt message` error:...

That is part of a larger Wireshark capture that's filtered down to just interactions with that port. I thought it'd contain less noise that way but maybe I accidentally filtered...

Actually, for the MySQL connection handshake it appears that the server is _supposed_ to send the first message: https://dev.mysql.com/doc/internals/en/initial-handshake.html

@GREsau There's really no good reason to strip excess newlines. By the CommonMark spec, these should already be treated as soft line breaks and handled during rendering: https://spec.commonmark.org/0.30/#softbreak IMO, stripping...

There's really no good reason to strip excess newlines. By the CommonMark spec, these should already be treated as soft line breaks and handled during rendering: https://spec.commonmark.org/0.30/#softbreak

If you're adding `from_str`, then `to_string` (or `display`) would be pretty neat. It'd be nice not to have to write a wrapper for that.