Error calling msg:from_header(): Data is not ASCII or UTF-8

get-help
21

Not sure if it’s related to this issue or not, but Kumo just had a panic while parsing something:

Feb 28 18:05:43 send kumod[1592]: 2024-02-28T18:05:43.493401Z ERROR localset-3 run{socket=PollEvented { io: Some(TcpStream { addr: 5.78.73.137:25, peer: 162.55.128.206:47106, fd: 47 }) }}:process: kumo_server_common::panic: panic at crates/mailparsing/src/strings.rs:94:13 - !?
Feb 28 18:05:43 send kumod[1592]:    0: <unknown>
Feb 28 18:05:43 send kumod[1592]:    1: <unknown>
Feb 28 18:05:43 send kumod[1592]:    2: <unknown>
Feb 28 18:05:43 send kumod[1592]:    3: <unknown>
...
Feb 28 18:05:43 send kumod[1592]:   31: <unknown>
Feb 28 18:05:43 send kumod[1592]:   32: <unknown>
Feb 28 18:05:43 send kumod[1592]: thread 'localset-3' panicked at crates/mailparsing/src/strings.rs:94:13:
Feb 28 18:05:43 send kumod[1592]: slice range 24..0 is invalid for "U3RhdHVzOiA1LjAuMA0KDQo="
Feb 28 18:05:43 send kumod[1592]: note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
22

hmm, not directly related to the changed code, but perhaps it opens the door for that

23

mailparsing: try to avoid illegal slice panic · KumoCorp/kumomta@0536542 · GitHub might resolve this. If you see it again, it would be great if you could save the message contents like you did previously; unfortunately you cannot catch a panic, so you’d need to speculatively capture all mail, which is a PITA

24

Thanks @free-spirited-yorksh, this issue is resolved. I haven’t seen any crashes since yesterday when I updated to the latest dev version.