Bugzilla

(In reply to Joshua Cranmer [:jcranmer] from comment #47)
> An interesting test to add to conduct (and add to our test suite) is
> replying inline to a non-UTF-8 message. A further test would be to reply
> inline to a message whose charset has been manually overridden by the UI.

Right, and, as I said in comment #32, we also need to cover the other code path, which is forward, "edit as new message", edit draft/template. You could have an old non-UTF-8 draft/template. Remember that composition has two very distinct code paths: One is reply and one is the other lot.

(In reply to Joshua Cranmer [:jcranmer] from comment #47)
> An interesting test to add to conduct (and add to our test suite) is
> replying inline to a non-UTF-8 message. A further test would be to reply
> inline to a message whose charset has been manually overridden by the UI.

Right, and, as I said in comment #32, we also need to cover the other code path, which is forward, "edit as new message", edit draft/template. You could have an old non-UTF-8 draft/template. Remember that composition has two very distinct code paths: One is reply and one is the other lot.

EDIT: Looks like there are some forwarding tests here: https://searchfox.org/comm-central/source/mail/test/browser/composition/browser_forwardUTF8.js which can be extended to non-UTF-8 messages.
There is also https://searchfox.org/comm-central/source/mail/test/browser/composition/browser_replyMultipartCharset.js which uses body-greek.eml. So you'd have to check what is covered and what is not and extend what is already there.