Commit messages in UTF-8 are garbled (BB-8715)
BitBucket does not display commit messages in UTF-8 correctly. It interprets them as Latin-1, which generates garbage. Consequently, the affected commit messages are unreadable.
Since it is very simple to determine whether a string is valid UTF-8, BitBucket can easily use UTF-8 for UTF-8 messages, and only fall back to something obsolete like Latin-1 if the message is not valid UTF-8.
Note that such a change has no effect on commit messages in pure ASCII, which is probably the vast majority.