bug#42904: [PATCH] Non-Unicode frame title crashes Emacs on macOS

["Mattias Engdegård" <mattiase@acm.org>]

17 aug. 2020 kl. 19.05 skrev Eli Zaretskii <eliz@gnu.org>:

> Well, it is always easier to special-case some use case, but we have
> general APIs for a reason.  In particular, having several similar but
> subtly different functions is confusing and causes mistakes.

The new function is much simpler and easier to use than encode_string_utf_8 precisely for that reason: to avoid confusion and mistakes, both of which I got in spades when trying to use it.

> And you seem to be saying that encode_string_utf_8 doesn't work as
> advertised, which means it should be fixed.

Actually I don't know exactly how it is supposed to work so I wouldn't even say that. It's probably fine code but it's not for me, not in this case.

> So I would prefer to use encode_string_utf_8 if reasonably practical.

Well, it doesn't seem to be reasonably practical. In order to fix a bug, I prefer not having to fix some unrelated but complex code, especially when it is unclear how and if that code really can and/or should be 'fixed', and exactly what that would entail.

Now if, after the proposed patch has been applied, someone wants to refactor so that string_to_valid_utf_8 disappears or becomes implemented in terms of something else, then that's perfectly fine, as long as the bug remains fixed.