Re: [PATCH] diff.c: use utf8_strwidth() instead of strlen() for display width

[Junio C Hamano <gitster@pobox.com>]

Patrick Steinhardt <ps@pks.im> writes:

> Also, do you know why we didn't use `utf8_strwidth()` right from the
> start? It would have saved the writer some time to just use
> `utf8_strwidth()` instead of writing a whole paragraph explaining that
> we should do it eventually. Makes me wonder whether there is anything
> else going on here.

I suspect that it is because it is not just that single strlen()
call.  The code assumed that byte count and display width were
interchangeable, and use of strlen() there was merely an example.
Starting there, the value returned by strlen() is treated as if it
were interchangeable with a display width, and then later used to
count how many bytes to trim from either end of the string so that
the trimmed string would eventually fit a given display width, which
means that the code to compute how much to trim (which is not that
strlen() the patch in question is touching) would have to compute
the reverse, i.e. "if we need to recover N display columns, how many
bytes do we need to trim from that string?"