Invalid location with utf-8 encoded files #11899
Description
(Not really a bug since the compiler doesn't officially support utf-8 encoded sources. But ..)
Reading #11736 and working on utf support in jsoo made me realize that while one can already use utf-8 in string literal and comments, the computed locations will be wrong. Computed column would be number of bytes instead of number codepoint.
I didn't read #11736 carefully but I suspect it would also result in bad locations.
Here is a small example:
We currently have
$ cat c.ml
let (s : string) = "npiπππ" ^ 2
$ ocamlc -c c.ml
File "c.ml", line 1, characters 33-34:
1 | let (s : string) = "npiπππ" ^ 2
^
Error: This expression has type int but an expression was expected of type
string
where it should be
$ cat c.ml
let (s : string) = "npiπππ" ^ 2
$ ocamlc -c c.ml
File "c.ml", line 1, characters 30-31:
1 | let (s : string) = "npiπππ" ^ 2
^
Error: This expression has type int but an expression was expected of type
string