Re: using utf8 strings in plots and printing

[Jordi Gutiérrez Hermoso]

On 6 December 2012 10:00, Olaf Till <address@hidden> wrote:
> On Thu, Dec 06, 2012 at 08:42:35AM -0500, Jordi Gutiérrez Hermoso wrote:
>> UTF-8 support in Octave is inexistent, and if it ever works, it's only
>> by accident. This includes strings
>
> do you think that support of Octave is necessary fur utf8 in strings?
> Isn't it enough that utf8 codes (provided by an utf8-enabled terminal
> or editor) are just contained in Octave strings, and that printing
> them to an utf8-terminal reproduces the correct symbol?

I haven't really studied the problem in detail, but I do believe
something has to change in the way we handle strings to account for
multi-byte characters. There are probably many locations in the code
where we make the implicit assumption that each character is a single
byte, and we have to track those down and deal with them.

If nothing else, consider that all strings in Octave are Fortran
arrays, so we would have to fix these Fortran arrays to hold a
different type so that indexing a multi-byte string indexing works as
expected.

- Jordi G. H.