[darcs-users] [darcs-devel] [patch639] Use utf-8 charset for darcs send in case of non-ascii ...

[Juliusz Chroboczek]

>> (Stephen: and why the bug is in Emacs/MULE, not in the UTF8_STRING
>> definition.

> Those of us working outside of the X dev community use the ICCCM, not
> the source, as the reference whenever possible.  UTF8_STRING was *not
> documented* in that standard at the time.

Someone is WRONG on the Internet.[1]

> The Chinese also have a "multiple locales in one locality" issue,
> though it's not as severe as the Japanese one.[...] generally they (or
> their tech staffs) know what they want to do about that, in the long
> run.  Their problem is OPI (Other People's Infrastructure).

Please try to focus at the issue at hand.  This is not about whether the
Chinese have any legacy mainframes using EBCDIC.  It is whether there
are any developers using Darcs who are running a non-UTF-8 locale by
default on their development machine.

-- Juliusz

[1] That's off-topic for this list, so I'll put my rant in a footnote.
    The bug actually was that MULE implemented COMPOUND_TEXT
    incorrectly, using a large subset of ISO 2022 instead of falling
    back to Unicode in extended ISO 2022 segments, as required by the
    ICCCM. That had nothing to do with UTF8_STRING, but we were blamed
    for it anyway.

    This issue was made worse by MULE not implementing UTF8_STRING,
    which, as you rightly note, was described in a document separate
    from the ICCCM[2].  My mistake, I guess, it was never published in
    GNU Texinfo format with a license longer than the document itself.

[2] http://www.pps.jussieu.fr/~jch/software/UTF8_STRING/UTF8_STRING.text