Re: cp932, cp949, c950 definitions

[Dave Love]

Jason Rumney <address@hidden> writes:

> Dave Love wrote:
>
>> The coding systems cp932, cp949, c950 have been added as aliases.  Are
>> these known to be correct according to a reliable reference?
>
> Obviously not judging by your reaction.

I'm not claiming any authority!  I just happened to remember
considering the issue and would be happy if it's trivial.  (With some
CJK charsets, it seems there isn't an authoritative translation to
unicodes, which is what's basically needed; I'm told jisx0208 and 0213
are cases in point.)

> I beleive the mappings from
> the standard encodings to the DOS/Windows codepages are all correct,
> but the codepages include extra characters that do not map back into
> the coding-systems they are aliases for.

You mean the CP versions are supersets of the character sets
corresponding to the coding systems they're aliased to?

> The reason for adding these aliases is to get clipboard support
> working out of the box on Windows easily. Even if the codepages are
> incomplete, it is better than the current situation of assuming
> latin-1 for the clipboard.

It might be possible to get them (more-or-less) right with
post-read/pre-write conversions if any additional characters are ones
Emacs can represent.  It would be good to avoid any incompatibility
with Emacs 22 anyhow.