Re: windows-XXXX and cpXXXX

[Eli Zaretskii]

> From: Kenichi Handa <address@hidden>
> CC: address@hidden, address@hidden
> Date: Wed, 11 Jan 2006 20:35:34 +0900
> 
> >>    cp437 cp720 cp737 cp775 cp850 cp851 cp852 cp855 cp857 cp860
> >>    cp861 cp862 cp863 cp864 cp865 cp866 cp869 cp874 cp1125
> 
> Here, I forgot to add cp1251 and cp9XXs.
> 
> >>  But it seems that windows-XXXX are quite frequently called
> >>  as cpXXXX(*).  If so, I'll register cpXXXXs as alises of the
> >>  corresponding windows-XXXXs, ok?
> 
> > Won't that increase the confusion, which is IMHO already too high,
> > between codepage.el and code-pages.el?  
> 
> They already support a coding system of the same name
> (e.g. cp720) in a different way.

Then I'd rather remove DOS codepages (as opposed to Windows codepages)
from code-pages.el, than add more cpNNN encodings to it.

> What kind of confusion does making aliases cp125[02345678] increase?

Confusion between codepage.el and code-pages.el.  They are different
and subtly incompatible, but define symbols that are almost identical.
As code-pages.el cannot be used in the MS-DOS port, we cannot throw
away codepage.el.  Thus, I think DOS codepages (whose names are cpNNN)
should be provided only by codepage.el.

> I didn't intend to change them.  My suggestion is just to
> make cp125[02345678] as an aliases of windows-125[02345678].

My concern would be how a user is to know which library of the two she
is using, or should use in a given situation.  I already saw several
confused users on help-gnu-emacs.  So I think we need to sanitize
these two libraries from each other's namespace.  Adding cpNNN aliases
would be a step in the wrong direction.

Alternatively, someone who has more time than I do could add to
code-pages.el what it is lacking now to fully support the MS-DOS port,
and then we could toss codepage.el and add the aliases you asked about
to code-pages.el.