Re: Unicodification of sources, part 1
Ar an fichiú lá de mí Meitheamh, scríobh Stephen J. Turnbull:
>> The 21.5 Unicode support is seriously twisted around
Windows
>> support and the whole Mule infrastructure.
Aidan> What? No. Its Windows support has always been irrelevant to
Aidan> me as a Unix user, it has never got in my way,
I'm not talking about using the code; I'm talking about developing it.
And I’ve been developing with it as much as or more than anyone, and its
Windows support has never got in my way.
>> SXEmacs doesn't want any Windows baggage, and
probably could
>> benefit from losing Mule, too. 90% of the Mule-related C code
>> seems to be devoted to variable-width character gymnastics.
Aidan> Much like Perl, I imagine. That’s not a disadvantage in
Aidan> itself.
We have serious problems with efficiency of text manipulations, and
nonnegligible ones with correctness in regexps and syntax tables, even,
at the moment. Presumably SXEmacs would inherit them if they ported the
current setup.
They already have what Mule problems we have (with the added spice of no
concrete plans for supporting Unicode!), and have even fewer people
interested in fixing them.
Redisplay has needed substantial work for years. Xft is merely
grafted
in, and hanging on by a scrap of skin. The obscurity of current Mule code
is an obstacle to working with all of these. Mule is everywhere on the
critical path to a release for us.
... because you and me are the only people that use its features, or that
have a real interest in doing so. Not because of the quality of the code, in
itself, if the Mule issues were routinely affecting monolingual English
speakers they would have been resolved long ago.
--
Santa Maradona, priez pour moi!