[zile-devel] The post-2.0 future: Unicode

[Reuben Thomas]

Although it's still early days for zile 2.0, I thought you might be 
interested to know my plans post-2.0. For me, the most important missing 
feature at the moment is Unicode support. This is tricky to add while 
retaining the goal that zile be a really minimal program, and in 
particular, while remaining fully Emacs-compatible, because of course we 
can't duplicate Emacs's complicated machinery for dealing with different 
character encodings.

I am open to suggestions on this; here's a sketch of my initial, rather 
vague idea.

First, I intend to support only "raw" and "wide-character" editing, and 
nothing else. This won't help users of other encodings (in particular 
those who use entrenched non 8-bit encodings such as ISO 2022), but is 
the simplest solution that will help the most people.

Secondly, the support will be based as far as possible on the standard C 
library, meaning that on platforms without wide character support, zile 
won't have it either. This avoids ad-hoc code, and also avoids bloating 
zile.

Thirdly, there should be auto-detection of file type, plus a manual 
override. Zile should as far as possible Do The Right Thing, but 
obviously not in all circumstances.

The main code that will have to be written, therefore, will be UTF-8 
detection, encoding and decoding; other than that, we can rely on the C 
library and terminal drivers.

--
http://rrt.sc3d.org/ | sad, a.  the efforts of musical debutantes (Bierce)