over the past few days i've checked in a bunch-o-code
that improves upon our unicode support/commitment. in
particular:
[] we now have a plain-text UTF8 importer and exporter
(these use the suffix .utf8) (the .txt i/e still
expects latin-1 text.)
[] i removed bob's temporary patch for #512 where we
had to fold fancy punctation (dashes, quotes, etc
(in the range 0x2000-0x2030) into us-ascii. this
mainly affects the RTF importer.
[] removed (on unix) and rewrote (on win32) all logic
dealing with the CharWidths array -- this was preventing
us from laying out any non-latin-1 characters.
[] opened up all non-latin-1 keybindings. let me know
if any of these are broken.
[] removed all latin-1 restrictions on the various
importers/exporters.
[] improved font cacheing strategy on win32.
with all of this, we are now closer to actually being
able to deal with unicode -- but we aren't there yet.
there are several problems remaining that we need help
with:
[1] on win95/98, these pseudo-OS's do not support
unicode. so, we need to deal with various CodePage
issues.
[2] on unix, i don't think we have any unicode fonts,
so some weird (wrong) things get drawn when a non-Latin-1
character is entered -- it's correct in the document, but
just drawn wrong. this may be a font issue or it may be
in the code which accesses the .pfa/.pfb files.
[3] on WinNT, we probably need to switch to use ToUnicode()
rather than ToAscii() in the keyboard handler and experiment
some.
enjoy,
jeff
