Code-cleaned ReadTMWdefinitionFile which was an awful mess. Did
this along the way when implementing TCC 8-bit support.
Unicode<->TMW conversion of whole document now applies to footnotes
and endnotes.
Now (per spec) use 00A0 space rather than 0020 (= ASCII 32) for EWTS _ .
Implemented [] nesting rule for F9. It already worked for F10,
fortuitously.
Optimize out most calls to MakeRange*Script (recently introduced to
work around Word Unicode font-change bug); this improves performance
and also decreases the number of useless font-change events on the
Undo list.
"combination" -> "deprecated" in Unicode definition reader.
for experimental Unicode extensions. It's now clear that Unicode is
going in the trunk -- in fact I'm close to a first beta release -- so
I'm clobbering the "Unicode hackup.dot" file and checking in its
latest version here.
The changes:
Unicode support. Extensive remodularization and code changes in most
parts of the system, far too much to detail here. 2.1 has roughly a
third more code than 2.0!
Per request of Cathy Cantwell, make F9 on body text that includes
footnotes/endnotes convert text of said notes.
In support of the previous, various changes to make Tibetan work in
footnotes (and endnotes) generally.
Pronunciation function now talks about "foreign" words rather than
"Sanskrit"; former is more generally correct.
Fixed prehistoric bug (goes back to Than's version 1.1) that I just
found: F6 went to end of document if you hit Cancel.
Fixed SourceForge bug 941183: literal curly quotes in VBA screwed up
Chinese Word. (Need to get Jeff Wu to verify fix -- I couldn't
duplicate bug.)
Fixed SourceForge bug 986847: double quote doesn't curlify. I
"fixed" this by removing the double-quote binding. That's not the
optimal fix, but the only thing that's non-optimal about it is that
when you are typing EWTS, double-quote self-inserts rather than
telling you that it's not a valid EWTS character.
Fixed SourceForge bug 986850 (Interlineal failing when loaded from
Startup, due to an instance of the Word bug where Startup-loading
doesn't run initializations)
Fixed bug: in F9 conversion, \u allowed only lower-case (not
upper-case) alpha hex digits.
Fixed bug: in F9 conversion, \ was not interpreted as an escape
within [].
Fixed bug: \Z] did not return to TScript, but instead inserted the
].
Fixed bug: [ and ], when in Interlineal mode, now beep rather than
screwing things up.
Various uses of "English" replaced in the UI. (Trying to be less
ethnocentric; we have lots of users whose primary language is not
English.)
CVS is whining about the GPL containing "escape characters" (presumably
control-L), but when I try to add it as binary, it complains that it's actually text.
I hope this works.
Change tibwn.ini to overwrite -- previously it was set not to, which seems
completely wrong.
The Uni font is now set to overwrite; I hope this is the right thing. For some
reason the TMW fonts had been set no-overwrite. I'm hoping that is just
because I assumed they wouldn't change rather than because it doesn't
work to overwrite fonts or something.
(but I haven't run extensive tests, and probably will break in some cases).
Unicode->Wylie is not working. I'm checkpointing here because I'm
about to attack the Tibetan->Wylie code and I'm not sure my approach
is correct -- may need to back out. (Current theory is to recycle the TMW->
Wylie code with small amounts of internal conditionality on TMW vs.
Unicode, but I'm not positive this can be made to work.)
It would cleaner to do a fork of "WylieWord development copy.dot",
which is what this, but the wincvs doc doesn't cover the fork command
at all, and I don't want to risk confusion.
Per request from Cathy Cantwell, add option to do F11 interlineal
conversion without introducing tables, producing text only. An
unfortunate side-effect of this is that you get a one-time-only error
message complaining about the options file (whose format has changed
to store the value of this option).
Changed the interlineal line breaking code to put only shad-like
things at the end of lines, and not zla tse type things. (But: it no
longer breaks lines on zla tse type things at all; ideally it should
break on those but put them at the beginning of lines. On the other
hand, you'd never expect to see a zla tse type thing other than after
a shad type thing (or at the begnning of a text).)
printing. The page format and second-page-header were somehow
inherited from Matthew Kapstein's reader. This included line
numbering, non-standard margins, and the header. Changed back to Word
defaults.
Cantwell (SF RFE #1034292), that it stripped footnotes (and other
non-text items).
Per request of Cathy Cantwell (SF bug 1034688), make
HandleUserApostrophe respect Options.AutoFormatAsYouTypeReplaceQuotes.
I.e., typing a ' will no longer get curly single quotes if you have
turned off the "smart quotes" autocorrect option (which lots of
Tibetanists may have done, so a-chung comes out decently).
Fix PartOfTshegBar to not accept [, which it did, due to an
inexplicable Obiwan error. Clarify code to make Obiwan error less
likely.
Fix PartOfTshegBar to not accept H. This is bogus; H is a Sanskrit
letter, but tibwn.ini claims it is punctuation, and our code depends
on that.
Sean Something <knowone@zensearch.com> pointed out that a special ha
glyph (1,102) is supposed to be used solely in the case of hU.
Added code to do so.
Fixed bug: Sean Something <knowone@zensearch.com> and Christopher
Walker <cewalker@uchicago.edu> correctly reported that lnga did not
not stack when typed. (It gave "lang". But it worked correctly in
F9 conversion.) This was a side-effect of the "sara" fixes in
2.0b2. Fixed by invoking the consonant stack buffering code when we
get "ln".
Changed both parsers to allow ' (a-chung) as root, in order to handle
"'od" which Robert Walker <robertjwalker@onetel.net.uk> correctly
reported failed in Wylie->Tibetan conversion. (It also failed in the
opposite direction.)
Fix typo in comment at beginning of Wylie parser.
