Yuku free message boards

I asked nitro and he told me his original version of SIDE can also be used under the GPL now, he just didn't get around to rerelease it with the new license headers. So I did use code from the original SIDE, but I have permission from the author to do just that.

Oh, and I don't use the original slow drawing routines, I adapted your optimized ones to work with syntax highlighting and grayscale.

The current one is cr*p (it only colors operators differently, and not even all of them), but I want to add support for the following colorings:
* identifier (string of alphanumeric characters starting with an alphabetic character) -> black
* number (string of alphanumeric characters starting with a digit) -> light gray
* operator (everything else) -> dark gray
This should make sense for most languages, including C and assembly. Alphabetic characters would comprise letters including accented and greek letters. Digits would be 0-9, the (-) minus and $ (for hex numbers in ASM). Possibly the point too (for floats - it would also color move.l or mystruct.member as single identifiers, not sure whether this is a good idea).

Maybe there could also be true language-specific syntax highlighting, but there aren't many colors to choose from (we could use effects like bold or italic, but it most likely means additional hand-edited versions of the font since applying automated effects to a font this small would probably be completely unreadable), and it would quickly get computationally expensive to parse.

Then your contrast setting is messed up.

But this requires lots of parsing effort, I basically need to parse the entire document for that. (Or do you think counting the parentheses only within a logical line, i.e. whatever is enclosed between 2 CRs, is OK? That would still be computationally expensive, but not as much.)

One problem is that I can't really add too much too computationally expensive coloring features because the coloring is recomputed and the screen is redrawn each time you type a character, so the responsiveness would pretty quickly drop beyond what's acceptable if the parsing gets too expensive, especially on the TI-92+ with its larger screen.

New version, which now uses the event loop! The AMS event loop is used instead of SIDE's old main polling loop. (Things like the file name prompt still use SIDE's custom key polling though.) From a user's point of view, this means you can use event hooks, for example XtraKeys to enter accented characters on the TI-89. From a developer's point of view, it means I no longer have to warn about not following SIDE's structure when coding a similar app.

 2007-03-18 Kevin Kofler r4 + Completely drop support for PedroM and AMS 1 (MIN_AMS was already 204 anyway) + Clean up indentation, get rid of tabs + Fix strict-aliasing violation + Redirect trap #0 to fix idle() for grayscale + Use AMS event loop instead of polling + Use MLink format relocs

The download link is still the same: www.tigen.org/kevin.kofler/ti89prog/side-latest.zip

Fixed another regression from the conversion: turning the calculator off within SIDE didn't work properly.

 * 2007-03-19 Kevin Kofler r6 * + Hook trap #4 for APD, reset keyboard repeat there instead of turn_off() * + Use internal F-Line emulator and lower MIN_AMS to 200 * + Remove turn_off(), replace all uses with off() * + Remove handling of CM_APD and CM_OFF which are apparently never used * + Handle KEY_OFF and KEY_OFF2

I just fixed another bug (or missing feature depending on your point of view), reported by Joey Adams. It's an old bug this time, not a regression: '!' (and some other characters entered through DIAMOND+key) couldn't be entered on the TI-89 without entering the CHAR menu (i.e. not at all in the options dialog).

 * 2007-04-30 Kevin Kofler r8 * + Add support for DIAMOND+key=char combos on the TI-89