/\/\i've seen CHAIN7 as well(thoughi dont remember under what circumstances)

anyways, here's a screenie of the invisible characters thing:

the same thing tokenized:

and here's the file, in case you want too look:
file

and i found another small bug. if the cursor is on a line with characters and is followed by an empty enter line and then another line (with or without any characters) comme ça:

hello[cursor here]

this is the third line

pressing down will jump over the enter line to the third line. this does not happen when pressing up from below, however.

oh, and i remembered what it was i was forgetting about before. the xlib/celtic libraries and all that do not work either on the homescreen(which is very useful for accessing all those other pics they enable) or for basic builder apps like that one bowling game.

EDIT: i am deeply ashamed of that typo

Thanks for that, I'm fairly sure it's just one or two characters that are a problem.

How strange; I'll certainly look into that.

Hmm, I know about the homescreen thing, but I'm a bit baffled why the BasicBuilder thing doesn't work. Chances are I won't get a chance to do anything about that before the DCS7 release, though.

*bump*

Aight, let me see if I can marshal my thoughts here. These are the issues I still need to look at before the release of Doors CS 7.0 Release Candidate (RC) 1:

:: Bouncy [ON] key on hardware TI-84+s
:: LCD still glitchy on some TI-84+s even after tuning
:: Shmibs' character with improper width? See post above this one.
:: re: cursor re-appearing after iGo+Options: somewhere there's still a function that's either bcall(_RunIndicOn/Off) or ei'ing incorrectly.

I believe this covers everything? Did anyone tell me something that I've since forgotten?

I have found a bug with the tab functions.

First of all, put the mouse to the right of the start button, where the higher folders would be showing. Press the tab button, until the inverted part is under the mouse. left click, and you can see it. after that, move the mouse, and see the damage.

All fixed, thanks for that!

*bump*

:: [ON] bouncing on 84+/SEs ameliorated. TI-83+s tested to still not bounce with the new change.
:: Shmibs' character with improper width traced to DocDE7 and the $E1 character. Shmibs, just click at the very beginning of your document and press [DEL] to delete the $E1 character you have at the very beginning from clicking in the blank area of the Insert Symbol window. I'll fix that for the next release of DocDE7.
:: Fun diagrams of the text input characters in each mode added to the manual (see below). Two pages of exhaustive features also mention. If you check out the manual with Doors CS 7.0 Release Candidate 1, I'd appreciate any/all of you making sure I'd didn't miss any features that I implemented long ago and have forgotten about.

Remaining items:
:: LCD still glitchy on some TI-84+s even after tuning(??)
:: re: cursor re-appearing after iGo+Options: somewhere there's still a function that's either bcall(_RunIndicOn/Off) or ei'ing incorrectly.

Where there two "G's" on the first calc, if my getkey memory serves, both keys 44 & 54 are "G."

Word.


Sorry, I had to do it.

*facepalm*

Anyways, looking forward to DCS7RC1!

Cheers comicIDIOT, rthprog. I don't suppose either of you have a TI-84+ with a glitchy screen? So far two users, iirc, have said that the tune / ALCDFix option either didn't work at all for them or didn't work enough. I'm not really sure what I can do about that, though, other than increasing the delay that the routine that calculates the delay value produces. ThePenguin, I'd appreciate any feedback.

Okay, weird sum(13 bug.
According to the wiki, the format is sum(13,{X},{Y},{title},{icon},{itemlist},{labellist}
Watch the screenshot and see what happens. (I show you the source first, so be patient).

I specify X and Y to be 7 and 7, but it displays at 0 and 7. For some reason, putting another number between 13 and {X} solves this.
Also, you can see the homescreen getting really glitchy at the end. I can solve this by using ClrHome, but I shouldn't have to. It does go to the correct label, though

1) Regarding the title and icon: the wiki looks correct to me? You have F8888888F8 as your first string, and that's the title of your window. Your second string is F8F8F8F8F8, and you get a nice square for your icon.

2) Through some freak of coincidence, I independently noticed and fixed the coordinate issue yesterday, so you almost beat me to finding that. The problem was I was reusing a routine from the sum(7,N,... routines, and thus it was starting at the third numerical argument instead of the second (since for sum(7...) the second is the GUI item type)

3) I'm aware of that glitchiness thing and usually just recommend ClrHome. Do you think it would be worth it for me to trace down what exactly is going on there?

Can you make a CalcSys style change: I would like + and - scroll by one page like they always have, but make * and / scroll by 4 pages. This would speed up scrolling because hotspots and the the mouse aren't used as much while scrolling through a big list of programs. Thanks for considering this!

Kerm, on this page, http://dcs.cemetech.net/index.php?title=DCSB_Libs&oldid=7036 it says icon, then title, while on this page, http://dcs.cemetech.net/index.php?title=BasicLibs:GUIMenu&oldid=7173 it says title, icon.

Ah, thanks for noticing that. I fixed http://dcs.cemetech.net/index.php?title=DCSB_Libs to say title, icon, and also remembered to update the note on sum(11) to sum(11[,ULCD]) on the same page.

I think that's doable. I presume that if you're less than four pages from the start or end, scrolling backwards or forward should only bring you as far as the first or last page rather than wrapping around? Does everyone else agree on 4 at a time, or think some other value (3?) would be better?

Edit: Found and fixed the interrupt issue with the Options menu - Arc_Unarc was to blame. I wrapped a ld a,i \ push af \ ...\ pop af \ jp pe,... around the bcall to leave the interrupts as they were before Arc_Unarc. That leaves the elusive Tune failures, the GUIMenu homescreen glitches, and the multi-page scrolling.

Can you make it a set-able option?

Ugh, not easily; I'm already pushing it with the Options menu spacing. I had to truncate "[ON][PRGM]/[MODE] hooks" to "[ON] hooks" to even fit that. Couldn't we just decide on a nice amenable value between 3 and 4?

In that case, my vote is 4.

That's two votes for 4 and no votes for 3, since I'm staying impartial. Any other opinions?

The only thing that seems logical to me would be that it requires a different delay between data writes than between command writes.

Ok, here is a routine that does just that. It's nearly exactly the same as the last one I gave you except that it writes to port 11h instead of port 10h. It also does all the writing off screen so that you won't notice that it's actually doing anything.


Code: