I'm working on a version of TiEmu targeted for TIGCC debugging. This version has GDB (see http://sources.redhat.com/gdb/ ) and the Insight graphical GDB frontend (see http://sources.redhat.com/insight/ ) built in, allowing to debug C programs with TIGCC 0.96 Beta 4 or higher. See http://tigcc.ticalc.org for more information, screenshots and a download link (hosted on SourceForge because there's no way I can upload my huge files on lpg.ticalc.org or tigcc.ticalc.org, I would be way over quota ;-) ).

PS: Long term, the version with GDB support (tigcc-debugging-branch) is slated to become TiEmu 3.00.

I'm using TiEmu (debugging branch) to debug some assembly I'm writing, and it's absolutely awesome! Great work.

There are some small problems with (ironically) the graphical debugger, however. I've fixed several of them -- where do I submit my bugfixes?

Preferably to the patch tracker: http://sourceforge.net/ tracker/? group_id=23169&atid=377682
(remove the spaces, or use this one: http://tinyurl.com/bm2lx )

You're using an ancient snapshot of my branch! One of the patches you submitted is already in since the 20050506 snapshot and another one since the 20050515 snapshot, so you must still be using the very first 20050430 one! The current version is 20050522.

Oops. I'll fix that by Monday.

There's a new tiemu-tigcc-debugging snapshot (20050529) up at SourceForge:
* The new Win32 installer is there! (You don't need 7-Zip anymore, you get nice menu entries, and the installer can even fetch and install GTK+ automatically for you (after prompting for confirmation) if needed.) Please try it out. :-)
* fixed grayscale on AI5 (used by Patrick Davidson's games and by genlib)
* fixed several minor debugger flaws
* clicking on the HOME and CATALOG keys of the TI-89 Titanium skin with the mouse now works (fixed by Romain)
* screenshots in PDF and EPS formats added by Julien Blache
* more OS X bugs have been fixed, hopefully enough to make it work, but I'm still waiting for the feedback

The URL is the usual one:
http://sourceforge.net/ project/showfiles.php? group_id=23169 . I've hidden all the old snapshots so you'll have an easier time finding the correct one.

i know that this is off topic but....

does anyone out there know how i can change my basic programmes into flash apps....

thanks

Nameon

Use Basic Builder.

I can't see what's wrong with VTI, why it's "declining in usefulness"?

Well, there are some reasons:
* Emulation of the HW2 and (up to a certain point) the HW3 (yes, we don't emulate the new RTC and USB circuit of the HW3 calculators yet...)
* Emulation of the Voyage 200 and TI-89 Titanium calculators (as mentioned in the newsitem)
* (Optional) emulation of the execution protection
* More precise instruction emulation (prefetch etc.)
* More precise interrupt emulation (AI4 is a timer on VTI! It should actually be triggered on link transfers, which is what TiEmu does.)
* Program entry breakpoints and the handle window work on AMS 3 and on PedroM (The last VTI from Rusty doesn't even support them on AMS 2! Julien Muchembled added AMS 2 support, but it still doesn't work on AMS 3 or PedroM.)
* Slighly better disassembler
* Builtin C debugger (tigcc-debugging-branch only)
* Portability, as mentioned in the newsitem (Windows, Linux, OS X - the latter still being worked on)

TI's SDKs are always handy... although nobody bothers to compare them to programs that can emulate their ROMs. Not that your allowed to copy your ROMs, atleast not in Australia.

You can't copy your ROM in Australia?? It's legal in America.

Are you allowed to use the ROMs that TI offers for download?

TI offers ROMs??? Last time I checked, I could only find link software, OS upgrades, and Flash Apps for my 83+. If they really have 83+ Roms, PLEASE PLEASE PLEASE SHOW ME!

TiEmu and VTI (for the 68k calcs) can emulate from a Flash upgrade. I don't know if any of the TI-83+ emulators can do that. (As far as I know, VTI doesn't support TI-83+ Flash upgrades.)

TilEm has supported app loading for many years, and since 0.970 supports OS upgrades as well. However, you need to have a copy of the boot code -- the last 16k of Flash, which isn't included in the OS upgrade -- to begin with. I haven't looked into exactly how much of it is necessary for normal operation, but you do need at least the initial hardware configuration code.

With that said, if you have a Windows machine handy, you can simply install Flashsim and use the provided CLC files. And for what it's worth, the 83+ SE boot code, with some slight modifications, seems to function properly as an 84+ boot code, so long as you don't try to use the USB.

If anyone wants to take the time to write a simple boot code that could be distributed under LGPL -- or similar permissive license -- I would be exceedingly grateful. I might do that myself at some point.

That's interesting. You can get away with not using a boot code (or almost - the "boot code" faked by TiEmu for Flash upgrades contains just a copy of the initial vectors, and the hardware parameter block, which is a small <30-byte structure) on the 68k calcs.

Yes, in case my comment above wasn't clear enough, if you download and install TI's SDK, you will find the files 'TI83Plus.clc', 'TI83ps.clc', and 'TI73.clc'. These are combined ROM+RAM state files, and you should be able to use them with VTI 3a or TilEm without a problem. VTI 2.5 may have trouble detecting them, in which case you need to truncate them to 512k (note that VTI 2.5 doesn't support the SE, and has poor support for the 73.)

If you have to truncate the files for VTi 2.5, what part of the CLCs do you delete?

Probably the end, since I would think that would be the part with the RAM data, which isn't needed.

"Truncate" means to cut off the end, and in computer-speak it's no different.

Actually, truncate just means "shorten."

You want the first 512k of the file, so you want to remove slightly more than 32k from the end.

and how do you do that?

truncate("TI83Plus.clc", 524288);

Yeah, that works too :)

If you were running a more sensible operating system, you would find that the necessary tool was built in. Since you're not, here are a few options:

Versions of dd for Windows (the command line is 'dd if=TI83Plus.clc of=TI83Plus.rom bs=1k count=512')
* http://unxutils.sourceforge.net/
* http://uranus.it.swin.edu.au/ ~jn/linux/rawwrite/dd.htm
* http://www.cygwin.com/
* http://www.google.com/ search?q=dd+Windows

Free UNIX-like systems you might enjoy:
* http://www.knopper.net/ knoppix/index-en.html
* http://www.damnsmalllinux.org/
* http://www.freesbie.org/
* http://www.google.com/ search?q=live+CD

Roll your own:

#include <stdio.h>

main() {
FILE *i, *o;
char d[1024];
int n;
i = fopen("TI83Plus.clc", "rb");
o = fopen("TI83Plus.rom", "wb");
for (n=0;n<512;n++) {
fread(d,1024,1,i);
fwrite(d,1024,1,o);
}
}

What's the RTC circuit?

Real-Time Clock

RTC = Real Time Clock

Every 1 second it increments a counter and sends an interrupt to the processor. This allows the OS to keep a clock. The RTC was added in HW2, and clock functionality in AMS 2.08.

But the RTC circuit is significantly different on HW2 and on HW3. On HW2, the only counter incremented in hardware is incremented every 8192 seconds, so it is useless in practice. Clock functionality is implemented in software through a timer interrupt (auto-int 3). On HW3, the counting is handled fully in hardware, and auto-int 3 has been reclaimed for USB functionality.

So the processor is woken up for a few dozen cycles every second even when the calculator is off in order to update the clock? That's not the way I would make the hardware. Handling it in hardware would likely lead to longer battery life---not much of a concern, given that during heavy usage my TI-89/HW2 gets about 5 weeks on a charge. It would be interesting to calculate exactly how much energy is used to update the clock.

I emailed you, but chello.at replied that Gmail is on their RBL.

Grrr, now that's really dumb... Some RBLs are listing Chello, now others (or maybe the same one?) are listing GMail... I guess some of these RBL admins really ought to get beaten with a cluestick!

Try mailing me at Kevin at tigcc.ticalc.org, maybe you'll get through that way.

Oh, and why are you e-mailing me in the first place? As I said above, you can submit your patches to the patch tracker on our SourceForge project, where they belong!

I've got some technical questions.

OK, they got through this time.

It also doesn't support the 83+SE or any of the SEs. calcs and version 3.0 isn't done. THey also need support for the Ti-Connect cable. Nothing supports that

Silverlink cable, not TI-Connect.

TI-Connect is the program, silverlink is the cable.

It will never be finished in that form.

I wish the release was retractable after the project came to a crashing halt.

Loss of initial source preevents even a release with the most basic fixes included which were accidentally left out.

corey

Really, the VTi source was lost? If that were true, couldn't you decompile a VTi build?

Decompiling is a LOT harder than you think - it's not like you can just run "easy_decompile.exe" and it'll generate a nice C source file from a program...

If you are looking for a TI8x emulator (including 84+), you should take a look at TiLEM (http://lpg.ticalc.org). This is another open-source emulator for Z80 based hand-helds.

It has been written in C & GTK like TiEmu. It's targetted for Linux but should be portable to Windows without too much pain of ass !

I'd just like to add that a Windows port is mostly done at this point. See the above link; that version is slightly outdated at this point but is mostly functional. (It's also a fairly large download, as it includes the GTK+ DLLs as well.)

An actual release is on its way...

Cool ;-)

Can't open the zip file...WinZip shows it's corrupted.

what about a mac os x port?

pkg-config doesn't like my ibook, so i can't make it myself...

maybe if i didn't need pkg-config to do it it would work...my gtk works (or appears to do so at any rate)

what about for ti-83+?