[Featured][Dev] PindurTI (Best Thread 2005)

Post by **qarnos** » Fri 14 Jul, 2006 10:21 am

AndySoft wrote:Then how do you send a bit to begin with?

Just alternate one of the lines. Or you could "handshake" first by raising/lowering both lines or something like that. Since both calcs "know" nothing has been sent yet, it should be fine.

Liazon · Post by **Liazon** » Fri 14 Jul, 2006 12:02 pm

You mean stuff gets sent only bits at a time? Then how long does it take to send a byte?

Post by **qarnos** » Fri 14 Jul, 2006 12:16 pm

Liazon wrote:You mean stuff gets sent only bits at a time? Then how long does it take to send a byte?

Stuff always get sent one bit at a time...

EDIT: Let me elaborate.

You can only ever send one bit at a time on the calcs, regarless of the protocol, since the lines can be either "high" (1) or "low" (0).

My idea allows you to:

a) let the receiver acklowledge receipt of the bit which was sent; AND
b) send a bit of it's own riding on top of the acklowledge signal.

The basic idea is, we have two lines, allowing a possible 4 different states we can be in. So why waste any of those states?

Timendus · Post by **Timendus** » Sat 15 Jul, 2006 11:00 pm

You mean like:

toggle tip status = 0 and ack
toggle ring status = 1 and ack
toggle both lines = ack and send nothing

That looks a lot like a less secure version of Ti's protocol. And besides, you can't use low level acknowledgements if you want to link with more than two calculators.

Post by **qarnos** » Sun 16 Jul, 2006 2:39 am

Timendus wrote:You mean like:

toggle tip status = 0 and ack
toggle ring status = 1 and ack
toggle both lines = ack and send nothing

That looks a lot like a less secure version of Ti's protocol. And besides, you can't use low level acknowledgements if you want to link with more than two calculators.

I'll take your word for it - this is not my area of expertise!

CoBB · Post by **CoBB** » Sun 16 Jul, 2006 6:47 am

Guys, could you move this discussion elsewhere? I always come here expecting some PTI related post.

Post by **qarnos** » Sun 16 Jul, 2006 7:29 am

CoBB wrote:Guys, could you move this discussion elsewhere? I always come here expecting some PTI related post.

Actually, I meant to ask something but I forgot all about it! Sorry about that.

I was wondering if you have any plans to make some kind of profiler for the calcs? That's one thing I would *love* to see but is sadly not present in any emulator I have seen.

CoBB · Post by **CoBB** » Sun 16 Jul, 2006 10:18 am

The idea has been brought up already, even if in a bit more elementary form, on a per-instruction basis. With some support from assemblers it could be made even better. There will be something for sure.

CoBB · Post by **CoBB** » Sun 23 Jul, 2006 1:02 pm

Okay, it seems that the long standing bug with sending archived files has never really existed. Simply sending those two magic bytes after the name seems to work perfectly by itself. The problem is elsewhere, most probably connected to force loading apps and hence messing up the flash. When a ROM image has this error, even on-calc archiving fails to work, giving an 'ERR: BAD ADDRESS".

Does the same happen with TilEm?

KermMartian · Post by **KermMartian** » Sun 23 Jul, 2006 3:24 pm

I have a request. I've noticed that for some reason PTI does not show the calc's serial number/ID on the "about" screen, even if I've specifically dumped the ROM from my calculator. Does this mean it doesn't have the serial number embedded in the rom, or that PTI just doesn't show it? I need the calculator to know its serial number for a program I'm working on.

pacHa · Post by **pacHa** » Mon 24 Jul, 2006 10:01 am

I thought it was in the hardware of the calc, and thus not in the ROM...

CoBB · Post by **CoBB** » Mon 24 Jul, 2006 11:49 am

There's definitely no way to access any such information that's not contained in the ROM in the case of the 82 and the 83. I don't know about the pluses.

KermMartian · Post by **KermMartian** » Mon 24 Jul, 2006 1:40 pm

CoBB: But suppose I have the information from my calculator derived from the [2nd][+][1] screen? I looked around on DetachedSolutions and they found that the ID number is stored encrypted in MD5 on one of the privileged flash pages. I also know for a fact that ROM images contain that page, since I had to dump the page from my calc in order to create my ROM from an OS upgrade with Rom8x. Anyway, even if there's a way that the user can type in the number and have it be used, that would be awesome. If not, I could always improvise by interrupting my program and manually setting the RAM where it stores the ID retrieved using _GetSerial. Oh, and are there any plans for editing the RAM via the hex display?

Post by **tr1p1ea** » Mon 24 Jul, 2006 2:31 pm

Well regardless of whether or not the information is stored on a flash page, it still must be derived from hardware to begin with.

KermMartian · Post by **KermMartian** » Mon 24 Jul, 2006 3:12 pm

Ah, very true. Well, I suppose that would have to be worked out then. But wouldn't dumping the ROM include that data? Or perhaps on a RAM restart, the calc rereads the hardware to get the information, hence why it's not working in PTI?