I'm going to try to see if I can get through Ti-Cares to the developers. (No one's ever successfully done this done this right?)
Here is what I sent:
Spoiler For Spoiler:
OS bug fixes Please pass this on to the developers. I'm part of the developer community for these calculators and we have always wanted to make contact with the creators to help them make the calculators run better.
Occasionally, when trying to ungroup a variable, the user will be greeted with an Err:Version. I have found the culprit of this error and I have even patched it in OS's 1.19, 2.43, 2.53, and 2.55. I was never going to email you guys because I thought these calculators were done, but with the release of 2.55 I can see that that is not true. I would post the code to fix it here, but I'm not sure if this email will even get to the right place. If you reply, I can give you a lot better explanation via email. But the gist is that most of the problems come from when the archive header spans a page boundary. There are two problem spots, the first causes the group to span three pages, which makes it unrecoverable, the second just errors when you try to ungroup it.
Also, sometimes when you go to unarchive a variable, it will give an Err:Bad Address. I have fixed this one in 2.43, 2.53, and 2.55. It is triggered when the data for a variable is on a separate page than the header. It can be remedied by a garbage collect, but your average user doesn't know that.
Like I said earlier, if I can get a response from a developer to this, you will have made the entire Ti community happy. I can also show you the specifics on how to fix these so it will be as simple as copy and paste.
With new OS's from TI come the old well documented errors. I don't know why TI doesn't feel like fixing these. Err:Version was fixed by NOPing some bytes in one area and shifting some bytes in another. Err:Bad Address was fixed by shifting some bytes to fix a beginner asm mistake. But they're fixed.
Also, this patches PTT. Both of the resets require STAT to be pressed as well. So 8+2+STAT and 8+5+STAT. It also patches Press-To-Test to not delete all your precious groups and now UP+DOWN+ON disables PTT.
I'll update this again later to make the RAM clear screen look like the classic ones. But I think I'm done for today.
If you want to make your own changes, or just to see my awesome new patcher , check out the source. (just remove all those z's)
Edit: Oh, btw. This takes ages to run. So don't give up hope. It also has an undo button.
You wouldn't want to do that. The extra RAM page is pretty volatile. The OS kills it in a few spots and a lot of programs kill it too. So it would work, but your programs wouldn't last long. It's especially bad because when you only have one extra page, every single attempt to use an extra page gets redirected to that one, whereas if you had all 6, you could at least rely on page 7 existing for some small period of time.
Well, the old bugs still remain. I guess only the real "bug" is the fact that mathprint can only be so big. The Omnicalc bugs are technically omnicalc's fault.
I just sent it to WabbitEmu. The first thing I noticed is that they expanded the OS onto page $73 (I knew they were out of space). Which means that the $70-$73 sector is now out of bounds to be screwed with. Which in turn, I believe, means that Ti-Boy would corrupt your OS.
Oh yeah, it's a Unicode text file because it uses some special characters like → and ►. I thought stuff like that was saved in a file header or something, but I guess not. Is there a certain way to open a text file in Unicode format that I should tell people about?
Try making the first few characters things that only exist in Unicode. I remember something like this with typing in "bush hid the facts"
Don't forget about ARCHUTIL by Dr.Dnar. It fit's in with the Calcsys method and is way easier assuming your programs are not in the same sector as the corrupted thing.
Also, I have found that removing all of the batteries is not necessary, the ON + battery has always given me a ram clear and saves a lot of time as opposed to leaving a battery out for 15 minutes. And, I think before you reformat the flash chip, you should have people try to reinstall an OS. Reinstalling an OS doesn't delete the archive so they would still have their stuff if all goes well.
Pages $08-$0B ought to be the swap sector, so just $0C-$15 should suffice. However, you should probably dump $08-$0C and $16-$17 as well, and your certificate. That should be enough to reconstruct your ROM (along with $1F from rom8x) as long as your OS isn't corrupted. Good luck!
Btw, you can't dump the certificate without unlocking flash first.
This seems rather unfair. All of the other players (or bots?) all have like 30,000 games. Still very interesting though.
To quote myself, I think I figured this game out. When you stop playing, the game uses your player to fight other real people. It still updates your stats which is why the bots have so many games played and all have at least one stat at 100. Also, the reason that some of the bots have so many games is that the view by rank is broken and shows the same bots every time.
Dr.Dnar actually wrote a program to recover deleted things in the archive. Here it is. If it encounters a problem and says archive corrupted, you, I, or someone can change the code to make it persist. (I remember Dr.Dnar saying that was one problem he didn't fix.)
I have probably recovered about 20 programs with this, so it definitely works.
Show Posts
, check out the source. (just remove all those z's)
