PolyPatch84 released

[DJ Omnimaga]

Critor has released a package called PolyPatch84, containing every TI-84 Plus OS patches made so far by the TI community. He plans to update it everytime a new one will be released. The various patches made so far were mostly to correct some bugs by TI, disable teacher modes and application signature checking. This includes

-For glitch fix: When using a lone If instruction inside a For() loop with the closing parhentesis missing and the If condition being false, the For loop ran much slower than normal and this slowdown sometimes persisted after exiting the loop, due to a memory leak, and persisted until you exit the program with ON. This affects every OS. Brandon Wilson made a patch fixing this in OS 2.41. PolyPatch84 adds 2.40 and 2.43 fixes.

-ERR:BAD ADDRESS fix: This bug occured when archiving a file in a Flash page with only 11 to 18 bytes left in it. Upon unarchiving it, the file was not archived at all, but it would take RAM anyway. Multiple attempts at unarchiving it caused your RAM to decrease more and more. The only way to recover this RAM was a RAM reset. This could cause potential data loss, especially for BASIC and Axe programmers. This affects all 15 MHz calculator models including the TI-83 Plus Silver Edition. The patch fixing this was made by ThePenguin77 and  is compatible with OS 2.43 and 2.53MP. PolyPatch adds compatibility for all other 84+ OSes including 0.46.

-ERR:VERSION: This bug occurs when you try to group a large file and the current Flash page is almost full. A bug will cause the grouped file to be saved on two Flash pages at once and prevent you from ungrouping files that crosses a sector, if not the entire group at all. ThePenguin77 patch fixes this in OS 2.43 and 2.55MP and PolyPatch84 adds support for other 84+ OSes.

-Self-test group deletion fix: MODE+ALPHA+S is supposed to clear the RAM, but not the archive. However, a bug in all TI-84 Plus OSes causes group files (8xg) to be deleted too. Critor added 2.41 and 2.55MP OS support to ThePenguin patch, which supported 2.43 and 2.53MP.

-Sequential graphing bug: on older OSes, the X value prompt on the graph screen is corrupted and will show garbage. BrandonW made a patch that fixes this in OS 2.21 and 2.22. PolyPatch84 adds 0.46 support.

-xLIB incompatibility: For people who still use xLIB, it will not run on OS 2.53 MP or higher. BrandonW made a patch that fixes that compatibility issue.

-Teacher reset shortcuts: In newer 84+ OSes, there are shortcuts for teachers (ON+8+5 and ON+8+2) which lets them quickly reset student calculators. BrandonW patch disable those.

-Press-to-test (PTT) mode: By OS 2.40, TI added another teacher mode to the TI-84 Plus series that disables the PRGM and APPS menu as well as various other functions on the calculator during tests. To get out of this mode, the student must connect his calculator to another one and send a variable to it, else he can say goodbye to his calculator content. Mapar007 originally released a patch on ticalc.org to disable this, but BrandonW released another one for all OSes that adds a fake PTT screen too.

-Application signing: To install an application on a TI-84 Plus, it needs to be signed with the 512 bits 010A RSA key first. Unsigned apps and those that are either not signed properly or are for the TI-73 will not be accepted by the calculator. BrandonW made a patch that removes this protection.

Eventually, TI-83 Plus patches might be included. Remember to backup everything you have before attempting to use any of these patches and read the instructions first (which are included in English)!

PolyPatch84 package can be downloaded on TI-BANK here and the original news is located here.

[Munchor]

This looks cool, so we don't have to get all of them separately

EDIT: Wow DJ, nice translation, that must've been some hard work, thanks!

[DJ Omnimaga]

Lol thanks, I couldn't find the original patches anymore so I just translated and summed up the news. I also mentionned about Mapar007 original PTT patch.

[thepenguin77]

Crosspost from Ti Bank. (I know DJ just translated so this isn't as important.)

Sorry I don't speak French. So you'll get the google translate version of this.

On the Err:Bad Address, the file gets saved, but it just cannot be unarchived.

Also, the groups are deleted as part of the Press to Test. When you disable the programs, that is when the groups are deleted, not when you do test mode. And error version makes the group span three pages, not two.

But other than that, great job compiling all of the patches.

[critor]

Hi! Thank you for the topic.

Please report any abnormal bug/problem.


If you think I've forgotten to credit somebody in the news or in the doc, please tell me.

I don't want to "steal" the work belonging to others. I'd just like to make it more popular and easier to use for most of us.

[leafy]

Hooray I can finally play Shift!

[critor]

By the way, thank you ThePenguin77 for telling me what was incorrect.
I'll update/edit as soon as possible

Thanks again.


Some note about the 1st post.

Brandon's PTT patch doesn't work fully with OS 2.55MP. I does disable the 8+5+ON/8+2+ON resets, but not the access to the PTT mode.
The fix was adapted for OS 2.55MP in PolyPatch84.


As you can see, severall ports are still missing. It's because the problematic code was different on those OSes. Either I could not find it, either I could not patch it.

* Brandon's patch for the "for(" glitch still has to be ported to OSes 2.55MP, 2.53MP, 2.30 and lower. I could find the code, but it was slightly different and I couldn't manage to mod it successfully.

* The fix for the self-test & groups has to be ported to OSes 2.40 and lower. I could not find the code at all.

* The "ERR:Version" fix has to be ported to OSes 2.30 and lower. I could not find the code at all.

* Brandon's fix for the Xlib bug has still to be ported to OS 2.55MP. I could find the code, but the ROM page is completly full! I cannot insert Brandon's additional code...


If anybody wants to help...

[DJ Omnimaga]

Ah ok thanks for clarifying guys! Good luck porting the remaining patches for 2.55 MP.

[jnesselr]

I'm surprised at how many patches we've needed to use.

[DJ Omnimaga]

One thing I wonder: Is it safe to install every patch at once? Will conflicts occur?

[jnesselr]

One thing I wonder: Is it safe to install every patch at once? Will conflicts occur?

I doubt it.  Unless two patches try and modify the same piece of code, it should be fine.

[critor]

I didn't have to mod the same code twice in different ways.

Most patches are independent and should not trigger conflicts (PTT has nothing to do with the for loop for example).


For the 3 patches related to groups & memory, I don't know.
I had no problem during the tests, but of course that does not prove anything and only a long term use can tell...


That's why I asked to report abnormal crashes.

For exemple, it's "normal" for the 2.53MP / 2.55MP to crash sometimes when being asked for 1+1 (not joking)

With OS 2.43 and lower, that would be "strange"...

[jnesselr]

I didn't have to mod the same code twice in different ways.

Most patches are independent and should not trigger conflicts (PTT has nothing to do with the for loop for example).


For the 3 patches related to groups & memory, I don't know.
I had no problem during the tests, but of course that does not prove anything and only a long term use can tell...


That's why I asked to report abnormal crashes.

For exemple, it's "normal" for the 2.53MP / 2.55MP to crash sometimes when being asked for 1+1 (not joking)

With OS 2.43 and lower, that would be "strange"...

I installed a TI-OS. Does this count as a bug?

No, but seriously 1+1 crashes? how?

[critor]

That's why I asked to report abnormal crashes.

For exemple, it's "normal" for the 2.53MP / 2.55MP to crash sometimes when being asked for 1+1 (not joking)

With OS 2.43 and lower, that would be "strange"...

I installed a TI-OS. Does this count as a bug?

No, but seriously 1+1 crashes? how?

OSes 2.53MP/2.55MP have severall bugs. Only one of them (the Xlib bug) is identified and patched for now.

They are not stable and you don't need assembly to crash a 2.53MP/2.55MP OS.

So yes, it can crash during any calculation (including 1+1) or input.

For example, they seem to crash during calculations when the RAM is running low. But there may be other non-asm ways to crash them.

[Happybobjr]

mine crashes the most with the use of sigma.