Re: A92: Severe TI-92 bug

[Prev][Next][Index][Thread]

Re: A92: Severe TI-92 bug


At 12:00 AM 5/24/98 -0700, fondacci christophe wrote:
>Excuse me, I thought everybody knows because I have read it in a book
>(well a french book actually), but I didn't discovered that bug....

Interesting...what is this book called? Who wrote it? Does it mention any
other bugs?

>That's right, but this is not the correct procedure.
>So if some people don't know this bug, I must explain the correct
>process :
>      1- Stopic pict,0,1,0,1 (this is very important)
>      2- Then, you can choose between :
>            Rclpic pict,1,1    (which not crash your calc each time)
>            Rplcpic pict,1,1   (which crash your calc almost every time)

Thank you. That worked. This is definately a real bug.

However the stored pic is NOT a "picture of 1 pixel". It is a picture with
a width of 0 and a height of 1. You aren't even supposed to be allowed to
create a picture with these dimensions. In fact, "StoPic pict,0,0,0,1" will
not work, but for some reason using a non-zero value as the column will
allow it to work. (Remeber, the syntax is "StoPic
VAR[,PXLROW,PXLCOL][,WIDTH,HEIGHT]".)

So, it would seem that RclPic and RplcPic cannot handle pictures with a
width of 0.

I got some interesting results by varying the height:
 StoPic pict,0,1,0,100
 RplcPic pict,0,0
that gave me a neat barcode pattern on the graph screen...didn't crash the
calc, but possibly corrupted memory...(haven't checked yet)

This is definately something to add to my TI-bugs text file! Maybe I will
finally release a new version of it. :-)

>We did the self test, because we were trying to scan this bug and to
>see when and why it happens so we have done several self tests before
>each rplcpic that might have crashed the calc. And one time, it
>crashed the calc, we run the self test (rom) and instead of returning
>to the menu, the calc displays the Test fail error at the top of the
>screen
>
>   I know this is very strange, and there is also another strange
>thing : when he removed the rom module (2.1 and he had a 1.12 int
>rom), and performed the rom test, everything was alright !!!

He should ask TI to replace his ROM 2.1 module. It sounds like it is a bad
unit. Possibly it could have been damaged by the increased bus activity
during the infinite loop that was triggered after the crash. A good ROM
module should not be damaged by this sort of thing.

>   Every time I tried this procedure on my calc, it crashed the calc,
>but the rom was OK...
>
>   He tried to hold [)] and [(-)] but it did nothing.
>He had also tried to remove the 5 batterreies an entire day -> same
>result ...

I guess that would be consistent with having a defective ROM module.


---
David Ellsworth
davidell@earthling.net
IRC: eXocomp
ICQ: 2300673
References: