Re: A92: Recovering from Crashes on the 92+
[Prev][Next][Index][Thread]
Re: A92: Recovering from Crashes on the 92+
At 23:12 1998-08-25 -0400, you wrote:
>
>Might it be possible to write a program like that that you could run just by
>having the exec jsr to the beginning of the program, and then rts back
when it
>was done? I don't see why you would be unable to call a program like that in
>the User Data Archive, as it's on the ROM, and you can call ROM functions.
>But I haven't done anything at all with assembly on my 92+ (too busy with
>other projects, mostly), so I could be totally wrong about this.
Data in the Archive can't be accessed directly, it has to be copied into
the real RAM before you can use it. The 92+ has the processor set up to
access 256Kb of RAM, and must therefore use I/O ports to read from/write to
the archive. Someone please prove me wrong on this one ;-)
What TI should have done is to make the archive a folder which could
contain its own set of sub-folders (kinda like a second "main") with a VAT
entry stored in the archive. This could have a fixed starting point in the
archive, or a fixed point containing a pointer to it. A more transparent
(and slower) solution would be for the OS to mix these two when showing
them to the user in f.x. the VAR-LINK window, making the archive more
integrated with the real RAM. Of course, they may still do this - that's
what the FLASH ROM is for, right? The archive resetting issue was one of
the complaints the beta testers had (according to an email reply I got from
TI-CARES)...
Bottom line: if we're ever going to be able to have anything like an
on-calc assembler or any such complex software on the calc we have need a
way to retain some data on the calc (in this case sources and the assembler
itself) when the assembled code crashes.
Niklas Brunlid - http://www.efd.lth.se/~e96nbr
PQF Quote follows:
- "Outside! What's it like?"
- "Well -- It's sort of big"
-- (Terry Pratchett, Truckers)
Follow-Ups:
References: