USAG v1.4 for 49 series
Gaak
------------------------
* improved: "PICT" for common commands (GOR, REPL, STO, SIZE,...).
* improved: extended for commands like "FOR", "THEN, "WHILE",...
* added: input supports command, {command}, "command", {"command"}.
* added: support to ALG mode.
* added: key [+/-] and menu [LANG+].
* added: support to 48gII.
* removed: support to current font.
* removed: freeze for some commands.
- Gaak -
Robert Tiismus
Thanks for the great little utility!
This last one (1.4) has on my v.2.10-7 HP50G an unfortunate side effect
:( When using it on single type argument command (like CLKADJ or
->KEYTIME), and pressing afterwards any key, the calculator
warmstarts... I downloaded the 1.3 from hpcalc.org for checking. 1.3
runs flawlessly.
Robert Tiismus
ZZ
Gaak
Uhm?, not for me. :-(
Testing on ROM 2.09 (50g) and ROM 2.10-7 (Emu-49G) runs flawlessly.
Commands like CLKADJ and ->KEYTIME are commands with common
structure...
More details?:
- Input mode (remember that ALG and RPN are supported now).
- Command mode (lists, strings, command).
Thanks for report bugs ;-)
- Gaak -
Robert Tiismus
Thank you for asking! I tried again. I did "ON-A-F NO", then turned on
the RPN mode, downloaded the program and tried again:
1: { CLKADJ }
usag.hp
It displayed the correct information, but when I pressed ANY KEY, I saw
very spectacular random pattern on the screen and "Try To Recover
Memory?" afterwards. "No" was unresponsible but "Yes" functioned.
Is it possible that my unit is faulty? I got it only half a year ago and
have not had much time to use it.
Robert Tiismus
Gaak
> Thank you for asking! I tried again. I did "ON-A-F NO", then turned on
> the RPN mode, downloaded the program and tried again:
> 1: { CLKADJ }
> usag.hp
>
> It displayed the correct information, but when I pressed ANY KEY, I saw
> very spectacular random pattern on the screen and "Try To Recover
> Memory?" afterwards. "No" was unresponsible but "Yes" functioned.
>
> Is it possible that my unit is faulty? I got it only half a year ago and
> have not had much time to use it.
>
> Robert Tiismus
Maybe another buggy object?... libraries with bad $CONFIG?...
Please, delete your 'STARTUP' variable from HOME and after Warmstart
press and hold [BS]... and retry again USAG.
Sorry for the inconvenience :-(
- Gaak -
Robert Tiismus
I erased all libraries, did again "ON-A-F NO". Bu still the same
behavior... Then I downgraded ROM to version 2.09 (from HP site), and it
behaved! The bug was back after upgrading to 2.10-7. I think it is
somehow related to LANGUAGE selection, because after every crash, the
"LANGUAGE->" command returns "5.". I tried to look into program, to try
to debug it, but it was inside CODE object. Is it this way for
aesthetical reasons? That this way the stack is not cluttered with
EXTERNAL's. Anyhow, I got it disassembled by this:
<< ->H DUP SIZE 33 SWAP SUB "D9D20D9D20" SWAP + H-> ->s2 >>
But could not find debugger like Jazz library had for HP48 series. So
again no avail :(
Robert Tiismus
Gaak
Good news for you... your calculator is ok, don't worry ;-)
Upgrading my 50g to 2.10-7 and debugging (for you), I found the bug,
many thanks.
The Flash Pointer 7 8E (aka ^WRITEMENU) available on "all" ROM
versions, is wrong on 2.10-7.
^WRITEMENU (2.10-7)
::
DUPTYPEREAL?
case
( ... )
;
^WRITEMENU (another version like 1.18, 1.19-6, .. 2.09)
::
ZERO SIX
ZERO_DO
( ... )
;
^WRITEMENU is called for commands that requires "One Argument" (-
>KEYTIME), "No Arguments" (MEM), and metas (->LIST).
Obviously, ROM 2.10-7 is beta :-S
Sorry for the inconvenience, again... :-(
- Gaak -
Robert Tiismus
>
> Good news for you... your calculator is ok, don't worry ;-)
>
> Upgrading my 50g to 2.10-7 and debugging (for you), I found the bug,
> many thanks.
> The Flash Pointer 7 8E (aka ^WRITEMENU) available on "all" ROM
> versions, is wrong on 2.10-7.
>
> ^WRITEMENU (2.10-7)
> ::
> DUPTYPEREAL?
> case
> ( ... )
> ;
>
> ^WRITEMENU (another version like 1.18, 1.19-6, .. 2.09)
> ::
> ZERO SIX
> ZERO_DO
> ( ... )
> ;
>
> ^WRITEMENU is called for commands that requires "One Argument" (-
>> KEYTIME), "No Arguments" (MEM), and metas (->LIST).
> Obviously, ROM 2.10-7 is beta :-S
>
> Sorry for the inconvenience, again... :-(
>
> - Gaak -
I already realized that 2.09 is the way to go and downgraded the ROM
version. Thank you! And don't be sorry, Mea Culpa for use of the
uncertified ROM version!
Robert Tiismus
Veli-Pekka Nousiainen
is there a real+beta ROM version compatible
1,4,1 USAG available somewhere?
thanks Gaak
"Robert Tiismus" <rob...@example.com> wrote in message
news:g84oqi$q5e$1...@aioe.org...
John H Meyers
> The Flash Pointer 7 8E (aka ^WRITEMENU)
> available on "all" ROM versions,
> is wrong on 2.10-7
There's more to it, because there was not just one "2.10";
there must have been a "regular" ROM (never released by HP),
and there was an experimental "Geometry/Spreadsheet" ROM
(made for math teachers in France?)
which was apparently derived from that,
and you are probably using the latter.
The problem appears to be that some of the FPTR "slots"
in bank 7 seem to have been "borrowed" (or re-arranged)
in the "Geometry" ROM, no longer matching their original functions
(or at least not matching what "extable" says),
while the "regular" orignal 2.10 ROM was no doubt actually fine.
I think you would have to ask BP about how that ROM was created,
to find out how these substitutions occurred,
and what else might be affected by them;
for example, was there not enough room to add Geometry+Spreadsheet,
and were other original functions (or just their bank table pointers)
replaced or re-arranged instead?
> NOT_WRITEMENU (2.10-7G? wrong function in bank table?)
(at 7:42797 ? in Geometry ROM)
> ::
> DUPTYPEREAL?
> case
> (...)
( clearly CAS symbolic code here, called from "symbolic" functions )
> ^WRITEMENU (another version like 1.18, 1.19-6, .. 2.09)
(at 7:461C0 ? in "regular" 2.10, called from ^MENUext and other sensible places)
> ::
> ZERO SIX
> ZERO_DO
( this is correct code )
> ;
> ^WRITEMENU is called for commands that requires "One Argument"
> (->KEYTIME), "No Arguments" (MEM), and metas (->LIST).
> Obviously, ROM 2.10-7 is beta :-S
It would seem that only the altered "Geometry" version has been rearranged,
evidently so obscurely that it doesn't seem to have been noticed before.
Evidently _you_ call ^WRITEMENU (using its original FPTR 7 8E)
from USAG, although the commands themselves (in ROM) do not;
perhaps you could choose another function
(or copy the original code) to become free of this dependence,
and thus to be compatible even with the "Geometry" ROM?
It would be nice if the very caring HP (new?) management
would care to either post the improved "regular" 2.10 ROM themselves,
or let someone else post it "unofficially,"
like 1.19-6 for old 49G, which was never acknowledged by HP,
but which was clearly improved over HP's "official" 1.18
Otherwise the altered "Geometry" version of 2.10
might remain the only one found,
and might prove incompatible with some "bank 7 dependent" SysRPL/ML,
just as apparently in this case.
[r->] [OFF]
Veli-Pekka Nousiainen
HP (Cyrille) just steel one more bank to the system
and/or
even another one for the extable 2
50g+ could have twice the flash and 90 MHz speed
and -5V/+5V real RS-232C
plus full HPGCC compatibility
"we can patch it up, baby"
VPN
"John H Meyers" <jhme...@nomail.invalid> wrote in message
news:op.ufyc1...@miu.edu...
username localhost
On Aug 15, 8:02 pm, "John H Meyers" <jhmey...@nomail.invalid> wrote:
> On Fri, 15 Aug 2008 13:03:32 -0500, Gaak wrote:
> > The Flash Pointer 7 8E (aka ^WRITEMENU)
> > available on "all" ROM versions,
> > is wrong on 2.10-7
>
> There's more to it, because there was not just one "2.10";
> there must have been a "regular" ROM (never released by HP),
> and there was an experimental "Geometry/Spreadsheet" ROM
> (made for math teachers in France?)
> which was apparently derived from that,
> and you are probably using the latter.
>
> The problem appears to be that some of the FPTR "slots"
> in bank 7 seem to have been "borrowed" (or re-arranged)
> in the "Geometry" ROM, no longer matching their original functions
> (or at least not matching what "extable" says),
> while the "regular" orignal 2.10 ROM was no doubt actually fine.
>
> I think you would have to ask BP about how that ROM was created,
> to find out how these substitutions occurred,
> and what else might be affected by them;
> for example, was there not enough room to add Geometry+Spreadsheet,
> and were other original functions (or just their bank table pointers)
> replaced or re-arranged instead?
I just checked.
Bank 7 contains 13 commands, and then the contents of library 222.
In the geometry rom, somehow after the FTR 7 C (the 13th command)
apparently 22 garbage entries got inserted into the the bank table.
This messed up all later bank 7 flash pointers (so FPTR 7 8E became
FPTR 7 A4, etc) BP should probably be able to fix this.
The issue did not impact any code compiled into bank 7 because the
compiler/linker replaced the symbols with the flashpointers that had a
slot incremented by 22 compared to the normal location.
Unfortunately, the extable/extable2 use the original flashpointers, as
do any existing programs that used those flashpointers.
BP, could you please generate a new rom that does not have those bogus
entries (or if they are not completely bogus that have those entries
in slots at the end of the banks' flashpointer table)? Thanks.
Oh, BP, I also have a few unrelated things to ask.
I presume you must have had access to what would be the official 2.10
if HP had released it (if you did not have access to it, that would
make basing your geometry rom on it a bit difficult i would imagine) .
Any chance you you leak it? No? Ok. Just wanted to ask. :-D
(BTW, the posted cas2user.s file for the geometry app was clearly not
used to compile the posted rom, as CASRESERVED62 and CASRESERVED63 in
the posted rom are RCLVX and STOVX. That means that two of those
pseudo-userrpl commands found in CASRESERVED64-99 must have been
removed. Which two where they? (Just curious).)
Gaak
Version 1.5 [18.08.2008]
------------------------
* improved: language correction and optimization.
* improved: "¿ ?" -> "Undefined Name" (with language support).
* fixed: bug ROM beta 2.10-7 with ^WRITEMENU.
Cheers.
- Gaak -