Rpl emulator new update

[ji3m]

(05-24-2017 11:54 PM)mfleming Wrote:  I take it for granted from your various posts that you rely on the underlying Prime functions rather than writing your own routines to match those in the HP48. I thought going through the Owners Manual would be a more disciplined approach to just banging on the keys to see what happens At any rate, it would result in me trying all of the HP48 functions, skipping the unimplemented graphics commands.

<Your post does raise a question though: Do you expect the Prime to be in CAS or Home mode when starting the simulator, or does it matter? >

>>It doesnt matter at all. When I drop out of RPL i use both windows to help >>debugging. Other than this pgm I dont use the prime since it
>>doesnt have a sensible rpn mode. Just use my go48g+ for simple things.
>>I see i implemented some graphics ops but i can't remembet how they work.
>>Thats for another day.


At any rate, here's some feedback using the 22May2017 version of the software on the physical Prime using the Owners Manual, Chapter 1 for guidance.

*Functions UPPER and LOWER work correctly
>> Tim W. pointed out that the bug seems to be just in the Android version. I fixed >>myUpper and myLower to work regardless by detecting the error.

*The ON key doesn't cancel a command line
>>The ESC cancels command. ON breaks the simulation since
>> its a Prime meta key

*Shift-Clear clears the stack
>> As designed

* SHIFT-Sto with value and variable name on stack works. An alpha string entered without quotes is assumed to be a variable name if it doesn't match a command name (The HP48 wants single-quotes around a name; not necessary here, I think)

>> The 48 puts ticks on unbound ids. Here they arent necessary.
>> RED shift , variable stores to tos. BLUE shift var recalls it.

*Can press SHIFT-singlequotes key then a variable softkey to put the variable name in quotes for the purge command
>> This is the desired method to get a quoted id just likev48

*The various RPN examples and quoted arithmetic expression examples on pages 1-18 to 1-20 work
>>what doc?

*Algebraic expressions are in exact format on the stack. The approx function (from D-key menu) or the a b/c key will convert to the decimal result
>>and BLUE shift executes 'exact (tos)'

*In algebraic mode, the X-squared key returns what looks like 2-space-caret rather than caret-2. The key works fine in immediate mode

>>I was trying to fix this today. I thought i had a flag indicating
>>algebraic entry but i cant find it. So it works for rpn. But very mysteriously
>>If you type 'x 2^' then hit enter 'x^2' appears on the stack. Like magic.
>> All algebraics are kept internally in rpn format (hit the NUM key to see that).
>>The '...' form is just for viewing so 'x 2^' gets "fixed".

*The Define function (SHIFT-D) doesn't map to the HP48 DEF command
>> Many keys are not implemented. But define is unnecessary since the form
>> <| -> ... '...' |> 'nam' sto does the job.

I've yet to go through all the function keys on the keypad, but I had no problems running through the introductory examples (other than HP48 functions with no Prime equivalent). I've worked a little bit with variables and directories, and will probably focus more on that next.
>>I hate there is no "prev" key so i have to walk arod the directory.

The code is still a bit fragile. I've seen mistakes do everything from bounce me into the debugger to resetting the Prime (fun!) More later...

>> the error handling is poor since you get caught in error loops and have to
>> restart. Todays version is a lot better (see new attacment for 25may above)
>> The nasty errors occur in decompile for the stack display.
>> The cheap fix today was to drop the top stack element when i trap
>>an error. This has calmed things down quite a bit. Also switching display to intetnal >>mode (num key) stops the decompiler and associated errors.

On the "reseting the prime" issue there is no way my code can do
that. But its a large complex program and the prime itself is very unstable.
The android version crashes frequently no matter what im doing.

~Mark

I haven't done anthing with the code since last year but i'm remembering
a lot in the last few days.

Im not tying to reproduce the 48-50 with regard to specific functions but just to have a true rpn-algebraic cas with RPL style code. Most of the cas functions
can be user rpn or algebraic style except the ones with variable number of arguments.
Some can work with a list of args rpn style.