Parsing mathematics seems to be problematic.

[Devilmarkus]

Hi coders!
I want to know the "real wanted result" of a simple ASM function:

Code Select Expand

DS &3318-&3301+4*14

I used 3 emulators with 3 different results!!! 

WinAPE:

Code Select Expand

WinAPE Z80 Assembler V1.0.13

000001  0000  00 00 00 00   DS &3318-&3301+4*14
        0004  00 00 00 00
        ....
        0174  00 00 00 00
        0178  00 00

378x "00" !!!

WinCPC:

Code Select Expand

000001 8000  00 00 00 00                        ds      &3318-&3301+4*14
000001 8004  00 00 00 00
....
000001 8040  00 00 00 00
000001 8044  00 00 00 00
000001 8048  00 00 00 00
000001 804C  00 00 00   

79x "00"!!! (most exact because * should be calculated first but why 79?)

JavaCPC:

Code Select Expand

000001 0000  DS &3318-&3301+4*14 
                        00 00 00 00
                        00 00 00 00
                        .......
                        00 00 00 00
Assembled in 0.047s from &0000 to &0126 length is &0126

294x "00"

(When I enable exact math parsing I get 73x "00")

So which is the correct value now???

Please help...

[Octoate]

The correct value is - of course - the mathmatical one. Which means multiplication before addition / substraction.
This means that the result is DS &3318-&3301+&38 = DS &4F.

[Devilmarkus]

The correct value is - of course - the mathmatical one. Which means multiplication before addition / substraction.
This means that the result is DS &3318-&3301+&38 = DS &4F.

Maxam creates this: (like WinApe)

but why?

[Devilmarkus]

Got it now....
My new parser had too much old code inside
Bug was: all & were removed in String before parsing it... 

Now I get the same results like WinApe (Or if I use exact maths I get the same result than WinCPC)

[Executioner]

The correct value is - of course - the mathmatical one. Which means multiplication before addition / substraction.
This means that the result is DS &3318-&3301+&38 = DS &4F.

That's only correct if you have operator precedence, which Maxam doesn't, and therefore WinAPE doesn't for compatability. All Maxam/WinAPE expressions are evaluated left-to-right regardless of operator.

[Devilmarkus]

That's only correct if you have operator precedence, which Maxam doesn't, and therefore WinAPE doesn't for compatability. All Maxam/WinAPE expressions are evaluated left-to-right regardless of operator.

I have this optional.
Left-to-right is "default" and also operator predence + other operations, too when precision is true in my parser.
View Source

[Devilmarkus]

Thanks to the improved parser I can compile most files properly now...
Example screenshot:
Assembled original source for "Synergy II" demo

[Devilmarkus]

I build a Scientific calculator in java as applet:

Scientific Calculator Applet

Have fun....
(It uses the math parser I made for JavaCPC)

[Gryzor]

Hmmm! Can you compile it to run as an exe? I'm having troubles here at work...

Any screenies? If it's CPC-themed it'd be super fun to use it at work!!!

For the time being I'm looking for a Sharp PC14xx programmable calculator (since they broke into my car and stole the last one), anyone might have one they wanna part with?

[Devilmarkus]

You can download the .jar file and double-click it... should work.
Or, if you don't have java installed, you'll get no cookies.

[Screen for Gryzor]

[Gryzor]

Heh It looks cute, but it ain't no CPC!

I did try executing the jar (duh), but it doesn't work - it's as if the Java bin is not in the system paths for some reason... weird.

[Devilmarkus]

Heh It looks cute, but it ain't no CPC!

I know. But also if it dont look like a CPC , the sun will keep rotating around the earth

Edit: Sorry, I forgot, the earth is like a disc

[Devilmarkus]

I hope you like my sarkasm...
In fact, the calculator was only a tool I coded to test my parser.

Then I thought "Hey this could be useful as single app"
(I'll also add it as JavaCPC Desktop tool)