Spin2

Rayman · 2020-02-11 14:30

Sorry, one more thing... Is there a list of the Spin2 operators somewhere?

I thought that "=>" was going to change to ">=", like C++, but I think Fastspin didn't make this change, so I wonder about Spin2...

Also, I'm getting an error on this:

value := ||(value+x)

So, I think absolute value operator has changed...

I found the 2018 proposed operators here: https://forums.parallax.com/discussion/166416/spin2-operator-syntax/p17
Seems || is now ABS...

Rayman · 2020-02-11 14:37

BTW: Maybe it is a good idea to keep variables separate...
After changing "x" to "x99" in my assembly code, I got an error here:

PUB  CreateP1TvColors()|i,x,p,y,r,g,b 'create colors used by graphics demo, starting at index #128
    repeat i from 0 to 14
        x:=((i+4)&$F)*$10 + $C 'mid brightness colors that start and green and run through yellow, red, purple, blue and then back to green

        repeat j from 0 to 256

The variable "j" was being used without being declared as a local variable.
So, I guess it was using the long declared in the assembly block.

Actually, once that assembly driver is started, maybe this is perfectly fine.
Perhaps I should have gone the other route and just not defined i,j,k,x,y etc as local and just left them in the assembly code

ersmith · 2020-02-11 14:41

Rayman wrote: »
This syntax is not recognized by Fastspin:
    pinclear(av_base_pin_ addpins 4)
Wouldn't that be better with commas?

"addpins" is one of the new operators, and it's recognized by fastspin 4.1.3. "av_basepin_ addpins 4" is the same as "av_basepin_ | (4<<6)". This may be a case where having a lot of operators obscures what's going on.

Rayman · 2020-02-11 14:51

I'm glad to see chip using 297 MHz clock in these demos.
That's what I'm using, but I spent an hour or so deliberating whether or not this is a good idea...
I might have to take a heat gun to it and see if this is going to work if the AC goes out during summer time...

Rayman · 2020-02-11 15:13

Is "result" no longer a Spin keyword?
I copied over a "DEC" routine from Spin1 that has this in it:

      result~~                                          'flag non-zero found
    elseif result or i == 1

Gives me an error in Spin2....

Rayman · 2020-02-11 15:21

Yeah! I got my 1551 ling long driver file to compile.

Now, I have to work on the top file...
Spin2 does not like how I include "smartserial" in my OBJ section:

OBJ
    VGA : "1080p_TileDriver_3n.spin2"  'Note:  You must set the fclk constant in this file to match _clkfreq
    gr  : "Graphics2b.spin2"
    usb1: "OneCogKbM_rja2b"   'Note:  You must change basepin and _FCLKFREQ settings in this file to match your setup
    ser: "spin/SmartSerial"

It also doesn't like me putting the ".spin2" in the filenames...

Rayman · 2020-02-11 15:37

Is there no CogNew in Spin2?
I get an error on this:

okay := cog := cognew(@loop, @command) + 1

I see the example does this:

okay := cog := coginit(16, @driver, @background0_) + 1

I guess the "16" means use the next free cog?

JRetSapDoog · 2020-02-11 15:59

Rayman wrote: »

Sorry, one more thing... Is there a list of the Spin2 operators somewhere?

You've probably found them by now, but I saw a .txt file attachment with the operators in Chip's 2020-2-9 15:28:40 post a couple pages back. And I think that file is in the latest zip file, too.

Rayman · 2020-02-11 17:50

Thanks, I think Chip just added this and some other things to the latest .zip file.

Does spin2 support preprocessor commands?
I'm getting an error in garryj's USB code on this line:

#ifndef USB_PORTB ' PortA is the default.

Rayman · 2020-02-11 18:10

Minor note: FastSpin will let you redefine a constant later on but Spin will not...
I had two instances of this:

USB_V2MODE     = %1_11011_0 + 1 << 16 ' In P2 silicon v2 all USB smart pin modes are consolidated to %11011.

This might be trouble:

' USB CRC constants:
        USB5_POLY      = %0_0101 >< 5           ' USB5 polynomial is reflected when calculating CRC
        USB5_RESIDUAL  = %0_1100 >< 5           ' Expected CRC5 residual value when checking received data
        USB16_POLY     = $8005 >< 16            ' USB16 polynomial is reflected when calculating CRC
        USB16_RESIDUAL = $800d >< 16            ' Expected CRC16 residual value when checking received data

Hope REV is same as ><

Rayman · 2020-02-11 18:20

I'm not sure why I'm getting this error in the USB code (attached).
I guess you can't use . labels in hubexec?

Strange... I replace the DJNZ with sub wz and if_nz jmp and its ok with that. These are the same, right?

                'djnz    htmp2, #.loop  'PNUT doesn't like this?
                sub     htmp2,#1 wz
        if_nz   jmp     #.loop

ersmith · 2020-02-11 18:21

REV is not the same as ><. "x REV y" is the same as "x >< (y+1)", if I remember correctly.

Rayman · 2020-02-11 18:31

It's not liking "dirl_" either:

    dirl_(USB_EVENT_REPO)                               ' DIR bit low puts smart pin in reset mode
    wrpin_(SP_REPO1_MODE, USB_EVENT_REPO)               ' Set "long repository" mode to act as an event mailbox
    dirh_(USB_EVENT_REPO)                               ' Enable the event mailbox smart pin (will raise IN)

What to do here?

garryj · 2020-02-11 18:33

Edit: "x REV (y - 1)" is the same as "x >< y".

ersmith · 2020-02-11 18:37

"pinfloat" will set the direction to low (like dirl_) and "pinlow" will set the direction to high (like dirh_). Use just plain "wrpin" instead of "wrpin_". fastspin 4.1.3 should accept all of those.

Rayman · 2020-02-11 18:55

Thanks, it took that.

Spin2 doesn't seem to like this:

PUB str(s = string("")) | c

Doesn't like waitx() either:

waitx(1000) ' wait to settle

Think this might be waitct

Rayman · 2020-02-11 19:05

Anybody know how to do inline assembly in Spin2?
ASM..ENDASM doesn't seem to work...

Rayman · 2020-02-11 19:11

Skipping Smartserial integration for now...

I thought it would compile now, but I get this (attached image).
No idea what to do about that...

cgracey · 2020-02-11 20:38

Rayman wrote: »

Anybody know how to do inline assembly in Spin2?
ASM..ENDASM doesn't seem to work...

PUB go() | i

  repeat

	org
_ret_	add	i,#1
	end

    pinwrite(56 addpins 7, !i)
    waitms(100)

cgracey · 2020-02-11 20:41

Rayman wrote: »
Thanks, it took that.

Spin2 doesn't seem to like this:
PUB str(s = string("")) | c
Doesn't like waitx() either:
waitx(1000) ' wait to settle
Think this might be waitct

In Spin2, there are three waits:

WAITCT(ct_target)
WAITUS(microseconds)
WAITMS(milliseconds)

There is also:

POLLCT(ct_target) 'returns true if past ct_target

cgracey · 2020-02-11 20:44

Rayman wrote: »

It's not liking "dirl_" either:

    dirl_(USB_EVENT_REPO)                               ' DIR bit low puts smart pin in reset mode
    wrpin_(SP_REPO1_MODE, USB_EVENT_REPO)               ' Set "long repository" mode to act as an event mailbox
    dirh_(USB_EVENT_REPO)                               ' Enable the event mailbox smart pin (will raise IN)

What to do here?

There are register variables, same as in PASM: DIRA/DIRB/OUTA/OUTB/INA/INB.

WRPIN/WXPIN/WYPIN/etc can be used in Spin2, as well as in PASM.

Rayman · 2020-02-11 23:06

Here are some files to reproduce the "Invalid Object File" error.

Graphics2c_Spin2.spin2 compiles OK
The barebones file that calls it "GraphicsTest1a.spin2" gives this error message...

cgracey · 2020-02-11 23:12

Rayman, I'm looking it over now. I'm getting the same error, of course. I'll figure out the problem.

cgracey · 2020-02-11 23:27

Okay. It's finding constant symbols that start with underscores, and for some reason I disallowed this in the object interface. I can't remember why. Let me change this...

cgracey · 2020-02-11 23:35

Rayman, here is PNut_v34b. This compiles your program:

https://drive.google.com/file/d/1HWnhswX2fd3W9H1gYSFqNeWToBMO7N7T/view?usp=sharing

Rayman · 2020-02-12 01:04

Is "test" reserved or something? Got a "expected unique variable name" error on this:

 long  test

Rayman · 2020-02-12 01:10

I guess you can't do this:

b := data.byte[0]

or this

x := ~byte[@data][1]           '~data.byte[1]

Also, looks like to call a function that returns multiple results, you shouldn't put them in parenthesis like this does:

(ptr, val) := usb1.getDebugInfo()

Rayman · 2020-02-12 01:39

Interesting... I commented out a "case" and now it gives me a "block is empty" error

    other:
      'ser.printf("Event%u: %u\n", portnum, event)

Rayman · 2020-02-12 01:45

All right, I finally got it to compile my 1080p code...
Didn't work though.

First problem seems to be with the clock setting...
Got VGA output after changing this part:

{
CON  'RJA:  new for real P2 - you can use different xdiv and xmul to set clock frequency:  /10*125 -> 250 MHz
  _XTALFREQ     = 20_000_000                                    ' crystal frequency
  _XDIV         = 20                                            ' crystal divider to give 1MHz
  _XMUL         = 297                                          ' crystal / div * mul
  _XDIVP        = 1                                             ' crystal / div * mul /divp to give _CLKFREQ (1,2,4..30)
  _XOSC         = %10                                  'OSC    ' %00=OFF, %01=OSC, %10=15pF, %11=30pF
  _XSEL         = %11                                   'XI+PLL ' %00=rcfast(20+MHz), %01=rcslow(~20KHz), %10=XI(5ms), %11=XI+PLL(10ms)
  _XPPPP        = ((_XDIVP>>1) + 15) & $F                       ' 1->15, 2->0, 4->1, 6->2...30->14
  _CLOCKFREQ    = _XTALFREQ / _XDIV * _XMUL / _XDIVP            ' internal clock frequency                
  _SETFREQ      = 1<<24 + (_XDIV-1)<<18 + (_XMUL-1)<<8 + _XPPPP<<4 + _XOSC<<2  ' %0000_000e_dddddd_mmmmmmmmmm_pppp_cc_00  ' setup  oscillator
  _ENAFREQ      = _SETFREQ + _XSEL                                             ' %0000_000e_dddddd_mmmmmmmmmm_pppp_cc_ss  ' enable oscillator

  _clkmode = _SETFREQ
  _clkfreq = _CLOCKFREQ
  
  ''NOTE:  You must manually change clock freq in both mouse and VGA drivers to match the above.
  
}
  
CON _clkfreq = 297_000_000

and also removing this:

    'First, set the clock
    'clkset(_clkmode, _clkfreq)
    'waitct(clkfreq*3)'waitcnt(clkfreq + cnt*3)

cgracey · 2020-02-12 02:30

Rayman, there are ways of achieving all these things in the PNut version of Spin2.

byte[base][index] 'look up a byte

To set up the clock, you only need to specify:

CON _clkfreq = 250_000_000 'assumes 20MHz crystal if no _xtlfreq specified
CON _xtlfreq = 16_000_000 'if your crystal frequency is not 20MHz, specify _xtlfreq

The compiler will pick the optimal PLL setting for you. You can see what it is in your code by using the constant clkmode_ and clkfreq_. Also Ctrl-L shows it.

And TEST is a PASM instruction, so it is a reserved word.

Until I get the documentation together for Spin2, there will be some ambiguities.

Spin2

Comments