Effectus suggestions and bug reports

zbyti · June 26, 2020

@tebe does the shorter code in Mad Pascal cause to generate a shorter code in MADS?

Edited June 26, 2020 by zbyti

zbyti · June 26, 2020

-nc           Effectus only translate source to Mad Pascal

I think -t would be more appropriate, what do you think? It seems to me that the -nc I have chosen is confusing because of the -c flag.

funkheld · June 26, 2020

Hi good afternoon.

what kind of task does that have in eff?

SET $E=$2000
SET $491=$2000

if I change that on :

SET $E=$8000
SET $491=$8000

I don't see any change in the asm program.

greeting


; Effectus example
; - - - - - - - - - - - - - - - - - -
; Display list interrupt (DLI) demo

SET $E=$2000
SET $491=$2000

BYTE ARRAY DL=[$70$70$70$42$00$40$2$2$2$2$F0$2$2$2$2$2$70$70$70$41$00$20]
BYTE ARRAY DLI=[$48$A9$DE$8D$0A$D4$8D$18$D0$68$40]
; 48        PHA
; A9 DE     LDA #$DE
; 8D 0A D4  STA $D40A   ;WSYNC
; 8D 18 D0  STA $D018   ;COLPF2
; 68        PLA
; 40        RTI

BYTE
  CH=$2FC,
  NMIEN=$D40E

CARD
  VDSLST=$200,
  SDLSTL=$230

PROC MAIN()
  SDLSTL=DL
  VDSLST=DLI
  NMIEN=$C0

  WHILE CH=$FF DO OD
  GRAPHICS(0)
RETURN

zbyti · June 26, 2020

@funkheld

Those are compiler directives not implemented yet.

$E may be useful but $491 make no sense in Effectus.

funkheld · June 26, 2020

but why is there no error message?

hello, thanks.

Gury · June 26, 2020

Error handling is not yet directly implemented in Effectus. It only shows Mad Pascal errors when they arise. But those errors are only produced by wrong usage of Effectus code or the statement syntax not yet supported.

Gury · June 26, 2020

3 hours ago, zbyti said:
@Gury
-nc           Effectus only translate source to Mad Pascal
I think -t would be more appropriate, what do you think? It seems to me that the -nc I have chosen is confusing because of the -c flag.

Yes, it is a good idea.

zbyti · June 26, 2020

[BUG ERROR FOR 0.5.0]

I started to change my test code and received this error.

NOTE: the code below does not have the correct values (and other stuff) in the DL array because I did not complete the changes due to a translation error.

SBYTE ARRAY TEXT="NNNNNNNNNNNNNNNNNNNNtajemnice@atariNNNNNNNNNNNNNNNNNNNN"
BYTE ARRAY DL=[$70$70$70$56$0$0$41$0$0]

BYTE
  CH=$2FC,
  HSCROL=$D404
CARD SDLSTL=$230

PROC WAIT=*(BYTE F)[$18$65$14$C5$14$D0$FC$60]

PROC MAIN()
  BYTE I=$CA

  DL(4)=PEEK(TEXT)   ; LSB of TEXT
  DL(5)=PEEK(TEXT+1) ; for test, proper is to get MSB

  SDLSTL=DL

  WHILE CH=$FF DO
    I=8
    WHILE I DO
      I==-1
      HSCROL=I
      WAIT(1)
    OD
    DL(4)==+1
    IF DL(4)=$23+$C THEN DL(4)=$C FI ; instead of $C we need DL(4)=PEEK(TEXT)
  OD
  GRAPHICS(0)
RETURN

DL(4)=PEEK(TEXT)
DL(5)=PEEK(TEXT+1)

; goes to

DL := Peek(4]=PEEK[TEXT);
DL := Peek(5]=PEEK[TEXT+1);

Hardcoded working version:

SBYTE ARRAY TEXT="NNNNNNNNNNNNNNNNNNNNtajemnice@atariNNNNNNNNNNNNNNNNNNNN"
BYTE ARRAY DL=[$70$70$70$56$0$0$41$0$0]

BYTE
  CH=$2FC,
  HSCROL=$D404
CARD SDLSTL=$230

PROC WAIT=*(BYTE F)[$18$65$14$C5$14$D0$FC$60]

PROC MAIN()
  BYTE I=$CA

  DL(4)=0   ; LSB
  DL(5)=$20 ; MSB

  SDLSTL=DL

  WHILE CH=$FF DO
    I=8
    WHILE I DO
      I==-1
      HSCROL=I
      WAIT(1)
    OD
    DL(4)==+1
    IF DL(4)=$23 THEN DL(4)=0 FI
  OD
  GRAPHICS(0)
RETURN

Nice to see in the future Action! syntax https://github.com/zbyti/a8-action-scramble/blob/master/src/ANTIC.ACT

SET $E=DLADDR
PROC DL=*()[
  $70             ; EMPTY ROW
  $42 TEXT        ; FIRST TEXT ROW (GR.0)
  $D2 SCREEN      ; STARTSCREEN (GR.0) WITH HSCROL AND DLI
  $12 $92 $12 $92 $12 $92 $12 $92 $12 $92 $12 $12 $12
  $12 $12 $12 $12 $12 $92 $92 $92 $92 $92 $92 $92
  $41 DL          ; JUMP TO BEGINNING OF DL
]

Edited June 26, 2020 by zbyti
working version

Gury · June 26, 2020

Okej, good analysys. Future will bring new fixed version

Great work on Scramble program, btw.

zbyti · June 26, 2020

  PROC TEST=*(BYTE CURSOR,BACK,BORDER,X,Y,UPDOWN)
  [
    $8E 710  ; BACKGROUND COLOR
    $8C 712  ; BORDER COLOR
    $8D 752  ; CURSOR VISIBILITY 
    $A5 $A5 $8D 755  ; CHARACTERS UPSIDE DOWN? 
    $A5 $A3 $8D 85 0  ; COLUMN FOR TEXT
    $A5 $A4 $8D 84 0  ; ROW FOR TEXT
    $60]

;-------

procedure TESTProc(CURSOR : byte; BACK : byte; BORDER : byte; X : byte; Y : byte; UPDOWN : byte); assembler;
  asm
  {
  lda CURSOR
  ldx BACK
  ldy BORDER
  mva X $a3
  mva Y $a4
  mva UPDOWN $a5
    .by $8E $C6 $02 $8C $C8 $02 $8D $F0 $02 $A5 $A5 $8D $F3 $02 $A5 $A3 $8D $55 $00 $A5 $A4 $8D $54 $00 $60 
  };
end;

I don't see the .eff source but I have a question...

This is obviously Action! way to handle the parameters (registers and $Ax) but this is a proper way in Effectus?

This is ~~implicit~~ zero page usage and I don't like it for this reason well... not so secret since we use explicite =*

Maybe let's leave this job to MP/MADS compilers and yours -z flags?

Passing 3 byte data through registers it's fine, but... @tebe don't do that during his optimizations?

;---------------------------------------------

EDIT:

OK, my bad... It's necessary to keep of the compatibility with PROC/FUNC =* written in the machine code.

Edited June 27, 2020 by zbyti
my bad

Gury · June 27, 2020

No problem The main goal is to maintain as much compatibility as possible. We can make some exceptions with parameters, but we should make them less than it is necessary.

zbyti · June 27, 2020

Okay :]

I just used =* mainly because this produce in Action! faster and direct code, instead of many parameters I preferred to use globals, so I forgot about the need for compatibility.

If the readme.txt file contains information about reserving $Ax that's enough, in general it can be very useful.

It will be fun if you extend Effectus with: {romoff}, downto and then you implement include we will be able to rewrite some things from blibs.

Edited June 27, 2020 by zbyti
add info

zbyti · June 27, 2020

~~[BUG ERROR FOR 0.5.0]~~ fixed https://github.com/Gury8/effectus/pull/6

BYTE ARRAY
  DL=[$70$70$70$52$00$40$52$60$40$52$C0$40$41$0$0]

BYTE
  HSCROL=$D404,
  RANDOM=$D20A

CARD
  SDLSTL=$230,
  SCREENI=[$4000]

PROC WAIT=*(BYTE F)[$18$65$14$C5$14$D0$FC$60]

PROC MAIN()
  BYTE
    HSCROLI=$CA,
    A,B,C

  SDLSTL=DL
  HSCROLI=$F

  DO
    IF HSCROLI=$B THEN
      A=(RANDOM&15)+33
      B=(RANDOM&15)+33
      C=(RANDOM&15)+33

      POKE(SCREENI,A)
      POKE(SCREENI+$60,B)
      POKE(SCREENI+$C0,C)

      POKE(SCREENI+$30,A)
      POKE(SCREENI+$30+$60,B)
      POKE(SCREENI+$30+$C0,C)

      DL(4)==+1
      DL(7)==+1
      DL($A)==+1

      SCREENI==+1

      IF DL(4)=$30 THEN
        DL(4)=0
        DL(7)=$60
        DL($A)=$C0
        SCREENI=$4000
      FI

      HSCROLI=$F
    FI

    HSCROL=HSCROLI HSCROLI==-1
    WAIT(1)
  OD
RETURN

A=(RANDOM&15)+33
B=(RANDOM&15)+33
C=(RANDOM&15)+33

; goes to

A := (RANDOM AND 15);
B := (RANDOM AND 15);
C := (RANDOM AND 15);

until 0 = 1; why not until false?

nescroll.eff nescroll.pas nescroll.xex

Edited June 28, 2020 by zbyti
fixed

zbyti · June 27, 2020

[BUG ERROR FOR 0.5.0]

var dl : array [0..0] of byte absolute $4000;

begin
	dl[0] := $80;
	dl[256] := $f0;
	repeat until false;
end.

byte array dl=$4000

proc main()
  dl(0)=$80
  dl(256)=$f0
  do od
return

goes to ranged array

// Effectus auto-generated Mad Pascal source code listing
program testPrg;

uses
  Crt;

var
  dl : array[0..255] of byte absolute $4000;

procedure MAINProc;
begin  // 1
  dl[0] := $80;
  dl[256] := $f0;
  repeat
  until 0 = 1;
end;  // 4

begin
  MAINProc;
end.

Edited June 27, 2020 by zbyti
divide the code

zbyti · June 27, 2020

[BUG ERROR FOR 0.5.0]

byte array dl=$6000, dl_temp=[$12$13$15]

proc main()
  dl=dl_temp
  do od
return

goes to

// Effectus auto-generated Mad Pascal source code listing
program testPrg;

uses
  Crt;

var
  dl : array[0..0] of byte = ($6000);
  dl_temp : array[0..2] of byte = ($12, $13, $15);

procedure MAINProc;
begin  // 1
  dl := dl_temp;
  repeat
  until 0 = 1;
end;  // 4

begin
  MAINProc;
end.

This is correct in some strange way Action! put values in dl_temp (in memory), Mad Pascal make a copy of that and put to dl, generally this works.

Edited June 28, 2020 by zbyti
this is not a critical bug but I think it should be fixed

zbyti · June 27, 2020

We need more convenient way then hardcoding to determine LMS for DL in Effectus next version

SBYTE ARRAY TEXT="NNNeffectus@rulezNNN"
BYTE ARRAY DL=[$70$70$70$70$67$1$20$41$0$0]

BYTE
  CH=$2FC,
  VSCROL=$D405
CARD SDLSTL=$230

PROC WAIT=*(BYTE F)[$18$65$14$C5$14$D0$FC$60]

PROC MAIN()
  BYTE I=$E0

  I=$F
  SDLSTL=DL
  WHILE CH=$FF DO
    WHILE I DO
      I==-1
      VSCROL=I
      WAIT(3)
    OD
    WAIT(50)
    WHILE I<$F DO
      I==+1
      VSCROL=I
      WAIT(2)
    OD
  OD
  GRAPHICS(0)
RETURN

vscroll.eff vscroll.xex

Edited June 28, 2020 by zbyti

funkheld · June 27, 2020

hello, why is the "N" in the text?

greeting

Edited June 27, 2020 by funkheld

funkheld · June 27, 2020

I want the memory from 2000-3fffhex.
how can you start the program at 4000hex with effectus?

thank you.
greeting

zbyti · June 27, 2020

@funkheld

1. "N" is a "." in GR.5

2. For now you can't directly do that. Use Effectus -nc flag and compile .pas with: mp -code:address

Edited June 27, 2020 by zbyti

funkheld · June 27, 2020

hello, thanks for the info.

why is the Effectus -nc flags ?

greeting.

Edited June 27, 2020 by funkheld

zbyti · June 27, 2020

@funkheld https://atariage.com/forums/topic/306222-effectus-suggestions-and-bug-reports/?do=findComment&comment=4572284

funkheld · June 27, 2020

compile .pas with: mp -code:address

not function !

mp -code:3000 vscroll.pas

greeting

Edited June 27, 2020 by funkheld

zbyti · June 27, 2020

@funkheld I suggest You resolve these issues in a different thread, this one is for bugs and suggestions

zbyti · June 28, 2020

            else if System.Pos('{AND}', params[1]) > 0 then begin
              SplitStr(params[1], '{AND}', aList);
              params[1] := aList[0] + ' AND ' + aList[1];
            end

@Gury for me this approach is very limited... I try to fix this "detail" but it's looks like this chunk need some serious refactor, better I leave it to You.

EDIT1: maybe I try fix only this:

(RANDOM&15)+33
; goes to
(RANDOM{AND}15){AND}33

EDIT2: ok, i fix that: core.pas line 845 https://github.com/Gury8/effectus/pull/6

                temp04 += paramsEx[i];
                if i < High(paramsEx) then begin
                  temp04 += op[i+1];
                end;

Edited June 28, 2020 by zbyti
code chunk

funkheld · June 28, 2020

Hi good afternoon.

how is the source of effectus compiled with freepascal?

thank you.

Effectus suggestions and bug reports

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Join the conversation

Recently Browsing 0 members