This is the only missing part of Plov, as for now: bit operations. OK, you could do the same with assignment operations but this is easier on the eye for the seasoned microcontroller programmer. For this to work reasonably I had to introduce one new first degree keyword (BIT) and three second degree keywords (SET, CLEAR and TOGGLE). Plov018 accepts the following source:

PROGRAM test31

CONSTANTS  Ser0 = 101
	   Ser1 = 105
	   DTR0 = 102
	   DTR1 = 118
	   wdt  =   5
END

VARIABLES  a b cc ddd  END

BEGIN
   BIT SET	Ser0 wdt + 1
   BIT CLEAR	1001 7
   BIT TOGGLE	ddd  wdt
END test31

The general syntax is

	BIT <operation> port bit

The operation is one of

SET
CLEAR
TOGGLE

both the 'port' and 'bit' parameters can be any valid Plov expression. Punctuation is not required. Errorchecking is such that the internal state machine doesn't loose too much synchronisation.
For this all to function, three new PALO words were required:

SET-BIT
CLEAR-BIT
TOGGLE-BIT

Here I had to make a trade-off. The trailers "-BIT" are identical for all three keywords. So in an earlier version I left them off. But that took its toll on the (human) readability of the PALO source. I considered using "SET BIT" with a space, but that would require the ALTO backend to fetch a redundant word. So I decided to go for the hyphen. Now the ALTO parser will fetch the word in one bite. Good enough.

Below are the changes and additions. First one new errormessage:

	35 : InOut.WriteString ("SET, CLEAR or TOGGLE expected")

Next the addition to StatementSequence:

	ELSIF  Strings.StrEq (token, "BIT")  	THEN  isBit

Then the actual implementation procedure:

PROCEDURE isBit;

BEGIN
   GetSymbol;
   IF  Strings.StrEq (token, "SET")	   THEN
      GetSymbol;
      Expression;
      Expression;
      CG ("SET-BIT");
      CG ("")
   ELSIF  Strings.StrEq (token, "CLEAR")   THEN
      GetSymbol;
      Expression;
      Expression;
      CG ("CLEAR-BIT");
      CG ("")
   ELSIF  Strings.StrEq (token, "TOGGLE")  THEN
      GetSymbol;
      Expression;
      Expression;
      CG ("TOGGLE-BIT");
      CG ("")
   ELSE
      ErrorMessage (35);
      Skip2LF;
      GetSymbol
   END
END isBit;

And last, the addition to Init:

	IF  StoreSymbol ("BIT")	 = FALSE  THEN  ErrorMessage (0)  END;

BIT must be in the symbol table with the attribute 'Keyword' since it is part of StatementSequence.

Here is the test program:

PROGRAM test31

CONSTANTS  Const	= 101
	   Ser1		= 105
	   DTR0		= 102
	   DTR1 	= 118
	   wdt  	=   5
END

VARIABLES  var1 var2 var3  END

BEGIN
   BIT SET	Const var3
   BIT TOGLE	var1  wdt
   BIT CLEAR	1001  7
END test31

And this is what the console looks like after compiling:

PROGRAM test31

CONSTANTS  Const	= 101
	   Ser1		= 105
	   DTR0		= 102
	   DTR1 	= 118
	   wdt  	=   5
END

VARIABLES  var1 var2 var3  END

BEGIN
   BIT SET	Const var3
   BIT TOGLE	
@@  -- SET, CLEAR or TOGGLE expected in line    14
var1  wdt
   BIT CLEAR	1001  7
END test31

1 errors found.

An error in one of my sources. That's a new experience. Let's try to fix the missing 'G' in 'TOGGLE' and run it again:

PROGRAM test31

CONSTANTS  Const	= 101
	   Ser1		= 105
	   DTR0		= 102
	   DTR1 	= 118
	   wdt  	=   5
END

VARIABLES  var1 var2 var3  END

BEGIN
   BIT SET	Const var3
   BIT TOGGLE	var1  wdt
   BIT CLEAR	1001  7
END test31

No errors found.

That's more like it! Now let's have a look at the PALO source:

# PROGRAM test31

VARIABLE	var1
VARIABLE	var2
VARIABLE	var3

LABEL MAINLOOP

STORE	101
FETCH var3
SET-BIT

FETCH var1
STORE	5
TOGGLE-BIT

STORE	1001
STORE	7
CLEAR-BIT

LABEL EXITMAINLOOP
# Done #

It works!

Next. I think I won't be able to postpone it any longer. Next will be a first attempt on an ALTO (i.e. a code generator) for the ATmega128 MCU (MicroController Unit). Look in the ETT section to see the hardware that ultimately will be used as the target system.

Plov019 will not be made. The next release will start with Plov020 and it will mainly be fixes and changes to make the ALTO backend run smoother, if any. At this moment, I am still struggling with the concept of the index register. Most MCU's have several index registers so arbitrary cells in RAM or EEPROM can be addresses and acted upon. I'm not sure how to deal with this.
Yet, I have no choice than implementing it. It is this, or start thinking about array parameters. And that's just one bridge too far, for the time being.

Page created on 3 October 2008 and

Page equipped with FroogleBuster technology