Plov004 : running compiler tests

Plov004 has seen the light of day. It is many times more stable than Plov003 used to be, 48 hours ago. Yet, there is still some room for improvement.

Plov004 : compiling source 'test1'

Testing source : 

PROGRAM Number1

BEGIN
END Number1
Compilation results in : 
jan@beryllium:~/modula/Plov$ Plov004 <test1
PROGRAM Number1 BEGIN   END     Number1


Number1 Program type
RETURN  Keyword type
EXIT    Keyword type
LOOP    Keyword type
BEGIN   Keyword type
END     Keyword type
IF      Keyword type
THEN    Keyword type
ELSIF   Keyword type
ODD     Keyword type
PROCEDURE       Keyword type
PROGRAM 	Keyword type
CONSTANTS       Keyword type
VARIABLES       Keyword type
jan@beryllium:~/modula/Plov$
It ends well, so all is well! This is no surprise, since the source of Plov003 was already altered to cope with test1.

Plov004 : compiling source 'test2'

Testing source : 

PROGRAM Number1

CONSTANTS  pi = 314
	   ee =	272
	   END
BEGIN
END Number1
Compilation results in : 
jan@beryllium:~/modula/Plov$ Plov004 <test2
PROGRAM Number1 CONSTANTS       pi      =       314     ee      =       272     END     BEGIN   END     Number1


Number1 	Program type
RETURN  	Keyword type
EXIT    	Keyword type
LOOP    	Keyword type
BEGIN   	Keyword type
END     	Keyword type
IF      	Keyword type
THEN    	Keyword type
ELSIF   	Keyword type
ODD     	Keyword type
PROCEDURE       Keyword type
PROGRAM 	Keyword type
CONSTANTS       Keyword type
VARIABLES       Keyword type
pi      	Constant type =   314
ee      	Constant type =   272
jan@beryllium:~/modula/Plov$
As you can see, the symbol table is printed after every run and also the output of GetSymbol is printed atop. This is caused by the following piece of code: 
PROCEDURE GetSymbol;

BEGIN
   IF  InOut.EOF () = FALSE  THEN
      InOut.ReadString (token)
   END;
   InOut.WriteString (token); InOut.Write (11C);   InOut.WriteBf
END GetSymbol;
InOut.WriteString (token); takes care of dumping the newly read token to screen. InOut.Write (11C); writes a TAB character to screen, inserting spaces until the cursor X-position is at a multiple of 8. InOut.WriteBf flushes the buffer NOW. This is handy since this is only done after a LineFeed is issued. But if the program gets caught in an endless loop there will be no final LF and hence also nothing to see on screen.

Plov004 : compiling source 'test3'

Testing source : 

PROGRAM Number1

CONSTANTS  pi = 314
	   ee =	272
	   END

VARIABLES  A 
	   B 
	   C 
	   Def 
	   GhYh 
	   J123
	   NH4NO3
	   END

BEGIN
END Number1
Compilation results in : 
jan@beryllium:~/modula/Plov$ Plov004 <test3
PROGRAM Number1 CONSTANTS       pi      =       314     ee      =       272     END     VARIABLES       A       B
C   Def      GhYh    J123    NH4NO3  END     BEGIN   END     Number1

Number1 	Program type
RETURN  	Keyword type
EXIT    	Keyword type
LOOP    	Keyword type
BEGIN   	Keyword type
END     	Keyword type
IF      	Keyword type
THEN    	Keyword type
ELSIF   	Keyword type
ODD     	Keyword type
PROCEDURE       Keyword type
PROGRAM 	Keyword type
CONSTANTS       Keyword type
VARIABLES       Keyword type
pi		Constant type =   314
ee      	Constant type =   272
A       	Variable type
B       	Variable type
C       	Variable type
Def     	Variable type
GhYh    	Variable type
J123    	Variable type
NH4NO3  	Variable type
jan@beryllium:~/modula/Plov$
Fine as well!

Plov004 : compiling source 'test4'

Testing source : 

PROGRAM Number1

CONSTANTS  pi = 314
	   ee =	272
	   END

VARIABLES  A 
	   B 
	   C 
	   Def 
	   GhYh 
	   J123
	   NH4NO3
	   END

PROCEDURE  AddItUp

BEGIN
END AddItUp

BEGIN
   AddItUp
END Number1
Compilation results in : 
jan@beryllium:~/modula/Plov$ Plov004 <test4
PROGRAM Number1 CONSTANTS       pi      =       314     ee      =       272     END     VARIABLES       A       B
C   Def      GhYh    J123    NH4NO3  END     PROCEDURE       AddItUp BEGIN   END     AddItUp BEGIN   AddItUp END
Number1

Number1 	Program type
RETURN  	Keyword type
EXIT    	Keyword type
LOOP    	Keyword type
BEGIN   	Keyword type
END     	Keyword type
IF      	Keyword type
THEN    	Keyword type
ELSIF   	Keyword type
ODD     	Keyword type
PROCEDURE       Keyword type
PROGRAM 	Keyword type
CONSTANTS       Keyword type
VARIABLES       Keyword type
pi      	Constant type =   314
ee      	Constant type =   272
A       	Variable type
B       	Variable type
C       	Variable type
Def     	Variable type
GhYh    	Variable type
J123    	Variable type
NH4NO3  	Variable type
AddItUp 	Procedure type
jan@beryllium:~/modula/Plov$
A rather complex source with Constant, Variable and Procedrue definitions.

Plov004 : compiling source 'fout1'

Testing source : 

PROGRAM Number1

CONSTANTS  pi = 314
	   ee =	272
	   END

VARIABLES  A 
	   B 
	   C 
	   0Def 
	   GhYh 
	   J123
	   NH4NO3
	   END

PROCEDURE  Diff

BEGIN
END Diff

PROCEDURE  AddItUp

BEGIN
END AddItUp

BEGIN
   AddItUp
   Diff
END Number1
Compilation results in : 
jan@beryllium:~/modula/Plov$ Plov004 <fout1
PROGRAM Number1 CONSTANTS       pi      =       314     ee      =       272     END     VARIABLES       A       B
C   0Def     Illegal identifier name : 0Def
GhYh    J123    NH4NO3  END     PROCEDURE       Diff    BEGIN   END     Diff    PROCEDURE       AddItUp BEGIN
END     AddItUp      BEGIN   AddItUp Diff    END     Number1
The symbol table has been omitted here. The Token Dump (above) gives the clues. Variable '0Def' starts with a zero so it is illegal. The error is flagged and compilation continues as if nothing had happened!

Plov004 : compiling source 'fout2'

Testing source : 

PROGRAM Number1

CONSTANTS  pi = 314
	   ee =	A272
	   272 = 122
	   END

VARIABLES  A 
	   B 
	   C 
	   0Def 
	   GhYh 
	   J123
	   NH4NO3
	   END

PROCEDURE  Diff

BEGIN
END Diff

PROCEDURE  AddItUp

BEGIN
END AddItUp

BEGIN
   AddItUp
   Diff
END Number1
Compilation results in : 
jan@beryllium:~/modula/Plov$ Plov004 <fout2
PROGRAM Number1 CONSTANTS       pi      =       314     ee      =       A272    Invalid digit in : A272
272     Illegal identifier name : 272
END     VARIABLES       A       B       C       0Def    Illegal identifier name : 0Def
GhYh    J123    NH4NO3  END     PROCEDURE       Diff    BEGIN   END     Diff    PROCEDURE       AddItUp BEGIN
END     AddItUp      BEGIN   AddItUp Diff    END     Number1
Three errors. Each one flagged. They do not interfere with eachother. Professor Wirth could be prooud of me!

Plov004 : compiling source 'test6'

Testing source : 

PROGRAM NH3production

CONSTANTS pi = 314159
	  ee = 271828
	  years = 2007
	  days = 31
	  END

VARIABLES temp 
	  press 
	  NH3
	  H2SO4
	  nitrogen
	  END

PROCEDURE Measure

BEGIN
END Measure

PROCEDURE Blow

BEGIN
END Blow

BEGIN
   LOOP
      Measure
      Blow
      EXIT
   END
END NH3production
Compilation results in : 
jan@beryllium:~/modula/Plov$ Plov004 <test6
PROGRAM NH3production   CONSTANTS       pi      =       314159  ee      =       271828  years   =       
2007	days    =   	31       END     VARIABLES       temp    press   NH3     H2SO4   nitrogen        
END     PROCEDURE	Measure BEGIN   END Measure  PROCEDURE       Blow    BEGIN   END     Blow
BEGIN   LOOP    Measure Blow    EXIT	END     END     NH3production


NH3production   Program type
RETURN  Keyword type
EXIT    Keyword type
LOOP    Keyword type
BEGIN   Keyword type
END     Keyword type
IF      Keyword type
THEN    Keyword type
ELSIF   Keyword type
ODD     Keyword type
PROCEDURE       Keyword type
PROGRAM Keyword type
CONSTANTS       Keyword type
VARIABLES       Keyword type
pi      Constant type = 314159
ee      Constant type = 271828
years   Constant type =  2007
days    Constant type =    31
temp    Variable type
press   Variable type
NH3     Variable type
H2SO4   Variable type
nitrogen        Variable type
Measure Procedure type
Blow    Procedure type
jan@beryllium:~/modula/Plov$
Even the LOOP/END/EXIT is processed!

Plov004 : compiling source 'fout3'

Testing source : 

PROGRAM NH3production

CONSTANTS pi = 314159
	  ee = 271828
	  years = 2007
	  days = 31
	  END

VARIABLES temp 
	  press 
	  NH3
	  H2SO4
	  nitrogen
	  END

PROCEDURE Measure

BEGIN
END Measure

PROCEDURE Blow

BEGIN
END Blow

BEGIN
   Measure
   Blow
   EXIT
END NH3production
Compilation results in : 
jan@beryllium:~/modula/Plov$ Plov004 <fout3
PROGRAM NH3production   CONSTANTS       pi      =       314159  ee      =       271828  years   =       2007
days    =   31       END     VARIABLES       temp    press   NH3     H2SO4   nitrogen        END     PROCEDURE
Measure BEGIN   END Measure  PROCEDURE       Blow    BEGIN   END     Blow    BEGIN   Measure Blow    EXIT    
EXIT without LOOP	END     NH3production
EXIT without LOOP. For the rest: nothing reported. Sounds fine to me.

Plov004 : compiling source 'test7'

Testing source : 

PROGRAM assignments

CONSTANTS nrOf = 12
	  PINcode = 1153
	  ZIPcode = 5012
	  END

VARIABLES aantal sum avg END

BEGIN
   aantal := nrOf
   sum := sum + PINcode
   avg := sum / aantal
END assignments
Compilation results in : 
jan@beryllium:~/modula/Plov$ Plov004 <test7
PROGRAM assignments     CONSTANTS       nrOf    =       12      PINcode =       1153    ZIPcode =
5012	END     VARIABLES    aantal  sum     avg     END     BEGIN   aantal  :=      nrOf    sum     
:=      sum     +	PINcode Error in StatementSequence	END     assignments
One error: in 'sum := sum + PINcode', the PINcode triggers an 'Error in StatementSequence'. And the next line 'avg := sum / aantal' is skipped altogether. I think the 'synchronize' function is to blame here.

Plov004 : compiling source 'test...'

The current version of Plov004 compiles all 10 test programs. The placement of GetSymbol is critical!

Page created on 7 October 2007 and

Page equipped with FroogleBuster technology