Pseudo-code Dictionaries
The pseudo-code generator uses a global dictionary, created when the program statement is executed (it actually exists once the Basic interpreter has been initialised).
There is also the local dictionary which is created at the start of a function or procedure block, and only exists until the end of function or end of procedure.
Pseudo-code Generation
This section look at the generation of pseudo-code from the abstract syntax tree. This is by far the most fun part of the interpreter, and certainly for myself as an assembler programmer, the most interesting. Pseudo-code can be best described as non-processor specific instructions that vaguely resemble assembler, but are at a more abstract level. The pesudo-code we will be using here does not deal with registers as such, but it does assume that we have a stack, and a single location we shall refer to as the accumulator. So, for example the Basic statement a% = a% + 3 would ultimately give us the pseudo-code shown below
load a%
addi 3
store a%
similarly the expression LEFT$(input$, 3) would produce the pseudo-code
stack 3
stack input$
loadi 2
call LEFT$
We shall consider each of the Basic statements in turn, and discuss the pseudo-code that is produced.
The DIM statement pseudocode
The DIM statement gives us the ability to create a one-, two-, or three-dimensional array. The pseudo-code that will be generated is as follows (assuming an array called array$).
For a one-dimensional array, 'DIM array$(20)'
stack 20
loadi 1
allocate
storeaddress array$
For a two-dimensional array, 'DIM array$(20, f%)'
stack f%
stacki 20
loadi 2
allocate
storeaddress array$
For a three-dimensional array, 'DIM array$(20, f%, p%)'
stack f%
stacki 20
stack p%
loadi 3
allocate
storeaddress array$
The FOR ... END FOR statement pseudocode
The FOR ... END FOR statement produces two blocks of pseudo-code, a prologue block and an epilogue block. For example if we take the statement FOR ix% = start% TO end% STEP stepper% ... END FOR, the prologue pseudo-code would be
load start%
store ix%
label_n load ix%
test_less end%
jump_false label_n+1
and the epilogue code would be
label_n+1 load ix%
add stepper%
store ix%
jump label_n
label_n+2 no_operation
Do not worry about the labels, as they will be explained more fully in Section IV of this tutorial.
The DO ... UNTIL statement pseudocode
The DO... UNTIL statement produces two blocks of pseudo-code, a prologue block and an epilogue block. For example if we take the statement DO ... UNTIL expression1% == expression2%, the prologue pseudo-code would be
label_n no_operation
and the epilogue code would be
label_n+1 load expression1%
test_equal expression2%
jump_false label_n
label_n+2 no_operation
The WHILE ... END WHILE statement pseudocode
The WHILE ... END WHILE statement produces two blocks of pseudo-code, a prologue block and an epilogue block. For example if we take the statement WHILE expression1% <> expression2% ... END WHILE, the prologue pseudo-code would be
label_n load expression1%
test_equal expression2%
jump_true label_n+2
and the epilogue code would be
label_n+1 jmp label_n label_n+2 no_operation
The PRINT statement pseudocode
The pseudo-code for the PRINT statement is simply
print
The accumulator will have been set up with the location of the item to be printed.
The INPUT statement pseudocode
The pseudo-code for the INPUT statement is simply
input
The accumulator will have been set up with the location of the item to be printed.
The IF ... ELSE ... END IF statement pseudocode
The IF ... END IF statement produces two blocks of pseudo-code, a prologue block and an epilogue block. For example if we take the statement IF expression1% == expression2% ... END IF, the prologue pseudo-code would be
label_n load expression1%
test_equal expression2%
jump_false label_n+1
and the epilogue code would be
label_n+1 no_operation
For an IF ... ELSE ... END IF statement produces trree blocks of pseudo-code, a prologue block, a mid-block and an epilogue block. For example if we take the statement IF expression1% == expression2% ... ELSE ... END IF, the prologue pseudo-code would be
label_n load expression1%
test_equal expression2%
jump_false label_n+1
the mid-block code would be
jmp label_n+2
label_n+1 no_operation
and the epilogue code would be
label_n+2 no_operation
The BREAK statement pseudocode
The pseudo-code for the BREAK statement is as follows.
jmp label_n+2.
The CONTINUE statement pseudocode
The pseudo-code for the CONTINUE statement is as follows.
jmp label_n+1.
The FUNCTION ... END FUNCTION statement pseudocode
The pseudo-code for the FUNCTION ... END FUNCTION statement has also a prologue and epilogue section. Let us say we have a function name$ which uses three parameters, first_name$, middle_name$ and surname$. The pseudo code for this function's prologue would be
label_n testi 3
jump_notequal label_n+1
unstack
store #return_address#
add_entry name$
dictionary
unstack
add_entry first_name$
store first_name$
unstack
add_entry middle_name$
store middle_name$
unstack
add_entry surname$
store surname$
stack #return_address#
and the epilogue would be
undictionary
return
label_n+1 error %1
The PROCEDURE ... END PROCEDURE statement pseudocode
The pseudo-code for the PROCEDURE ... END PROCEDURE statement has also a prologue and epilogue section. Let us say we have a procedure concatenate$ which uses two parameters, string1$ and string2$. The pseudo code for this procedure's prologue would be
label_n testi 2
jump_notequal label_n+1
unstack
store #return_address#
dictionary
add_entry concatenate$
unstack
add_entry string1$
store string1$
unstack
add_entry string2$
store string2$
stack #return_address#
and the epilogue would be
undictionary
return
label_n+1 error %1
You will notice there is considerable similarity between the function pseudo-code and the procedure pseudo-code. The difference lies in which dictionary, the name of the routine is placed. For a function, the routine name is place in the global dictionary, whereas in a procedure it is stored in the local dictionary, which means that it is removed (undictionary) when the procedure ends. During the procedure code block however, it can be used just as any other variable.
Expression Pesudo-instructions
All pesudo-instructions have an equivalient immediate operation (e.g. xori, subtracti, modulusi.
[i]bitwise operations:[/i]
and variable ; bitwise and operation on the accumulator and variable
or variable ; bitwise or operation on the accumulator and variable
xor variable ; bitwise xor operation on the accumulator and variable
not variable ; bitwise not operation on the accumulator and variable
[i]arithmetic operations:[/i]
add variable ; add variable to the accumulator
subract variable ; subtract variable from the accumulator
multiply variable ; multiply variable and the accumulator
divide variable ; divide accumulator by the variable
modulus variable ; divide accumulator by the variable and set modulus in accumulator
This completes part III of the "Writing an Interperter" tutorial.
MultiQuote
| 