$ stack setup && stack build$ stack exec looper -- -p test/whileTest.progA few example programs are included in the test directory.
Due to a bug in stack exec backspace may not work. You can just run the
executable directly from the .stack-work directory to sidestep this issue.
The interpreter supports running instructions backwards and forwards. It does
this by first translating a structured program tree containing high-level
constructs such as if and while into a list of instruction primitives such as
JMP and JMPF.
Running an instruction then becomes evaluating its side effects and determining how far it moves the current index into the instruction tape.
The environment contains a mapping of variable names to list of values that represent the history of a variable over time which facilitates reversing execution.
Going backwards amounts to popping a value off the list of executed instructions, reversing its effects if it is an assign instruction and then determining how far it moves the current index into the tape back.
Commands may be entered using either their first letter or their entire name.
Pressing enter with no input will either repeat the step or back command
depending on which one you ran last.
inspect/i(inspect the current environment)inspect [VARIABLE NAME]/i [VARIABLE NAME](inspect a single variable’s history)step/s(steps to the next instruction)back/b(steps back through the last instruction executed)location/l(shows the current instruction awaiting execution)
The main program runs an unused variable analysis on the program prior to execution. The user is shown all of the variables that have been assigned to somewhere in the program but never refernced again in any statements or expressions. This signals to the user writing the program that they can remove this dead code.
A test program has been included that demonstrates this feature and can be run with:
$ stack exec looper -- -p test/unusedTest.prog