Very simple Forth-like VM/Interpreter written in Elixir.
I have written it to experiment implementing a stack-based preemtive multitasking interpreter (and to play) with Elixir.
The package can be installed by adding forthvm to your list of dependencies in mix.exs:
def deps do
[
{:forthvm, "~> 0.5.0"}
]
endOr check hex.pm
Check hexdocs.pm
- ForthVM supports quoted strings like
"hello world" .. - ForthVM supports multiple concurrent isolated cores (Elixir processes) running concurrent Forth processes.
- ForthVM processes inside the same core can send messages to other processes and trigger execution of any defined word.
- ForthVM captures stdio and send corresponding messages to all registered processes (example, for UI stuff).
ForthVM
|- ForthVM.Supervisor: top-level supervisor. Only one of this can be started.
|- ForthVM.Registry: used to keep tab of core workers.
|- ForthVM.Subscriptions: used to subscribe to ForthVM events, like those from IOCapture.
|- ForthVM.IOCapture: collects all ForthVM outputs and dispatches to registered processes.
|- ForthVM.IOLogger: simple logger receiving messages from IOCapture.
|- ForthVM.Core.Supervisor: spawns Core workers, one for each Core.
|- ForthVM.Core.Worker: run ForthVM processes. Can spawn as many new processes as needed.
|- ...The easiest way to play with it is to use the very bearbones REPL:
mix replAnd inside it:
ForthVM REPL (v0.5.0)
>> 22 22 +
>> .
44
>> .s
>> s.
>> dictionary
Dictionary ([w] = word, [v] = variable, [c] = constant):
[w] ! ( x name -- ) - store value in variable
[w] & ( x y -- v ) - bitwise and
[w] ( ( -- ) - discard all tokens till ")" is fountd
[w] * ( x y -- n ) - multiplies x by y
[w] ** ( x y -- n ) - calculates pow of x by y
[w] */ ( x -- n ) - perform multiplication and divide result by x
[w] + ( y x -- n ) - sums y to x
[w] +! ( x name -- ) - increment variable by given value
[w] +loop ( inc -- ) - keep processing do_tokens till count < end_count, incrementing count by top value on the data stack
[w] - ( y x -- n ) - subtracts y from x
[w] -rot ( x y z -- z x y ) - rotate the top three stack entries, top goes on bottom
[w] . ( x -- ) - pops and prints the literal value on the top of the data_stack
[w] .s ( -- ) - prints the whole data_stack, without touching it
[w] / ( x y -- n ) - divides x by y
[w] /mod ( x y -- rem div ) - divides x by y, places divident and reminder on top of data stack
[w] 0< ( x -- bool ) - check if value is less than zero
[w] 0= ( x -- bool ) - check value is euqal to 0
[w] 0> ( x -- bool ) - check if value is greater than zero
[w] 1+ ( x -- n ) - adds 1 to x
[w] 1- ( x -- n ) - subtracts 1 to x
[w] 2drop ( x y -- ) - remove two elements from top of stack
[w] 2dup ( x y -- x y x y ) - duplicate two elements from top of stack
[w] 2over ( y2 x2 y1 x1 -- y2 x2 y1 x1 y2 x2) - swap top copules on top of stack
[w] 2swap ( y2 x2 y1 x1 -- y1 x1 y2 x2 ) - swap top copules on top of stack
[w] : ( -- ) - convert all tokens till ";" is found into a new word
[w] < ( x y -- bool ) - check if x is less than y
[w] << ( x y -- v ) - bitwise shift left
[w] <= ( x y -- bool ) - check if x is less than or equal to y
[w] <> ( x y -- bool ) - check two values are different. Works on different types
[w] = ( x y -- bool ) - check two values are equal. Works on different types
[w] > ( x y -- bool ) - check if x is greater than y
[w] >= ( x y -- bool ) - check if x is greater than or equal to y
[w] >> ( x y -- v ) - bitwise shift right
[w] >r ( data -- ) - move the top of the data stack to the return stack
[w] ? ( var -- ) - fetch a variable value and prints it
[w] ?dup ( x -- x x ) - duplicate element from top of stack if element value is truthly
[w] @ ( name -- ) - get value in variable
[w] @- ( x -- n ) - negates x
[w] ^ ( x y -- v ) - bitwise xor
[w] abort ( i * x -- ) - ( R: j * x -- ) empty the data stack and perform the function of QUIT, which includes emptying the return stack, without displaying a message.
[w] abort? ( flag i * x -- ) - ( R: j * x -- ) if flag is truthly empty the data stack and perform the function of QUIT, which includes emptying the return stack, displaying a message.
[w] abs ( x -- n ) - absolute of x
[w] and ( x y -- bool ) - logical and
[w] begin ( -- ) - start loop declaration
[w] constant ( x -- ) - create a new costant with name from next token and value from data stack
[w] cos ( x -- n ) - cos of x
[w] cr ( -- ) - emits a carriage return
[w] debug-disable ( -- ) - set process debug flag to false.
[w] debug-dump-word ( -- ) - prints the definition of the word specified in the next token.
[w] debug-enable ( -- ) - set process debug flag to true.
[w] depth ( -- x ) - get stack depth
[w] dictionary ( -- ) - ( -- ) print list of words in dictionary
[w] do ( end_count count -- ) - start loop declaration
[w] drop ( x -- ) - remove element from top of stack
[w] dup ( x -- x x ) - duplicate element from top of stack
[w] emit ( c -- ) - pops and prints (without cr) the binary of an ascii value on the top of the data_stack
[w] end ( -- ) - ( R: -- ) explicit process termination
[w] exit ( -- ) - ( -- ) explicit VM termination
[w] false ( -- bool ) - the false constant
[w] help ( -- ) - ( -- ) print description of dictionary's word/var/const specified as the next token
[w] i ( -- count ) - copy the top of a LOOP's return stack to the data stack
[w] if/else/than ( x -- ) - if x is truthly execute words before than, if falsly and else is specified, execute code before else
[w] include ( -- ) - include program file from filename specified in next token.
[w] inspect ( x -- ) - pops and prints the inspected value on the top of the data_stack
[w] j ( -- data ) - copy data from return stack after LOOP's definition to the data stack
[w] l[ ( x y z -- l ) - collect all tokens till ] is found and store on the data stack as a list
[w] loop ( -- ) - keep processing do_tokens till count < end_count, each step incrementing count by 1
[w] max ( x y -- n ) - miaximum between two values
[w] min ( x y -- n ) - minimum between two values
[w] not ( x -- bool ) - logical not
[w] or ( x y -- bool ) - logical or
[w] over (y x -- y x y) - copy second element on top of stack
[w] pi ( -- n ) - puts Pi value on top of stack
[w] puts ( x -- ) - pops and prints the literal value on the top of the data_stack
[w] r> ( -- data ) - move the top of the return stack to the data stack
[w] r@ ( -- data ) - copy the top of the return stack to the data stack
[w] rand ( -- n ) - puts random number on stack
[w] rot ( x y z -- y z x ) - rotate the top three stack entries, bottom goes on top
[w] send (message_data cnt ":"word process_id -- ) - ( -- ) sends a message to a process inside the current core. The message is handled by a word with same name minus the ":" prefix. Cnt is the number elements in the data stack to be included in the message.
[w] set-io-device ( name -- ) - set the current IO device to the value on the top of the data_stack
[w] sin ( x -- n ) - sin of x
[w] sleep ( x -- ) - sleep for given milliseconds
[w] sqrt ( x -- n ) - sqrt of x
[w] swap ( x y -- y x ) - swap top two elements on top of stack
[w] tan ( x -- n ) - tan of x
[w] true ( -- bool ) - the true constant
[w] until (bool -- ) - keep processing untill contition is truthly
[w] variable ( -- ) - create a new variable with name from next token
[w] | ( x y -- v ) - bitwise or
[w] ~ ( x -- v ) - bitwise not- tokenizer
- interpreter
- core words
- loops
- if-then-else
- include
- REPL
- sleep word
- list type definition
- VM
- multiple Forth processes running in a single VM
- process messages: sending messages to a process will call a matching word
- step-by-step debugger
- real Forth: define and handled immediate words
ForthVM is provided under the MIT license