This project contains a parser, static analyzer and interpreter for SQF (Arma scripting language), written in Python. It can be used to:
- syntax-check and static analyze SQF files and projects
- execute SQF on a limited virtual environment
One of the major bottlenecks of scripting in SQF is the time spent testing it in-game, by running the game.
Often, these scripts contain errors (missing ";", wrong usage of params
) that everyone would
love to find without restarting the mission.
This package allows to parse SQF to check for syntactic errors, wrong types, problems in variables scopes, and many more subtle issues of SQF.
>>> code = 'if (true) {1}'
>>> errors = sqf.analyzer.analyze(sqf.parser.parse(code))
>>> errors[0]
SQFParserError((1, 11), "'(true)' can't preceed '{1}' (missing ';'?)")
>>> code = 'private _y = _z'
>>> analyzer = sqf.scope_analyzer.interpret(sqf.parser.parse(code))
>>> analyzer.exceptions[0]
SQFWarning((1, 14), 'Local variable "_z" is not from this scope (not private)')
Behind the curtains, the analyzer uses a parser to convert SQF code in a set of statements:
>>> code = '[1, 2, 3]'
>>> sqf.parser.parse(code)
Statement([Statement([
Array([Statement([N(1)]), Statement([Space(), N(2)]), Statement([Space(), N(3)])])
])])
The interpreter is able to run scripts on an emulated (and limited) environment. The interpreter is obviously not intended to run Arma simulation; it is aimed for you, moder, run tests of your scripts (e.g. Unit Tests) without having to run the game.
from sqf.interpreter import interpret
interpreter, outcome = interpret('_x = [1, 2]; _y = _x; reverse _y;')
# outcome equals to "Nothing"
# interpreter['_y'] equals to Array([Number(2), Number(1)])
# interpreter['_x'] equals to Array([Number(2), Number(1)]) # <= _y is a reference
This code is written in Python 3 and has no dependencies. You can install it using
pip3 install sqflint
The code is heavily tested (coverage 98%+), and the tests
can be found in tests
. Run them using standard Python unittest:
python -m unittest discover
This package is compatible with known editors, and can be used to efficiently write SQF with them. See the respective projects for more details:
This code contains essentially 4 components, a tokenizer, a parser, analyzer and interpreter:
The interpreter is a class that executes parsed scripts. It receives a
Statement
and executes it as per described in the Arma 3' wiki.
To automatically initialise the interpreter and execute code, run interpret
:
>>> from sqf.interpreter import interpret
>>> script = '''
a = 0;
b = true;
for [{_i = 0}, {_i < 10 && b}, {_i = _i + 1}] do {
a = a + 1;
if (a >= 7) then {b = false}
}
'''
>>> interpreter, outcome = interpret(script)
>>> interpreter['a']
Number(7)
>>> outcome
Boolean(False)
The call interpreter['a']
returns the outermost Scope
of the Namespace
"missionNamespace"
, but you can use setVariable
and getVariable
to interact with other namespaces.
sqf.tests.test_interpreter
contains the tests of the implemented functionality.
The source is in sqf/interpreter.py
, the tests in tests/test_interpreter.py
.
The main loop of the interpreter is defined in sqf/interpreter.py
, and the
expressions it evaluates are defined in sqf/expressions.py
.
The analyzer consumes the result of the parser and checks for static errors.
The source is in sqf/analyzer.py
, the tests in tests/test_analyzer.py
.
The parser transforms a string into a nested Statement
, i.e.
a nested list of instances of types
, operators, and keywords defined in SQF.
For example,
>>> from sqf.parser import parser
>>> script = '_x=2;'
>>> result = parse(script)
>>> result
Statement([Statement([Variable('_x'), Keyword('='), Number(2)], ending=';')])
>>> script == str(result) # True
This rather convolved result
takes into account operator precedence and
the meaning of the different parenthesis ([]
, {}
, ()
).
To transform the script into tokens used in the parser, the tokenizer is called.
sqf.tests.test_parser
contains the tests.
The tokenizer transforms a string into a list of tokens split by the relevant tokens of SQF. E.g.
>>> from sqf.base_tokenizer import tokenize
>>> tokenize('/*_x = 1;*/')
['/*', '_x', ' ', '=', ' ', '1', ';', '*/']
The source can be found in sqf.base_tokenizer
.
This code is licenced under BSD.
This code was written by Lord Golias, [email protected]