Created
July 15, 2023 13:09
-
-
Save nat-n/068f084cbfcd76f5c32391fba65a7e8c to your computer and use it in GitHub Desktop.
An experimental parser for bash syntax in python
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| """ | |
| An experimental tokenizer for bash syntax. | |
| This is an alternative to shlex from the standard library for tokenizing posix | |
| style command line statements. | |
| On top of the word-splitting and quote removal functionality of shlex, this | |
| library also makes it possilbe to tell which glob characters are escaped, and | |
| preserves semantics for variable & arithmetic expansions. | |
| In scope: | |
| - quoting & escaping | |
| - parameter, arithmetic, command expansions | |
| - glob characters | |
| - redirects | |
| - control operations | |
| - comments | |
| - tilde expansions | |
| Out of scope: | |
| - history expansion | |
| - shell keywords (case, if, export, etc) | |
| - multiline constructs/multi-token grammar: control flow, functions, multiline groups | |
| - recursively tokenizing the content of command or arithmetic expansions | |
| - advanced parameter expansion, like ${param:-} | |
| - arrays | |
| - groups | |
| - advanced glob patterns (classes and pattern lists) | |
| """ | |
| from collections import namedtuple, deque | |
| from enum import Enum | |
| from io import StringIO | |
| from typing import IO, List, Optional, Tuple | |
| import re | |
| # characters that are escaped by glob:escape | |
| GLOB_INIT_CHARS = "*?[" | |
| PREFIX_CHARS = "\\$" | |
| METACHARACTERS = "|;&<>(){}" | |
| WHITESPACE = " \t\r\n" | |
| NON_RANGE_CHARS = WHITESPACE + "!`'\";&|()<>" | |
| VAR_INIT_CHARS = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz_" | |
| DIGITS = "0123456789" | |
| VAR_CHARS = VAR_INIT_CHARS + DIGITS | |
| SPECIAL_VARS = DIGITS + "!@#$*-?" | |
| def tokenize(cmd_line: str): | |
| return list(CommandLineTokenizer(StringIO(cmd_line))) | |
| class TokenType(Enum): | |
| WORD = 0 | |
| CONTROL_OPERATOR = 1 | |
| REDIRECT_OPERATOR = 2 | |
| COMMENT = 3 | |
| class TokenFeatureType(str, Enum): | |
| VARIABLE = "var" # TODO: rename PARAMETER | |
| COMMAND = "cmd" | |
| ARITHMETIC = "arth" | |
| TILDE = "tld" | |
| GROUP = "grp" | |
| class TokenFeature( | |
| namedtuple( | |
| "TokenFeature", | |
| ("type", "content", "start", "end", "quoted"), | |
| ) | |
| ): | |
| type: TokenFeatureType | |
| content: str | |
| start: int | |
| end: int | |
| quoted: bool | |
| class Token( | |
| namedtuple( | |
| "Token", | |
| ( | |
| "content", | |
| "includes_glob", | |
| "escaped_glob_indices", | |
| "expansions", | |
| "type", | |
| ), | |
| ) | |
| ): | |
| content: str | |
| includes_glob: bool | |
| escaped_glob_indices: Tuple[int, ...] | |
| expansions: Tuple[TokenFeature, ...] | |
| type: TokenType | |
| @staticmethod | |
| def operator(content: str): | |
| return Token(content, False, tuple(), tuple(), TokenType.CONTROL_OPERATOR) | |
| _expansion_types = { | |
| "(": (")", TokenFeatureType.GROUP), | |
| "{": ("}", TokenFeatureType.GROUP), | |
| "${": ("}", TokenFeatureType.VARIABLE), | |
| "$((": ("))", TokenFeatureType.ARITHMETIC), | |
| "$(": (")", TokenFeatureType.COMMAND), | |
| "`": ("`", TokenFeatureType.COMMAND), | |
| "$": ("", TokenFeatureType.VARIABLE), | |
| "~": ("", TokenFeatureType.TILDE), | |
| } | |
| def includes_glob(input_stream: IO[str]) -> bool: # TODO: unit test me | |
| """ | |
| Check if the input stream includes are valid glob sequence, such as: | |
| - * | |
| - ? | |
| - a valid simple range expression | |
| Escaping of glob characters is not considered. | |
| """ | |
| next_char = None | |
| range_sequence_len = 0 | |
| escape_active = False | |
| while next_char := input_stream.read(1): | |
| if not next_char: | |
| return False | |
| if next_char in "*?": | |
| return True | |
| if range_sequence_len: | |
| if not escape_active: | |
| if next_char == "]": | |
| if range_sequence_len == 1: | |
| range_sequence_len += 1 | |
| else: | |
| return True | |
| if next_char == '"': | |
| escape_active = True | |
| if next_char in NON_RANGE_CHARS: | |
| range_sequence_len = 0 | |
| continue | |
| range_sequence_len += 1 | |
| if next_char == "[": | |
| range_sequence_len = 1 | |
| class RedirectParseState: | |
| """ | |
| Here-document and here-string redirects are not supported. | |
| """ | |
| left_fd: str = "" | |
| operator: str = "" | |
| right_fd: str = "" | |
| def __init__(self, left_fd: str = "", operator: str = ""): | |
| self.left_fd = left_fd | |
| self.operator = operator | |
| self.right_fd = "" | |
| @property | |
| def valid(self): | |
| if self.left_fd and not self.left_fd.isdigit(): | |
| return False | |
| if not re.match(r"^\d*\-?$", self.right_fd): | |
| return False | |
| if self.left_fd and not (self.operator or self.right_fd): | |
| return True | |
| if self.operator in ("&>", "&>>") and not (self.left_fd or self.right_fd): | |
| return True | |
| if self.operator in (">", ">>", ">|", "<", "<>") and not self.right_fd: | |
| return True | |
| if self.operator in (">&", "<&"): | |
| return True | |
| return False | |
| def read(self) -> str: | |
| return self.left_fd + self.operator + self.right_fd | |
| def append(self, next_char: str) -> bool: | |
| """ | |
| This method assumes self will always be initialized with left_fd OR | |
| with operator as one of: "<" or ">" or "&>" | |
| If next_char is successfully appended to this token then return empty string. | |
| Otherwise return the most recent chars that are cannot be appended to this | |
| token. Usually this is just next_char but it can also more some digits. | |
| """ | |
| rejected = "" | |
| if self.operator in ("<&", ">&"): | |
| if not self.right_fd.endswith("-") and ( | |
| next_char.isdigit() or next_char == "-" | |
| ): | |
| self.right_fd += next_char | |
| elif next_char in " \t\n|;&<>": | |
| rejected = next_char | |
| else: | |
| # what we thought was closing digits is actually the next word | |
| rejected = self.right_fd + next_char | |
| self.right_fd = "" | |
| elif self.operator == ">": | |
| if next_char in ">&|": | |
| self.operator += next_char | |
| else: | |
| rejected = next_char | |
| elif self.operator == "<": | |
| if next_char in ">&": | |
| self.operator += next_char | |
| else: | |
| rejected = next_char | |
| elif self.operator == "&>": | |
| if next_char == ">": | |
| self.operator += next_char | |
| elif next_char in "|;&<>()": | |
| raise CommandLineTokenizeError( | |
| f"Syntax error near redirect operator {self.read()!r}" | |
| ) | |
| else: | |
| rejected = next_char | |
| elif not self.operator: | |
| if next_char.isdigit(): | |
| self.left_fd += next_char | |
| elif next_char in "<>": | |
| self.operator += next_char | |
| else: | |
| rejected = next_char | |
| else: | |
| rejected = next_char | |
| return rejected | |
| class TokenParseState: | |
| quote_active: str | |
| range_active: bool | |
| active_prefix: str | |
| scope_terminator: str | |
| redirect: Optional[RedirectParseState] | |
| paren_depth: int | |
| includes_glob: bool | |
| _expansion_scope_content: List[str] | |
| _parts: List[str] | |
| _escaped_glob_indices: List[int] | |
| _expansions: List[TokenFeature] | |
| _expansion_scope_start_index: Optional[int] = None | |
| expansion_type: Optional[str] = None | |
| def __init__(self): | |
| self.quote_active = "" | |
| self.range_active = False | |
| self.active_prefix = "" | |
| self.scope_terminator = "" | |
| self.redirect = None | |
| self.paren_depth = 0 | |
| self.includes_glob = False | |
| self._expansion_scope_content = [] | |
| self._parts = [] | |
| self._escaped_glob_indices = [] | |
| self._expansions = [] | |
| def read(self): | |
| if self.redirect: | |
| return self.redirect.read() | |
| # TODO: keep count of length in append instead of calling read just to measure? | |
| if self.is_empty: | |
| return "" | |
| if len(self._parts) > 1: | |
| self._consolidate() | |
| return self._parts[0] | |
| def get_token(self) -> Optional[Token]: | |
| if self.is_empty: | |
| return None | |
| print("T:", self.read()) | |
| return Token( | |
| content=self.read(), | |
| includes_glob=self.includes_glob, | |
| escaped_glob_indices=tuple(self._escaped_glob_indices), | |
| expansions=tuple(self._expansions), | |
| type=self._get_token_type(), | |
| ) | |
| def _get_token_type(self) -> TokenType: | |
| if self.redirect: | |
| return TokenType.REDIRECT_OPERATOR | |
| if self.is_comment: | |
| return TokenType.COMMENT | |
| return TokenType.WORD | |
| def _consolidate(self): | |
| self._parts = ["".join(self._parts)] | |
| @property | |
| def is_empty(self) -> bool: | |
| return ( | |
| not self._parts | |
| and not self.redirect | |
| and not self.active_prefix | |
| and not self.quote_active | |
| and not self.scope_terminator | |
| ) | |
| @property | |
| def escape_active(self) -> bool: | |
| return self.active_prefix == "\\" | |
| @property | |
| def is_comment(self) -> bool: | |
| return self.active_prefix == "#" | |
| @property | |
| def in_single_quotes(self) -> bool: | |
| return self.quote_active == "'" | |
| @property | |
| def in_double_quotes(self) -> bool: | |
| return self.quote_active == '"' | |
| @property | |
| def in_expansion(self) -> bool: | |
| return self.expansion_type is not None | |
| def init_redirect(self, digits: str = "", operator: str = ""): | |
| self.redirect = RedirectParseState(digits, operator) | |
| def cancel_redirect(self): | |
| """ | |
| This can be called when we parsed some digits then realized it's not a redirect. | |
| """ | |
| self.append(self.redirect.read()) | |
| self.redirect = None | |
| def start_expansion(self, opener: str, no_container: bool = False): | |
| assert not self.in_expansion | |
| self.append(opener, force_in_expansion=no_container) | |
| expansion_type = _expansion_types.get(opener, None) | |
| if expansion_type is None: | |
| raise TokenizerInternalError(f"Unknown scope opener {opener!r}") | |
| self.scope_terminator, self.expansion_type = expansion_type | |
| self._expansion_scope_start_index = len(self.read()) - len(opener) | |
| def end_expansion(self): | |
| assert self.in_expansion | |
| expansion_content = "".join(self._expansion_scope_content) | |
| self.append(self.scope_terminator) | |
| self._expansions.append( | |
| TokenFeature( | |
| type=self.expansion_type, | |
| content=expansion_content, | |
| start=self._expansion_scope_start_index, | |
| end=len(self.read()), | |
| quoted=self.in_double_quotes, | |
| ) | |
| ) | |
| self._expansion_scope_start_index = None | |
| self._expansion_scope_content = [] | |
| self.expansion_type = None | |
| self.scope_terminator = "" | |
| def cancel_expansion(self): | |
| assert self.in_expansion | |
| self._expansion_scope_start_index = None | |
| self._expansion_scope_content = [] | |
| self.expansion_type = None | |
| self.scope_terminator = "" | |
| def close_quotes(self): | |
| self.quote_active = "" | |
| # This just ensures we don't consider this token to be empty anymore | |
| self.append("") | |
| def append( | |
| self, char: str, escaped_glob: bool = False, force_in_expansion: bool = False | |
| ): | |
| if self.in_expansion or force_in_expansion: | |
| self._expansion_scope_content.append(char) | |
| self._parts.append(char) | |
| if escaped_glob: | |
| self._escaped_glob_indices.append(len(self.read()) - 1) | |
| class CommandLineTokenizer: | |
| pushback: List[str] = [] | |
| lookback: List[str] = [] | |
| def __init__(self, input_stream: IO[str]): | |
| self.input_stream = input_stream | |
| def parse_token(self): | |
| state = TokenParseState() | |
| next_char = None | |
| while True: | |
| if self.pushback: | |
| next_char = self.pushback.pop() | |
| print("p:", next_char) | |
| else: | |
| next_char = self.input_stream.read(1) | |
| if not next_char: | |
| # end of input | |
| if state.escape_active: | |
| raise CommandLineTokenizeError("No escaped character") | |
| if state.in_single_quotes or state.in_double_quotes: | |
| raise CommandLineTokenizeError("No closing quotation") | |
| if state.scope_terminator: | |
| raise CommandLineTokenizeError( | |
| f"No closing bracket in token {state.read()!r}" | |
| ) | |
| if state.in_expansion: | |
| for char in self.lookback: | |
| state.append(char) | |
| self.lookback.clear() | |
| state.end_expansion() | |
| if state.redirect: | |
| if state.redirect.valid: | |
| if not state.redirect.operator: | |
| state.cancel_redirect() | |
| else: | |
| raise CommandLineTokenizeError( | |
| f"Unexpected end of input in redirect: {state.read()!r}" | |
| ) | |
| if state.range_active: | |
| state.append("[") | |
| if self.lookback: | |
| self._cancel_lookback() | |
| if state.active_prefix: | |
| # TODO: this should vary depending on the prefix | |
| state.append(state.active_prefix) | |
| state.active_prefix = "" | |
| break | |
| if state.is_comment: | |
| if next_char == "\n": | |
| break | |
| state.append(next_char) | |
| elif state.range_active: | |
| if state.escape_active: | |
| self.lookback.append(next_char) | |
| state.escape_active = False | |
| elif next_char == "]" and self.lookback: | |
| # range expression complete | |
| state.includes_glob = True | |
| state.append("[") | |
| for char in self.lookback: | |
| state.append(char) | |
| self.lookback.clear() | |
| state.append(next_char) | |
| state.range_active = False | |
| elif next_char in NON_RANGE_CHARS: | |
| # range expression invalidated, pretend it never happened | |
| state.append("[", escaped_glob=True) | |
| self._cancel_lookback(next_char) | |
| state.range_active = False | |
| else: | |
| self.lookback.append(next_char) | |
| elif state.escape_active: | |
| if state.in_double_quotes and not next_char in '"$': | |
| state.append("\\") | |
| elif next_char in GLOB_INIT_CHARS: | |
| state.append(next_char, escaped_glob=True) | |
| else: | |
| state.append(next_char) | |
| state.active_prefix = "" | |
| elif state.in_single_quotes: | |
| if next_char == "'": | |
| state.close_quotes() | |
| elif next_char in GLOB_INIT_CHARS: | |
| state.append(next_char, escaped_glob=True) | |
| else: | |
| state.append(next_char) | |
| elif state.in_expansion: | |
| if state.scope_terminator == "}": | |
| if next_char == "}": | |
| state.end_expansion() | |
| else: | |
| state.append(next_char) | |
| elif state.scope_terminator == ")": | |
| # TODO: to do this properly we need to parse the sub command to tell | |
| # whether the closing bracket is quoted or escaped | |
| if next_char == ")": | |
| state.end_expansion() | |
| else: | |
| state.append(next_char) | |
| elif state.scope_terminator == "`": | |
| # N.B. we don't respect escaping of backticks inside backticks | |
| if next_char == "`": | |
| state.end_expansion() | |
| else: | |
| state.append(next_char) | |
| elif state.scope_terminator == "))": | |
| if next_char == ")": | |
| if state.paren_depth: | |
| state.paren_depth -= 1 | |
| state.append(next_char) | |
| elif self.lookback: | |
| if self.lookback[0] == ")": | |
| assert len(self.lookback) == 1 | |
| state.end_expansion() | |
| self.lookback.clear() | |
| else: | |
| self._cancel_lookback(next_char) | |
| else: | |
| self.lookback.append(next_char) | |
| continue | |
| elif next_char == "(": | |
| state.paren_depth += 1 | |
| state.append(next_char) | |
| elif next_char == "\\": | |
| state.active_prefix = next_char | |
| else: | |
| # TODO: verify if next_char is valid here? | |
| state.append(next_char) | |
| elif state.expansion_type == TokenFeatureType.VARIABLE: | |
| # This must be a variable name | |
| if next_char in VAR_CHARS: | |
| state.append(next_char) | |
| else: | |
| # We've gone off the end of the variable name, so we need to | |
| # start the loop again with the same same char. | |
| state.end_expansion() | |
| self.pushback.append(next_char) | |
| continue | |
| elif state.expansion_type == TokenFeatureType.TILDE: | |
| if next_char in "/ \n\t\n|;&<>(){}": | |
| # Reached the end of valid looking tilde expansion | |
| for char in self.lookback: | |
| state.append(char) | |
| self.lookback.clear() | |
| state.end_expansion() | |
| self.pushback.append(next_char) | |
| elif self.lookback: | |
| if self.lookback[0] in "+-": | |
| if next_char.isdigit(): | |
| self.lookback.append(next_char) | |
| else: | |
| self._cancel_lookback(next_char) | |
| state.cancel_expansion() | |
| elif next_char in VAR_CHARS or next_char in "-.": | |
| self.lookback.append(next_char) | |
| elif next_char in "+-" or next_char in VAR_CHARS: | |
| self.lookback.append(next_char) | |
| else: | |
| self._cancel_lookback(next_char) | |
| state.cancel_expansion() | |
| else: | |
| raise TokenizerInternalError( | |
| f"Unexpected expansion_type {state.expansion_type!r}" | |
| ) | |
| elif state.active_prefix == "$": | |
| if next_char == "(": | |
| state.active_prefix = "$(" | |
| else: | |
| state.active_prefix = "" | |
| if next_char == "{": | |
| state.start_expansion("${") | |
| elif next_char in SPECIAL_VARS: | |
| state.start_expansion("$") | |
| state.append(next_char) | |
| state.end_expansion() | |
| elif next_char in VAR_INIT_CHARS: | |
| state.start_expansion("$") | |
| state.append(next_char) | |
| else: | |
| # this $ has no meaning | |
| state.append("$") | |
| self.pushback.append(next_char) | |
| continue | |
| elif state.active_prefix == "$(": | |
| state.active_prefix = "" | |
| if next_char == "(": | |
| state.start_expansion("$((") | |
| else: | |
| state.start_expansion("$(") | |
| self.pushback.append(next_char) | |
| continue | |
| elif state.redirect: | |
| if rejected := state.redirect.append(next_char): | |
| if not state.redirect.operator: | |
| state.cancel_redirect() | |
| self.pushback.append(next_char) | |
| continue | |
| elif state.redirect.valid: | |
| for char in reversed(rejected): | |
| self.pushback.append(char) | |
| return state.get_token() | |
| else: | |
| raise CommandLineTokenizeError( | |
| f"Syntax error near redirect operator {state.read()!r}" | |
| ) | |
| elif state.active_prefix == "|": | |
| state.active_prefix = "" | |
| if next_char == "|": | |
| return Token.operator("||") | |
| elif next_char == "&": | |
| return Token.operator("|&") | |
| else: | |
| self.pushback.append(next_char) | |
| return Token.operator("|") | |
| elif state.active_prefix == "&": | |
| state.active_prefix = "" | |
| if next_char == "&": | |
| return Token.operator("&&") | |
| elif next_char == ">": | |
| state.init_redirect(operator="&>") | |
| else: | |
| self.pushback.append(next_char) | |
| return Token.operator("&") | |
| elif state.active_prefix == ";": | |
| if next_char == ";": | |
| state.active_prefix = ";;" | |
| elif next_char == "&": | |
| return Token.operator(";&") | |
| else: | |
| self.pushback.append(next_char) | |
| return Token.operator(";") | |
| elif state.active_prefix == ";;": | |
| if next_char == "&": | |
| return Token.operator(";;&") | |
| else: | |
| self.pushback.append(next_char) | |
| return Token.operator(";;") | |
| elif next_char == "`": | |
| state.start_expansion(next_char) | |
| elif state.in_double_quotes: | |
| if next_char == '"': | |
| state.close_quotes() | |
| elif next_char in PREFIX_CHARS: | |
| state.active_prefix = next_char | |
| elif next_char in GLOB_INIT_CHARS: | |
| state.append(next_char, escaped_glob=True) | |
| else: | |
| state.append(next_char) | |
| elif next_char in "><": | |
| if state.is_empty: | |
| state.init_redirect(operator=next_char) | |
| else: | |
| self.pushback.append(next_char) | |
| break | |
| elif next_char in "&|": | |
| if state.is_empty: | |
| # need to check the next char to know what this means | |
| state.active_prefix = next_char | |
| else: | |
| self.pushback.append(next_char) | |
| break | |
| elif next_char in WHITESPACE: | |
| if state.is_empty: | |
| # ignore leading whitespace | |
| continue | |
| else: | |
| # token complete | |
| break | |
| elif next_char in "'\"": | |
| # Open quotes and drop quote char | |
| state.quote_active = next_char | |
| elif next_char == "[": | |
| # This could be a range expression, but we need to look ahead | |
| # to parse the whole expression to be sure. This means validating | |
| # that there is a closing square bracked with valid contents. | |
| state.range_active = True | |
| elif next_char in "*?": | |
| state.includes_glob = True | |
| state.append(next_char) | |
| elif next_char in PREFIX_CHARS: | |
| state.active_prefix = next_char | |
| elif next_char == ";": | |
| if not state.is_empty: | |
| self.pushback.append(next_char) | |
| break | |
| state.active_prefix = ";" | |
| elif next_char == "~": | |
| if state.is_empty: | |
| # need to check the next char to know what this means | |
| state.start_expansion("~", no_container=True) | |
| else: | |
| state.append(next_char) | |
| elif next_char == "#": | |
| if not state.is_empty: | |
| self.pushback.append(next_char) | |
| break | |
| state.active_prefix = next_char | |
| state.append(next_char) | |
| elif next_char.isdigit() and state.is_empty: | |
| state.init_redirect(digits=next_char) | |
| else: | |
| # TODO: validate next_char is allowed out in the open like this | |
| # ... why wouldn't it be? | |
| # like () is usually a problem unless at start of token/line | |
| # ... need to count tokens? | |
| state.append(next_char) | |
| return state.get_token() | |
| def _cancel_lookback(self, next_char: str = ""): | |
| if next_char: | |
| self.pushback.append(next_char) | |
| for char in reversed(self.lookback): | |
| self.pushback.append(char) | |
| self.lookback.clear() | |
| def __iter__(self): | |
| return self | |
| def __next__(self): | |
| if token := self.parse_token(): | |
| return token | |
| raise StopIteration | |
| class CommandLineTokenizeError(RuntimeError): | |
| # TODO: keep track of where we are through the whole source for errors!! | |
| pass | |
| @property | |
| def message(self) -> str: | |
| return self.args[0] if self.args else "Failed to tokenize command line" | |
| class TokenizerInternalError(Exception): | |
| pass |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| import pytest | |
| import math | |
| from shell_tokenizer import ( | |
| tokenize, | |
| TokenType, | |
| TokenFeature, | |
| TokenFeatureType, | |
| CommandLineTokenizeError, | |
| ) | |
| VARIABLE = TokenFeatureType.VARIABLE | |
| COMMAND = TokenFeatureType.COMMAND | |
| ARITHMETIC = TokenFeatureType.ARITHMETIC | |
| TILDE = TokenFeatureType.TILDE | |
| def test_parse_empty_token(): | |
| tokens = tokenize("echo ''1 '' 2") | |
| assert_tokens( | |
| tokens, | |
| TokenDesc("echo"), | |
| TokenDesc("1"), | |
| TokenDesc(""), | |
| TokenDesc("2"), | |
| ) | |
| def test_parse_comments(): | |
| tokens = tokenize( | |
| """ | |
| echo hi! # la de da | |
| # |;&<>(){} # @!2 | |
| # | |
| ## | |
| # OK ;) \t | |
| """ | |
| ) | |
| assert_tokens( | |
| tokens, | |
| TokenDesc("echo"), | |
| TokenDesc("hi!"), | |
| TokenDesc("# la de da", type=TokenType.COMMENT), | |
| TokenDesc("# |;&<>(){} # @!2", type=TokenType.COMMENT), | |
| TokenDesc("#", type=TokenType.COMMENT), | |
| TokenDesc("##", type=TokenType.COMMENT), | |
| TokenDesc("# OK ;) \t", type=TokenType.COMMENT), | |
| ) | |
| def test_parse_glob_tokens(): | |
| # label unescaped ? * | |
| # label unescaped [ with valid content before ] | |
| tokens = tokenize( | |
| "echo h?llo/**/*.py * thing[123].txt " | |
| """ "h?llo/**/*.py" "*" "thing[123].txt" """ | |
| "'h?llo/**/*.py' '*' 'thing[123].txt' " | |
| "h\\?llo/\\*\\*/\\*.py \\* thing\\[123].txt " | |
| " h?ll'*' " | |
| " foo[ld] o[]r lo][l 2 " | |
| "[[] [[yep] [nope!] [still!*] " | |
| "[ha&]" | |
| ) | |
| # assert len(tokens) == 4 | |
| assert_tokens( | |
| tokens, | |
| TokenDesc("echo"), | |
| TokenDesc("h?llo/**/*.py", includes_glob=True), | |
| TokenDesc("*", includes_glob=True), | |
| TokenDesc("thing[123].txt", includes_glob=True), | |
| TokenDesc( | |
| "h?llo/**/*.py", | |
| includes_glob=False, | |
| escaped_glob_indices=( | |
| 1, | |
| 6, | |
| 7, | |
| 9, | |
| ), | |
| ), | |
| TokenDesc("*", includes_glob=False, escaped_glob_indices=(0,)), | |
| TokenDesc("thing[123].txt", includes_glob=False, escaped_glob_indices=(5,)), | |
| TokenDesc( | |
| "h?llo/**/*.py", | |
| includes_glob=False, | |
| escaped_glob_indices=( | |
| 1, | |
| 6, | |
| 7, | |
| 9, | |
| ), | |
| ), | |
| TokenDesc("*", includes_glob=False, escaped_glob_indices=(0,)), | |
| TokenDesc("thing[123].txt", includes_glob=False, escaped_glob_indices=(5,)), | |
| TokenDesc( | |
| "h?llo/**/*.py", | |
| includes_glob=False, | |
| escaped_glob_indices=( | |
| 1, | |
| 6, | |
| 7, | |
| 9, | |
| ), | |
| ), | |
| TokenDesc("*", includes_glob=False, escaped_glob_indices=(0,)), | |
| TokenDesc("thing[123].txt", includes_glob=False, escaped_glob_indices=(5,)), | |
| TokenDesc("h?ll*", includes_glob=True, escaped_glob_indices=(4,)), | |
| TokenDesc("foo[ld]", includes_glob=True), | |
| TokenDesc("o[]r", includes_glob=False, escaped_glob_indices=(1,)), | |
| TokenDesc("lo][l", includes_glob=False, escaped_glob_indices=(3,)), | |
| TokenDesc("2", includes_glob=False), | |
| TokenDesc("[[]", includes_glob=True), | |
| TokenDesc("[[yep]", includes_glob=True), | |
| TokenDesc("[nope!]", includes_glob=False, escaped_glob_indices=(0,)), | |
| TokenDesc("[still!*]", includes_glob=True, escaped_glob_indices=(0,)), | |
| TokenDesc("[ha", includes_glob=False, escaped_glob_indices=(0,)), | |
| TokenDesc("&", type=TokenType.CONTROL_OPERATOR, includes_glob=False), | |
| TokenDesc("]", includes_glob=False), | |
| ) | |
| def test_parse_arithmetic_expansions(): | |
| tokens = tokenize( | |
| """ | |
| echo "A$(( 1 + 1 ))B" '$(( 2**3 ))' A$((1&&1==( 1 || 1 )))B | |
| """ | |
| ) | |
| assert len(tokens) == 4 | |
| assert_tokens( | |
| tokens, | |
| TokenDesc("echo"), | |
| TokenDesc( | |
| "A$(( 1 + 1 ))B", | |
| expansions=( | |
| TokenFeature(ARITHMETIC, " 1 + 1 ", start=1, end=13, quoted=True), | |
| ), | |
| ), | |
| TokenDesc("$(( 2**3 ))", escaped_glob_indices=(5, 6)), | |
| TokenDesc( | |
| "A$((1&&1==( 1 || 1 )))B", | |
| expansions=( | |
| TokenFeature( | |
| ARITHMETIC, | |
| "1&&1==( 1 || 1 )", | |
| start=1, | |
| end=22, | |
| quoted=False, | |
| ), | |
| ), | |
| ), | |
| ) | |
| def test_parse_command_expansions(): | |
| tokens = tokenize( | |
| r""" | |
| echo "A$(ls "foo")B" '$(ls foo)' A$(ls foo)B | |
| """ | |
| ) | |
| assert len(tokens) == 4 | |
| assert_tokens( | |
| tokens, | |
| TokenDesc("echo"), | |
| TokenDesc( | |
| 'A$(ls "foo")B', | |
| expansions=( | |
| TokenFeature(COMMAND, 'ls "foo"', start=1, end=12, quoted=True), | |
| ), | |
| ), | |
| TokenDesc("$(ls foo)"), | |
| TokenDesc( | |
| "A$(ls foo)B", | |
| expansions=( | |
| TokenFeature(COMMAND, "ls foo", start=1, end=10, quoted=False), | |
| ), | |
| ), | |
| ) | |
| def test_parse_operator_tokens(): | |
| # NB we also parse `;;`, `;&`, `;;&` but we don't parse case statements so nevermind | |
| for op in ("&", "&&", "|", "||", "|&", ";"): | |
| tokens = tokenize(f"echo 1{op}echo 2 {op} echo 3") | |
| assert len(tokens) == 8 | |
| assert_tokens( | |
| tokens, | |
| TokenDesc("echo"), | |
| TokenDesc("1"), | |
| TokenDesc(op, type=TokenType.CONTROL_OPERATOR), | |
| TokenDesc("echo"), | |
| TokenDesc("2"), | |
| TokenDesc(op, type=TokenType.CONTROL_OPERATOR), | |
| TokenDesc("echo"), | |
| TokenDesc("3"), | |
| ) | |
| def test_parse_tilde_tokens(): | |
| tokens = tokenize( | |
| r""" | |
| echo ~ ~/ ~/foo ~user.name ~root/Library ~+ ~- ~12 ~+42 ~-11 ~-/foo ~+/foo x~ ~@ "~" | |
| """ | |
| ) | |
| assert len(tokens) == 16 | |
| assert_tokens( | |
| tokens, | |
| TokenDesc("echo"), | |
| TokenDesc( | |
| "~", | |
| expansions=(TokenFeature(TILDE, "~", start=0, end=1, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~/", | |
| expansions=(TokenFeature(TILDE, "~", start=0, end=1, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~/foo", | |
| expansions=(TokenFeature(TILDE, "~", start=0, end=1, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~user.name", | |
| expansions=( | |
| TokenFeature(TILDE, "~user.name", start=0, end=10, quoted=False), | |
| ), | |
| ), | |
| TokenDesc( | |
| "~root/Library", | |
| expansions=(TokenFeature(TILDE, "~root", start=0, end=5, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~+", | |
| expansions=(TokenFeature(TILDE, "~+", start=0, end=2, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~-", | |
| expansions=(TokenFeature(TILDE, "~-", start=0, end=2, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~12", | |
| expansions=(TokenFeature(TILDE, "~12", start=0, end=3, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~+42", | |
| expansions=(TokenFeature(TILDE, "~+42", start=0, end=4, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~-11", | |
| expansions=(TokenFeature(TILDE, "~-11", start=0, end=4, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~-/foo", | |
| expansions=(TokenFeature(TILDE, "~-", start=0, end=2, quoted=False),), | |
| ), | |
| TokenDesc( | |
| "~+/foo", | |
| expansions=(TokenFeature(TILDE, "~+", start=0, end=2, quoted=False),), | |
| ), | |
| TokenDesc("x~"), | |
| TokenDesc("~@"), | |
| TokenDesc("~"), | |
| ) | |
| def test_parse_backticks(): | |
| tokens = tokenize( | |
| r""" | |
| echo $WOO`goal or "${wut}" span`" and `more stuff` !" `yes` | |
| """ | |
| ) | |
| assert len(tokens) == 3 | |
| assert_token(tokens[0], "echo") | |
| assert_token( | |
| tokens[1], | |
| '$WOO`goal or "${wut}" span` and `more stuff` !', | |
| expansions=( | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="WOO", | |
| start=0, | |
| end=4, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=COMMAND, | |
| content='goal or "${wut}" span', | |
| start=4, | |
| end=27, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=COMMAND, | |
| content="more stuff", | |
| start=32, | |
| end=44, | |
| quoted=True, | |
| ), | |
| ), | |
| ) | |
| assert_token( | |
| tokens[2], | |
| "`yes`", | |
| expansions=( | |
| TokenFeature( | |
| type=COMMAND, | |
| content="yes", | |
| start=0, | |
| end=5, | |
| quoted=False, | |
| ), | |
| ), | |
| ) | |
| def test_parse_isolated_variables(): | |
| tokens = tokenize( | |
| """ | |
| echo $NAKED_UPPER $naked_lower_1 ${CURLIES1} ${curlies2} "$Quoted" "${QuotedBrackets}" '$SQuoted' '${SQuotedBrackets}' | |
| """ | |
| ) | |
| assert len(tokens) == 9 | |
| assert_token(tokens[0], "echo") | |
| assert_token( | |
| tokens[1], | |
| "$NAKED_UPPER", | |
| expansions=(TokenFeature(VARIABLE, "NAKED_UPPER", 0, 12, False),), | |
| ) | |
| assert_token( | |
| tokens[2], | |
| "$naked_lower_1", | |
| expansions=(TokenFeature(VARIABLE, "naked_lower_1", 0, 14, False),), | |
| ) | |
| assert_token( | |
| tokens[3], | |
| "${CURLIES1}", | |
| expansions=(TokenFeature(VARIABLE, "CURLIES1", 0, 11, False),), | |
| ) | |
| assert_token( | |
| tokens[4], | |
| "${curlies2}", | |
| expansions=(TokenFeature(VARIABLE, "curlies2", 0, 11, False),), | |
| ) | |
| assert_token( | |
| tokens[5], | |
| "$Quoted", | |
| expansions=(TokenFeature(VARIABLE, "Quoted", 0, 7, True),), | |
| ) | |
| assert_token( | |
| tokens[6], | |
| "${QuotedBrackets}", | |
| expansions=(TokenFeature(VARIABLE, "QuotedBrackets", 0, 17, True),), | |
| ) | |
| assert_token(tokens[7], "$SQuoted") | |
| assert_token(tokens[8], "${SQuotedBrackets}") | |
| tokens = tokenize("echo $") | |
| assert len(tokens) == 2 | |
| assert_token(tokens[0], "echo") | |
| assert_token(tokens[1], "$") | |
| def test_parse_escaped_variables(): | |
| tokens = tokenize( | |
| r""" | |
| echo \$NAKED_UPPER \\\$naked_lower_1 \\\\\${CURLIES1} \${curlies2} "\$Quoted" "\${QuotedBrackets}" '\$SQuoted' '\${SQuotedBrackets}' | |
| """ | |
| ) | |
| assert len(tokens) == 9 | |
| assert_token(tokens[0], "echo") | |
| assert_token(tokens[1], "$NAKED_UPPER") | |
| assert_token(tokens[2], "\\$naked_lower_1") | |
| assert_token(tokens[3], "\\\\${CURLIES1}") | |
| assert_token(tokens[4], "${curlies2}") | |
| assert_token(tokens[5], "$Quoted") | |
| assert_token(tokens[6], "${QuotedBrackets}") | |
| assert_token(tokens[7], "\\$SQuoted") | |
| assert_token(tokens[8], "\\${SQuotedBrackets}") | |
| def test_parse_special_variables(): | |
| tokens = tokenize( | |
| """ | |
| echo $FOO$0A$1A$2A$3A$4A$5A$6A$7A$8A$9A$$A$!A$*A$?A$-A$#A$${LOL} | |
| """ | |
| ) | |
| assert len(tokens) == 2 | |
| assert_token(tokens[0], "echo") | |
| assert_token( | |
| tokens[1], | |
| "$FOO$0A$1A$2A$3A$4A$5A$6A$7A$8A$9A$$A$!A$*A$?A$-A$#A$${LOL}", | |
| expansions=( | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="FOO", | |
| start=0, | |
| end=4, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="0", | |
| start=4, | |
| end=6, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="1", | |
| start=7, | |
| end=9, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="2", | |
| start=10, | |
| end=12, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="3", | |
| start=13, | |
| end=15, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="4", | |
| start=16, | |
| end=18, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="5", | |
| start=19, | |
| end=21, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="6", | |
| start=22, | |
| end=24, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="7", | |
| start=25, | |
| end=27, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="8", | |
| start=28, | |
| end=30, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="9", | |
| start=31, | |
| end=33, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="$", | |
| start=34, | |
| end=36, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="!", | |
| start=37, | |
| end=39, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="*", | |
| start=40, | |
| end=42, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="?", | |
| start=43, | |
| end=45, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="-", | |
| start=46, | |
| end=48, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="#", | |
| start=49, | |
| end=51, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="$", | |
| start=52, | |
| end=54, | |
| quoted=False, | |
| ), | |
| ), | |
| ) | |
| def test_parse_jumbled_variables(): | |
| tokens = tokenize( | |
| """ | |
| echo go$NAKED_UPPER$naked_lower_1>${CURLIES1}"${curlies2} $Quoted"ok''"${QuotedBrackets}"'$SQuoted'99'${SQuotedBrackets}' | |
| """ | |
| ) | |
| assert_token(tokens[0], "echo") | |
| assert_token( | |
| tokens[1], | |
| "go$NAKED_UPPER$naked_lower_1", | |
| expansions=( | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="NAKED_UPPER", | |
| start=2, | |
| end=14, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="naked_lower_1", | |
| start=14, | |
| end=28, | |
| quoted=False, | |
| ), | |
| ), | |
| ) | |
| assert_token(tokens[2], ">", type=2) | |
| assert_token( | |
| tokens[3], | |
| "${CURLIES1}${curlies2} $Quotedok${QuotedBrackets}$SQuoted99${SQuotedBrackets}", | |
| expansions=( | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="CURLIES1", | |
| start=0, | |
| end=11, | |
| quoted=False, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="curlies2", | |
| start=11, | |
| end=22, | |
| quoted=True, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="Quoted", | |
| start=23, | |
| end=30, | |
| quoted=True, | |
| ), | |
| TokenFeature( | |
| type=VARIABLE, | |
| content="QuotedBrackets", | |
| start=32, | |
| end=49, | |
| quoted=True, | |
| ), | |
| ), | |
| ) | |
| @pytest.mark.parametrize( | |
| "example", | |
| [ | |
| ">", | |
| "<", | |
| ">>", | |
| ">&", | |
| "<&", # semantically invalid without following digit | |
| "&>", | |
| "&>>", | |
| ">|", | |
| "<>", | |
| "1>", | |
| "1>>", | |
| "1>&", | |
| "1>|", | |
| "1<>", | |
| "1>&2", | |
| "1<&2", | |
| ">&2", | |
| "<&2", | |
| "1>&-", | |
| "1<&-", | |
| "1>&2-", | |
| "1<&2-", | |
| ">&2-", | |
| "<&2-", | |
| "11>", | |
| "11>>", | |
| "11>&", | |
| "11>|", | |
| "11<>", | |
| "11>&12", | |
| "11<&12", | |
| ">&12", | |
| "<&12", | |
| "11>&-", | |
| "11<&-", | |
| "11>&12-", | |
| "11<&12-", | |
| ">&12-", | |
| "<&12-", | |
| ], | |
| ) | |
| def test_tokenize_redirects(example): | |
| tokens = tokenize(f"echo foo {example} bar") | |
| assert len(tokens) == 4 | |
| assert_token(tokens[0], "echo") | |
| assert_token(tokens[1], "foo") | |
| assert_token(tokens[2], example, type=2) | |
| assert_token(tokens[3], "bar") | |
| @pytest.mark.parametrize( | |
| "example, equivalent", | |
| [ | |
| ("echo foo 1&> bar", "echo foo 1 &> bar"), | |
| ("echo foo 1&>bar", "echo foo 1 &> bar"), | |
| ("echo foo 1&>> bar", "echo foo 1 &>> bar"), | |
| ("echo foo 1&>>bar", "echo foo 1 &>> bar"), | |
| ("echo foo 1&>2 bar", "echo foo 1 &> 2 bar"), | |
| ("echo foo 1&>>2bar", "echo foo 1 &>> 2bar"), | |
| ("echo foo 1>2 bar", "echo foo 1> 2 bar"), | |
| ("echo foo 1>2bar", "echo foo 1> 2bar"), | |
| ("echo foo 1>>2 bar", "echo foo 1>> 2 bar"), | |
| ("echo foo 1>>2bar", "echo foo 1>> 2bar"), | |
| ("echo foo 1&>2 bar", "echo foo 1&> 2 bar"), | |
| ("echo foo 1&>2bar", "echo foo 1&> 2bar"), | |
| ("echo foo 1&>>2 bar", "echo foo 1&>> 2 bar"), | |
| ("echo foo 1&>>2bar", "echo foo 1&>> 2bar"), | |
| ("echo foo 1>|2 bar", "echo foo 1>| 2 bar"), | |
| ("echo foo 1>|2bar", "echo foo 1>| 2bar"), | |
| ("echo foo 1<>2 bar", "echo foo 1<> 2 bar"), | |
| ("echo foo 1<>2bar", "echo foo 1<> 2bar"), | |
| ("echo foo >2 bar", "echo foo > 2 bar"), | |
| ("echo foo >2bar", "echo foo > 2bar"), | |
| ("echo foo >>2 bar", "echo foo >> 2 bar"), | |
| ("echo foo >>2bar", "echo foo >> 2bar"), | |
| ("echo foo &>2 bar", "echo foo &> 2 bar"), | |
| ("echo foo &>2bar", "echo foo &> 2bar"), | |
| ("echo foo &>>2 bar", "echo foo &>> 2 bar"), | |
| ("echo foo &>>2bar", "echo foo &>> 2bar"), | |
| ("echo foo >|2 bar", "echo foo >| 2 bar"), | |
| ("echo foo >|2bar", "echo foo >| 2bar"), | |
| ("echo foo <>2 bar", "echo foo <> 2 bar"), | |
| ("echo foo <>2bar", "echo foo <> 2bar"), | |
| ("echo foo 2x> bar", "echo foo 2x > bar"), | |
| ("echo foo 2x>bar", "echo foo 2x > bar"), | |
| ("echo foo &< bar", "echo foo & < bar"), | |
| ("echo foo &<bar", "echo foo & < bar"), | |
| ], | |
| ) | |
| def test_tokenize_redirect_boundaries(example, equivalent): | |
| assert tokenize(example) == tokenize(equivalent) | |
| @pytest.mark.parametrize( | |
| "example, error_msg", | |
| [ | |
| ("echo foo &>| bar", "Syntax error near redirect operator '&>'"), | |
| ("echo foo &>< bar", "Syntax error near redirect operator '&>'"), | |
| ("echo foo &>& bar", "Syntax error near redirect operator '&>'"), | |
| ], | |
| ) | |
| def test_tokenize_redirect_errors(example, error_msg): | |
| with pytest.raises(CommandLineTokenizeError) as e_info: | |
| tokenize(example) | |
| assert e_info.value.args[0] == error_msg | |
| def test_tokens_with_digits(): | |
| tokens = tokenize(f"echo foo 888bar") | |
| assert_token(tokens[0], "echo") | |
| assert_token(tokens[1], "foo") | |
| assert_token(tokens[2], "888bar") | |
| tokens = tokenize(f"echo foo888 bar") | |
| assert_token(tokens[0], "echo") | |
| assert_token(tokens[1], "foo888") | |
| assert_token(tokens[2], "bar") | |
| tokens = tokenize(f"echo foo888bar") | |
| assert_token(tokens[0], "echo") | |
| assert_token(tokens[1], "foo888bar") | |
| tokens = tokenize(f"echo foo 888-bar") | |
| assert_token(tokens[0], "echo") | |
| assert_token(tokens[1], "foo") | |
| assert_token(tokens[2], "888-bar") | |
| tokens = tokenize(f"echo foo 888 bar") | |
| assert_token(tokens[0], "echo") | |
| assert_token(tokens[1], "foo") | |
| assert_token(tokens[2], "888") | |
| assert_token(tokens[3], "bar") | |
| def assert_token( | |
| token, | |
| content: str, | |
| includes_glob=False, | |
| escaped_glob_indices=tuple(), | |
| expansions=tuple(), | |
| type=0, | |
| ): | |
| assert token.content == content | |
| assert token.includes_glob == includes_glob | |
| assert token.escaped_glob_indices == escaped_glob_indices | |
| assert token.expansions == expansions | |
| assert token.type == TokenType(type) | |
| class TokenDesc: | |
| def __init__( | |
| self, | |
| content: str, | |
| includes_glob=False, | |
| escaped_glob_indices=tuple(), | |
| expansions=tuple(), | |
| type=0, | |
| ): | |
| self.content = content | |
| self.includes_glob = includes_glob | |
| self.escaped_glob_indices = escaped_glob_indices | |
| self.expansions = expansions | |
| self.type = type | |
| def assert_tokens(tokens, *token_descriptions): | |
| for index, desc in enumerate(token_descriptions): | |
| assert_token( | |
| tokens[index], | |
| content=desc.content, | |
| includes_glob=desc.includes_glob, | |
| escaped_glob_indices=desc.escaped_glob_indices, | |
| expansions=desc.expansions, | |
| type=desc.type, | |
| ) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment