wip cmd-type-dsl

This commit is contained in:
Michael Sippel 2023-10-30 15:30:42 +01:00
parent 63e06a2472
commit 1d860d1811
Signed by: senvas
GPG key ID: F96CF119C34B64A6
8 changed files with 50 additions and 646 deletions

View file

@ -1,115 +0,0 @@
use std::boxed::Box;
//<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\
#[derive(Debug, PartialEq)]
pub enum Command {
Simple {
assignments: Vec<Assignment>,
command_word: Word,
redirections: Vec<Redirection>
},
Pipeline(Vec<Command>),
Sequence(Vec<Command>),
ShortCircuitConjunction(Vec<Command>),
ShortCircuitDisjunction(Vec<Command>),
Negation(Box<Command>),
While {
condition: Box<Command>,
loop_body: Box<Command>
},
For {
varname: String,
sequence: Word,
loop_body: Box<Command>
},
If {
condition: Box<Command>,
then_branch: Box<Command>,
else_branch: Box<Command>
},
Case {
expr: Word,
cases: Vec<(Word, Command)>
},
Function {
name: String,
body: Box<Command>
}
}
//<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\
#[derive(Debug, PartialEq)]
pub struct Assignment {
pub name: String,
pub value: Word
}
#[derive(Debug, PartialEq)]
pub struct Word {
pub segments: Vec<WordSegment>
}
#[derive(Debug, PartialEq)]
pub enum WordSegment {
Tilde(String),
Literal(String),
Parameter(String, ParameterFormat),
Subshell(Command),
DoubleQuote(Word),
}
#[derive(Debug, PartialEq)]
pub enum ParameterFormat {
Normal,
Length,
Default(Word),
Assign(Word),
Error(Word),
Alt(Word),
Sub(ParamSubSide, ParamSubMode, Word),
}
#[derive(Debug, PartialEq)]
pub enum ParamSubMode {
Shortest, Longest
}
#[derive(Debug, PartialEq)]
pub enum ParamSubSide {
Prefix, Suffix
}
//<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\
#[derive(Debug, PartialEq)]
pub struct Redirection {
redirection_type: RedirectionType,
fd: u64,
target: Word
}
#[derive(Debug, PartialEq)]
pub enum RedirectionType {
File(FileRedirectionType),
Dup(DupRedirectionType),
Heredoc // '<<'
}
#[derive(Debug, PartialEq)]
pub enum FileRedirectionType {
In, // '<'
InOut, // '<>'
Out, // '>'
OutReplace, // '>|'
OutAppend, // '>>'
}
#[derive(Debug, PartialEq)]
pub enum DupRedirectionType {
In, // '<&'
Out // '>&'
}
//<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\

View file

@ -3,17 +3,12 @@ use std::{
boxed::Box boxed::Box
}; };
use crate::ast::Command; use crate::sh::ast::Command;
use laddertypes::*; use laddertypes::*;
pub struct Substitution(HashMap< String, CommandTypeExpr >); pub struct Substitution(HashMap< String, CommandTypeExpr >);
impl Substitution {
pub fn apply(&self, expr: &mut CommandTypeExpr) {
}
}
#[derive(Clone)]
pub enum CommandArgPattern { pub enum CommandArgPattern {
Literal(String), Literal(String),
Variable(String), Variable(String),
@ -23,39 +18,58 @@ pub enum CommandArgPattern {
Disjunction(Vec<CommandArgPattern>) Disjunction(Vec<CommandArgPattern>)
} }
#[derive(Clone)]
pub struct CommandPattern { pub struct CommandPattern {
name: String, name: String,
args: Vec<CommandArgPattern>, args: Vec<CommandArgPattern>,
env: Vec<(String, CommandTypeExpr)>, env: Vec<(String, CommandTypeExpr)>,
} }
#[derive(Clone)]
pub struct MatchCandidate {
at: usize,
expected: CommandPattern,
found: CommandTypeExpr,
}
#[derive(Clone)]
pub struct UnificationError( Vec<MatchCandidate> );
#[derive(Clone)]
pub enum CommandTypeExpr {
Type(TypeTerm),
Match(Box<CommandTypeExpr>, Vec<(CommandArgPattern, CommandTypeExpr)>)
}
impl CommandArgPattern { impl CommandArgPattern {
pub fn match_cmd(&self, cmd: &Command) -> Result<Substitution, UnificationError> { pub fn match_cmd(&self, cmd: &Command) -> Result<Substitution, UnificationError> {
Err(UnificationError(vec![])) Err(UnificationError(vec![]))
} }
} }
pub struct MatchCandidate {
at: usize,
expected: CommandPattern,
found: CommandTypeExpr,
}
pub struct UnificationError( Vec<MatchCandidate> );
pub enum CommandTypeExpr {
Parameter(String),
ParameterPack(String),
Char(char),
Match(Box<CommandTypeExpr>, Vec<(CommandArgPattern, CommandTypeExpr)>)
}
impl CommandTypeExpr { impl CommandTypeExpr {
pub fn eval(self) -> CommandTypeExpr { pub fn eval(self) -> Result<TypeTerm, CommandTypeExpr> {
match self { match self {
s=>s CommandTypeExpr::Type(typ) => Ok(typ),
CommandTypeExpr::Match(pattern, cases) => {
}
s=> Ok(s)
}
}
pub fn apply_subst(&mut self, subst: &Substitution) {
match self {
CommandTypeExpr::Type(typ) => {
self = CommandTypeExpr::Type(
typ.apply_substitution(|v: String| subst.get(v))
);
}
CommandTypeExpr::Match( pattern, cases ) => {
// todo
}
_ => {}
} }
} }
} }

View file

@ -1,9 +0,0 @@
use std::collections::HashMap;
//<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\
struct Environment {
variables: HashMap<String, String>
}
//<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\

View file

@ -1,44 +0,0 @@
use crate::ast::*;
//<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\
impl Word {
pub fn (&mut self, env: &Environment) {
for x in self.0.iter_mut() {
x.expand_tilde();
match x {
Word::Tilde => Word::Literal( env.get_home() ),
other => other,
}
}
}
pub fn expand(&self) -> Vec<String> {
let mut fields = Vec::new();
for seg in self.segments.iter() {
//
}
fields
}
pub fn split_field(&mut self) {
}
}
impl WordSegment {
pub fn split_field(&self) -> Word {
match self {
}
}
pub fn expand(&self) -> Word {
match
}
}
//<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\

View file

@ -6,11 +6,7 @@ use {
tiny_ansi::TinyAnsi tiny_ansi::TinyAnsi
}; };
mod ast; mod sh;
mod env;
mod parse;
//mod expand;
mod cmd_type_dsl; mod cmd_type_dsl;
//<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\ //<<<<>>>><<>><><<>><<<*>>><<>><><<>><<<<>>>>\\
@ -42,7 +38,7 @@ fn main() {
let stdin = std::io::stdin(); let stdin = std::io::stdin();
for line in std::io::BufReader::new(stdin).lines() { for line in std::io::BufReader::new(stdin).lines() {
if let Ok(line) = line { if let Ok(line) = line {
let cmd = parse::parse_cmd( &mut line.chars().peekable() ); let cmd = sh::parse::parse_cmd( &mut line.chars().peekable() );
eprintln!("parsed cmd: {:?}", cmd); eprintln!("parsed cmd: {:?}", cmd);
/* /*
let mut lex = parse::WordLexer::from( line.chars() ); let mut lex = parse::WordLexer::from( line.chars() );

View file

@ -1,438 +0,0 @@
use {
crate::ast::*,
std::iter::{Peekable},
};
#[derive(Debug, PartialEq)]
pub enum LexError {
UnexpectedEnd(Vec<Option<char>>),
UnexpectedToken(char),
InvalidFileRedirectionType
}
///! iterates chars until it finds some char in `delim`
pub struct DelimIter<'a, It>
where It: Iterator<Item = char> {
chars: &'a mut Peekable<It>,
delim: Vec<(Option<char>, bool)>
}
impl<'a, It> DelimIter<'a, It>
where It: Iterator<Item = char> {
fn new(chars: &'a mut Peekable<It>, delim: Vec<(Option<char>, bool)>) -> Self {
DelimIter { chars, delim }
}
fn new_whitespace(chars: &'a mut Peekable<It>) -> Self {
DelimIter::new(chars, vec![
(None, true),
(Some(' '), true),
(Some('\t'), true),
(Some('\n'), true)
])
}
fn new_shell_word(chars: &'a mut Peekable<It>) -> Self {
DelimIter::new(chars, vec![
(None, true),
(Some(' '), true),
(Some('\t'), true),
(Some('\n'), true),
(Some('|'), false),
(Some('&'), false),
(Some(';'), false),
(Some('\"'), false),
(Some('\''), false)
])
}
fn new_shell_word_or_assignment(chars: &'a mut Peekable<It>) -> Self {
DelimIter::new(chars, vec![
(None, true),
(Some(' '), true),
(Some('\t'), true),
(Some('\n'), true),
(Some('='), false),
(Some('|'), false),
(Some('&'), false),
(Some(';'), false),
(Some('\"'), false),
(Some('\''), false)
])
}
}
impl<'a, It> Iterator for DelimIter<'a, It>
where It: 'a + Iterator<Item = char> {
type Item = Result<char, LexError>;
fn next(&mut self) -> Option<Result<char, LexError>> {
for (delim, consume) in self.delim.iter() {
if self.chars.peek().cloned() == *delim {
if *consume {
self.chars.next();
}
return None;
}
}
match self.chars.next() {
Some(c) => Some(Ok(c)),
None => Some(Err(LexError::UnexpectedEnd(vec![])))
}
}
}
pub struct WordLexer<'a, It>
where It: 'a + Iterator<Item = char> {
chars: &'a mut Peekable<It>
}
impl<'a, It> WordLexer<'a, It>
where It: Iterator<Item = char> {
fn collect_until(&mut self, close: Option<char>) -> Result<String, LexError> {
DelimIter::new(&mut self.chars, vec![(close, true)])
.try_collect::<String>()
}
}
pub fn skip_whitespace<It>(chars: &mut Peekable<It>)
where It: Iterator<Item = char>
{
while let Some(c) = chars.peek() {
if c.is_whitespace() {
chars.next();
} else {
break;
}
}
}
pub fn parse_quoted<It>(chars: &mut Peekable<It>) -> Result<WordSegment, LexError>
where It: Iterator<Item = char>
{
assert_eq!( chars.next(), Some('\''));
let quoted = DelimIter::new(chars, vec![(Some('\''), true)]).try_collect::<String>();
match quoted {
Ok(s) => {
Ok(WordSegment::Literal(s))
},
Err(e) => Err(e)
}
}
pub fn parse_doublequoted<It>(chars: &mut Peekable<It>) -> Result<WordSegment, LexError>
where It: Iterator<Item = char>
{
assert_eq!( chars.next(), Some('\"'));
let quoted = DelimIter::new(chars, vec![(Some('\"'), true)]).try_collect::<String>();
match quoted {
Ok(s) => {
let word = Word {
segments: // fixme: handle spaces correctly -> create QuoteLexer
WordLexer { chars: &mut s.chars().peekable() }
.scan((), |_, x| x.ok())
.collect::<Vec<_>>()
};
Ok(WordSegment::DoubleQuote(word))
},
Err(e) => Err(e)
}
}
pub fn parse_word<It>(chars: &mut Peekable<It>) -> Result<Word, LexError>
where It: Iterator<Item = char>
{
Ok(Word {
segments: WordLexer{ chars }.try_collect::<Vec<_>>()?
})
}
impl std::str::FromStr for FileRedirectionType {
type Err = LexError;
fn from_str(s: &str) -> Result<FileRedirectionType, LexError> {
match s {
"<" => Ok(FileRedirectionType::In),
"<>" => Ok(FileRedirectionType::InOut),
">" => Ok(FileRedirectionType::Out),
">|" => Ok(FileRedirectionType::OutReplace),
">>" => Ok(FileRedirectionType::OutAppend),
_ => Err(LexError::InvalidFileRedirectionType)
}
}
}
pub fn parse_redirection<It>(chars: &mut Peekable<It>) -> Result<Redirection, LexError>
where It: Iterator<Item = char>
{
Err(LexError::InvalidFileRedirectionType)
// let name = DelimIterator::new(chars, vec!['<', '>']).collect::<String>();
}
pub fn parse_simple_cmd<It>(chars: &mut Peekable<It>) -> Result<Option<Command>, LexError>
where It: Iterator<Item = char>
{
let mut assignments = Vec::new();
let mut redirections = Vec::new();
if chars.peek() == None {
return Ok(None);
}
loop {
skip_whitespace(chars);
let mut name = DelimIter::new_shell_word_or_assignment(chars).try_collect::<String>()?;
match chars.peek().clone() {
Some('=') => {
chars.next();
let mut lex = WordLexer{ chars };
match lex.next() {
Some(Ok(value)) => {
assignments.push(Assignment { name, value: Word{ segments: vec![ value ] } });
},
Some(Err(e)) => {
return Err(e);
},
None => {
return Err(LexError::UnexpectedEnd(vec![]));
}
}
}
_ => {
let mut cmd_segments = WordLexer{ chars }.try_collect::<Vec<_>>()?;
cmd_segments.insert(0, WordSegment::Literal(name));
return Ok(Some(Command::Simple {
assignments,
command_word: Word { segments: cmd_segments },
redirections,
}));
}
}
}
}
pub fn parse_cmd<It>(chars: &mut Peekable<It>) -> Result<Option<Command>, LexError>
where It: Iterator<Item = char>
{
skip_whitespace(chars);
match chars.peek() {
Some('!') => {
chars.next();
if let Some(cmd) = parse_cmd(chars)? {
Ok(Some(Command::Negation(Box::new(cmd))))
} else {
Err(LexError::UnexpectedEnd(vec![]))
}
}
_ => {
if let Some(head) = parse_simple_cmd(chars)? {
skip_whitespace(chars);
match chars.peek() {
Some(';') => {
chars.next();
let tail = parse_cmd( chars ) ?;
match tail {
Some(Command::Sequence(mut s)) => {
s.insert(0, head);
Ok(Some(Command::Sequence(s)))
}
Some(tail) => {
Ok(Some(Command::Sequence(vec![ head, tail ])))
}
None => {
Ok(Some(head))
}
}
}
Some('|') => {
chars.next();
match chars.peek() {
Some('|') => {
chars.next();
let tail = parse_cmd( chars ) ?;
match tail {
Some(Command::ShortCircuitDisjunction(mut s)) => {
s.insert(0, head);
Ok(Some(Command::ShortCircuitDisjunction(s)))
}
Some(tail) => {
Ok(Some(Command::ShortCircuitDisjunction(vec![ head, tail ])))
}
None => {
Err(LexError::UnexpectedEnd(vec![Some('|')]))
}
}
}
_ => {
let tail = parse_cmd( chars ) ?;
match tail {
Some(Command::Pipeline(mut s)) => {
s.insert(0, head);
Ok(Some(Command::Pipeline(s)))
}
Some(tail) => {
Ok(Some(Command::Pipeline(vec![ head, tail ])))
}
None => {
Err(LexError::UnexpectedEnd(vec![]))
}
}
}
}
}
Some('&') => {
chars.next();
match chars.peek() {
Some('&') => {
chars.next();
let tail = parse_cmd( chars ) ?;
match tail {
Some(Command::ShortCircuitConjunction(mut s)) => {
s.insert(0, head);
Ok(Some(Command::ShortCircuitConjunction(s)))
}
Some(tail) => {
Ok(Some(Command::ShortCircuitConjunction(vec![ head, tail ])))
}
None => {
Err(LexError::UnexpectedEnd(vec![Some('&'), Some('&')]))
}
}
}
Some(c) => {
Err(LexError::UnexpectedToken(*c))
}
None => {
// todo:
// background job
Ok(Some(head))
}
}
}
Some(c) => {
Err(LexError::UnexpectedToken(*c))
}
None => {
Ok(Some(head))
}
}
} else {
Ok(None)
}
}
}
}
impl<'a, It> Iterator for WordLexer<'a, It>
where It: 'a + Iterator<Item = char> {
type Item = Result<WordSegment, LexError>;
fn next(&mut self) -> Option<Result<WordSegment, LexError>> {
skip_whitespace(self.chars);
match self.chars.peek().cloned() {
Some('|') => { None }
Some('&') => { None }
Some(';') => { None }
Some('~') => {
self.chars.next();
let user = DelimIter::new_whitespace(self.chars).collect();
match user {
Ok(user) => Some(Ok(WordSegment::Tilde(user))),
Err(e) => Some(Err(e))
}
}
Some('"') => { Some(parse_doublequoted(self.chars)) },
Some('\'') => { Some(parse_quoted(self.chars)) },
Some('$') => {
self.chars.next();
match self.chars.peek() {
Some('{') => {
self.chars.next();
match DelimIter::new(&mut self.chars, vec![(Some('}'), true)]).try_collect::<String>() {
Ok(s) => {
Some(Ok(WordSegment::Parameter(s, ParameterFormat::Normal)))
}
Err(e) => Some(Err(e))
}
}
Some('(') => {
self.chars.next();
let subcmd_str = DelimIter::new(&mut self.chars, vec![(Some(')'), true)]).try_collect::<String>();
match subcmd_str {
Ok(subcmd_str) => {
match parse_cmd(&mut subcmd_str.chars().peekable()) {
Ok(Some(subcmd)) => {
Some(Ok(WordSegment::Subshell(subcmd)))
}
Ok(None) => None,
Err(err) => Some(Err(err))
}
}
Err(err) => Some(Err(err))
}
}
_ => {
match DelimIter::new_whitespace(self.chars).collect() {
Ok(s) => {
Some(Ok(WordSegment::Parameter(s, ParameterFormat::Normal)))
}
Err(e) => Some(Err(e))
}
}
}
}
Some(c) => {
let s : Result<String, LexError> = DelimIter::new_shell_word(self.chars).collect();
match s {
Ok(s) => Some(Ok(WordSegment::Literal(s))),
Err(e) => Some(Err(e))
}
}
None => {
None
}
}
}
}
mod test {
use crate::parse::*;
#[test]
fn test_delim_iter() {
let mut cs = "test 1234".chars().peekable();
let mut lexer = DelimIter::new_shell_word(&mut cs);
assert_eq!(lexer.try_collect::<String>(), Ok(String::from("test")));
}
#[test]
fn test_word_lexer() {
let mut cs = "test 1234|test".chars().peekable();
{
let mut lexer = WordLexer{ chars: &mut cs };
assert_eq!(lexer.next(), Some(Ok(WordSegment::Literal(String::from("test")))));
assert_eq!(lexer.next(), Some(Ok(WordSegment::Literal(String::from("1234")))));
assert_eq!(lexer.next(), None);
}
assert_eq!(cs.next(), Some('|'));
{
let mut lexer = WordLexer{ chars: &mut cs };
assert_eq!(lexer.next(), Some(Ok(WordSegment::Literal(String::from("test")))));
assert_eq!(lexer.next(), None);
}
}
}