difftreelog
feat(parser) better descriptions for strings/ops
in: master
1 file changed
crates/jrsonnet-parser/src/lib.rsdiffbeforeafterboth1use peg::parser;2use std::{path::PathBuf, rc::Rc};3mod expr;4pub use expr::*;5pub use peg;67#[derive(Default)]8pub struct ParserSettings {9 pub loc_data: bool,10 pub file_name: Rc<PathBuf>,11}1213parser! {14 grammar jsonnet_parser() for str {15 use peg::ParseLiteral;1617 /// Standard C-like comments18 rule comment()19 = "//" (!['\n'][_])* "\n"20 / "/*" ("\\*/" / "\\\\" / (!("*/")[_]))* "*/"21 / "#" (!['\n'][_])* "\n"2223 rule single_whitespace() = quiet!{([' ' | '\r' | '\n' | '\t'] / comment())} / expected!("<whitespace>")24 rule _() = single_whitespace()*2526 /// For comma-delimited elements27 rule comma() = quiet!{_ "," _} / expected!("<comma>")28 rule alpha() -> char = c:$(['_' | 'a'..='z' | 'A'..='Z']) {c.chars().next().unwrap()}29 rule digit() -> char = d:$(['0'..='9']) {d.chars().next().unwrap()}30 rule end_of_ident() = !['0'..='9' | '_' | 'a'..='z' | 'A'..='Z']31 /// Sequence of digits32 rule uint() -> u64 = a:$(digit()+) { a.parse().unwrap() }33 /// Number in scientific notation format34 rule number() -> f64 = quiet!{a:$(uint() ("." uint())? (['e'|'E'] (s:['+'|'-'])? uint())?) { a.parse().unwrap() }} / expected!("<number>")3536 /// Reserved word followed by any non-alphanumberic37 rule reserved() = ("assert" / "else" / "error" / "false" / "for" / "function" / "if" / "import" / "importstr" / "in" / "local" / "null" / "tailstrict" / "then" / "self" / "super" / "true") end_of_ident()38 rule id() = quiet!{ !reserved() alpha() (alpha() / digit())*} / expected!("<identifier>")3940 rule keyword(id: &'static str)41 = ##parse_string_literal(id) end_of_ident()42 // Adds location data information to existing expression43 rule l(s: &ParserSettings, x: rule<Expr>) -> LocExpr44 = start:position!() v:x() end:position!() {loc_expr!(v, s.loc_data, (s.file_name.clone(), start, end))}4546 pub rule param(s: &ParserSettings) -> expr::Param = name:$(id()) expr:(_ "=" _ expr:expr(s){expr})? { expr::Param(name.into(), expr) }47 pub rule params(s: &ParserSettings) -> expr::ParamsDesc48 = params:param(s) ** comma() comma()? {49 let mut defaults_started = false;50 for param in ¶ms {51 defaults_started = defaults_started || param.1.is_some();52 assert_eq!(defaults_started, param.1.is_some(), "defauld parameters should be used after all positionals");53 }54 expr::ParamsDesc(Rc::new(params))55 }56 / { expr::ParamsDesc(Rc::new(Vec::new())) }5758 pub rule arg(s: &ParserSettings) -> expr::Arg59 = name:$(id()) _ "=" _ expr:expr(s) {expr::Arg(Some(name.into()), expr)}60 / expr:expr(s) {expr::Arg(None, expr)}61 pub rule args(s: &ParserSettings) -> expr::ArgsDesc62 = args:arg(s) ** comma() comma()? {63 let mut named_started = false;64 for arg in &args {65 named_started = named_started || arg.0.is_some();66 assert_eq!(named_started, arg.0.is_some(), "named args should be used after all positionals");67 }68 expr::ArgsDesc(args)69 }70 / { expr::ArgsDesc(Vec::new()) }7172 pub rule bind(s: &ParserSettings) -> expr::BindSpec73 = name:$(id()) _ "=" _ expr:expr(s) {expr::BindSpec{name:name.into(), params: None, value: expr}}74 / name:$(id()) _ "(" _ params:params(s) _ ")" _ "=" _ expr:expr(s) {expr::BindSpec{name:name.into(), params: Some(params), value: expr}}75 pub rule assertion(s: &ParserSettings) -> expr::AssertStmt76 = keyword("assert") _ cond:expr(s) msg:(_ ":" _ e:expr(s) {e})? { expr::AssertStmt(cond, msg) }7778 pub rule whole_line() -> &'input str79 = str:$((!['\n'][_])* "\n") {str}80 pub rule string_block() -> String81 = "|||" (!['\n']single_whitespace())* "\n"82 empty_lines:$(['\n']*)83 prefix:[' ' | '\t']+ first_line:whole_line()84 lines:("\n" {"\n"} / [' ' | '\t']*<{prefix.len()}> s:whole_line() {s})*85 [' ' | '\t']*<, {prefix.len() - 1}> "|||"86 {let mut l = empty_lines.to_owned(); l.push_str(first_line); l.extend(lines); l}87 pub rule string() -> String88 = "\"" str:$(("\\\"" / "\\\\" / (!['"'][_]))*) "\"" {unescape::unescape(str).unwrap()}89 / "'" str:$(("\\'" / "\\\\" / (!['\''][_]))*) "'" {unescape::unescape(str).unwrap()}90 / "@'" str:$(("''" / (!['\''][_]))*) "'" {str.replace("''", "'")}91 / "@\"" str:$(("\"\"" / (!['"'][_]))*) "\"" {str.replace("\"\"", "\"")}92 / string_block()9394 pub rule field_name(s: &ParserSettings) -> expr::FieldName95 = name:$(id()) {expr::FieldName::Fixed(name.into())}96 / name:string() {expr::FieldName::Fixed(name.into())}97 / "[" _ expr:expr(s) _ "]" {expr::FieldName::Dyn(expr)}98 pub rule visibility() -> expr::Visibility99 = ":::" {expr::Visibility::Unhide}100 / "::" {expr::Visibility::Hidden}101 / ":" {expr::Visibility::Normal}102 pub rule field(s: &ParserSettings) -> expr::FieldMember103 = name:field_name(s) _ plus:"+"? _ visibility:visibility() _ value:expr(s) {expr::FieldMember{104 name,105 plus: plus.is_some(),106 params: None,107 visibility,108 value,109 }}110 / name:field_name(s) _ "(" _ params:params(s) _ ")" _ visibility:visibility() _ value:expr(s) {expr::FieldMember{111 name,112 plus: false,113 params: Some(params),114 visibility,115 value,116 }}117 pub rule obj_local(s: &ParserSettings) -> BindSpec118 = keyword("local") _ bind:bind(s) {bind}119 pub rule member(s: &ParserSettings) -> expr::Member120 = bind:obj_local(s) {expr::Member::BindStmt(bind)}121 / assertion:assertion(s) {expr::Member::AssertStmt(assertion)}122 / field:field(s) {expr::Member::Field(field)}123 pub rule objinside(s: &ParserSettings) -> expr::ObjBody124 = pre_locals:(b: obj_local(s) comma() {b})* "[" _ key:expr(s) _ "]" _ ":" _ value:expr(s) post_locals:(comma() b:obj_local(s) {b})* _ forspec:forspec(s) others:(_ rest:compspec(s) {rest})? {125 let mut compspecs = vec![CompSpec::ForSpec(forspec)];126 compspecs.extend(others.unwrap_or_default());127 expr::ObjBody::ObjComp(expr::ObjComp{128 pre_locals,129 key,130 value,131 post_locals,132 compspecs,133 })134 }135 / members:(member(s) ** comma()) comma()? {expr::ObjBody::MemberList(members)}136 pub rule ifspec(s: &ParserSettings) -> IfSpecData137 = keyword("if") _ expr:expr(s) {IfSpecData(expr)}138 pub rule forspec(s: &ParserSettings) -> ForSpecData139 = keyword("for") _ id:$(id()) _ keyword("in") _ cond:expr(s) {ForSpecData(id.into(), cond)}140 pub rule compspec(s: &ParserSettings) -> Vec<expr::CompSpec>141 = s:(i:ifspec(s) { expr::CompSpec::IfSpec(i) } / f:forspec(s) {expr::CompSpec::ForSpec(f)} ) ** _ {s}142 pub rule local_expr(s: &ParserSettings) -> LocExpr143 = l(s,<keyword("local") _ binds:bind(s) ** comma() _ ";" _ expr:expr(s) { Expr::LocalExpr(binds, expr) }>)144 pub rule string_expr(s: &ParserSettings) -> LocExpr145 = l(s, <s:string() {Expr::Str(s.into())}>)146 pub rule obj_expr(s: &ParserSettings) -> LocExpr147 = l(s,<"{" _ body:objinside(s) _ "}" {Expr::Obj(body)}>)148 pub rule array_expr(s: &ParserSettings) -> LocExpr149 = l(s,<"[" _ elems:(expr(s) ** comma()) _ comma()? "]" {Expr::Arr(elems)}>)150 pub rule array_comp_expr(s: &ParserSettings) -> LocExpr151 = l(s,<"[" _ expr:expr(s) _ comma()? _ forspec:forspec(s) _ others:(others: compspec(s) _ {others})? "]" {152 let mut specs = vec![CompSpec::ForSpec(forspec)];153 specs.extend(others.unwrap_or_default());154 Expr::ArrComp(expr, specs)155 }>)156 pub rule number_expr(s: &ParserSettings) -> LocExpr157 = l(s,<n:number() { expr::Expr::Num(n) }>)158 pub rule var_expr(s: &ParserSettings) -> LocExpr159 = l(s,<n:$(id()) { expr::Expr::Var(n.into()) }>)160 pub rule if_then_else_expr(s: &ParserSettings) -> LocExpr161 = l(s,<cond:ifspec(s) _ keyword("then") _ cond_then:expr(s) cond_else:(_ keyword("else") _ e:expr(s) {e})? {Expr::IfElse{162 cond,163 cond_then,164 cond_else,165 }}>)166167 pub rule literal(s: &ParserSettings) -> LocExpr168 = l(s,<v:(169 keyword("null") {LiteralType::Null}170 / keyword("true") {LiteralType::True}171 / keyword("false") {LiteralType::False}172 / keyword("self") {LiteralType::This}173 / keyword("$") {LiteralType::Dollar}174 / keyword("super") {LiteralType::Super}175 ) {Expr::Literal(v)}>)176177 pub rule expr_basic(s: &ParserSettings) -> LocExpr178 = literal(s)179180 / string_expr(s) / number_expr(s)181 / array_expr(s)182 / obj_expr(s)183 / array_expr(s)184 / array_comp_expr(s)185186 / l(s,<keyword("importstr") _ path:string() {Expr::ImportStr(PathBuf::from(path))}>)187 / l(s,<keyword("import") _ path:string() {Expr::Import(PathBuf::from(path))}>)188189 / var_expr(s)190 / local_expr(s)191 / if_then_else_expr(s)192193 / l(s,<keyword("function") _ "(" _ params:params(s) _ ")" _ expr:expr(s) {Expr::Function(params, expr)}>)194 / l(s,<assertion:assertion(s) _ ";" _ expr:expr(s) { Expr::AssertExpr(assertion, expr) }>)195196 / l(s,<keyword("error") _ expr:expr(s) { Expr::ErrorStmt(expr) }>)197198 rule slice_part(s: &ParserSettings) -> Option<LocExpr>199 = e:(_ e:expr(s) _{e})? {e}200 pub rule slice_desc(s: &ParserSettings) -> SliceDesc201 = start:slice_part(s) ":" pair:(end:slice_part(s) step:(":" e:slice_part(s){e})? {(end, step.flatten())})? {202 let (end, step) = if let Some((end, step)) = pair {203 (end, step)204 }else{205 (None, None)206 };207208 SliceDesc { start, end, step }209 }210211 rule expr(s: &ParserSettings) -> LocExpr212 = start:position!() a:precedence! {213 a:(@) _ "||" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Or, b))}214 --215 a:(@) _ "&&" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::And, b))}216 --217 a:(@) _ "|" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::BitOr, b))}218 --219 a:@ _ "^" _ b:(@) {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::BitXor, b))}220 --221 a:(@) _ "&" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::BitAnd, b))}222 --223 a:(@) _ "==" _ b:@ {loc_expr_todo!(Expr::Apply(224 el!(Expr::Intrinsic("equals".into())),225 ArgsDesc(vec![Arg(None, a), Arg(None, b)]),226 true227 ))}228 a:(@) _ "!=" _ b:@ {loc_expr_todo!(Expr::UnaryOp(UnaryOpType::Not, el!(Expr::Apply(229 el!(Expr::Intrinsic("equals".into())),230 ArgsDesc(vec![Arg(None, a), Arg(None, b)]),231 true232 ))))}233 --234 a:(@) _ "<" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Lt, b))}235 a:(@) _ ">" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Gt, b))}236 a:(@) _ "<=" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Lte, b))}237 a:(@) _ ">=" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Gte, b))}238 a:(@) _ keyword("in") _ b:@ {loc_expr_todo!(Expr::Apply(239 el!(Expr::Intrinsic("objectHasEx".into())), ArgsDesc(vec![Arg(None, b), Arg(None, a), Arg(None, el!(Expr::Literal(LiteralType::True)))]),240 true241 ))}242 --243 a:(@) _ "<<" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Lhs, b))}244 a:(@) _ ">>" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Rhs, b))}245 --246 a:(@) _ "+" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Add, b))}247 a:(@) _ "-" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Sub, b))}248 --249 a:(@) _ "*" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Mul, b))}250 a:(@) _ "/" _ b:@ {loc_expr_todo!(Expr::BinaryOp(a, BinaryOpType::Div, b))}251 a:(@) _ "%" _ b:@ {loc_expr_todo!(Expr::Apply(252 el!(Expr::Intrinsic("mod".into())), ArgsDesc(vec![Arg(None, a), Arg(None, b)]),253 false254 ))}255 --256 "-" _ b:@ {loc_expr_todo!(Expr::UnaryOp(UnaryOpType::Minus, b))}257 "!" _ b:@ {loc_expr_todo!(Expr::UnaryOp(UnaryOpType::Not, b))}258 "~" _ b:@ { loc_expr_todo!(Expr::UnaryOp(UnaryOpType::BitNot, b)) }259 --260 a:(@) _ "[" _ s:slice_desc(s) _ "]" {loc_expr_todo!(Expr::Apply(261 el!(Expr::Intrinsic("slice".into())),262 ArgsDesc(vec![263 Arg(None, a),264 Arg(None, s.start.unwrap_or_else(||el!(Expr::Literal(LiteralType::Null)))),265 Arg(None, s.end.unwrap_or_else(||el!(Expr::Literal(LiteralType::Null)))),266 Arg(None, s.step.unwrap_or_else(||el!(Expr::Literal(LiteralType::Null)))),267 ]),268 true,269 ))}270 a:(@) _ "." _ s:$(id()) {loc_expr_todo!(Expr::Index(a, el!(Expr::Str(s.into()))))}271 a:(@) _ "[" _ s:expr(s) _ "]" {loc_expr_todo!(Expr::Index(a, s))}272 a:(@) _ "(" _ args:args(s) _ ")" ts:(_ keyword("tailstrict"))? {loc_expr_todo!(Expr::Apply(a, args, ts.is_some()))}273 a:(@) _ "{" _ body:objinside(s) _ "}" {loc_expr_todo!(Expr::ObjExtend(a, body))}274 --275 e:expr_basic(s) {e}276 "(" _ e:expr(s) _ ")" {loc_expr_todo!(Expr::Parened(e))}277 } end:position!() {278 let LocExpr(e, _) = a;279 LocExpr(e, if s.loc_data {280 Some(ExprLocation(s.file_name.clone(), start, end))281 } else {282 None283 })284 }285 / e:expr_basic(s) {e}286287 pub rule jsonnet(s: &ParserSettings) -> LocExpr = _ e:expr(s) _ {e}288 }289}290291pub type ParseError = peg::error::ParseError<peg::str::LineCol>;292pub fn parse(str: &str, settings: &ParserSettings) -> Result<LocExpr, ParseError> {293 jsonnet_parser::jsonnet(str, settings)294}295296#[macro_export]297macro_rules! el {298 ($expr:expr) => {299 LocExpr(std::rc::Rc::new($expr), None)300 };301}302303#[cfg(test)]304pub mod tests {305 use super::{expr::*, parse};306 use crate::ParserSettings;307 use std::path::PathBuf;308 use std::rc::Rc;309310 macro_rules! parse {311 ($s:expr) => {312 parse(313 $s,314 &ParserSettings {315 loc_data: false,316 file_name: Rc::new(PathBuf::from("/test.jsonnet")),317 },318 )319 .unwrap()320 };321 }322323 mod expressions {324 use super::*;325326 pub fn basic_math() -> LocExpr {327 el!(Expr::BinaryOp(328 el!(Expr::Num(2.0)),329 BinaryOpType::Add,330 el!(Expr::BinaryOp(331 el!(Expr::Num(2.0)),332 BinaryOpType::Mul,333 el!(Expr::Num(2.0)),334 )),335 ))336 }337 }338339 #[test]340 fn multiline_string() {341 assert_eq!(342 parse!("|||\n Hello world!\n a\n|||"),343 el!(Expr::Str("Hello world!\n a\n".into())),344 );345 assert_eq!(346 parse!("|||\n Hello world!\n a\n|||"),347 el!(Expr::Str("Hello world!\n a\n".into())),348 );349 assert_eq!(350 parse!("|||\n\t\tHello world!\n\t\t\ta\n|||"),351 el!(Expr::Str("Hello world!\n\ta\n".into())),352 );353 assert_eq!(354 parse!("|||\n Hello world!\n a\n |||"),355 el!(Expr::Str("Hello world!\n a\n".into())),356 );357 }358359 #[test]360 fn slice() {361 parse!("a[1:]");362 parse!("a[1::]");363 parse!("a[:1:]");364 parse!("a[::1]");365 parse!("str[:len - 1]");366 }367368 #[test]369 fn string_escaping() {370 assert_eq!(371 parse!(r#""Hello, \"world\"!""#),372 el!(Expr::Str(r#"Hello, "world"!"#.into())),373 );374 assert_eq!(375 parse!(r#"'Hello \'world\'!'"#),376 el!(Expr::Str("Hello 'world'!".into())),377 );378 assert_eq!(parse!(r#"'\\\\'"#), el!(Expr::Str("\\\\".into())),);379 }380381 #[test]382 fn string_unescaping() {383 assert_eq!(384 parse!(r#""Hello\nWorld""#),385 el!(Expr::Str("Hello\nWorld".into())),386 );387 }388389 #[test]390 fn string_verbantim() {391 assert_eq!(392 parse!(r#"@"Hello\n""World""""#),393 el!(Expr::Str("Hello\\n\"World\"".into())),394 );395 }396397 #[test]398 fn imports() {399 assert_eq!(400 parse!("import \"hello\""),401 el!(Expr::Import(PathBuf::from("hello"))),402 );403 assert_eq!(404 parse!("importstr \"garnish.txt\""),405 el!(Expr::ImportStr(PathBuf::from("garnish.txt")))406 );407 }408409 #[test]410 fn empty_object() {411 assert_eq!(parse!("{}"), el!(Expr::Obj(ObjBody::MemberList(vec![]))));412 }413414 #[test]415 fn basic_math() {416 assert_eq!(417 parse!("2+2*2"),418 el!(Expr::BinaryOp(419 el!(Expr::Num(2.0)),420 BinaryOpType::Add,421 el!(Expr::BinaryOp(422 el!(Expr::Num(2.0)),423 BinaryOpType::Mul,424 el!(Expr::Num(2.0))425 ))426 ))427 );428 }429430 #[test]431 fn basic_math_with_indents() {432 assert_eq!(parse!("2 + 2 * 2 "), expressions::basic_math());433 }434435 #[test]436 fn basic_math_parened() {437 assert_eq!(438 parse!("2+(2+2*2)"),439 el!(Expr::BinaryOp(440 el!(Expr::Num(2.0)),441 BinaryOpType::Add,442 el!(Expr::Parened(expressions::basic_math())),443 ))444 );445 }446447 /// Comments should not affect parsing448 #[test]449 fn comments() {450 assert_eq!(451 parse!("2//comment\n+//comment\n3/*test*/*/*test*/4"),452 el!(Expr::BinaryOp(453 el!(Expr::Num(2.0)),454 BinaryOpType::Add,455 el!(Expr::BinaryOp(456 el!(Expr::Num(3.0)),457 BinaryOpType::Mul,458 el!(Expr::Num(4.0))459 ))460 ))461 );462 }463464 /// Comments should be able to be escaped465 #[test]466 fn comment_escaping() {467 assert_eq!(468 parse!("2/*\\*/+*/ - 22"),469 el!(Expr::BinaryOp(470 el!(Expr::Num(2.0)),471 BinaryOpType::Sub,472 el!(Expr::Num(22.0))473 ))474 );475 }476477 #[test]478 fn suffix() {479 // assert_eq!(parse!("std.test"), el!(Expr::Num(2.2)));480 // assert_eq!(parse!("std(2)"), el!(Expr::Num(2.2)));481 // assert_eq!(parse!("std.test(2)"), el!(Expr::Num(2.2)));482 // assert_eq!(parse!("a[b]"), el!(Expr::Num(2.2)))483 }484485 #[test]486 fn array_comp() {487 use Expr::*;488 assert_eq!(489 parse!("[std.deepJoin(x) for x in arr]"),490 el!(ArrComp(491 el!(Apply(492 el!(Index(el!(Var("std".into())), el!(Str("deepJoin".into())))),493 ArgsDesc(vec![Arg(None, el!(Var("x".into())))]),494 false,495 )),496 vec![CompSpec::ForSpec(ForSpecData(497 "x".into(),498 el!(Var("arr".into()))499 ))]500 )),501 )502 }503504 #[test]505 fn reserved() {506 use Expr::*;507 assert_eq!(parse!("null"), el!(Literal(LiteralType::Null)));508 assert_eq!(parse!("nulla"), el!(Var("nulla".into())));509 }510511 #[test]512 fn multiple_args_buf() {513 parse!("a(b, null_fields)");514 }515516 #[test]517 fn infix_precedence() {518 use Expr::*;519 assert_eq!(520 parse!("!a && !b"),521 el!(BinaryOp(522 el!(UnaryOp(UnaryOpType::Not, el!(Var("a".into())))),523 BinaryOpType::And,524 el!(UnaryOp(UnaryOpType::Not, el!(Var("b".into()))))525 ))526 );527 }528529 #[test]530 fn infix_precedence_division() {531 use Expr::*;532 assert_eq!(533 parse!("!a / !b"),534 el!(BinaryOp(535 el!(UnaryOp(UnaryOpType::Not, el!(Var("a".into())))),536 BinaryOpType::Div,537 el!(UnaryOp(UnaryOpType::Not, el!(Var("b".into()))))538 ))539 );540 }541542 #[test]543 fn double_negation() {544 use Expr::*;545 assert_eq!(546 parse!("!!a"),547 el!(UnaryOp(548 UnaryOpType::Not,549 el!(UnaryOp(UnaryOpType::Not, el!(Var("a".into()))))550 ))551 )552 }553554 #[test]555 fn array_test_error() {556 parse!("[a for a in b if c for e in f]");557 // ^^^^ failed code558 }559560 #[test]561 fn can_parse_stdlib() {562 parse!(jrsonnet_stdlib::STDLIB_STR);563 }564565 // From source code566 /*567 #[bench]568 fn bench_parse_peg(b: &mut Bencher) {569 b.iter(|| parse!(jrsonnet_stdlib::STDLIB_STR))570 }571 */572}