1use std::{
4 any::type_name,
5 borrow::Borrow,
6 collections::{HashMap, HashSet},
7 env::{current_dir, var},
8 error::Error,
9 fmt::{self, Debug, Display, Write as _},
10 fs::{self, create_dir_all, read_to_string, File},
11 hash::Hash,
12 io::Write,
13 path::{Path, PathBuf},
14 sync::Mutex,
15};
16
17use bincode::Encode;
18use cfgrammar::{
19 header::{
20 GrmtoolsSectionParser, Header, HeaderError, HeaderErrorKind, HeaderValue, Namespaced,
21 Setting, Value,
22 },
23 markmap::MergeBehavior,
24 span::{Location, Span},
25};
26use glob::glob;
27use lazy_static::lazy_static;
28use lrpar::{
29 diagnostics::{DiagnosticFormatter, SpannedDiagnosticFormatter},
30 CTParserBuilder, LexerTypes,
31};
32use num_traits::{AsPrimitive, PrimInt, Unsigned};
33use proc_macro2::TokenStream;
34use quote::{format_ident, quote, ToTokens, TokenStreamExt};
35use regex::Regex;
36
37use crate::{DefaultLexerTypes, LRNonStreamingLexer, LRNonStreamingLexerDef, LexFlags, LexerDef};
38
39const RUST_FILE_EXT: &str = "rs";
40
41const ERROR: &str = "[Error]";
42const WARNING: &str = "[Warning]";
43
44lazy_static! {
45 static ref RE_TOKEN_ID: Regex = Regex::new(r"^[a-zA-Z_][a-zA-Z_0-9]*$").unwrap();
46 static ref GENERATED_PATHS: Mutex<HashSet<PathBuf>> = Mutex::new(HashSet::new());
47}
48
49#[non_exhaustive]
50pub enum LexerKind {
51 LRNonStreamingLexer,
52}
53
54impl<T: Clone> TryFrom<&Value<T>> for LexerKind {
55 type Error = cfgrammar::header::HeaderError<T>;
56 fn try_from(it: &Value<T>) -> Result<LexerKind, Self::Error> {
57 match it {
58 Value::Flag(_, loc) => Err(HeaderError {
59 kind: HeaderErrorKind::ConversionError(
60 "LexerKind",
61 "Expected `LexerKind` found bool",
62 ),
63 locations: vec![loc.clone()],
64 }),
65 Value::Setting(Setting::Num(_, loc)) => Err(HeaderError {
66 kind: HeaderErrorKind::ConversionError(
67 "LexerKind",
68 "Expected `LexerKind` found numeric",
69 ),
70 locations: vec![loc.clone()],
71 }),
72 Value::Setting(Setting::String(_, loc)) => Err(HeaderError {
73 kind: HeaderErrorKind::ConversionError(
74 "LexerKind",
75 "Expected `LexerKind` found string",
76 ),
77 locations: vec![loc.clone()],
78 }),
79 Value::Setting(Setting::Constructor {
80 ctor:
81 Namespaced {
82 namespace: _,
83 member: (_, loc),
84 },
85 arg: _,
86 }) => Err(HeaderError {
87 kind: HeaderErrorKind::ConversionError(
88 "LexerKind",
89 "Expected `LexerKind` found constructor",
90 ),
91 locations: vec![loc.clone()],
92 }),
93 Value::Setting(Setting::Unitary(Namespaced {
94 namespace,
95 member: (member, member_loc),
96 })) => {
97 if let Some((ns, loc)) = namespace {
98 if ns.to_lowercase() != "lexerkind" {
99 return Err(HeaderError {
100 kind: HeaderErrorKind::ConversionError(
101 "LexerKind",
102 "Expected namespace `LexerKind`",
103 ),
104 locations: vec![loc.clone()],
105 });
106 }
107 }
108 if member.to_lowercase() != "lrnonstreaminglexer" {
109 return Err(HeaderError {
110 kind: HeaderErrorKind::ConversionError(
111 "LexerKind",
112 "Unknown `LexerKind` Variant",
113 ),
114 locations: vec![member_loc.clone()],
115 });
116 }
117
118 Ok(LexerKind::LRNonStreamingLexer)
119 }
120 }
121 }
122}
123
124#[derive(Clone, PartialEq, Eq, Debug)]
126#[non_exhaustive]
127pub enum Visibility {
128 Private,
130 Public,
132 PublicSuper,
134 PublicSelf,
136 PublicCrate,
138 PublicIn(String),
140}
141
142impl ToTokens for Visibility {
143 fn to_tokens(&self, tokens: &mut TokenStream) {
144 tokens.extend(match self {
145 Visibility::Private => quote!(),
146 Visibility::Public => quote! {pub},
147 Visibility::PublicSuper => quote! {pub(super)},
148 Visibility::PublicSelf => quote! {pub(self)},
149 Visibility::PublicCrate => quote! {pub(crate)},
150 Visibility::PublicIn(data) => {
151 let other = str::parse::<TokenStream>(data).unwrap();
152 quote! {pub(in #other)}
153 }
154 })
155 }
156}
157
158#[derive(Clone, Copy, PartialEq, Eq, Debug)]
162#[non_exhaustive]
163pub enum RustEdition {
164 Rust2015,
165 Rust2018,
166 Rust2021,
167}
168
169struct QuoteOption<T>(Option<T>);
175
176impl<T: ToTokens> ToTokens for QuoteOption<T> {
177 fn to_tokens(&self, tokens: &mut TokenStream) {
178 tokens.append_all(match self.0 {
179 Some(ref t) => quote! { ::std::option::Option::Some(#t) },
180 None => quote! { ::std::option::Option::None },
181 });
182 }
183}
184
185struct QuoteTuple<T>(T);
188
189impl<A: ToTokens, B: ToTokens> ToTokens for QuoteTuple<(A, B)> {
190 fn to_tokens(&self, tokens: &mut TokenStream) {
191 let (a, b) = &self.0;
192 tokens.append_all(quote!((#a, #b)));
193 }
194}
195
196struct QuoteToString<'a>(&'a str);
198
199impl ToTokens for QuoteToString<'_> {
200 fn to_tokens(&self, tokens: &mut TokenStream) {
201 let x = &self.0;
202 tokens.append_all(quote! { #x.to_string() });
203 }
204}
205
206struct ErrorString(String);
208impl fmt::Display for ErrorString {
209 fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
210 let ErrorString(s) = self;
211 write!(f, "{}", s)
212 }
213}
214impl fmt::Debug for ErrorString {
215 fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
216 let ErrorString(s) = self;
217 write!(f, "{}", s)
218 }
219}
220impl Error for ErrorString {}
221
222pub struct CTLexerBuilder<'a, LexerTypesT: LexerTypes = DefaultLexerTypes<u32>>
225where
226 LexerTypesT::StorageT: Debug + Eq + Hash + ToTokens,
227 usize: num_traits::AsPrimitive<LexerTypesT::StorageT>,
228{
229 lrpar_config: Option<Box<dyn Fn(CTParserBuilder<LexerTypesT>) -> CTParserBuilder<LexerTypesT>>>,
230 lexer_path: Option<PathBuf>,
231 output_path: Option<PathBuf>,
232 lexerkind: Option<LexerKind>,
233 mod_name: Option<&'a str>,
234 visibility: Visibility,
235 rust_edition: RustEdition,
236 rule_ids_map: Option<HashMap<String, LexerTypesT::StorageT>>,
237 allow_missing_terms_in_lexer: bool,
238 allow_missing_tokens_in_parser: bool,
239 warnings_are_errors: bool,
240 show_warnings: bool,
241 header: Header<Location>,
242 #[cfg(test)]
243 inspect_lexerkind_cb: Option<Box<dyn Fn(LexerKind) -> Result<(), Box<dyn Error>>>>,
244}
245
246impl CTLexerBuilder<'_, DefaultLexerTypes<u32>> {
247 pub fn new() -> Self {
249 CTLexerBuilder::<DefaultLexerTypes<u32>>::new_with_lexemet()
250 }
251}
252
253impl<'a, LexerTypesT: LexerTypes<LexErrorT = crate::LRLexError> + 'static>
254 CTLexerBuilder<'a, LexerTypesT>
255where
256 LexerTypesT::StorageT:
257 'static + Debug + Eq + Hash + PrimInt + Encode + TryFrom<usize> + Unsigned + ToTokens,
258 usize: AsPrimitive<LexerTypesT::StorageT>,
259{
260 pub fn new_with_lexemet() -> Self {
277 let mut header = Header::new();
278 header.set_default_merge_behavior(MergeBehavior::Ours);
279 CTLexerBuilder {
280 lrpar_config: None,
281 lexer_path: None,
282 output_path: None,
283 lexerkind: None,
284 mod_name: None,
285 visibility: Visibility::Private,
286 rust_edition: RustEdition::Rust2021,
287 rule_ids_map: None,
288 allow_missing_terms_in_lexer: false,
289 allow_missing_tokens_in_parser: false,
290 warnings_are_errors: false,
291 show_warnings: true,
292 header,
293 #[cfg(test)]
294 inspect_lexerkind_cb: None,
295 }
296 }
297
298 pub fn lrpar_config<F>(mut self, config_func: F) -> Self
317 where
318 F: 'static + Fn(CTParserBuilder<LexerTypesT>) -> CTParserBuilder<LexerTypesT>,
319 {
320 self.lrpar_config = Some(Box::new(config_func));
321 self
322 }
323
324 pub fn lexer_in_src_dir<P>(mut self, srcp: P) -> Result<Self, Box<dyn Error>>
342 where
343 P: AsRef<Path>,
344 {
345 if !srcp.as_ref().is_relative() {
346 return Err(format!(
347 "Lexer path '{}' must be a relative path.",
348 srcp.as_ref().to_str().unwrap_or("<invalid UTF-8>")
349 )
350 .into());
351 }
352
353 let mut lexp = current_dir()?;
354 lexp.push("src");
355 lexp.push(srcp.as_ref());
356 self.lexer_path = Some(lexp);
357
358 let mut outp = PathBuf::new();
359 outp.push(var("OUT_DIR").unwrap());
360 outp.push(srcp.as_ref().parent().unwrap().to_str().unwrap());
361 create_dir_all(&outp)?;
362 let mut leaf = srcp
363 .as_ref()
364 .file_name()
365 .unwrap()
366 .to_str()
367 .unwrap()
368 .to_owned();
369 write!(leaf, ".{}", RUST_FILE_EXT).ok();
370 outp.push(leaf);
371 Ok(self.output_path(outp))
372 }
373
374 pub fn lexer_path<P>(mut self, inp: P) -> Self
378 where
379 P: AsRef<Path>,
380 {
381 self.lexer_path = Some(inp.as_ref().to_owned());
382 self
383 }
384
385 pub fn output_path<P>(mut self, outp: P) -> Self
390 where
391 P: AsRef<Path>,
392 {
393 self.output_path = Some(outp.as_ref().to_owned());
394 self
395 }
396
397 pub fn lexerkind(mut self, lexerkind: LexerKind) -> Self {
399 self.lexerkind = Some(lexerkind);
400 self
401 }
402
403 pub fn mod_name(mut self, mod_name: &'a str) -> Self {
407 self.mod_name = Some(mod_name);
408 self
409 }
410
411 pub fn visibility(mut self, vis: Visibility) -> Self {
413 self.visibility = vis;
414 self
415 }
416
417 pub fn rust_edition(mut self, edition: RustEdition) -> Self {
420 self.rust_edition = edition;
421 self
422 }
423
424 pub fn rule_ids_map<T: std::borrow::Borrow<HashMap<String, LexerTypesT::StorageT>> + Clone>(
429 mut self,
430 rule_ids_map: T,
431 ) -> Self {
432 self.rule_ids_map = Some(rule_ids_map.borrow().to_owned());
433 self
434 }
435
436 pub fn build(mut self) -> Result<CTLexer, Box<dyn Error>> {
456 let lexerp = self
457 .lexer_path
458 .as_ref()
459 .expect("lexer_path must be specified before processing.");
460 let outp = self
461 .output_path
462 .as_ref()
463 .expect("output_path must be specified before processing.");
464
465 {
466 let mut lk = GENERATED_PATHS.lock().unwrap();
467 if lk.contains(outp.as_path()) {
468 return Err(format!("Generating two lexers to the same path ('{}') is not allowed: use CTLexerBuilder::output_path (and, optionally, CTLexerBuilder::mod_name) to differentiate them.", &outp.to_str().unwrap()).into());
469 }
470 lk.insert(outp.clone());
471 }
472 let lex_src = read_to_string(lexerp)
473 .map_err(|e| format!("When reading '{}': {e}", lexerp.display()))?;
474 let lex_diag = SpannedDiagnosticFormatter::new(&lex_src, lexerp);
475 let mut header = self.header;
476 let (parsed_header, _) = GrmtoolsSectionParser::new(&lex_src, false)
477 .parse()
478 .map_err(|es| {
479 let mut out = String::new();
480 out.push_str(&format!(
481 "\n{ERROR}{}\n",
482 lex_diag.file_location_msg(" parsing the `%grmtools` section", None)
483 ));
484 for e in es {
485 out.push_str(&indent(" ", &lex_diag.format_error(e).to_string()));
486 out.push('\n');
487 }
488 ErrorString(out)
489 })?;
490 header.merge_from(parsed_header)?;
491 header.mark_used(&"lexerkind".to_string());
492 let lexerkind = match self.lexerkind {
493 Some(lexerkind) => lexerkind,
494 None => {
495 if let Some(HeaderValue(_, lk_val)) = header.get("lexerkind") {
496 LexerKind::try_from(lk_val)?
497 } else {
498 LexerKind::LRNonStreamingLexer
499 }
500 }
501 };
502 #[cfg(test)]
503 if let Some(inspect_lexerkind_cb) = self.inspect_lexerkind_cb {
504 inspect_lexerkind_cb(lexerkind)?
505 }
506 let (lexerdef, lex_flags): (LRNonStreamingLexerDef<LexerTypesT>, LexFlags) =
507 match lexerkind {
508 LexerKind::LRNonStreamingLexer => {
509 let lex_flags = LexFlags::try_from(&mut header)?;
510 let lexerdef = LRNonStreamingLexerDef::<LexerTypesT>::new_with_options(
511 &lex_src, lex_flags,
512 )
513 .map_err(|errs| {
514 let mut out = String::new();
515 out.push_str(&format!(
516 "\n{ERROR}{}\n",
517 lex_diag.file_location_msg("", None)
518 ));
519 for e in errs {
520 out.push_str(&indent(" ", &lex_diag.format_error(e).to_string()));
521 out.push('\n');
522 }
523 ErrorString(out)
524 })?;
525 let lex_flags = lexerdef.lex_flags().cloned();
526 (lexerdef, lex_flags.unwrap())
527 }
528 };
529
530 let ct_parser = if let Some(ref lrcfg) = self.lrpar_config {
531 let mut closure_lexerdef = lexerdef.clone();
532 let mut ctp = CTParserBuilder::<LexerTypesT>::new().inspect_rt(Box::new(
533 move |yacc_header, rtpb, rule_ids_map, grm_path| {
534 let owned_map = rule_ids_map
535 .iter()
536 .map(|(x, y)| (&**x, *y))
537 .collect::<HashMap<_, _>>();
538 closure_lexerdef.set_rule_ids(&owned_map);
539 yacc_header.mark_used(&"test_files".to_string());
540 let test_glob = yacc_header.get("test_files");
541 match test_glob {
542 Some(HeaderValue(_, Value::Setting(Setting::String(test_files, _)))) => {
543 let path_joined = grm_path.parent().unwrap().join(test_files);
544 for path in
545 glob(&path_joined.to_string_lossy()).map_err(|e| e.to_string())?
546 {
547 let path = path?;
548 let input = fs::read_to_string(&path)?;
549 let l: LRNonStreamingLexer<LexerTypesT> =
550 closure_lexerdef.lexer(&input);
551 for e in rtpb.parse_noaction(&l) {
552 Err(format!("parsing {}: {}", path.display(), e))?
553 }
554 }
555 Ok(())
556 }
557 Some(_) => Err("Invalid value for setting 'test_files'".into()),
558 None => Ok(()),
559 }
560 },
561 ));
562 ctp = lrcfg(ctp);
563 let ct_parser = ctp.build()?;
564 self.rule_ids_map = Some(ct_parser.token_map().to_owned());
565 Some(ct_parser)
566 } else {
567 None
568 };
569
570 let mut lexerdef = Box::new(lexerdef);
571 let unused_header_values = header.unused();
572 if !unused_header_values.is_empty() {
573 return Err(
574 format!("Unused header values: {}", unused_header_values.join(", ")).into(),
575 );
576 }
577
578 let (missing_from_lexer, missing_from_parser) = match self.rule_ids_map {
579 Some(ref rim) => {
580 let owned_map = rim
582 .iter()
583 .map(|(x, y)| (&**x, *y))
584 .collect::<HashMap<_, _>>();
585 let (x, y) = lexerdef.set_rule_ids_spanned(&owned_map);
586 (
587 x.map(|a| a.iter().map(|&b| b.to_string()).collect::<HashSet<_>>()),
588 y.map(|a| {
589 a.iter()
590 .map(|(b, span)| (b.to_string(), *span))
591 .collect::<HashSet<_>>()
592 }),
593 )
594 }
595 None => (None, None),
596 };
597
598 let mut has_unallowed_missing = false;
599 let err_indent = " ".repeat(ERROR.len());
600 if !self.allow_missing_terms_in_lexer {
601 if let Some(ref mfl) = missing_from_lexer {
602 if let Some(ct_parser) = &ct_parser {
603 let grm = ct_parser.yacc_grammar();
604 let token_spans = mfl
605 .iter()
606 .map(|name| {
607 ct_parser
608 .yacc_grammar()
609 .token_span(*grm.tokens_map().get(name.as_str()).unwrap())
610 .expect("Given token should have a span")
611 })
612 .collect::<Vec<_>>();
613
614 let yacc_diag = SpannedDiagnosticFormatter::new(
615 ct_parser.grammar_src(),
616 ct_parser.grammar_path(),
617 );
618
619 eprintln!("{ERROR} these tokens are not referenced in the lexer but defined as follows");
620 eprintln!(
621 "{err_indent} {}",
622 yacc_diag.file_location_msg("in the grammar", None)
623 );
624 for span in token_spans {
625 eprintln!(
626 "{}",
627 yacc_diag.underline_span_with_text(
628 span,
629 "Missing from lexer".to_string(),
630 '^'
631 )
632 );
633 }
634 eprintln!();
635 } else {
636 eprintln!("{ERROR} the following tokens are used in the grammar but are not defined in the lexer:");
637 for n in mfl {
638 eprintln!(" {}", n);
639 }
640 }
641 has_unallowed_missing = true;
642 }
643 }
644 if !self.allow_missing_tokens_in_parser && self.show_warnings {
645 if let Some(ref mfp) = missing_from_parser {
646 let error_prefix = if self.warnings_are_errors {
647 ERROR
648 } else {
649 WARNING
650 };
651 let err_indent = " ".repeat(error_prefix.len());
652 let mut outs = Vec::new();
653 outs.push(format!("{error_prefix} these tokens are not referenced in the grammar but defined as follows"));
654 outs.push(format!(
655 "{err_indent} {}",
656 lex_diag.file_location_msg("in the lexer", None)
657 ));
658 for (_, span) in mfp {
659 let error_contents = lex_diag.underline_span_with_text(
660 *span,
661 "Missing from parser".to_string(),
662 '^',
663 );
664 outs.extend(error_contents.lines().map(|s| s.to_string()));
665 }
666
667 for s in outs {
668 if !self.warnings_are_errors && std::env::var("OUT_DIR").is_ok() {
669 println!("cargo:warning={}", s)
670 } else {
671 eprintln!("{}", s);
672 }
673 }
674
675 has_unallowed_missing |= self.warnings_are_errors;
676 }
677 }
678 if has_unallowed_missing {
679 fs::remove_file(outp).ok();
680 panic!();
681 }
682
683 let mod_name = match self.mod_name {
684 Some(s) => s.to_owned(),
685 None => {
686 let mut stem = lexerp.to_str().unwrap();
691 loop {
692 let new_stem = Path::new(stem).file_stem().unwrap().to_str().unwrap();
693 if stem == new_stem {
694 break;
695 }
696 stem = new_stem;
697 }
698 format!("{}_l", stem)
699 }
700 };
701 let mod_name = format_ident!("{}", mod_name);
702 let mut lexerdef_func_impl = {
703 let LexFlags {
704 allow_wholeline_comments,
705 dot_matches_new_line,
706 multi_line,
707 octal,
708 posix_escapes,
709 case_insensitive,
710 unicode,
711 swap_greed,
712 ignore_whitespace,
713 size_limit,
714 dfa_size_limit,
715 nest_limit,
716 } = lex_flags;
717 let allow_wholeline_comments = QuoteOption(allow_wholeline_comments);
718 let dot_matches_new_line = QuoteOption(dot_matches_new_line);
719 let multi_line = QuoteOption(multi_line);
720 let octal = QuoteOption(octal);
721 let posix_escapes = QuoteOption(posix_escapes);
722 let case_insensitive = QuoteOption(case_insensitive);
723 let unicode = QuoteOption(unicode);
724 let swap_greed = QuoteOption(swap_greed);
725 let ignore_whitespace = QuoteOption(ignore_whitespace);
726 let size_limit = QuoteOption(size_limit);
727 let dfa_size_limit = QuoteOption(dfa_size_limit);
728 let nest_limit = QuoteOption(nest_limit);
729
730 quote! {
732 let mut lex_flags = ::lrlex::DEFAULT_LEX_FLAGS;
733 lex_flags.allow_wholeline_comments = #allow_wholeline_comments.or(::lrlex::DEFAULT_LEX_FLAGS.allow_wholeline_comments);
734 lex_flags.dot_matches_new_line = #dot_matches_new_line.or(::lrlex::DEFAULT_LEX_FLAGS.dot_matches_new_line);
735 lex_flags.multi_line = #multi_line.or(::lrlex::DEFAULT_LEX_FLAGS.multi_line);
736 lex_flags.octal = #octal.or(::lrlex::DEFAULT_LEX_FLAGS.octal);
737 lex_flags.posix_escapes = #posix_escapes.or(::lrlex::DEFAULT_LEX_FLAGS.posix_escapes);
738 lex_flags.case_insensitive = #case_insensitive.or(::lrlex::DEFAULT_LEX_FLAGS.case_insensitive);
739 lex_flags.unicode = #unicode.or(::lrlex::DEFAULT_LEX_FLAGS.unicode);
740 lex_flags.swap_greed = #swap_greed.or(::lrlex::DEFAULT_LEX_FLAGS.swap_greed);
741 lex_flags.ignore_whitespace = #ignore_whitespace.or(::lrlex::DEFAULT_LEX_FLAGS.ignore_whitespace);
742 lex_flags.size_limit = #size_limit.or(::lrlex::DEFAULT_LEX_FLAGS.size_limit);
743 lex_flags.dfa_size_limit = #dfa_size_limit.or(::lrlex::DEFAULT_LEX_FLAGS.dfa_size_limit);
744 lex_flags.nest_limit = #nest_limit.or(::lrlex::DEFAULT_LEX_FLAGS.nest_limit);
745 let lex_flags = lex_flags;
746 }
747 };
748 {
749 let start_states = lexerdef.iter_start_states();
750 let rules = lexerdef.iter_rules().map(|r| {
751 let tok_id = QuoteOption(r.tok_id);
752 let n = QuoteOption(r.name().map(QuoteToString));
753 let target_state =
754 QuoteOption(r.target_state().map(|(x, y)| QuoteTuple((x, y))));
755 let n_span = r.name_span();
756 let regex = QuoteToString(&r.re_str);
757 let start_states = r.start_states();
758 quote! {
763 Rule::new(::lrlex::unstable_api::InternalPublicApi, #tok_id, #n, #n_span, #regex.to_string(),
764 vec![#(#start_states),*], #target_state, &lex_flags).unwrap()
765 }
766 });
767 lexerdef_func_impl.append_all(quote! {
769 let start_states: Vec<StartState> = vec![#(#start_states),*];
770 let rules = vec![#(#rules),*];
771 });
772 }
773 let lexerdef_ty = match lexerkind {
774 LexerKind::LRNonStreamingLexer => {
775 quote!(::lrlex::LRNonStreamingLexerDef)
776 }
777 };
778 lexerdef_func_impl.append_all(quote! {
780 #lexerdef_ty::from_rules(start_states, rules)
781 });
782
783 let mut token_consts = TokenStream::new();
784 if let Some(rim) = self.rule_ids_map {
785 let mut rim_sorted = Vec::from_iter(rim.iter());
786 rim_sorted.sort_by_key(|(k, _)| *k);
787 for (name, id) in rim_sorted {
788 if RE_TOKEN_ID.is_match(&name) {
789 let tok_ident = format_ident!("N_{}", name.to_ascii_uppercase());
790 let storaget =
791 str::parse::<TokenStream>(type_name::<LexerTypesT::StorageT>()).unwrap();
792 let tok_const = quote! {
794 #[allow(dead_code)]
795 pub const #tok_ident: #storaget = #id;
796 };
797 token_consts.extend(tok_const)
798 }
799 }
800 }
801 let token_consts = token_consts.into_iter();
802 let out_tokens = {
803 let lexerdef_param = str::parse::<TokenStream>(type_name::<LexerTypesT>()).unwrap();
804 let mod_vis = self.visibility;
805 quote! {
807 #mod_vis mod #mod_name {
808 use ::lrlex::{LexerDef, Rule, StartState};
809 #[allow(dead_code)]
810 pub fn lexerdef() -> #lexerdef_ty<#lexerdef_param> {
811 #lexerdef_func_impl
812 }
813
814 #(#token_consts)*
815 }
816 }
817 };
818 let unformatted = out_tokens.to_string();
820 let mut outs = String::new();
821 let timestamp = env!("VERGEN_BUILD_TIMESTAMP");
825 write!(outs, "// lrlex build time: {}\n\n", quote!(#timestamp),).ok();
826 outs.push_str(
827 &syn::parse_str(&unformatted)
828 .map(|syntax_tree| prettyplease::unparse(&syntax_tree))
829 .unwrap_or(unformatted),
830 );
831 if let Ok(curs) = read_to_string(outp) {
835 if curs == outs {
836 return Ok(CTLexer {
837 missing_from_lexer,
838 missing_from_parser,
839 });
840 }
841 }
842 let mut f = File::create(outp)?;
843 f.write_all(outs.as_bytes())?;
844 Ok(CTLexer {
845 missing_from_lexer,
846 missing_from_parser,
847 })
848 }
849
850 #[deprecated(
857 since = "0.11.0",
858 note = "Please use lexer_in_src_dir() and build() instead"
859 )]
860 #[allow(deprecated)]
861 pub fn process_file_in_src(
862 self,
863 srcp: &str,
864 ) -> Result<(Option<HashSet<String>>, Option<HashSet<String>>), Box<dyn Error>> {
865 let mut inp = current_dir()?;
866 inp.push("src");
867 inp.push(srcp);
868 let mut outp = PathBuf::new();
869 outp.push(var("OUT_DIR").unwrap());
870 outp.push(Path::new(srcp).parent().unwrap().to_str().unwrap());
871 create_dir_all(&outp)?;
872 let mut leaf = Path::new(srcp)
873 .file_name()
874 .unwrap()
875 .to_str()
876 .unwrap()
877 .to_owned();
878 write!(leaf, ".{}", RUST_FILE_EXT).ok();
879 outp.push(leaf);
880 self.process_file(inp, outp)
881 }
882
883 #[deprecated(
901 since = "0.11.0",
902 note = "Please use lexer_in_src_dir() and build() instead"
903 )]
904 pub fn process_file<P, Q>(
905 mut self,
906 inp: P,
907 outp: Q,
908 ) -> Result<(Option<HashSet<String>>, Option<HashSet<String>>), Box<dyn Error>>
909 where
910 P: AsRef<Path>,
911 Q: AsRef<Path>,
912 {
913 self.lexer_path = Some(inp.as_ref().to_owned());
914 self.output_path = Some(outp.as_ref().to_owned());
915 let cl = self.build()?;
916 Ok((
917 cl.missing_from_lexer().map(|x| x.to_owned()),
918 cl.missing_from_parser()
919 .map(|x| x.iter().map(|(n, _)| n.to_owned()).collect::<HashSet<_>>()),
920 ))
921 }
922
923 pub fn allow_missing_terms_in_lexer(mut self, allow: bool) -> Self {
926 self.allow_missing_terms_in_lexer = allow;
927 self
928 }
929
930 pub fn allow_missing_tokens_in_parser(mut self, allow: bool) -> Self {
934 self.allow_missing_tokens_in_parser = allow;
935 self
936 }
937
938 pub fn warnings_are_errors(mut self, flag: bool) -> Self {
941 self.warnings_are_errors = flag;
942 self
943 }
944
945 pub fn show_warnings(mut self, flag: bool) -> Self {
948 self.show_warnings = flag;
949 self
950 }
951
952 pub fn allow_wholeline_comments(mut self, flag: bool) -> Self {
960 let key = "allow_wholeline_comments".to_string();
961 self.header.insert(
962 key,
963 HeaderValue(
964 Location::Other("CTLexerBuilder".to_string()),
965 Value::Flag(flag, Location::Other("CTLexerBuilder".to_string())),
966 ),
967 );
968 self
969 }
970
971 pub fn dot_matches_new_line(mut self, flag: bool) -> Self {
976 let key = "dot_matches_new_line".to_string();
977 self.header.insert(
978 key,
979 HeaderValue(
980 Location::Other("CTLexerBuilder".to_string()),
981 Value::Flag(flag, Location::Other("CTLexerBuilder".to_string())),
982 ),
983 );
984 self
985 }
986
987 pub fn multi_line(mut self, flag: bool) -> Self {
992 let key = "multi_line".to_string();
993 self.header.insert(
994 key,
995 HeaderValue(
996 Location::Other("CTLexerBuilder".to_string()),
997 Value::Flag(flag, Location::Other("CTLexerBuilder".to_string())),
998 ),
999 );
1000 self
1001 }
1002
1003 pub fn posix_escapes(mut self, flag: bool) -> Self {
1008 let key = "posix_escapes".to_string();
1009 self.header.insert(
1010 key,
1011 HeaderValue(
1012 Location::Other("CTLexerBuilder".to_string()),
1013 Value::Flag(flag, Location::Other("CTLexerBuilder".to_string())),
1014 ),
1015 );
1016 self
1017 }
1018
1019 pub fn octal(mut self, flag: bool) -> Self {
1024 let key = "octal".to_string();
1025 self.header.insert(
1026 key,
1027 HeaderValue(
1028 Location::Other("CTLexerBuilder".to_string()),
1029 Value::Flag(flag, Location::Other("CTLexerBuilder".to_string())),
1030 ),
1031 );
1032 self
1033 }
1034
1035 pub fn swap_greed(mut self, flag: bool) -> Self {
1040 let key = "swap_greed".to_string();
1041 self.header.insert(
1042 key,
1043 HeaderValue(
1044 Location::Other("CTLexerBuilder".to_string()),
1045 Value::Flag(flag, Location::Other("CTLexerBuilder".to_string())),
1046 ),
1047 );
1048 self
1049 }
1050
1051 pub fn ignore_whitespace(mut self, flag: bool) -> Self {
1056 let key = "ignore_whitespace".to_string();
1057 self.header.insert(
1058 key,
1059 HeaderValue(
1060 Location::Other("CTLexerBuilder".to_string()),
1061 Value::Flag(flag, Location::Other("CTLexerBuilder".to_string())),
1062 ),
1063 );
1064 self
1065 }
1066
1067 pub fn unicode(mut self, flag: bool) -> Self {
1072 let key = "unicode".to_string();
1073 self.header.insert(
1074 key,
1075 HeaderValue(
1076 Location::Other("CTLexerBuilder".to_string()),
1077 Value::Flag(flag, Location::Other("CTLexerBuilder".to_string())),
1078 ),
1079 );
1080 self
1081 }
1082
1083 pub fn case_insensitive(mut self, flag: bool) -> Self {
1088 let key = "case_insensitive".to_string();
1089 self.header.insert(
1090 key,
1091 HeaderValue(
1092 Location::Other("CTLexerBuilder".to_string()),
1093 Value::Flag(flag, Location::Other("CTLexerBuilder".to_string())),
1094 ),
1095 );
1096 self
1097 }
1098
1099 pub fn size_limit(mut self, sz: usize) -> Self {
1104 let key = "size_limit".to_string();
1105 self.header.insert(
1106 key,
1107 HeaderValue(
1108 Location::Other("CTLexerBuilder".to_string()),
1109 Value::Setting(Setting::Num(
1110 sz as u64,
1111 Location::Other("CTLexerBuilder".to_string()),
1112 )),
1113 ),
1114 );
1115 self
1116 }
1117
1118 pub fn dfa_size_limit(mut self, sz: usize) -> Self {
1123 let key = "dfa_size_limit".to_string();
1124 self.header.insert(
1125 key,
1126 HeaderValue(
1127 Location::Other("CTLexerBuilder".to_string()),
1128 Value::Setting(Setting::Num(
1129 sz as u64,
1130 Location::Other("CTLexerBuilder".to_string()),
1131 )),
1132 ),
1133 );
1134 self
1135 }
1136
1137 pub fn nest_limit(mut self, lim: u32) -> Self {
1142 let key = "nest_limit".to_string();
1143 self.header.insert(
1144 key,
1145 HeaderValue(
1146 Location::Other("CTLexerBuilder".to_string()),
1147 Value::Setting(Setting::Num(
1148 lim as u64,
1149 Location::Other("CTLexerBuilder".to_string()),
1150 )),
1151 ),
1152 );
1153 self
1154 }
1155
1156 #[cfg(test)]
1157 pub fn inspect_lexerkind(
1158 mut self,
1159 cb: Box<dyn Fn(LexerKind) -> Result<(), Box<dyn Error>>>,
1160 ) -> Self {
1161 self.inspect_lexerkind_cb = Some(cb);
1162 self
1163 }
1164}
1165
1166pub struct CTLexer {
1168 missing_from_lexer: Option<HashSet<String>>,
1169 missing_from_parser: Option<HashSet<(String, Span)>>,
1170}
1171
1172impl CTLexer {
1173 fn missing_from_lexer(&self) -> Option<&HashSet<String>> {
1174 self.missing_from_lexer.as_ref()
1175 }
1176
1177 fn missing_from_parser(&self) -> Option<&HashSet<(String, Span)>> {
1178 self.missing_from_parser.as_ref()
1179 }
1180}
1181
1182pub fn ct_token_map<StorageT: Display + ToTokens>(
1207 mod_name: &str,
1208 token_map: impl Borrow<HashMap<String, StorageT>>,
1209 rename_map: Option<&HashMap<&str, &str>>,
1210) -> Result<(), Box<dyn Error>> {
1211 let mut outs = String::new();
1215 let timestamp = env!("VERGEN_BUILD_TIMESTAMP");
1216 let mod_ident = format_ident!("{}", mod_name);
1217 write!(outs, "// lrlex build time: {}\n\n", quote!(#timestamp),).ok();
1218 let mut token_map_sorted = Vec::from_iter(token_map.borrow().iter());
1221 token_map_sorted.sort_by_key(|(k, _)| *k);
1222 let tokens = &token_map_sorted
1223 .into_iter()
1224 .map(|(k, id)| {
1225 let name = match rename_map {
1226 Some(rmap) => *rmap.get(k.as_str()).unwrap_or(&k.as_str()),
1227 _ => k,
1228 };
1229 let tok_ident = format_ident!("T_{}", name.to_ascii_uppercase());
1230 let storaget = str::parse::<TokenStream>(type_name::<StorageT>()).unwrap();
1231 quote! {
1233 pub const #tok_ident: #storaget = #id;
1234 }
1235 })
1236 .collect::<Vec<_>>();
1237 let unformatted = quote! {
1240 mod #mod_ident {
1241 #![allow(dead_code)]
1242 #(#tokens)*
1243 }
1244 }
1245 .to_string();
1246 let out_mod = syn::parse_str(&unformatted)
1247 .map(|syntax_tree| prettyplease::unparse(&syntax_tree))
1248 .unwrap_or(unformatted);
1249 outs.push_str(&out_mod);
1250 let mut outp = PathBuf::from(var("OUT_DIR")?);
1251 outp.push(mod_name);
1252 outp.set_extension("rs");
1253
1254 if let Ok(curs) = read_to_string(&outp) {
1258 if curs == outs {
1259 return Ok(());
1260 }
1261 }
1262
1263 let mut f = File::create(outp)?;
1264 f.write_all(outs.as_bytes())?;
1265 Ok(())
1266}
1267
1268fn indent(indent: &str, s: &str) -> String {
1279 format!("{indent}{}\n", s.trim_end_matches('\n')).replace('\n', &format!("\n{}", indent))
1280}
1281
1282#[cfg(test)]
1283mod test {
1284 use std::fs::File;
1285 use std::io::Write;
1286
1287 use super::{CTLexerBuilder, LexerKind};
1288 #[test]
1289 fn test_grmtools_section_lexerkind() {
1290 let lexerkinds = [
1291 "LRNonStreamingLexer",
1292 "lrnonstreaminglexer",
1293 "LexerKind::lrnonstreaminglexer",
1294 "lexerkind::LRNonStreamingLexer",
1295 ];
1296 for (i, kind) in lexerkinds.iter().enumerate() {
1297 let lex_src = format!(
1298 "
1299%grmtools{{lexerkind: {}}}
1300%%
1301. ;
1302",
1303 kind
1304 );
1305 let lex_path = format!(
1306 "{}/test_grmtools_section_lexerkind_{}.l",
1307 env!("OUT_DIR"),
1308 i
1309 );
1310 let mut l_file = File::create(lex_path.clone()).unwrap();
1311 l_file.write_all(lex_src.as_bytes()).unwrap();
1312 CTLexerBuilder::new()
1313 .output_path(format!("{}.rs", lex_path.clone()))
1314 .lexer_path(lex_path.clone())
1315 .inspect_lexerkind(Box::new(move |lexerkind| {
1316 assert!(matches!(lexerkind, LexerKind::LRNonStreamingLexer));
1317 Ok(())
1318 }))
1319 .build()
1320 .unwrap();
1321 }
1322 }
1323}