1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953
//! **This module is deprecated by the [crate::zc_parser] module** and is only kept for
//! compatibility with existing parsers.
//!
//! This module implements the parsing routines that uses the state machine
//! generated by rustlr. The main structure here is [RuntimeParser].
//! All parsing functions are organized around the [RuntimeParser::parse_base]
//! function, which implements the basic LR parsing algorithm. This function
//! expects dynamic [Lexer] and [ErrHandler] trait-objects.
//! This module provides generic
//! parsing and parser-training routines that use stdio for interface, but
//! the [ErrHandler] trait allows custom user interfaces to be build separately.
#![allow(dead_code)]
#![allow(unused_variables)]
#![allow(non_snake_case)]
#![allow(non_camel_case_types)]
#![allow(unused_parens)]
#![allow(unused_mut)]
#![allow(unused_assignments)]
#![allow(unused_doc_comments)]
#![allow(unused_imports)]
use std::fmt::Display;
use std::default::Default;
use std::collections::{HashMap,HashSet,BTreeSet};
use std::io::{self,Read,Write,BufReader,BufRead};
use std::cell::{RefCell,Ref,RefMut};
use std::hash::{Hash,Hasher};
use std::any::Any;
use std::fs::File;
use std::io::prelude::*;
use std::path::Path;
use std::mem;
use crate::{Stateaction,iserror};
#[cfg(feature = "generator")]
use crate::Statemachine;
use crate::{LBox,LRc};
use crate::Stateaction::*;
use crate::{lbup,lbdown,lbget};
#[cfg(feature = "legacy-parser")]
use crate::{Lexer,Lextoken};
/// this structure is only exported because it is required by the generated parsers.
/// There is no reason to use it in other programs.
#[cfg(feature = "legacy-parser")]
#[derive(Clone)]
pub struct RProduction<AT:Default,ET:Default> // runtime rep of grammar rule
{
pub lhs: &'static str, // left-hand side nonterminal of rule
pub Ruleaction : fn(&mut RuntimeParser<AT,ET>) -> AT, //parser as arg
}
#[cfg(feature = "legacy-parser")]
impl<AT:Default,ET:Default> RProduction<AT,ET>
{
pub fn new_skeleton(lh:&'static str) -> RProduction<AT,ET>
{
RProduction {
lhs : lh,
Ruleaction : |p|{<AT>::default()},
}
}
}//impl RProduction
pub struct Stackelement<AT:Default>
{
#[cfg(feature = "legacy-parser")]
pub si : usize, // state index
pub value : AT, // semantic value (don't clone grammar symbols)
//pub line: usize, // line and column
//pub column: usize,
}
/// this is the structure created by the generated parser. The generated parser
/// program will contain a make_parser function that returns this structure.
/// Most of the pub items are, however, only exported to support the operation
/// of the parser, and should not be accessed directly. Only the functions
/// [RuntimeParser::parse], [RuntimeParser::report], [RuntimeParser::abort]
/// and [RuntimeParser::error_occurred] should be called directly
/// from user programs. Only the field [RuntimeParser::exstate] should be accessed
/// by user programs.
#[cfg(feature = "legacy-parser")]
pub struct RuntimeParser<AT:Default,ET:Default>
{
/// this is the "external state" structure, with type ET defined by the grammar.
/// The semantic actions associated with each grammar rule, which are written
/// in the grammar, have ref mut access to the RuntimeParser structure, which
/// allows them to read and change the external state object. This gives
/// the parsers greater flexibility and capability, including the ability to
/// parse some non-context free languages. See
/// [this sample grammar](<https://cs.hofstra.edu/~cscccl/rustlr_project/ncf.grammar>).
/// The exstate is initialized to ET::default().
pub exstate : ET, // external state structure, usage optional
/// used only by generated parser: do not reference
pub RSM : Vec<HashMap<&'static str,Stateaction>>, // runtime state machine
// do not reference
//pub Expected : Vec<Vec<&'static str>>,
/// do not reference
pub Rules : Vec<RProduction<AT,ET>>, //rules with just lhs and delegate function
////// this value should be set through abort or report
stopparsing : bool,
/// do not reference
pub stack : Vec<Stackelement<AT>>, // parse stack
// pub recover : HashSet<&'static str>, // for error recovery
pub resynch : HashSet<&'static str>,
pub Errsym : &'static str,
err_occurred : bool,
pub linenum : usize,
pub column : usize,
pub src_id : usize,
report_line : usize,
// training : bool,
// pub trained: HashMap<(usize,String),String>,
/// Hashset containing all grammar symbols (terminal and non-terminal). This is used for error reporting and training.
pub Symset : HashSet<&'static str>,
}//struct RuntimeParser
#[cfg(feature = "legacy-parser")]
impl<AT:Default,ET:Default> RuntimeParser<AT,ET>
{
/// this is only called by the make_parser function in the machine-generated
/// parser program. *Do not call this function in other places* as it
/// only generates a skeleton.
pub fn new(rlen:usize, slen:usize) -> RuntimeParser<AT,ET>
{ // given number of rules and number states
let mut p = RuntimeParser {
RSM : Vec::with_capacity(slen),
//Expected : Vec::with_capacity(slen),
Rules : Vec::with_capacity(rlen),
stopparsing : false,
exstate : ET::default(),
stack : Vec::with_capacity(1024),
Errsym : "",
err_occurred : false,
linenum : 0,
column : 0,
src_id : 0,
report_line : 0,
resynch : HashSet::new(),
//added for training
//training : false,
//trained : HashMap::new(),
Symset : HashSet::with_capacity(64),
};
for _ in 0..slen {
p.RSM.push(HashMap::with_capacity(16));
//p.Expected.push(Vec::new());
}
return p;
}//new
/// this function can be called from with the "semantic" actions attached
/// to grammar production rules that are executed for each
/// "reduce" action of the parser.
pub fn abort(&mut self, msg:&str)
{
eprintln!("\n!!!Parsing Aborted: {}",msg);
self.err_occurred = true;
self.stopparsing=true;
}
/// may be called from grammar semantic actions to report error.
/// this report function will print to stdout.
pub fn report(&mut self, errmsg:&str)
{ // linenum must be set prior to call
if (self.report_line != self.linenum || self.linenum==0) {
// eprint!("ERROR on line {}, column {}:\n{}\n",self.linenum,self.column,tokenizer.current_line());
eprintln!("ERROR on line {}, column {}: {}",self.linenum,self.column,errmsg);
self.report_line = self.linenum;
}
else {
eprint!(" {} ",errmsg);
}
self.err_occurred = true;
}// report
/// this function is only exported to support the generated code
pub fn bad_pattern(&mut self,pattern:&str) -> AT
{
let msg = format!("pattern {} failed to bind to stacked values\n",pattern);
self.report(&msg);
AT::default()
}
/// sets an index that index source information, such as the source file
/// when compiling multiple sources. This information must be maintained externally.
/// The source id will also be passed on to the [LBox] and [LRc] smartpointers by
/// the [RuntimeParser::lb] function.
pub fn set_src_id(&mut self, id:usize)
{ self.src_id =id; }
//called to simulate a shift
fn errshift(&mut self, sym:&str) -> bool
{
let csi = self.stack[self.stack.len()-1].si; // current state
let actionopt = self.RSM[csi].get(sym);
if let Some(Shift(ni)) = actionopt {
self.stack.push(Stackelement{si:*ni,value:AT::default()}); true
}
else {false}
}
// this is the LR parser shift action: push the next state, along with the
// value of the current lookahead token onto the parse stack, returns the
// next token
fn shift(&mut self, nextstate:usize, lookahead:Lextoken<AT>, tokenizer:&mut dyn Lexer<AT>) -> Lextoken<AT>
{
self.stack.push(Stackelement{si:nextstate,value:lookahead.value});
self.nexttoken(tokenizer)
}
fn reduce(&mut self, ri:&usize)
{
let rulei = &self.Rules[*ri];
let ruleilhs = rulei.lhs; // &'static : Copy
let val = (rulei.Ruleaction)(self); // calls delegate function
let newtop = self.stack[self.stack.len()-1].si;
let goton = self.RSM[newtop].get(ruleilhs).unwrap();
if let Stateaction::Gotonext(nsi) = goton {
self.stack.push(Stackelement{si:*nsi,value:val});
// DO NOT CHANGE LOOKAHEAD AFTER REDUCE!
}// goto next state after reduce
else {
self.report("state transition table corrupted: no suitable action after reduce");
self.stopparsing=true;
}
}//reduce
/// can be called to determine if an error occurred during parsing. The parser
/// will not panic.
pub fn error_occurred(&self) -> bool {self.err_occurred}
fn nexttoken(&self, tokenizer:&mut dyn Lexer<AT>) -> Lextoken<AT>
{
if let Some(tok) = tokenizer.nextsym() {tok}
else { Lextoken{sym:"EOF".to_owned(), value:AT::default()} }
}
/// *this function is equivalent to [RuntimeParser::parse_stdio_train]*.
pub fn parse_train(&mut self, tokenizer:&mut dyn Lexer<AT>, parserfile:&str) -> AT
{
self.parse_stdio_train(tokenizer,parserfile)
}//parse_train
/// creates a [LBox] smart pointer that includes line/column/src information;
/// should be called from the semantic actions of a grammar rule, e.g.
///```ignore
/// E --> E:a + E:b {PlusExpr(parser.lb(a),parser.lb(b))}
///```
pub fn lb<T>(&self,e:T) -> LBox<T> { LBox::new(e,self.linenum,self.column /*,self.src_id*/) }
/// creates a `LBox<dyn Any>`, which allows attributes of different types to
/// be associated with grammar symbols. Use in conjuction with [LBox::downcast], [LBox::upcast] and the [lbdown], [lbup] macros.
pub fn lba<T:'static>(&self,e:T) -> LBox<dyn Any> { LBox::upcast(LBox::new(e,self.linenum,self.column /*,self.src_id*/)) }
/// similar to [RuntimeParser::lb], but creates a [LRc] instead of [LBox]
pub fn lrc<T>(&self,e:T) -> LRc<T> { LRc::new(e,self.linenum,self.column /*,self.src_id*/) }
/// similar to [RuntimeParser::lba] but creates a [LRc]
pub fn lrca<T:'static>(&self,e:T) -> LRc<dyn Any> { LRc::upcast(LRc::new(e,self.linenum,self.column /*,self.src_id*/)) }
}// impl RuntimeParser
#[cfg(not(feature = "legacy-parser"))]
#[cfg(feature = "generator")]
impl Statemachine {
pub fn write_verbose(&self, filename:&str)->Result<(),std::io::Error> {
panic!("verbose feature not supported without `legacy-parser` option");
}
pub fn writeparser(&self, filename:&str)->Result<(),std::io::Error> {
panic!("feature not supported without `legacy-parser` option");
}
} // for compatibility
//////////////////////////////////////////////////////////////////
//////////////////////////////////////////////////////////////////
//// new version of write_fsm:
#[cfg(all(feature = "generator", feature = "legacy-parser"))]
impl Statemachine
{
pub fn writeparser(&self, filename:&str)->Result<(),std::io::Error>
{
let mut fd = File::create(filename)?;
write!(fd,"//Parser generated by rustlr\n
#![allow(unused_variables)]
#![allow(non_snake_case)]
#![allow(non_camel_case_types)]
#![allow(unused_parens)]
#![allow(unused_mut)]
#![allow(unused_imports)]
#![allow(unused_assignments)]
#![allow(irrefutable_let_patterns)]
extern crate rustlr;
use rustlr::{{RuntimeParser,RProduction,Stateaction,decode_action}};\n")?;
write!(fd,"{}\n",&self.Gmr.Extras)?; // use clauses
// write static array of symbols
write!(fd,"const SYMBOLS:[&'static str;{}] = [",self.Gmr.Symbols.len())?;
for i in 0..self.Gmr.Symbols.len()-1
{
write!(fd,"\"{}\",",&self.Gmr.Symbols[i].sym)?;
}
write!(fd,"\"{}\"];\n\n",&self.Gmr.Symbols[self.Gmr.Symbols.len()-1].sym)?;
// position of symbols must be inline with self.Gmr.Symhash
// record table entries in a static array
let mut totalsize = 0;
for i in 0..self.FSM.len() { totalsize+=self.FSM[i].len(); }
write!(fd,"const TABLE:[u64;{}] = [",totalsize)?;
// generate table to represent FSM
let mut encode:u64 = 0;
for i in 0..self.FSM.len() // for each state index i
{
let row = &self.FSM[i];
for key in row.keys()
{ // see function decode for opposite translation
let k = *key; //*self.Gmr.Symhash.get(key).unwrap(); // index of symbol
encode = ((i as u64) << 48) + ((k as u64) << 32);
match row.get(key) {
Some(Shift(statei)) => { encode += (*statei as u64) << 16; },
Some(Gotonext(statei)) => { encode += ((*statei as u64) << 16)+1; },
Some(Reduce(rulei)) => { encode += ((*rulei as u64) << 16)+2; },
Some(Accept) => {encode += 3; },
_ => {encode += 4; }, // 4 indicates Error
}//match
write!(fd,"{},",encode)?;
} //for symbol index k
}//for each state index i
write!(fd,"];\n\n")?;
// must know what absyn type is when generating code.
let ref absyn = self.Gmr.Absyntype;
let ref extype = self.Gmr.Externtype;
write!(fd,"pub fn make_parser() -> RuntimeParser<{},{}>",absyn,extype)?;
write!(fd,"\n{{\n")?;
// write code to pop stack, assign labels to variables.
write!(fd," let mut parser1:RuntimeParser<{},{}> = RuntimeParser::new({},{});\n",absyn,extype,self.Gmr.Rules.len(),self.FSM.len())?;
// generate rules and Ruleaction delegates, must pop values from runtime stack
write!(fd," let mut rule = RProduction::<{},{}>::new_skeleton(\"{}\");\n",absyn,extype,"start")?;
for i in 0..self.Gmr.Rules.len()
{
write!(fd," rule = RProduction::<{},{}>::new_skeleton(\"{}\");\n",absyn,extype,self.Gmr.Rules[i].lhs.sym)?;
write!(fd," rule.Ruleaction = |parser|{{ ")?;
let mut k = self.Gmr.Rules[i].rhs.len();
//form if-let labels and patterns as we go...
let mut labels = String::from("(");
let mut patterns = String::from("(");
while k>0
{
let gsym = &self.Gmr.Rules[i].rhs[k-1];
let gsymtype = gsym.gettype(&self.Gmr);
if gsym.label.len()>1 && gsym.label.find('@').is_some() { // if-let pattern
let atindex = gsym.label.find('@').unwrap();
let varlab = if atindex>0 {gsym.label[0..atindex].to_string()}
else {format!("_vflab_{}",k-1)};
if atindex>0 {labels.push_str("&mut ");}
labels.push_str(&varlab); labels.push(',');
patterns.push_str(&gsym.label[atindex+1..]); patterns.push_str(",");
write!(fd," let mut {}=",&varlab)?;
}
else if gsym.label.len()>0 { // simple pattern, no need for if-let
if gsymtype.len()>=3 && &gsymtype[0..3]=="mut" {
write!(fd," let mut {}:{}=",&gsym.label,absyn)?;
}
else {write!(fd," let {}:{}=",&gsym.label,absyn)?;}
}
write!(fd,"parser.stack.pop()")?;
if gsym.label.len()>0 { write!(fd,".unwrap().value; ")?;}
else {write!(fd,"; ")?;}
k -= 1;
}// for each symbol on right hand side of rule
// form if-let clause
let defaultaction = format!("return <{}>::default();}}",absyn);
let mut semaction = &self.Gmr.Rules[i].action; //string that ends with }
if semaction.len()<=1 {semaction = &defaultaction;}
if labels.len()<2 { write!(fd,"{};\n",semaction.trim_end())?; } //empty pattern
else { // write an if-let
labels.push(')'); patterns.push(')');
write!(fd,"\n if let {}={} {{ {} else {{parser.bad_pattern(\"{}\")}} }};\n",&patterns,&labels,semaction.trim_end(),&patterns)?;
}// if-let semantic action
/*
while k>0 // k-1 indexes backwards rhs of grammar rule
{
let gsym = &self.Gmr.Rules[i].rhs[k-1];
if gsym.label.len()>0 && &gsymtype[0..3]=="mut"
{ write!(fd," let mut {}:{}=",gsym.label,absyn)?; }
else if gsym.label.len()>0
{ write!(fd," let {}:{}=",gsym.label,absyn)?; }
write!(fd,"parser.stack.pop()")?;
if gsym.label.len()>0 { write!(fd,".unwrap().value; ")?;}
else {write!(fd,"; ")?;}
k -= 1;
} // for each symbol on right hand side of rule
let mut semaction = &self.Gmr.Rules[i].action; //this is a string
if semaction.len()>1 {write!(fd,"{};\n",semaction.trim_end())?;}
else {write!(fd," return <{}>::default();}};\n",absyn)?;}
*/
write!(fd," parser1.Rules.push(rule);\n")?;
}// for each rule
write!(fd," parser1.Errsym = \"{}\";\n",&self.Gmr.Errsym)?;
// resynch vector
for s in &self.Gmr.Resynch {write!(fd," parser1.resynch.insert(\"{}\");\n",s)?;}
// generate code to load RSM from TABLE
write!(fd,"\n for i in 0..{} {{\n",totalsize)?;
write!(fd," let symi = ((TABLE[i] & 0x0000ffff00000000) >> 32) as usize;\n")?;
write!(fd," let sti = ((TABLE[i] & 0xffff000000000000) >> 48) as usize;\n")?;
write!(fd," parser1.RSM[sti].insert(SYMBOLS[symi],decode_action(TABLE[i]));\n }}\n\n")?;
// write!(fd,"\n for i in 0..{} {{for k in 0..{} {{\n",rows,cols)?;
// write!(fd," parser1.RSM[i].insert(SYMBOLS[k],decode_action(TABLE[i*{}+k]));\n }}}}\n",cols)?;
write!(fd," for s in SYMBOLS {{ parser1.Symset.insert(s); }}\n\n")?;
write!(fd," load_extras(&mut parser1);\n")?;
write!(fd," return parser1;\n")?;
write!(fd,"}} //make_parser\n\n")?;
////// Augment!
write!(fd,"fn load_extras(parser:&mut RuntimeParser<{},{}>)\n{{\n",absyn,extype)?;
write!(fd,"}}//end of load_extras: don't change this line as it affects augmentation\n")?;
Ok(())
}//writeparser
//////////////
///////////////// non-binary version (no augmentation) //////////////////
pub fn write_verbose(&self, filename:&str)->Result<(),std::io::Error>
{
let mut fd = File::create(filename)?;
write!(fd,"//Parser generated by rustlr\n
#![allow(unused_variables)]
#![allow(non_snake_case)]
#![allow(non_camel_case_types)]
#![allow(unused_parens)]
#![allow(unused_mut)]
#![allow(unused_assignments)]
extern crate rustlr;
use rustlr::{{RuntimeParser,RProduction,Stateaction}};\n")?;
write!(fd,"{}\n",&self.Gmr.Extras)?; // use clauses
let ref absyn = self.Gmr.Absyntype;
let ref extype = self.Gmr.Externtype;
write!(fd,"pub fn make_parser() -> RuntimeParser<{},{}>",absyn,extype)?;
write!(fd,"\n{{\n")?;
// write code to pop stack, assign labels to variables.
write!(fd," let mut parser1:RuntimeParser<{},{}> = RuntimeParser::new({},{});\n",absyn,extype,self.Gmr.Rules.len(),self.FSM.len())?;
// generate rules and Ruleaction delegates, must pop values from runtime stack
write!(fd," let mut rule = RProduction::<{},{}>::new_skeleton(\"{}\");\n",absyn,extype,"start")?;
for i in 0..self.Gmr.Rules.len()
{
write!(fd," rule = RProduction::<{},{}>::new_skeleton(\"{}\");\n",absyn,extype,self.Gmr.Rules[i].lhs.sym)?;
write!(fd," rule.Ruleaction = |parser|{{ ")?;
let mut k = self.Gmr.Rules[i].rhs.len();
while k>0
{
let gsym = &self.Gmr.Rules[i].rhs[k-1];
let gsymtype = gsym.gettype(&self.Gmr);
if gsym.label.len()>0 && &gsymtype[0..3]=="mut"
{ write!(fd," let mut {}:{}=",gsym.label,absyn)?; }
else if gsym.label.len()>0
{ write!(fd," let {}:{}=",gsym.label,absyn)?; }
write!(fd,"parser.stack.pop()")?;
if gsym.label.len()>0 { write!(fd,".unwrap().value; ")?;}
else {write!(fd,"; ")?;}
k -= 1;
} // for each symbol on right hand side of rule
let mut semaction = &self.Gmr.Rules[i].action; //this is a string
//if semaction.len()<1 {semaction = "}}";}
//if al>1 {semaction = semaction.substring(0,al-1);}
if semaction.len()>1 {write!(fd,"{};\n",semaction.trim_end())?;}
else {write!(fd," return <{}>::default();}};\n",absyn)?;}
write!(fd," parser1.Rules.push(rule);\n")?;
}// for each rule
write!(fd," parser1.Errsym = \"{}\";\n",&self.Gmr.Errsym)?;
// resynch vector
for s in &self.Gmr.Resynch {write!(fd," parser1.resynch.insert(\"{}\");\n",s)?;}
for i in 0..self.FSM.len()
{
let row = &self.FSM[i];
for key in row.keys()
{
write!(fd," parser1.RSM[{}].insert(\"{}\",Stateaction::{:?});\n",i,self.Gmr.symref(*key),row.get(key).unwrap())?;
} //for each string key in row
}//for each state index i
write!(fd," return parser1;\n")?;
write!(fd,"}} //make_parser\n")?;
Ok(())
}//write_verbose
} // impl Statemachine
/*
//// independent function
fn iserror(actionopt:&Option<&Stateaction>) -> bool
{
match actionopt {
None => true,
Some(Error(_)) => true,
_ => false,
}
}//iserror
*/
///////////////////////////////////////////////////////////////////////////
////// reimplementing the parsing algorithm more modularly, with aim of
////// allowing custom parsers
//////////// errors should compile a report
#[cfg(feature = "legacy-parser")]
impl<AT:Default,ET:Default> RuntimeParser<AT,ET>
{
/// this function is used to invoke the generated parser returned by
/// the generated parser program's make_parser function. *This function
/// is equivalent to [RuntimeParser::parse_stdio]*.
pub fn parse(&mut self, tokenizer:&mut dyn Lexer<AT>) -> AT
{
self.parse_stdio(tokenizer)
}
/// Error recovery routine of rustlr, separate from error_reporter.
/// This function will modify the parser and lookahead symbol and return
/// either the next action the parser should take (if recovery succeeded)
/// or None if recovery failed.
pub fn error_recover<'t>(&mut self, lookahead:&mut Lextoken<AT>,tokenizer:&mut dyn Lexer<AT>) -> Option<Stateaction>
{
let mut erraction = None;
///// prefer to ue Errsym method
if self.Errsym.len()>0 {
let errsym = self.Errsym;
// lookdown stack for state with trainsiton on Errsym
// but that could be current state too (start at top)
let mut k = self.stack.len(); // offset by 1 because of usize
let mut spos = k+1;
while k>0 && spos>k
{
let ksi = self.stack[k-1].si;
erraction = self.RSM[ksi].get(errsym);
if let None = erraction {k-=1;} else {spos=k;}
}//while k>0
if spos==k { self.stack.truncate(k); } // new current state revealed
// run all reduce actions that are valid before the Errsym:
while let Some(Reduce(ri)) = erraction // keep reducing
{
//self.reduce(ri); // borrow error- only need mut self.stack
let rulei = &self.Rules[*ri];
let ruleilhs = rulei.lhs; // &'static : Copy
let val = (rulei.Ruleaction)(self); // calls delegate function
let newtop = self.stack[self.stack.len()-1].si;
let gotonopt = self.RSM[newtop].get(ruleilhs);
match gotonopt {
Some(Gotonext(nsi)) => {
self.stack.push(Stackelement{si:*nsi,value:val});
},// goto next state after reduce
_ => {self.abort("recovery failed"); },
}//match
// end reduce
let tos=self.stack[self.stack.len()-1].si;
erraction = self.RSM[tos].get(self.Errsym);
} // while let erraction is reduce
// remaining defined action on Errsym must be shift
if let Some(Shift(i)) = erraction { // simulate shift errsym
self.stack.push(Stackelement{si:*i,value:AT::default()});
// keep lookahead until action is found that transitions from
// current state (i). but skipping ahead without reducing
// the error production is not a good idea
while let None = self.RSM[*i].get(&lookahead.sym[..]) {
if &lookahead.sym[..]=="EOF" {break;}
*lookahead = self.nexttoken(tokenizer);
}//while let
// either at end of input or found action on next symbol
erraction = self.RSM[*i].get(&lookahead.sym[..]);
} // if shift action found down under stack
}//errsym exists
// at this point, if erraction is None, then Errsym failed to recover,
// try the resynch symbol method next ...
if iserror(&erraction) && self.resynch.len()>0 {
while &lookahead.sym!="EOF" &&
!self.resynch.contains(&lookahead.sym[..]) {
self.linenum = tokenizer.linenum(); self.column = tokenizer.column();
*lookahead = self.nexttoken(tokenizer);
}//while
if &lookahead.sym!="EOF" {
// look for state on stack that has action defined on next symbol
self.linenum = tokenizer.linenum(); self.column = tokenizer.column();
*lookahead = self.nexttoken(tokenizer); // skipp err-causing symbol
}
let mut k = self.stack.len()-1; // offset by 1 because of usize
let mut position = 0;
while k>0 && erraction==None
{
let ksi = self.stack[k-1].si;
erraction = self.RSM[ksi].get(&lookahead.sym[..]);
if let None=erraction {k-=1;}
}//while k>0 && erraction==None
match erraction {
None => {}, // do nothing, whill shift next symbol
_ => { self.stack.truncate(k);},//pop stack
}//match
}// there are resync symbols
// at this point, if erraction is None, then resynch recovery failed too.
// only action left is to skip ahead...
let mut eofcx = 0;
while iserror(&erraction) && eofcx<1 { //skip input
self.linenum = tokenizer.linenum(); self.column = tokenizer.column();
*lookahead = self.nexttoken(tokenizer);
if &lookahead.sym=="EOF" {eofcx+=1;}
let csi =self.stack[self.stack.len()-1].si;
erraction = self.RSM[csi].get(&lookahead.sym[..]);
}// skip ahead
match erraction {
None => None,
Some(act) => Some(*act),
}//return match
}//error_recover function
}//impl RuntimeParser 2
/////////////////////////////////////////////////////////////////////////
/////////////// new approach using more flexible trait object
/// A trait object that implements ErrHandler is expected by the [RuntimeParser::parse_base]
/// function, which implements the basic LR parsing algorithm using the
/// generated state machine. The struct [StandardReporter] is provided as
/// the default ErrHandler that uses standard I/O as interface and has the
/// ability to train the parser. But other implementations of the trait
/// can be created that use different interfaces, such as a graphical IDE.
#[cfg(feature = "legacy-parser")]
pub trait ErrHandler<AT:Default,ET:Default> // not same as error recovery
{
fn err_reporter(&mut self, parser:&mut RuntimeParser<AT,ET>, lookahead:&Lextoken<AT>, erropt:&Option<Stateaction>);
fn report_err(&self, parser:&mut RuntimeParser<AT,ET>, msg:&str) { parser.report(msg) }
// fn training_mode(&self, parser:&RuntimeParser<AT,ET>) -> bool {false}
// fn interactive_mode(&self, parser:&RuntimeParser<AT,ET>) -> bool {false}
}// ErrHandler trait
/*
The structure here is a bit strange. The script file is written to in
interactive training mode and read from in script-training mode. However,
the actual modification of the parser file is done after the training, by
the augmenter module. Thus there's another wrapper function that's needed
besides the creation of the right kind of StandardReporter.
*/
/// these objects implement the [ErrHandler] trait. They are used in the
/// [RuntimeParser::parse_stdio], [RuntimeParser::parse_stdio_train] and
/// [RuntimeParser::train_from_script] functions.
pub struct StandardReporter
{
pub training : bool,
// pub interactive : bool, scriptinopt==None
pub trained : HashMap<(usize,String),String>,
pub scriptinopt: Option<BufReader<File>>, // for training from script
pub scriptoutopt: Option<File>, // created during interactive training
}
impl StandardReporter
{
/// creates default standard reporter, used by parse_stdio (does not train)
pub fn new() -> StandardReporter
{
StandardReporter {
training:false, trained:HashMap::new(), scriptinopt:None, scriptoutopt:None,}
}
/// creates a stdio error handler with interactive training, takes as
/// argument parser file name, to create script for future retraining.
pub fn new_interactive_training(existingparser:&str) -> StandardReporter
{
let outfile = format!("{}_script.txt", existingparser);
let mut fout = File::create(outfile).expect("failed to create training script file");
let _ = write!(fout,"# Rustlr training script for {}\n\n",existingparser);
StandardReporter {
training:true, trained:HashMap::with_capacity(8),
scriptoutopt:Some(fout), scriptinopt:None,}
}
/// creates a stdio error handler that trains (non-interactively) from
/// a previously created script. It's the user's responsibility to match
/// the script file with the input source.
pub fn new_script_training(existingparser:&str,scriptfile:&str) -> StandardReporter
{
let fin = BufReader::new(File::open(scriptfile).expect("failed to open training script file"));
StandardReporter {
training:true, trained:HashMap::with_capacity(32),
scriptoutopt:None,
scriptinopt:Some(fin), }
}
// augment_train implemented in augmenter.rs
}//impl StandardReporter
#[cfg(feature = "legacy-parser")]
impl<AT:Default,ET:Default> ErrHandler<AT,ET> for StandardReporter
{
// this function will be able to write training script to file
fn err_reporter(&mut self, parser:&mut RuntimeParser<AT,ET>, lookahead:&Lextoken<AT>, erropt:&Option<Stateaction>)
{
let mut wresult:std::io::Result<()> = Err(std::io::Error::new(std::io::ErrorKind::Other,"")); // dummy
// known that actionop is None or Some(Error(_))
let cstate = parser.stack[parser.stack.len()-1].si; // current state
let mut actionopt = if let Some(act)=erropt {Some(act)} else {None};
let lksym = &lookahead.sym[..];
// is lookahead recognized as a grammar symbol?
// if actionopt is NONE, check entry for ANY_ERROR
if parser.Symset.contains(lksym) {
if let None=actionopt {
actionopt = parser.RSM[cstate].get("ANY_ERROR");
}
}// lookahead is recognized grammar sym
else {
actionopt = parser.RSM[cstate].get("ANY_ERROR");
}// lookahead is not a grammar sym
let errmsg = if let Some(Error(em)) = &actionopt {
format!("unexpected symbol {}, ** {} ** ..",lksym,em.trim())
} else {format!("unexpected symbol {} .. ",lksym)};
parser.report(&errmsg);
if self.training { ////// Training mode
let csym = lookahead.sym.clone();
let mut inp = String::from("");
if let None=self.scriptinopt { // interactive mode
if let Some(outfd1) = &self.scriptoutopt {
let mut outfd = outfd1;
print!("\n>>>TRAINER: if this message is not adequate (for state {}), enter a replacement (default no change): ",cstate);
let rrrflush = io::stdout().flush();
if let Ok(n) = io::stdin().read_line(&mut inp) {
if inp.len()>5 && parser.Symset.contains(lksym) {
print!(">>>TRAINER: should this message be given for all unexpected symbols in the current state? (default yes) ");
let rrrflush2 = io::stdout().flush();
let mut inp2 = String::new();
if let Ok(n) = io::stdin().read_line(&mut inp2) {
if inp2.trim()=="no" || inp2.trim()=="No" {
wresult = write!(outfd,"{}\t{}\t{} ::: {}\n",parser.linenum,parser.column,&csym,inp.trim());
self.trained.insert((cstate,csym),inp);
}
else {// insert for any error
wresult = write!(outfd,"{}\t{}\t{} ::: {}\n",parser.linenum,parser.column,"ANY_ERROR",inp.trim());
self.trained.insert((cstate,String::from("ANY_ERROR")),inp);
}
}// read ok
}// unexpected symbol is grammar sym
else if inp.len()>5 && !parser.Symset.contains(lksym) {
wresult = write!(outfd,"{}\t{}\t{} ::: {}\n",parser.linenum,parser.column,"ANY_ERROR",inp.trim());
self.trained.insert((cstate,String::from("ANY_ERROR")),inp);
}
}// process user response
}}// interactive mode
else { // training from script mode (non-interactive)
if let Some(brfd) = &mut self.scriptinopt {
let mut scin = brfd;
let mut readn = 0;
while readn < 1
{
inp = String::new();
match scin.read_line(&mut inp) {
Ok(n) if n>1 && &inp[0..1]!="#" && inp.trim().len()>0 => {readn=n;},
Ok(n) if n>0 => { readn=0; }, // keep reading
_ => {readn = 1; } // stop - this means End of Stream
}//match
if readn>1 { // read something
let inpsplit:Vec<&str> = inp.split_whitespace().collect();
if inpsplit.len()>4 && inpsplit[3].trim()==":::" {
let inline = inpsplit[0].trim().parse::<usize>().unwrap();
let incolumn = inpsplit[1].trim().parse::<usize>().unwrap();
let insym = inpsplit[2].trim();
if parser.linenum==inline && parser.column==incolumn {
if &csym==insym || insym=="ANY_ERROR" {
let posc = inp.find(":::").unwrap()+4;
println!("\n>>>Found matching entry from training script for {}, error message: {}",insym,&inp[posc..]);
self.trained.insert((cstate,String::from(insym)),String::from(&inp[posc..]));
} // unexpected symbol match
}// line/column match
}//inpsplit check
}// valid training line read
}//while readn<2
}}//training from script mode
}//if training //// END TRAINING MODE
}// standardreporter function
}// impl ErrHandler for StandardReporter
//////////////// temporary: live side by side with parse_core (tobe replaced)
#[cfg(feature = "legacy-parser")]
impl<AT:Default,ET:Default> RuntimeParser<AT,ET>
{
/// Core parser (temporarily lives side by side with parse_core) that
/// takes dynamic trait objects for lexical scanning and err_reporting.
/// This design makes it possible to create a custom error reporting
/// interface.
pub fn parse_base(&mut self, tokenizer:&mut dyn Lexer<AT>, err_handler:&mut dyn ErrHandler<AT,ET>) -> AT
{
self.stack.clear();
self.err_occurred = false;
let mut result = AT::default();
self.stack.push(Stackelement {si:0, value:AT::default()});
self.stopparsing = false;
let mut action = Stateaction::Error("");
let mut lookahead = Lextoken{sym:"EOF".to_owned(),value:AT::default()};
if let Some(tok) = tokenizer.nextsym() {lookahead=tok;}
else {self.stopparsing=true;}
while !self.stopparsing
{
self.linenum = tokenizer.linenum(); self.column=tokenizer.column();
let currentstate = self.stack[self.stack.len()-1].si;
let mut actionopt = self.RSM[currentstate].get(lookahead.sym.as_str());
let actclone:Option<Stateaction> = match actionopt {
Some(a) => Some(*a),
None => None,
};
if iserror(&actionopt) { // either None or Error
if !self.err_occurred {self.err_occurred = true;}
err_handler.err_reporter(self,&lookahead,&actclone);
//err_reporter(self,&lookahead,&actclone);
match self.error_recover(&mut lookahead,tokenizer) {
None => { self.stopparsing=true; break; }
Some(act) => {action = act;},
}//match
}// iserror
else { action = actclone.unwrap(); }
match &action {
Shift(nextstate) => {
lookahead = self.shift(*nextstate,lookahead,tokenizer);
},
Reduce(rulei) => { self.reduce(rulei); },
Accept => {
self.stopparsing=true;
if self.stack.len()>0 {result = self.stack.pop().unwrap().value;}
else {self.err_occurred=true;}
},
_ => {}, // continue
}//match action
}// main parse loop
return result;
}//parse_base
///provided generic parsing function that reports errors on std::io. This
///function is equivalent to [RuntimeParser::parse].
pub fn parse_stdio(&mut self, tokenizer:&mut dyn Lexer<AT>) -> AT
{
let mut stdeh = StandardReporter::new();
self.parse_base(tokenizer,&mut stdeh)
}//parse_stdio
///Parses in interactive training mode with provided path to parserfile.
///The parser file will be modified and a training script file will be
///created for future retraining after grammar is modified. This function
///is equivalent to [RuntimeParser::parse_train].
///
/// When an error occurs, the parser will
/// ask the human trainer for an appropriate error message: it will
/// then insert an entry into its state transition table to
/// give the same error message on future errors of the same type.
/// If the error is caused by an unexpected token that is recognized
/// as a terminal symbol of the grammar, the trainer can select to
/// enter the entry
/// under the reserved ANY_ERROR symbol. If the unexpected token is
/// not recognized as a grammar symbol, then the entry will always
/// be entered under ANY_ERROR. ANY_ERROR entries for a state will match
/// all future unexpected symbols for that state: however, entries for
/// valid grammar symbols will still override the generic entry.
///
/// Example: with the parser for this [toy grammar](https://cs.hofstra.edu/~cscccl/rustlr_project/cpm.grammar), parse_train can run as follows:
///```ignore
/// Write something in C+- : cout << x y ;
/// ERROR on line 1, column 0: unexpected symbol y ..
/// >>>TRAINER: is this error message adequate? If not, enter a better one: need another <<
/// >>>TRAINER: should this message be given for all unexpected symbols in the current state? (default yes) yes
///```
/// (ignore the column number as the lexer for this toy language does not implement it)
///
/// parse_train will then produced a [modified parser](https://cs.hofstra.edu/~cscccl/rustlr_project/cpmparser.rs) as specified
/// by the filename (path) argument. When the augmented parser is used, it will
/// give a more helpful error message:
///```
/// Write something in C+- : cout << x endl
/// ERROR on line 1, column 0: unexpected symbol endl, ** need another << ** ..
///```
///
/// parse_stdio_train calls parse_stdio, which uses stdin/stdout for user interface.
/// Parsing in interactive training mode also produces a [training script file](http://cs.hofstra.edu/~cscccl/rustlr_project/cpmparser.rs_script.txt) which can
/// be used to re-train a parser using [RuntimeParser::train_from_script].
/// This is useful after a grammar is modified with extensions to a language.
pub fn parse_stdio_train(&mut self, tokenizer:&mut dyn Lexer<AT>, parserfile:&str) -> AT
{
let mut stdtrainer = StandardReporter::new_interactive_training(parserfile);
let result = self.parse_base(tokenizer,&mut stdtrainer);
if let Err(m) = stdtrainer.augment_training(parserfile) {
eprintln!("Error in augmenting parser: {:?}",m)
}
return result;
}//parse_stdio_train
/// trains parser from a [training script](https://cs.hofstra.edu/~cscccl/rustlr_project/cpmparser.rs_script.txt)
/// created by interactive training. This
/// is intended to be used after a grammar has been modified and the parser
/// is regenerated with different state numbers. It is the user's
/// responsibility to keep consistent the parser file, script file, and sample
/// input that was used when the script was created. The script contains
/// the line and column numbers of each error encountered, along with either
/// the unexpected symbol that caused the error, or the reserved ANY_ERROR
/// symbol if the error message is to be applied to all unexpected symbols.
/// These entries must match, in sequence, the errors encountered during
/// retraining - it is therefore recommended that the same tokenizer be used
/// during retraining so that the same line/column information are given.
/// The trainer will augment the parser (parserfile) with new Error
/// entries, overriding any previous ones. It is also recommended that the
/// user examines the "load_extras" function that appears at the end of
/// the [augmented parser](https://cs.hofstra.edu/~cscccl/rustlr_project/cpmparser.rs).
/// The train_from_script function does not return
/// a value, unlike [RuntimeParser::parse_stdio] and [RuntimeParser::parse_stdio_train].
pub fn train_from_script(&mut self, tokenizer:&mut dyn Lexer<AT>, parserfile:&str, scriptfile:&str)
{
let mut stdtrainer = StandardReporter::new_script_training(parserfile,scriptfile);
let result = self.parse_base(tokenizer,&mut stdtrainer);
if let Err(m) = stdtrainer.augment_training(parserfile) {
eprintln!("Error in augmenting parser: {:?}",m)
}
if !self.err_occurred {println!("no errors encountered during parsing");}
}//train_from_script
}// 3rd impl RuntimeParser
//// for uniformity
/// These structures are what's on the parse stack.
pub struct StackedItem<AT:Default> // replaces Stackelement
{
pub si : usize, // state index
pub value : AT, // semantic value (don't clone grammar symbols)
pub line: usize, // line and column
pub column: usize,
}
impl<AT:Default> StackedItem<AT>
{
pub fn new(si:usize,value:AT,line:usize,column:usize) -> StackedItem<AT>
{ StackedItem{si,value,line,column} }
/// converts the information in a stacked item to an [LBox] enclosing
/// the abstract syntax value along with starting line and column numbers
pub fn lbox(self) -> LBox<AT> // no longer used
{ LBox::new(self.value,self.line,self.column) }
}