1ANTLR 3.4 Release 2July 18, 2011 3 4Terence Parr, parrt at cs usfca edu 5ANTLR project lead and supreme dictator for life 6University of San Francisco 7 8CHANGES 9 10July 18, 2011 -- release 3.4 11 12* Added tree method insertChild(int i, Object t). 13 14July 14, 2011 15 16* Added BaesTree.freshenParentAndChildIndexesDeeply() to recursively 17 walk tree and set ptrs. 18 19July 6, 2011 20 21* reset() for token stream didn't skip initial off-channel tokens. 22 23July 5, 2011 24 25* Sam fixes rare infinite loop upon error in nextToken(). 26* rewrites weren't pulled from syntactic predicates. 27 28June 29, 2011 29 30* Sam noticed CommonToken.getText() cached substring pulled from input, which 31 defeated purpose of pointing into input char array. Altered to not cache. 32 Should reduce memory footprint. 33 34June 24, 2011 35 36* empty alts with actions didn't have EPSILON dummy alt node. 37 38June 19, 2011 39 40* Udo noticed that Parser.getMissingSymbol() didn't set invented token's input 41 42June 8, 2011 43 44* Added inContext(String context) back into TreeParser. 45 46April 21, 2011 47 48* Updated for ST v4.0.2 (setting iterateAcrossValues = true as instance var) 49* Needed throws condition for delegatedRules. 50 51April 20, 2011 (Sam Harwell) 52 53* Implement the 'throwsSpec' feature of parser rules for the Java target 54* Improve algorithm for SemanticContext Boolean predicate reduction 55 56April 13, 2011 57 58* Unmangled region names in STViz hiearchy tree display. 59* Removed conversion timeout thing again 60 61April 11, 2011 62 63* Added option -Xconversiontimeout back in. Turns out we hit NFA conversion 64 time landmine occasionally with huge grammars; fails over to backtracking 65 (if turned on) if it can't make DFA. 66 67March 29 - April 10, 2011 68 69* Uses ST v4 now!!! Massive change. Only updated Java target so far. 70 Ripped out ST v3 usage to remove indirect dependency on ANTLR v2. 71 72March 28, 2011 73 74* Sam Harwell ported all v2 grammars to v3! 75 76March 3, 2011 77 78* left-recursion pattern off until all targets catch up 79 80* ANTLRCore.sti no longer used; removed from all targets. 81 82* Adding type, text terminal options 83 84* Replaced hetero arg with terminalOptions arg in all templates that reference hetero 85 (this is the class name / node type used for TOKEN<NODETYPE> references 86 in grammar). Widespread but trivial changes to all targets. hetero is 87 now terminalOptions.node. Can also get terminalOptions.type and 88 terminalOptions.text from dictionary of options. 89 90* Fixed mispelling of license in BSD license headers 91 92March 3, 2011 93 94* Add tree, getTree() to ParserRuleReturnScope to do away with specific ones like: 95 public static class rewrite_template_args_return extends ParserRuleReturnScope { 96 CommonTree tree; 97 public Object getTree() { return tree; } 98 }; 99 Removed these special classes if it's just AST; keep if they have defined "returns" 100 values 101 102February 26, 2011 103 104* All finally {} have comment now to suppress warning. 105 106* removed ; from {;} blank method 107 108* Added @SuppressWarnings({"all"}) to front of each 109 generated class. 110 111* -print wasn't always showing ^ and ! in grammar 112 113* Added java-left-recur/Java.g example. 114 115* left-recursion pattern detection handles backtracking mode now 116 117February 25, 2011 118 119* -Xmaxinlinedfastates went to 60 from 10 for forcing prediction in left- 120 recursive expression rules to stay in rule; preds use a parameter. 121 122* trees know where they came from now start..stop tokens; todo: use for better err handling. 123 124* Got immediate left-recursion working for rules. Added TestLeftRecursion.java 125 126February 21, 2011 127 128* Fixed http://www.antlr.org/jira/browse/ANTLR-437 thanks to Vladislav Kuzkokov. 129 added unit test. 130 131February 17, 2011 132 133* Add -language L option to override language=L option in grammar. Same 134 grammar can yield multiple parsers in different languages now. 135 136February 10, 2011 137 138* Added method to generated Java code to return the array of delegates; e.g., 139 import Sub1, Sub2; 140 yields: 141 142 public Parser[] getDelegates() { 143 return new Parser[] {gSub1, gSub2}; 144 } 145 146January 25, 2011 147 148* Improve error messages for no viable alt parse exceptions 149 150January 20, 2011 151 152* TokenRewriteStream had dead code; also updated insertBefore followed by 153 replace or delete. If input is abc and I did insertBefore(2,"y"), where 154 'c' is index 2, then did delete of 2 previously defined functionality 155 was to ignore the insert. that's weird; fixed to keep insert. Also 156 Delete special case of replace (text==null): 157 D.i-j.u D.x-y.v | boundaries overlap => combine to max(min)..max(right) 158 159December 12, 2010 160 161* Send couldBacktrack now to enterDecision in debug protocol 162 163December 4, 2010 164 165* TreeWizard ctor needed a default tree adapator. 166 167November 29, 2010 -- ANTLR v3.3 168 169November 23, 2010 170 171* CodeGenerator.loadLanguageTarget is now static and available to load 172 targets so we can ask them questions during analysis. 173 174* Fixed and added unit test 175 http://www.antlr.org/jira/browse/ANTLR-370 176 http://www.antlr.org/jira/browse/ANTLR-375 177 178November 23, 2010 179 180* Added source name to syntax error msgs 181 182October 20, 2010 183 184Added boolean couldBacktrack to enterDecision in dbg interface. Breaks AW 185interface and other tools! [BREAKS BACKWARD COMPATIBILITY] 186 187October 17, 2010 188 189* Missing -trace in help msg 190 191November 22, 2010 192 193* Added GrammarAST: public int getCharPositionInLine() { return getColumn()-1; } 194 and Grammar.getHasDelegates() for C# guys 195 196October 16, 2010 197 198* Doesn't write profile data to file anymore; emits decision data to stderr 199 200October 14, 2010 201 202* Make OrderedHashSet have deterministic iteration 203 204July 20, 2010 205 206* greedy=true option shuts off nondeterminism warning. 207 208* code gen for AST and -profile didn't compile. had useless line: 209 210 proxy.setTreeAdaptor(adap); 211 212 213July 17, 2010 214 215* Removed conversion timeout failsafe; no longer needed. 216 217* Stats updated to be correct for -report. 218 219June 10, 2010 220 221* added toArray in OrderedHashSet to make addAll calls get same order for DFA edges and possibly code gen in some areas. 222 223June 5, 2010 224 225* Added -Xsavelexer 226 227May 24, 2010 228 229* lexerStringRef was missing elementIndex attribute. i='import' didn't work 230 in lexer. Altered all target stg files. Set in codegen.g 231 232* output=AST, rewrite=true for tree rewriters broken. nextNode for subtree 233 streams didn't dup node, it gave whole tree back. 234 235March 17, 2010 236 237* Added MachineProbe class to make it easier to highlight ambig paths in 238 grammar. More accurate than DecisionProbe; retrofitted from v4. 239 240February 20, 2010 241 242* added range to TokenStream and implementors: 243 /** How far ahead has the stream been asked to look? The return 244 * value is a valid index from 0..n-1. 245 */ 246 int range(); 247 248* added new method to get subset of tokens to buffered token streams: 249 public List get(int start, int stop); 250 251February 15, 2010 252 253* Refs to other tokens in a lexer rule didn't get its line/charpos right. 254 altered Java.stg. 255 256January 31, 2010 257 258* Creating token from another token didn't copy input stream in CommonToken. 259 makes sense to copy too; i don't think anybody relies on it being null after 260 a copy. We might want to know where token came from. 261 262January 26, 2009 263 264* TreeParser.getMissingSymbol() use CommonTree instead of using 265 adaptor.create() 266 267December 8, 2009 268 269* Instead of sharing Token.EOF_TOKEN, I'm now creating EOF tokens so I can set the char position for better error messages. 270 271December 5, 2009 272 273* Fixed bug in TreeVisitor when rewrites altered number of children. Thanks to Chris DiGiano. 274 275* added new buffered on-demand streams: BufferedTokenStream. Renamed CommonTokenStream to LegacyCommonTokenStream and made new one as subclass of BufferedTokenStream. 276 277November 3, 2009 278 279* Added org.antlr.runtime.UnbufferedTokenStream. Was trivial and works! 280 281November 1, 2009 282 283* Couldn't properly reuse parser state; ctor reset the state; fixed. 284 Parser(TokenStream input, RecognizerSharedState state) 285 286* LookaheadStream<T> used some hardcoded Object return types for LT, etc... 287 uses T now. 288 289September 23, 2009 -- ANTLR v3.2 290 291September 21, 2009 [Jim Idle] 292 293* Added new options for tool invocation to control the points at which the code 294 generator tells the target code to use its equivalent of switch() instead of 295 inline ifs. 296 -Xmaxswitchcaselabels m don't generate switch() statements for dfas 297 bigger than m [300] 298 -Xminswitchalts m don't generate switch() statements for dfas smaller 299 than m [3] 300* Upgraded -X help output to include new optins and provide the default 301 settings, as well as provide units for those settings that need them. 302 303* Change the C Target to overide the deafults for the new settings to 304 generate the most optimizable C code from the modern C compiler point of 305 view. This is essentially to always use swtich statements unless there 306 is absolutely no other option. C defaults are to use 1 for minimum and 307 3000 for maximum number of alts that trigger switch(). This results in 308 object code that is 30% smaller and up to 20% faster. 309 310April 23, 2009 311 312* Added reset to TreeNodeStream interface. 313 314April 22, 2009 315 316* Fixed ANTLR-374. Was caused by moved of grammars. %foo() stuff didn't work 317 318April 9, 2009 319 320* .g and .g3 file extensions work again. 321* introduced bug in 3.1.3: gives exception not error msg upon 322 missing .g file 323 324March 26, 2009 325 326* Made ctor in TreeRewriter and TreeFilter call this not super. 327 328March 21, 2009 329 330* Added ctor to RecognizerSharedState to allow cloning it. 331 332March 17, 2009 -- ANTLR v3.1.3 333 334* improved ANTLRv3.g to handle <...> element options 335 336March 15, 2009 337 338* Fixed ANTLR-389. Strip didn't ignore options in subrules; also seemed 339 to demand stdin. 340 341March 15, 2009 342 343* ANTLR always sorts incoming grammar list by dependency. For example, 344 If W.g depends on tokens from P.g then P.g is done first even if 345 W.g mentioned first on command line. It does not ignore any files you 346 specify the commandline. If you do *.g and that includes some 347 imported grammars, it will run antlr on them. 348 349* -make option prevents ANTLR from running on P.g if P older than 350 generated files. 351 352* Added org.antlr.tool.GrammarSpelunker to build a faster dependency 353 checker (what grammars depend on etc...). Totally independent of any 354 ANTLR code; easy to pull into other tools. 355 356* Added org.antlr.misc.Graph, a general graph with nodes 357 containing an Object payload. It knows how to do a topological sort 358 on the nodes. 359 360March 10, 2009 361 362* Added associativity token option to support upcoming special expression 363 parsing. Added rule option strategy=precedence also 364 365March 1, 2009 366 367* Changed ANTLRWorks debug port from 49153 to 49100. Apparently we change the port in 368 ANTLRWorks to 49100 in 1.2 but forgot to do so in the ANTLR targets. 369 370START CHANGES FOR TREE FILTER MODE (pulled from dev branch) 371 372This feature will be announced in 3.2, but I am integrating from my development branch now into the mainline so target developers have a chance to implement. We might release 3.1.3 bug fix release before 3.2. 373 374* CommonTreeNodeStream -> BufferedTreeNodeStream. Now, 375 CommonTreeNodeStream is completely unbuffered unless you are 376 backtracking. No longer making a list of all nodes before tree parsing. 377 378* Added tree grammar filter=true mode. 379 380 Altered templates: 381 Java.stg: added filterMode to genericParser and treeParser. 382 This required a change to ANTLRCore.sti 383 Defined a default parameter in treeParser to set the superclass 384 to TreeFilter for tree grammar with filter=true. It sets 385 superclass to TreeRewriter if filter=true and output=AST. 386 Other them that, I only had to change ANTLR itself a little bit. 387 Made filter mode valid for tree grammars and have it automatically set 388 the necessary elements: @synpredgate, backtrack=true, rewrite=true 389 (if output=AST). Added error message for detecting conflicting 390 options. 391 392* Added misc.FastQueue and TestFastQueue: 393 A queue that can dequeue and get(i) in O(1) and grow arbitrarily large. 394 A linked list is fast at dequeue but slow at get(i). An array is 395 the reverse. This is O(1) for both operations. 396 397* Added tree.TreeIterator, a generator that walks a doubly linked tree. 398 The nodes must know what index they are. It's an Iterator but 399 remove() is not supported. Returns navigation nodes always: 400 UP, DOWN, EOF. 401 402* Added misc.LookaheadStream: A lookahead queue that knows how 403 to mark/release locations in the buffer for backtracking purposes. 404 I hope to use for both tree nodes and tokens. Just implement 405 nextElement() to say how to get next node or token. 406 407END CHANGES FOR TREE FILTER MODE 408 409February 23, 2009 -- ANTLR v3.1.2 410 411February 18, 2009 412 413* Added org.antlr.tool.Strip (reads from file arg or stdin, emits to stdout) 414 to strip actions from a grammar. 415 416February 4, 2009 417 418* Added CommonTree.setUnknownTokenBoundaries(). Sometimes we build trees 419 in a grammar and some of the token boundaries are not set properly. 420 This only matters if you want to print out the original text associated 421 with a subtree. Check this out rule: 422 423 postfixExpression 424 : primary ('.'^ ID)* 425 ; 426 427 For a.b.c, we get a '.' that does not have the token boundaries set. 428 ANTLR only sets token boundaries for subtrees returned from a rule. 429 SO, the overall '.' operator has the token boundaries set from 'a' 430 to 'c' tokens, but the lower '.' subtree does not get the boundaries 431 set (they are -1,-1). Calling setUnknownTokenBoundaries() on 432 the returned tree sets the boundaries appropriately according to the 433 children's token boundaries. 434 435January 22, 2009 436 437* fixed to be listeners.add(listener); in addListener() of DebugEventHub.java 438 439January 20, 2009 440 441* Removed runtime method: mismatch in BaseRecognizer and TreeParser. Seems 442 to be unused. Had to override method recoverFromMismatchedToken() in 443 TreeParser to get rid of single token insertion and deletion for 444 tree parsing because it makes no sense with all of the up-and-down nodes. 445 446* Changed JIRA port number from 8888 to no port spec (aka port 80) and all 447 refs to it in this file. 448 449* Changed BaseTree to Tree typecase in getChild and toStringTree() and 450 deleteChild() to make more generic. 451 452December 16, 2008 453 454* Added -verbose cmd-line option and turned off standard header 455 and list of read files. Silent now without -verbose. 456 457November 24, 2008 458 459* null-ptr protected getParent and a few others. 460 461* Added new ctor to CommonTreeNodeStream for walking subtrees. Avoids 462 having to make new serialized stream as it can reuse overall node stream 463 buffer. 464 465November 20, 2008 466 467* Updated BaseTest to isolate tests better. 468 469November 17, 2008 470 471* BaseTreeAdaptor.getType() was hosed; always gave 0. Thanks to Sam Harwell. 472 473November 8, 2008 474 475* Added methods to BaseRecognizer: 476 public void setBacktrackingLevel(int n) { state.backtracking = n; } 477 /** Return whether or not a backtracking attempt failed. */ 478 public boolean failed() { return state.failed; } 479 480November 5, 2008 481 482* Tweaked traceIn/Out to say "fail/succeeded" 483 484* Bug in code gen for tree grammar wildcard list label x+=. 485 486* Use of backtrack=true anywhere in grammar causes backtracking sensitive 487 code to be generated. Actions are gated etc... Previously, that only 488 happened when a syntactic predicate appeared in a DFA. But, we need 489 to gate actions when backtracking option is set even if no decision 490 is generated to support filtering of trees. 491 492October 25, 2008 493 494* Fixed debug event socket protocol to allow spaces in filenames. 495 496* Added TreeVisitor and TreeVisitorAction to org.antlr.runtime.tree. 497 498October 22, 2008 499 500* Added inContext() to TreeParser. Very useful for predicating 501 tree grammar productions according to context (their parent list). 502 Added new TestTreeContext unit tests (15). 503 504 /** Check if current node in input has a context. Context means sequence 505 * of nodes towards root of tree. For example, you might say context 506 * is "MULT" which means my parent must be MULT. "CLASS VARDEF" says 507 * current node must be child of a VARDEF and whose parent is a CLASS node. 508 * You can use "..." to mean zero-or-more nodes. "METHOD ... VARDEF" 509 * means my parent is VARDEF and somewhere above that is a METHOD node. 510 * The first node in the context is not necessarily the root. The context 511 * matcher stops matching and returns true when it runs out of context. 512 * There is no way to force the first node to be the root. 513 */ 514 public boolean inContext(String context) {...} 515 516* Added 3 methods to Tree interface [BREAKS BACKWARD COMPATIBILITY] 517 518 /** Is there is a node above with token type ttype? */ 519 public boolean hasAncestor(int ttype); 520 521 /** Walk upwards and get first ancestor with this token type. */ 522 public Tree getAncestor(int ttype); 523 524 /** Return a list of all ancestors of this node. The first node of 525 * list is the root and the last is the parent of this node. 526 */ 527 public List getAncestors(); 528 529October 21, 2008 530 531* Updated unit tests to be correct for \uFFFE->\uFFFF change 532 533* Made . in tree grammar look like ^(. .*) to analysis, though ^(. foo) 534 is illegal (can't have . at root). Wildcard is subtree or node. 535 Fixed bugs: 536 http://www.antlr.org/browse/ANTLR-248 537 http://www.antlr.org/browse/ANTLR-344 538 539October 1, 2008 -- ANTLR v3.1.1 540 541September 8, 2008 542 543* Labels on tokens, rules carry into synpreds now so semantic predicates work. 544 This didn't work since labels were stripped in the synpred and they weren't 545 defined in the generated method. 546 547 a : x=A z=a {$x.text.equals($z.text)}? A 548 | y=A a A A 549 ; 550 551September 3, 2008 552 553* Made a REV static variable in Tool so that we can change the rev for 554 daily builds. 555 556* Made \uFFFF a valid character. Token types are 32-bit clean using -1 557 not 0x0000FFFF as -1 so it should be okay. Label.java: 558 public static final int MIN_CHAR_VALUE = '\u0000'; 559 public static final int MAX_CHAR_VALUE = '\uFFFF'; 560 561August 30, 2008 562 563* Changed messages in en.stg so that TOKEN_NONDETERMINISM correctly 564 indicates when actions hid semantic predicates. 565 566August 15, 2008 567 568* Tweaked build properties and build.xml 569 570August 13, 2008 571 572* Fixed ANTLR-314; 3.1 introduced a problem with list labels += 573 574August 12, 2008 -- ANTLR v3.1 575 576* Added JavaScript target 577 578August 7, 2008 579 580* an NFA target of EOF predicate transition in DFA cause an exception in 581 getPredicatesPerNonDeterministicAlt(). 582 583* Kay Roepke found a nasty bug when debugging AST-constructing 584 composite recognizers. If the input state was null to the constructor, 585 super class constructor created a new parser state object. 586 Later, though we passed the argument state not this.state 587 to the delegate constructors, forcing them to share a different 588 state objects! Changed state to this.state in Dbg.stg constructors. 589 590* Ack. messed up debug/AST. Have to set proxy's tree adaptor; it's 591 a circular ref. Just an ASTDbg.stg change. 592 593August 4, 2008 594 595* superClass works now for lexers 596 597* Made Grammar.defineNamedAction propogate header actions down to all 598 delegates if root grammar; regardless of lexer/parser scope. 599 600* Rejiggered AST templates to propogate changes to tree adaptor 601 for delegate grammars. Fixes ANTLR-302 602 603August 4, 2008 604 605* FOLLOW set computations altered constant FOLLOW bit sets. 606 607* Added (...) are all predicate evaluations. 608 609* Extra init code for tree parser nonrewrite mode removed. 610 611* Added empty child list check in becomeRoot 612 613August 3, 2008 614 615* Was using RuleReturnScope not Rulename_return for list labels in tree 616 parser. 617 618* Didn't set _last in tree parser for rule ref track stuff (rewrite=true) 619 620August 2, 2008 621 622* Benjamin found another rewrite engine bug. 623 624July 30, 2008 625 626* CommonTreeNodeStream / CommonTokenStream did not reset properly. 627 628July 29, 2008 629 630* Fixed another bug in TokenRewriteStream; didn't like inserts after end. 631 632July 28, 2008 633 634* Fixed bug in TokenRewriteStream.toString(start,stop); it ignored 635 parameters. ;) 636 637July 17, 2008 638 639* allow qualified type names in hetero <...> options like T<a.b.c.Node> 640 641July 5, 2008 642 643* treeLevel not set for setBlock alts; added unit test 644 645July 3, 2008 646 647* Fixed ANTLR-267. parse tree added nodes during backtracking and 648 cyclic DFAs. tracks hidden tokens too now. Added toInputString() to 649 get text back including hidden tokens. Shows <epsilon> for rules 650 that match nothing. 651 652June 26, 2008 653 654* Added gParent ptr that points to immediate parent grammar. E.g., 655 // delegators 656 public MParser gM; 657 public M_S gS; 658 public M_S gParent = gS; // NEW 659 660* Grammar imports didn't set all the delegate pointers...lots of imported 661 grammars would cause a null ptr exception. Fixes ANTLR-292. 662 663June 25, 2008 664 665* List labels in tree construction didn't always track the tree; sometimes 666 had a rule result structure. 667 668June 4, 2008 669 670* Improved unit testing so that each test suite executes and builds grammars 671 in a separate temporary directory. This means they can execute concurrently. 672 Also seem to be a problem with my class path during execution. Moved 673 tmpdir for ahead of standard CLASSPATH. 674 675* By virtue of an improvement to StringTemplate, output newlines 676 in generated files should be normalized to whatever your host uses. 677 678June 3, 2008 679 680* Restrict legality of grammar options; for example you cannot use output option 681 in lexer anymore. 682 683June 2, 2008 684 685* Throw illegal arg exception upon invalid TokenRewriteStream ops. Rewrote 686 core of engine. Slightly different operation. Added many more unit tests. 687 6883.1b1 - May 20, 2008 689 690May 11, 2008 691 692* rewrite=true, output=AST for tree grammar was not working. Altered trees were not 693 propagated back up the rule reference chain. Required a number of mods to 694 ASTTreeParser.stg. Added unit tests. 695 696May 10, 2008 697 698* [BACKWARD INCOMPATIBLE if you override match()] 699 I had turned off single token insertion and deletion because I could not figure 700 out how to work with trees and actions. Figure that out and so I turned it back on. 701 match() returns Object matched now (parser, tree parser) so we can set labels 702 on token refs properly after single token ins/del error recovery. Allows actions 703 and tree construction to proceed normally even though we recover in the middle of 704 an alternative. Added methods for conjuring up missing symbols: getMissingSymbol(). 705 706* refactored BaseRecognizer error handling routines 707 708* Single token error recovery was not properly taking into consideration EOF. 709 710* ANTLR no longer tries to recover in tree parsers inline using single node deletion or insertion; throw exception. Trees should be well formed as they are not created by users. 711 712* Added empty constructors to the exception classes that did not have them so that ANTLRWorks can create the exceptions. 713 714* Made debug tree adaptor deal with tokens conjured up during error recovery. 715 716* Removed extra location() debug element that was emitted. 717 718May 8, 2008 719 720* ANTLR didn't update line/col to DFA map for AW. 721 722May 6-7, 2008 723 724* Insufficiently covered (with semantic predicates) alt warnings are now emitted before 725 nondeterminisms so it's clear the nondeterminism is a result of insufficient preds. 726 727* Improved insufficiently covered alt warnings from: 728 warning(203): T.g:2:3: The following alternatives are insufficiently covered with predicates: 1 729 to: 730 warning(203): T.g:2:3: Input B is insufficiently covered with predicates at loca 731tions: alt 1: line 3:15, alt 2: line 2:9 732 733* Improved nondeterminism warning to have: 734 Semantic predicates were present but were hidden by actions. 735parser grammar U; 736a : (A B)? ; 737b : X a {p1}? A B | Y a {a1} {p2}? A B | Z a ; 738 739To create the prediction DFA for the optional sub rule in 'a', ANTLR must find all references to 'a' to determine what can follow. A B can follow 'a' in the first two alts rule 'b'. To resolve the conflict between matching A B immediately in the sub rule and exiting rule 'a' to match it in 'b', ANTLR looks for predicates. In this case, there are two predicates that indicate the semantic context in which the surrounding alternatives are valid. The problem is that one of the predicates is hidden by an action. It took me 1.5 days, but I've finally have gotten ANTLR to properly track the insufficiently covered alternatives. Further, I have gotten it to tell you precisely where the uncovered predicates are even if they are simply hidden by actions. I have also updated all of the nondeterminism warnings so that it tells you if there was a predicate but one hidden by an action (this could be a separate condition from insufficiently covered predicates). here are your messages from ANTLR: 740 741ANTLR Parser Generator Version 3.1b1 (??) 1989-2007 742warning(203): U.g:2:5: Input such as "A B" is insufficiently covered with predicates at locations: alt 2: line 3:38 at B 743Semantic predicates were present but were hidden by actions. 744warning(200): U.g:2:5: Decision can match input such as "A B" using multiple alternatives: 1, 2 745As a result, alternative(s) 2 were disabled for that input 746Semantic predicates were present but were hidden by actions. 747 748* Fixed issue where 749r41 750 : (INT -> INT) ( ('+' i=INT) -> ^($i $r41) )* ';' 751 ; 752still warned about $r41 being ambig. 753 754* actions are now added to the NFA. 755 756* Fixed ANTLR-222. ANTLR now ignores preds after actions. 757 758May 5, 2008 759 760* Fixed ANTLR-235 by backing out a change from 12-31-07. 761 762* Fixed ANTLR-249; I include semantic context again in closure busy signal. 763 764May 3, 2008 765 766* Fixed ANTLR-208. Looks in library or in -o output path. antlr -o foo T.g U.g where U needs T.tokens won't work unless we look in foo too. fixed. 767 768* Refactored assign.types.g to move methods to a class called AssignTokenTypesBehavior. 769 770* Fixed ANTLR-207. Lexers importing vocabs didn't see ';'=4 type aliases in .tokens. 771 772* Fixed ANTLR-228. Couldn't use wildcard in alts with AST rewrites. 773 774May 2, 2008 775 776* Fixed ANTLR-230; can use \' now in action. 777 778* Scope attributes no longer have a stack depth check on front. If you ref $r::a when r has not invoked you, then you get an exception not a default value. Back to the way 3.0.1 worked. 779 780* $channel was a global variable in 3.0.1 unlike $type which did not affect an invoking lexer rule. Now it's local too. Only $type and $channel are ever set with regularity. Setting those should not affect an invoking lexer rule as in the following should work: 781 782 X : ID WS? '=' ID ; // result is X on normal channel 783 WS : ' '+ {$channel = HIDDEN; } ; 784 785 STRING : '"' (ESC|.)* '"' ; // result is STRING not ESC 786 787 FLOAT : INT '.' INT? ; // should be FLOAT 788 INT : Digit+ ; 789 fragment 790 Digit : '0'..'9' ; 791 792* Fixed bug in interpreter regarding (...)* loops 793 794May 1, 2008 795 796* Fixed ANTLR-202. These now give warnings about ambig ref to $a. 797 a : ID a -> $a | INT ; 798 and 799 a : A a {$a.text} | B ; 800 801April 30, 2008 802 803* Fixed ANTLR-237. updated -depend to know about imported grammars. 804$ java org.antlr.Tool -depend -lib foo T.g 805 ANTLR Parser Generator Version 3.1b1 (??) 1989-2007 806 T.g: foo/Java.g 807 TParser.java : T.g 808 T.tokens : T.g 809 TLexer.java : T.g 810 T_Java : T.g 811 812April 29, 2008 813 814* Fixed ANTLR-217; scope A,B,C; didn't work 815 816* Fixed ANTLR-224; ! or ^ on item in alt with rewrite gave exception 817 818* Added token options to terminals: ID<node=V; foo="Big bob"> etc... 819 node is default so you can do ID<V> for hetero tree types. most common. 820 821April 17, 2008 822 823* Use default msg if unknown recog type in getErrorMessage(): 824 String msg = e.getMessage(); 825 826April 14, 2008 827 828* %x.y = foo; was not working in @members section 829 830March 29, 2008 831 832* Import couldn't handle A imports B imports C. 833 834March 27, 2008 835 836* Added get/setInputStream to Token interface and affected classes. 837 838February 26, 2008 839 840* made fillBuffer public in CommonTreeNodeStream so we can add trees 841 to stream for interactive interpreters. 842 843February 14, 2008 844 845* Fixed a bug in the code generation where tree level 0 was used 846 no matter what to rewrite trees in tree grammars. added unit test 847 848* Fixed ANTLR-221. exceptions were generated when using 849 AST construction operators and no output=AST option. 850 851February 13, 2008 852 853* Improved error msgs for unreachable alts and tokens. 854 855February 11-12, 2008 856 857* Fixed ANTLR-219. 858 It looks like the AST construction code for sets was totally messed up. 859 This was for not only the new tree parser AST construction, but also 860 the regular tree construction for parsers. I had to introduce templates 861 in the ASTTreeParser.stg file to deal with this. added unit tests: 862 TestTreeGrammarRewriteAST.testSetMatchNoRewrite(), 863 testSetMatchNoRewriteLevel2(), testSetMatchNoRewriteLevel2Root(). 864 Had to add template matchRuleBlockSet() 865 to differentiate between a regular set in one that is an entire rule. 866 If it is an entire rule, it has to set the return value, retval.tree. 867 868* Fixed ANTLR-220. 869 Made TreeAdaptor dupNode and dupTree events emit debugging events 870 so AW could see tree node duplications. 871 872February 4, 2008 873 874* BACKWARD INCOMPATIBILITY 875 Added getSourceName to IntStream and TokenSource interfaces and also the 876 BaseRecognizer. Have to know where char come from for error messages. 877 Widespread change, but a trivial one. 878 879January 17, 2008 880 881* Interpreter throws FailedPredicateException now when it sees a predicate; 882 before it was silently failing. I'll make it work one of these days. ;) 883 884January 12, 2008 885 886* Copy ctor not copying start and stop for common token. Fixes ANTLR-212 887 888* Removed single token insertion and deletion for tokens, sets. 889 Required a change to the code generation for matchSet() template 890 and a tweak inside the BaseRecognizer. To engage this again is easy, 891 just override mismatch() to call mismatchRecover(). I changed it to simply 892 throw an exception. 893 894* Added syntaxError recognizer state var so you can easily tell if 895 a recognizer failed. Added getNumberOfSyntaxErrors() to recognizers. 896 897* Added doc for the error node stuff: 898 http://www.antlr.org/wiki/display/ANTLR3/Tree+construction 899 900* Fixed ANTLR-193 901 902* Added recognizer methods to answer questions about current mismatched 903 token error. Useful now since i don't automatically recover inline 904 to such errors (I throw exception): 905 mismatchIsUnwantedToken(IntStream input, int ttype) 906 mismatchIsMissingToken(IntStream input, BitSet follow) 907 908* Added UnwantedTokenException and MissingTokenException to make 909 match() problems more precise in case you want to catch differently. 910 Updated getErrorMessage() to be more precise. Says: 911 912 line 2:9 missing EQ at '0' 913 914 now instead of 915 916 line 2:9 mismatched input '0' expecting EQ 917 918 Input "x=9 9;" gives 919 920 line 3:8 extraneous input '9' expecting ';' 921 922 When very confused, "x=9 for;", you still get old mismatched message: 923 924 line 3:8 extraneous input 'for' expecting ';' 925 line 3:11 mismatched input ';' expecting '(' 926 927* Added unit tests to TestAutoAST and copied to TestRewriteAST with 928 suitable rewrites to ensure basic error node insertion works. 929 930January 11, 2008 931 932* Adding errorNode to TreeAdaptor and various debug 933 events/listeners. Had to add new class runtime.tree.CommonErrorNode 934 to hold all the goodies: input stream, start/stop objects. 935 936* Tweaked CommonTree.getType() to return INVALID_TOKEN_TYPE 937 instead of literal 0 (same thing at moment though). 938 939* Updated ANTLRWorks to show error nodes in tree as much as I could; Jean 940 will get to rest of it. 941 942January 9-10, 2008 943 944* Continued work on debugging/profiling composite grammars. 945 946* Updated debug protocol for debugging composite grammars. enter/exit 947 rule needs grammar to know when to flip display in AW. 948 949* Fixed ANTLR-209. ANTLR consumed 2 not 1 char to recover in lexer. 950 951* Added two faqs instead of making changes to antlr runtime about 952 lexer error handling: 953 http://www.antlr.org/wiki/pages/viewpage.action?pageId=5341230 954 http://www.antlr.org/wiki/pages/viewpage.action?pageId=5341217 955 956January 1-8, 2008 957 958* Making debugging/profiling work with composite grammars. 959 960* Updated ANTLRWorks so it works still for noncomposite grammars. 961 962* two new examples: import and composite-java (the java example grammar 963 broken up into multiple pieces using import). 964 965* Worked on composite grammars. Had to refactor a lot of code to make 966 ANTLR deal with one meta grammar made up of multiple grammars. I 967 thought I had it sort of working back in August. Yes, but barely. Lots 968 of work to do it seemed. Lots of clean up work. Many new unit tests 969 in TestCompositeGrammars. Had to add new error messages warning about 970 conflicting tokens inherited from multiple grammars etc... 971 972 TOKEN_ALIAS_CONFLICT(arg,arg2) ::= 973 "cannot alias <arg>; string already assigned to <arg2>" 974 TOKEN_ALIAS_REASSIGNMENT(arg,arg2) ::= 975 "cannot alias <arg>; token name already assigned to <arg2>" 976 TOKEN_VOCAB_IN_DELEGATE(arg,arg2) ::= 977 "tokenVocab option ignored in imported grammar <arg>" 978 INVALID_IMPORT(arg,arg2) ::= 979 "<arg.grammarTypeString> grammar <arg.name> cannot import <arg2.grammarTypeString> grammar <arg2.name>" 980 IMPORTED_TOKENS_RULE_EMPTY(arg,arg2) ::= 981 "no lexer rules contributed to <arg> from imported grammar <arg2>" 982 IMPORT_NAME_CLASH(arg,arg2) ::= 983 "combined grammar <arg.name> and imported <arg2.grammarTypeString> grammar <arg2.name> both generate <arg2.recognizerName>; import ignored" 984 985 This stuff got really really complicated. Syntactic predicate names even 986 had to be scoped per grammar so they don't conflict. 987 988* When using subrules like (atom->atom) to set result tree, it was not 989 properly setting result (early enough). Future code got null for 990 $rule.tree. 991 992December 31, 2007 993 994* Added the start of a semantic predicate computation for LL(1) to 995 solve a problem with slow grammar analysis even with k=1 due to 996 predicates. Then I realized the problem with that grammar was 997 elsewhere. Semantic context really shouldn't be used when 998 preventing closure recomputation (May 2008 I discovered I was 999 wrong--you do need it). The predicates became huge even though the 1000 reduced value would be no different. The analyzer seems faster now 1001 that I am not testing predicate values all the time. Further it may 1002 terminate sooner just due to reduced closure recursion. 1003 1004* Moved FIRST/FOLLOW computations to a separate class LL1Analyzer to 1005 tidy up. 1006 1007* ANTLR lexer allowed octal escapes, but they didn't work. ;) Rather than 1008 fix, I'm removing. Use '\uxxxx' to get even 8 bit char values: \u00xx. 1009 1010December 29, 2007 1011 1012* Fixed ANTLR-206. I wasn't avoiding analyzing decisions in 1013 left-recursive rules. 1014 1015* Had to add hetero arg to all tokenRef*() templates. Added _last 1016 local var to track last child so we can do replaceChildren() during 1017 AST rewrite mode for tree grammars. Should be useful later for .text 1018 property. Ack, hetero arg is on lots of templates. :( Moved 1019 ruleCleanUp() template into ASTTreeParser and ASTParser groups. 1020 1021* added noRewrite() template (to Java.stg) so we can insert code during 1022 rewrite mode to return original tree if no rewrite. Might be useful 1023 for token rewrites later. For templates too? 1024 1025* Had to add if !rewriteMode around tree construction in tree parser 1026 templates. 1027 1028* Harald Muller pointed out that we need to use <initValue(attr.type)> 1029 in our tests for null token/rule property references. For int types 1030 we need 0 not null. (p!=null?p.line:0). Changed scopeAttributeRef, 1031 ruleLabelRef. Also changed the known typed attributes like 1032 lexerRuleLabelPropertyRef_line to yield 0 upon null rule ref to 1033 be consistent with case when we don't know the type. Fixes ANTLR-195. 1034 Added testTypeOfGuardedAttributeRefIsCorrect test and reset expected 1035 output for 13 tests that now "fail". 1036 1037December 28, 2007 1038 1039* added polydiff example (Java target) 1040 1041* added "int" property for token and lexer rule refs. super convenient. E.g., 1042 a : b=INT {int x = $b.int;} ; 1043 1044December 27, 2007 1045 1046* Changed -Xnoinlinedfa to -Xmaxinlinedfastates m where m is 1047 maximum number of states a DFA can have before ANTLR avoids 1048 inlining it. Instead, you get a table-based DFA. This 1049 affectively avoids some acyclic DFA that still have many states 1050 with multiple incident edges. The combinatorial explosion smacks 1051 of infinite loop. Fixes ANTLR-130. 1052 1053* [...] are allowed in args now but ] must be escaped as \]. E.g., 1054 a[String[\] ick, int i] : ... ; 1055 And calling a rule: foo[x[i\], 34] 1056 Fixes ANTLR-140. 1057 1058* Fixed ANTLR-105. Target.getTargetStringLiteralFromANTLRStringLiteral() 1059 escaped " that were already escaped. 1060 1061* target's can now specify how to encode int as char escape. Moved 1062 DFA.encodeIntAsCharEscape to Target. 1063 1064* Bug in runtime.DFA. If a special state (one with predicate) failed, it 1065 tried to continue (causing out of range exception due to state = -1) 1066 instead of reporting error. 1067 1068* If -dfa with combined grammar T.g, builds T.dec-*.dot and TLexer.dec-*.dot 1069 1070* Fix ANTLR-165. 1071 Generate TParser.java and TLexer.java from T.g if combined, else 1072 use T.java as output regardless of type. 1073 BACKWARD INCOMPATIBILITY since file names change. 1074 I changed the examples-v3/java to be consistent. Required XML.g -> 1075 XMLLexer.java and fuzzy/Main.java change. 1076 1077* Fix ANTLR-169. Deletes tmp lexer grammar file. 1078 1079December 25, 2007 1080 1081* Fixed ANTLR-111. More unit tests in TestAttributes. 1082 1083December 25, 2007 1084 1085* Dangling states ("decision cannot distinguish between alternatives 1086 for at least one input sequence") is now an error not a warning. 1087 1088* Added sample input sequence that leads to dangling DFA state, one 1089 that cannot reach an accept state. ANTLR ran into a case where 1090 the same input sequence reaches multiple locations in the NFA 1091 (and so not nondeterministic), but analysis ran out of further 1092 NFA states to look for more input. Commonly at EOF target states. 1093 Now says: 1094 1095 error(202): CS.g:248:95: the decision cannot distinguish between alternative(s) 1,2 for input such as "DOT IDENTIFIER EOF" 1096 1097 Also fixed bug where dangling states did not resolve to stop states. 1098 1099* Fixed ANTLR-123 1100 1101December 17-21, 2007 1102 1103* k=1 doesn't prevent backtracking anymore as in 1104 (options {k=1;}:'else' statement)? 1105 if backtrack=true for overall grammar. Set to false in subrule. 1106 1107* Optimized the analysis engine for LL(1). Doesn't attempt LL(*) unless 1108 LL(1) fails. If not LL(1) but autobacktracking but no other kind of 1109 predicate, it also avoids LL(*). This is only important for really 1110 big 4000 line grammars etc... 1111 1112* Lots of code clean up 1113 1114December 16, 2007 1115 1116* Yet more Kay pair programming. Saved yet more RAM; 15% by 1117 wacking NFA configurations etc in each DFA state after DFA construction. 1118 1119* Overall we drop from 2m49s to 1m11s for a huge 4000 line TSQL grammar 1120 with k=*. Only needs -Xconversiontimeout 2000 now not 1121 -Xconversiontimeout 5000 too. With k=1, it's 1m30s down to 40s. 1122 1123December 15, 2007 1124 1125* Working with Kay Roepke, we got about 15% speed improvement in 1126 overall ANTLR exec time. Memory footprint seems to be about 50% 1127 smaller. 1128 1129December 13-14, 2007 1130 1131* I abort entire DFA construction now when I see recursion in > 1 alt. 1132 Decision is non-LL(*) even if some pieces are LL(*). Safer to bail 1133 out and try with fixed k. If user set fixed k then it continues because 1134 analysis will eventually terminate for sure. If a pred is encountered 1135 and k=* and it's non-LL(*), it aborts and retries at k=1 but does NOT 1136 emit an error. 1137 1138* Decided that recursion overflow while computing a lookahead DFA is 1139 serious enough that I should bail out of entire DFA computation. 1140 Previously analysis tried to keep going and made the rules about 1141 how analysis worked more complicated. Better to simply abort when 1142 decision can't be computed with current max stack (-Xm option). 1143 User can adjust or add predicate etc... This is now an error 1144 not a warning. 1145 1146* Recursion overflow and unreachable alt is now a fatal error; no code gen. 1147 The decision will literally not work. 1148 1149* Cleaned up how DFA construction/analysis aborts due to non-LL(*) and 1150 overflow etc... Throws exceptions now, which cleans up a bunch of IF 1151 checks etc... Very nice now. Exceptions: 1152 analysis/AnalysisRecursionOverflowException.java 1153 analysis/AnalysisTimeoutException.java 1154 analysis/NonLLStarDecisionException.java 1155 1156* ErrorManager.grammarWarning() counted them as errors not warnings. 1157 1158* Unreachable alt warnings are now errors. 1159 1160* The upshot of these changes is that I fixed ANTLR-178 and did 1161 lots of refactoring of code handling analysis failure. 1162 1163December 11, 2007 1164 1165* Could not deal with spaces, oddly enough in arg lists: 1166 grammar Bad; 1167 a : A b["foo", $A.text] ; 1168 b[String x, String y] : C ; 1169 1170October 28, 2007 1171 1172* Made ANTLR emit a better error message when it cannot write the 1173 implicit lexer file from a combined grammar. Used to say "cannot open 1174 file", now says "cannot write file" and gives backtrace. 1175 1176September 15, 2007 1177 1178add getCharStream to Lexer. 1179 1180September 10, 2007 1181 1182* Added {{...}} forced action executed even during backtracking. 1183 1184September 9, 2007 1185 1186* r='string' in lexer got a duplicate label definition. 1187 1188August 21, 2007 1189 1190* $scope::variable refs now check for empty stack so that expr == null if 1191 $scope has an empty stack. Works for $scope[...]::variable too. Nice! 1192 1193August 20, 2007 1194 1195* Added reset() to CommonTreeNodeStream, token stream too 1196 1197* Made refs to rule/token properties use ?: to avoid null ptr exception. 1198 $label.st now is label!=null?label.st:null. Updated TestAttributes. 1199 This is useful not only for optional rule/token refs, but also during 1200 error recovery. If ID is not matched, $ID.text won't cause a null ptr. 1201 1202August 20, 2007 1203* Fixed ANTLR-177: hashCode/equals not consistent for label 1204 Fixed bug where Rule was compared to string; introduced from dev branch 1205 1206August 15, 2007 -- Got rough draft of the grammar import working. 1207 Submit to dev and then integrate into mainline. 1208 1209 All file changes/additions: 1210 1211 README.txt # edit 1212 CHANGES.txt # add 1213 Factored out the changes from the readme. 1214 1215 runtime/Java/src/org/antlr/runtime/BaseRecognizer.java # edit 1216 runtime/Java/src/org/antlr/runtime/DFA.java # edit 1217 runtime/Java/src/org/antlr/runtime/Lexer.java # edit 1218 runtime/Java/src/org/antlr/runtime/Parser.java # edit 1219 runtime/Java/src/org/antlr/runtime/debug/DebugParser.java # edit 1220 runtime/Java/src/org/antlr/runtime/tree/TreeParser.java # edit 1221 Factored state fields into RecognizerSharedState 1222 object. You will see a lot of things like 1223 state.errorRecovery = false; 1224 runtime/Java/src/org/antlr/runtime/RecognizerSharedState.java # add 1225 Shares all recognizer state variables including lexer even though 1226 these are superfluous to parsers and tree parsers. There 1227 was a casting issue that I could not resolve. 1228 1229 src/org/antlr/Tool.java # edit 1230 Broke a part Grammar.setGrammarContent() into 1231 parseAndBuildAST() and analyzeGrammar() to make the grammar 1232 import work. I needed to be able to look at the trees for 1233 imported grammars before analyzing them and building DFA. Added 1234 use of the CompositeGrammar object and handling of multiple 1235 delegate grammars. Changed decision DFA DOT file names to 1236 include the grammar name. 1237 1238 src/org/antlr/analysis/DFA.java # edit 1239 Just tweaked to use generics, updated a comment. 1240 1241 src/org/antlr/analysis/DecisionProbe.java # edit 1242 Just tweaked to use generics. 1243 1244 src/org/antlr/analysis/NFA.java # edit 1245 NFA now span multiple grammars and so I moved the NFAs state 1246 tracking to the composite grammar object. 1247 1248 src/org/antlr/analysis/NFAState.java # edit 1249 Added some null checking and made a field public. 1250 1251 src/org/antlr/analysis/NFAToDFAConverter.java # edit 1252 Changed a method call to directly access a field. 1253 1254 src/org/antlr/analysis/RuleClosureTransition.java # edit 1255 Instead of using a rule index, which does not span multiple 1256 grammars, the transition object now attracts a pointer to 1257 the actual Rule definition object. 1258 1259 src/org/antlr/analysis/SemanticContext.java # edit 1260 Tweaked to use a field instead of a method 1261 1262 src/org/antlr/codegen/ActionTranslator.g # edit 1263 src/org/antlr/codegen/ActionTranslatorLexer.java # edit 1264 Tweaked to use new runtime and they changed method name. 1265 1266 src/org/antlr/codegen/CodeGenerator.java # edit 1267 Tweaked comments. 1268 1269 src/org/antlr/codegen/codegen.g # edit 1270 Added import grammar syntax and altered rule atom to pass a 1271 scope around so that grammar.rule works. Caution this 1272 feature is used internally by ANTLR and is not meant to be 1273 used by users at this point. 1274 1275 src/org/antlr/codegen/templates/ANTLRCore.sti # edit 1276 Added scope to all ruleref template interfaces. 1277 1278 src/org/antlr/codegen/templates/Java/Java.stg # edit 1279 Grammars can now import other grammars, which I implemented 1280 using a delegation pointer to the other grammar(s). So if 1281 grammar A imports grammars B and C, then the generated 1282 recognizer for A must have delegation pointers to BParser 1283 and CParser objects. These are now fields: 1284 1285 // delegates 1286 <grammar.delegates: 1287 {g|public <g.name>Lexer <g:delegateName()>;}; separator="\n"> 1288 1289 Also, B and C must have back pointers to the delegator so 1290 that they can refer to rules that have been overridden. 1291 This is a mechanism akin to static inheritance: 1292 1293 // delegators 1294 <grammar.delegators: 1295 {g|public <g.name>Lexer <g:delegateName()>;}; separator="\n"> 1296 1297 This file also has a lot of changes so that state variables 1298 now are state.backtracking instead of the implied 1299 this.backtracking. 1300 1301 The file also refers to grammar.delegatedRules attribute 1302 which is the list of Rule objects for which you must 1303 generate manual delegation. This amounts to a stub whereby 1304 rule foo's method foo() simply calls X.foo() if foo is not 1305 defined inside the delegator. 1306 1307 You will notice that the ruleref templates now take a scope 1308 so that I can have implicit rule Tokens referred to 1309 delegate.Tokens rule in a delegate grammar. This is the way 1310 I do lexer grammar imports. 1311 1312 I added a template called delegateName which uses the 1313 grammar name to compute a delegate name if the user does not 1314 specify a label in the import statement such as: 1315 1316 import x=X; 1317 1318 Oh, note that rule reference templates all receive a Rule 1319 object now instead of the simple rule name as the 'rule' 1320 attribute. You will see me doing <rule.name> instead of 1321 <name> now. 1322 1323 src/org/antlr/codegen/templates/Java/Dbg.stg # edit 1324 Changes mirroring the constructor and field stuff from 1325 Java.stg. Part of this is a cut and paste because of a bug 1326 in ST. 1327 1328 src/org/antlr/codegen/templates/Java/AST.stg # edit 1329 src/org/antlr/codegen/templates/Java/ASTParser.stg # edit 1330 src/org/antlr/codegen/templates/Java/ASTTreeParser.stg # edit 1331 Just added the scope attribute. 1332 1333 src/org/antlr/test/BaseTest.java # edit 1334 Added functionality to support testing composite grammars. 1335 execLexer() 1336 1337 src/org/antlr/test/TestAttributes.java # edit 1338 Tweak to deal with shared recognizer state. 1339 1340 src/org/antlr/test/TestCompositeGrammars.java # add 1341 Start of my unit tests. 1342 1343 src/org/antlr/tool/CompositeGrammar.java # add 1344 src/org/antlr/tool/CompositeGrammarTree.java # add 1345 Tracks main grammar and all delegate grammars. Tracks unique 1346 NFA state numbers and unique token types. This keeps a tree 1347 of grammars computed from the import/delegation chain. When 1348 you want to look up a rule, it starts at the root of the 1349 tree and does a pre-order search to find the rule. 1350 1351 src/org/antlr/tool/ActionAnalysis.g # edit 1352 src/org/antlr/tool/ActionAnalysisLexer.java # edit 1353 1354 src/org/antlr/tool/AttributeScope.java # edit 1355 Updated to use generics in one place. 1356 1357 src/org/antlr/tool/DOTGenerator.java # edit 1358 Updated to indicate when nonlocal rules are referenced. 1359 1360 src/org/antlr/tool/ErrorManager.java # edit 1361 Added some error messages for import grammars; I need more. 1362 1363 src/org/antlr/tool/FASerializer.java # edit 1364 Tweaked to use a field not method. 1365 1366 src/org/antlr/tool/Grammar.java # edit 1367 This is where most of the meat is for the grammar import 1368 stuff as you can imagine. I factored out the token type 1369 tracking into the CompositeGrammar object. I added code to 1370 the addArtificialMatchTokensRule method so that it includes 1371 references to all delegate lexer Tokens rules. Altered the 1372 rule lookup stuff so that it knows about delegate grammars. 1373 1374 src/org/antlr/tool/GrammarAST.java # edit 1375 src/org/antlr/tool/GrammarAnalysisAbortedMessage.java # edit 1376 src/org/antlr/tool/GrammarReport.java # edit 1377 src/org/antlr/tool/NonRegularDecisionMessage.java # edit 1378 Made enclosing rule visible as field. 1379 1380 src/org/antlr/tool/GrammarSanity.java # edit 1381 General cleanup and addition of generics. 1382 1383 src/org/antlr/tool/Interpreter.java # edit 1384 Reference fields instead of methods. 1385 1386 src/org/antlr/tool/NFAFactory.java # edit 1387 General cleanup and use of Rule object instead of rule 1388 index. 1389 1390 src/org/antlr/tool/NameSpaceChecker.java # edit 1391 A little bit of cleanup and changes to use either the local 1392 or globally visible rule. Added code to check that scopes 1393 are valid on scoped rule references. again this is an 1394 internal feature, not to be used by users. 1395 1396 src/org/antlr/tool/RandomPhrase.java # edit 1397 Tweaked. 1398 1399 src/org/antlr/tool/Rule.java # edit 1400 Added field imported. Removed some unused methods by 1401 commenting them out. Made toString() more expressive. 1402 1403 src/org/antlr/tool/antlr.g # edit 1404 src/org/antlr/tool/antlr.print.g # edit 1405 src/org/antlr/tool/assign.types.g # edit 1406 src/org/antlr/tool/buildnfa.g # edit 1407 src/org/antlr/tool/define.g # edit 1408 Added syntax for import statement. assign.types.g is the 1409 grammar that invokes Grammar.importGrammar(). 1410 1411 src/org/antlr/tool/templates/messages/languages/en.stg # edit 1412 Added error messages. 1413 1414 Added 1415 1416 CHANGES.txt 1417 runtime/Java/src/org/antlr/runtime/RecognizerSharedState.java 1418 src/org/antlr/test/TestCompositeGrammars.java 1419 src/org/antlr/tool/CompositeGrammar.java 1420 src/org/antlr/tool/CompositeGrammarTree.java 1421 14223.0.1 - August 13, 2007 1423 1424[See target pages on the wiki for more information on the non-Java targets] 1425 1426August 7, 2007 1427 1428* added escaping of double quotes in DOTTreeGenerator 1429 1430July 22, 2007 1431 1432* fixed dynamic scope implementation in lexers. They were not creating new scope 1433 entries on the stack. Unsupported feature! 1434 1435July 30, 2007 1436 1437* float return values were initalized to 0.0 not 0.0f in java. 1438 1439July 28, 2007 1440 1441* Sam Ellis points out an init var bug in ANTLRReaderStream. 1442 1443July 27, 2007 (done in dev branch) 1444 1445* Moved token type index stuff from CommonTreeNodeStream to TreeWizard 1446 1447* Added getChildren to BaseTree. 1448 1449* Added heterogeneous tree functionality; rewrite for parser/tree parser 1450 and auto AST constr. for parser. 1451 1452 org/antlr/runtime/tree/RewriteRuleElementStream.java 1453 org/antlr/runtime/tree/RewriteRuleNodeStream.java 1454 org/antlr/runtime/tree/RewriteRuleTokenStream.java 1455 Renamed method next() and re-factor things to have more 1456 specific methods: nextToken, nextNode, nextTree. 1457 1458 codegen/codegen.g 1459 Updated to include new <NodeType> AST structure for 1460 token references. Pushed hereto attribute into 1461 all tokenRef* templates. 1462 codegen/templates/Java/AST.stg 1463 Factored out a few templates: 1464 createImaginaryNode(tokenType,hetero,args) 1465 createRewriteNodeFromElement(token,hetero,args) 1466 Converted a lot of stream next() calls to more specific 1467 nextToken, nextNode, nextTree per above. 1468 codegen/templates/Java/ASTParser.stg 1469 Added createNodeFromToken template and re-factored creation 1470 sites to use that template. Added hetero attribute. 1471 codegen/templates/Java/ASTTreeParser.stg 1472 Added createRewriteNodeFromElement template and re-factored. 1473 1474 test/TestHeteroAST.java 1475 New file. Unit tests to test new hetero tree construction. 1476 test/TestRewriteAST.java 1477 Fixed test. Nil single-node trees no longer return nil; 1478 They return null. 1479 1480 tool/ErrorManager.java 1481 tool/templates/messages/languages/en.stg 1482 Added error message: 1483 HETERO_ILLEGAL_IN_REWRITE_ALT(arg) ::= 1484 "alts with rewrites can't use heterogeneous types left of ->" 1485 1486 tool/antlr.g 1487 tool/antlr.print.g 1488 tool/assign.types.g 1489 tool/buildnfa.g 1490 tool/define.g 1491 Added syntax for <NodeType> to token references. 1492 Altered AST structure rippled through different phases. 1493 1494July 24, 2007 1495 1496* Deleted DoubleLinkTree.java; CommonTree does that now. 1497 1498July 23, 2007 1499 1500* template group outputFile; changed rewrite arg to rewriteMode. 1501 1502* added rewrite mode for tree parser build AST. 1503 1504July 22, 2007 1505 1506* Kay fixed dynamic scope implementation in lexers. They were not 1507 creating new scope entries on the stack. This is an UNSUPPORTED feature. 1508 1509* added getParent and getChildIndex to TreeAdaptor. Added 1510 implementation to CommonTree. It's just too useful having those 1511 parent and child indexes available for rewriting etc... I tried 2x 1512 to make an implementation of tree rewriting w/o this and the 1513 constraints just made it too expensive and complicated. Have to 1514 update adaptors to set parent, child index values. Updated Tree 1515 interface and BaseTree also. Should only affect target developers 1516 not users. Well, unless they impl Tree. 1517 1518* dupNode (via ctor) of CommonTree didn't copy start/stop token indexes. 1519 1520TARGET DEVELOPERS WARNING -- AST.stg split with some functionality 1521 going into ASTParser.stg then I added 1522 ASTTreeParser.stg. CodeGenerator 1523 assumes new subgroups exist. 1524 1525July 20, 2007 1526 1527* Added AST construction for tree parsers including -> rewrite rules. 1528 Rewrite mode (rewrite=true) alters the tree in place rather than 1529 constructing a whole new tree. Implementation notes: 1530 1531 org/antlr/runtime/tree/Tree.java 1532 Add methods for parent and child index functionality. 1533 Also added freshenParentAndChildIndexes() which you can use 1534 to ensure that all double linking is set up right after you 1535 manipulate the tree manually. The setChild preteens etc. do 1536 the proper thing so you shouldn't need this. 1537 Added replaceChildren() to support tree rewrite mode in tree parsers 1538 org/antlr/runtime/tree/BaseTree.java 1539 Updated to set parent and child index stuff. Added replaceChildren 1540 method etc... It still only has a list of children as sole field 1541 but calls methods that subclasses can choose to implement such as 1542 CommonTree. 1543 org/antlr/runtime/tree/CommonTree.java 1544 Added parent and childIndex fields to doubly link. 1545 org/antlr/runtime/tree/TreeAdaptor.java 1546 Added methods for new parent and child index functionality. 1547 Also added method for rewrite mode in tree parsers: 1548 replaceChildren(Object parent, int startChildIndex, 1549 int stopChildIndex, Object t); 1550 Added setChild and deleteChild methods 1551 org/antlr/runtime/tree/BaseTreeAdaptor.java 1552 Moved dupTree here from BaseTree. 1553 Updated rulePostProcessing to deal with parent and child index. 1554 Added setChild and deleteChild implementations 1555 org/antlr/runtime/tree/CommonTreeAdaptor.java 1556 Added methods to deal with the parent and child index for a node. 1557 1558 org/antlr/runtime/tree/CommonTreeNodeStream.java 1559 Removed token type index and method fillReverseIndex etc... 1560 Probably will move into the tree wizard in the future. 1561 Changed call/seek stack implementation to use IntArray 1562 Added replaceChildren interface. 1563 org/antlr/runtime/tree/TreeNodeStream.java 1564 Added replaceChildren. 1565 org/antlr/runtime/tree/UnBufferedTreeNodeStream.java 1566 Added replaceChildren method but no implementation 1567 1568 codegen/templates/ANTLRCore.sti 1569 Changed rewrite to a better name: rewriteMode 1570 Added tree level argument to alt, tree so that auto AST 1571 construction can occur while recognizing in tree parsers. 1572 1573 codegen/templates/Java/AST.stg 1574 Split template group: added two subclasses to handle different 1575 functionality for normal parsing and tree parsing + AST 1576 construction. Tree parsers default behavior is to dup tree 1577 not construct another. Added ASTParser.stg and 1578 ASTTreeParser.stg to handle auto AST construction during 1579 recognition for the two different parser types. I just copied 1580 the token, rule, set, wildcard templates to the subclasses. 1581 The rewrite templates are still in AST.stg. I factored out the 1582 node creation so that the same rewrite templates can be used 1583 for both parsing and tree parsing. 1584 1585 codegen/templates/Java/ASTParser.stg 1586 The templates needed to build trees with auto construction 1587 during parsing. 1588 codegen/templates/Java/ASTTreeParser.stg 1589 The templates needed to build trees with auto construction 1590 during tree parsing. 1591 codegen/templates/Java/Java.stg 1592 genericParser now has rewriteElementType (Note or Token) so 1593 that the rewrite streams know what kind of elements are inside 1594 during rewrite rule tree construction. 1595 codegen/templates/Java/ST.stg 1596 rewrite attribute name change to rewriteMode 1597 1598 org/antlr/runtime/debug/DebugTreeAdaptor.java 1599 org/antlr/runtime/debug/DebugTreeNodeStream.java 1600 Updated to handle new interfaces 1601 1602 test/BaseTest.java 1603 Added test rig update to handle AST construction by tree parsers. 1604 All tree construction runs automatically test sanity of parent 1605 and child indexes. 1606 test/TestTreeGrammarRewriteAST.java 1607 test/TestTreeNodeStream.java 1608 test/TestTrees.java 1609 new file; tests the new parent and child index stuff in trees. 1610 1611July 19, 2007 1612 1613* implemented new unique ID; GC was causing non unique hash codes. Debugging 1614 tree grammars was messing up. 1615 1616* got tree rewrites working in tree grammars. It builds a completely new 1617 tree from old tree; i.e., you get two trees in memory. W/o a rewrite 1618 rule, the input for that rule is duplicated and returned. -> w/o elements 1619 to the right means don't return anything; i.e., delete. Ooops...way 1620 harder than I thought. Real implementation notes above. 1621 1622INCOMPATIBILITY WARNING -- templates have changed; must regen output from 1623 grammars. Runtime libraries have also changed. 1624 Debug event listener interface has changed also. 1625 1626July 17, 2007 1627 1628* Added line/charposition to node socket events and event dump so 1629 we have more info during tree parsing. Only works if your 1630 tree adaptor returns a value Token object from getToken(treenode) 1631 with line/col set. Refactored consumeNode/LN to use deserializeNode(). 1632 1633* Fixed mismatched tree node exceptions; for imaginary nodes, it said 1634 "missing null". Now prints the token type we found. 1635 1636* Cleaned up exception stuff. MismatchedTreeNodeException was setting 1637 line/col, but only RecognitionException should do that. 1638 1639* If imaginary token gets a mismatch, there is no line info. Search 1640 backwards in stream if input node stream supports to find last 1641 node with good line/col info. E.g., 1642 1643ANTLRv3Tree.g: node from after line 156:72 mismatched tree node: EOA expecting <UP> 1644 1645 which used to be: 1646 1647ANTLRv3Tree.g: node from line 0:0 mismatched tree node: null expecting <UP> 1648 1649* mismatched tree node exceptions were not sent to the debug event stream. 1650 Due to a type being slightly different on recoverFromMismatchedToken() 1651 in DebugTreeParser. Was calling BaseRecognizer version not subclass. 1652 Now we get: 1653 1654 9459: Recognition exception MismatchedTreeNodeException(0!=0) 1655 1656* List labels were not allowed as root nodes in tree rewrites like 1657 ^($listlabel ...). Had to add a template to AST.stg: 1658 1659 /** Gen ^($label ...) where label+=... */ 1660 rewriteTokenListLabelRefRoot ::= rewriteTokenLabelRefRoot 1661 1662 1663July 16, 2007 1664 1665* fixed nextNode in RewriteRuleSubtreeStream was dup'ing too much, 1666 screwing up debug event stream. Also there was a bug in how 1667 the rewrite tree stream stuff decided to dup nodes. 1668 1669* fixed bug in LT for tree parsing; text was not transmitted properly; 1670 only single words worked. 1671 1672* made decision for rule put line/col on colon not first token of first alt. 1673 1674* remote ProxyToken now emits token index for easier debugging when looking 1675 at AW's event stream. For example, the @5 here is the token index: 1676 1677 31 Consume hidden [ /<64>,channel=99,30:7, @5] 1678 1679* same is true for consume nodes now: 1680 1681 25586 Consume node [')'/, <44>, 4712040,@1749] 25 1682 1683 When debugging tree parsers, it helps to track errors when you know 1684 what corresponding input symbol created this tree node. 1685 1686* Changed debug events associated with trees quite a bit. Passes nodes around 1687 now rather than text, type, unique IDs etc... Mostly affects internal stuff. 1688 Target developers will have some work in their runtime to do to match 1689 this change. :( BUT, there is only a slight tweak in the Dbg.stg 1690 and ASTDbg.stg templates. 1691 Interface just didn't make sense as is. If you turn on debugging, and 1692 want to track a node creation, you want the node pointer not its ID, 1693 text, etc... 1694 Added ProxyTree for passing across socket. Has line/charpos and tokenIndex 1695 1696July 15, 2007 1697 1698* added null ptr protection in CommonTreeAdaptor. 1699 1700July 14, 2007 1701 1702* null child in TreeAdaptor does nothing now. Changed interface and 1703 implementation. Changed DebugTreeAdaptor to not fire events on null add 1704 as well. 1705 1706July 12, 2007 1707 1708* added get method for the line/col to DFA map in Grammar.java 1709 1710July 7, 2007 1711 1712* fixed wrong order of test for exceptions in Lexer.getErrorMessage() 1713 1714June 28, 2007 1715 1716* Added ability to set the port number in the constructor for the debug parser. 1717 1718June 5, 2007 1719 1720* Changed (hidden) option -verbose to -Xnfastates; this just prints out the NFA states along each nondeterministic path for nondeterminism warnings. 1721 1722May 18, 2007 1723 1724* there were some dependencies with org.antlr.* that I removed from 1725 org.antlr.runtime.* 1726 17273.0 final - May 17, 2007 1728 1729May 14, 2007 1730 1731* Auto backtracking didn't work with ! and ^ suffixes on first element 1732 of an alt. 1733 1734* Auto backtracking didn't work with an action as first element. 1735 1736May 10, 2007 1737 1738* turn off the warning about no local messages: 1739 no such locale file org/antlr/tool/templates/messages/languages/ru.stg retrying with English locale 1740 1741May 5, 2007 1742 1743* moving org.antlr.runtime to runtime/Java/src/org/... Other target 1744 source / libs are under runtime/targetname. 1745 1746May 4, 2007 1747 1748* You could not use arguments on a token reference that was a route in a 1749 tree rewrite rule like -> ^(ID[args] ...). 1750 1751May 3, 2007 1752 1753* Fixed ANTLR-82. Actions after the root were considered part of 1754 an optional child. They were not always executed. Required a change 1755 to the ANTLRCore.sti interface for tree() template. 1756 1757May 2, 2007 1758 1759* Fixed ANTLR-117. Wasn't building decisions properly for subrules in 1760 syntactic predicates. 1761 1762April 22, 2007 1763 1764* Made build.xml ref all jars in antlr lib. Thanks to Miguel Ping. 1765 1766* Fixed ANTLR-11 1767 1768* Now labels on ranges and such in lexer work properly. 1769 1770* ActionAnalysisLexer was in wrong package. 1771 1772April 21, 2007 1773 1774* Pushing a huge update that fixes: 1775 http://www.antlr.org/browse/ANTLR-112 1776 http://www.antlr.org/browse/ANTLR-110 1777 http://www.antlr.org/browse/ANTLR-109 1778 http://www.antlr.org/browse/ANTLR-103 1779 http://www.antlr.org/browse/ANTLR-97 1780 http://www.antlr.org/browse/ANTLR-113 1781 http://www.antlr.org/browse/ANTLR-66 1782 http://www.antlr.org/browse/ANTLR-98 1783 http://www.antlr.org/browse/ANTLR-24 1784 http://www.antlr.org/browse/ANTLR-114 1785 http://www.antlr.org/browse/ANTLR-5 1786 http://www.antlr.org/browse/ANTLR-6 1787 1788 Basically, I gutted the way AST rewrites work. MUCH better. 1789 1790* Fixed lots of little label issues in the lexer. Couldn't do x+=ID 1791 in lexer, for example. Fixed ANTLR-114, ANTLR-112 1792 1793* Isolated EOT transition in lexer generated dangling else clause. 1794 Fixed ANTLR-113. 1795 1796April 17, 2007 1797 1798* Fixed a major problem with gated semantic predicates. Added more 1799 unit tests. 1800 1801* Fixed bug in cyclic DFA with syntactic predicates. Wasn't rewinding 1802 properly. Further, mark() in token stream did not fill buffer so 1803 when you rewound back to last marker index was -1 not 0. At same time 1804 I fixed ANTLR-103. Syn preds evaluated only once now. 1805 1806* Altered code gen file writing so it writes directly to a file 1807 instead of building a big string and then writing that out. Should 1808 be faster and much less memory intensive. 1809 1810* Fixed so antlr writes files to correct location again. See: 1811 1812http://www.antlr.org/wiki/pages/viewpage.action?pageId=1862 1813 18143.0b7 - April 12, 2007 1815 1816April 10, 2007 1817 1818* Allows -> {...} actions now when building ASTs. Fixed ANTLR-14. 1819 1820* Allows ! on sets and wildcard now during output=AST option. Fixed ANTLR-17. 1821 1822* Fixed ANTLR-92 bug. Couldn't use sets with -> tree construction. 1823 1824* No lexer rule for a token type is now a warning. 1825 1826* Fixed set labels in lexer; ANTLR-60 bug 1827 1828* Fixed problem with duplicate state variable definitions in switch-case 1829 1830April 9, 2007 1831 1832* Gated predicates didn't work properly in cyclic DFA. 1833 1834April 7, 2007 1835 1836* Couldn't have more than one set per rule it seems. Fixed. 1837 1838April 3, 2007 1839 1840* Fix a problem in my unused label optimization. Added new 1841 pass over actions to examine them. 1842 1843* RuleReturnScope has method back: 1844 /** Has a value potentially if output=template; Don't use StringTemplate 1845 * type as it then causes a dependency with ST lib. 1846 */ 1847 public Object getTemplate() { return null; } 1848 1849March 30, 2007 1850 1851* Fixed ANTLR-8. Labels to rules w/o return values caused compile errors. 1852 1853* Fixed ANTLR-89; semantic predicates in lexer sometimes 1854 caused exception in code gen. 1855 1856* Fixed ANTLR-36; remove runtime dependency with ST 1857 1858March 29, 2007 1859 1860* Over last few days, I've gutted how ANTLR handles sets of chars or 1861 tokens. I cleaned up a lot of stuff in the grammars and added lots 1862 of unit tests. 1863 1864March 26, 2007 1865 1866* CommonTreeNodeStream didn't push correctly; couldn't handle very 1867 deeply nested trees. 1868 1869* Fixed bug that E : 'a' 'b' ; made E be seen as an alias of 'a'. 1870 1871March 22, 2007 1872 1873* Working with Egor Ushakov from Sun Optimization / NetBeans team I 1874 made all the Java lexer transition tables static w/o screwing up 1875 ability to reference semantic predicates etc... Only changed Java.stg 1876 1877* cached text string in CommonToken.getText(); saves on repeated calls; 1878 Java mode. 1879 1880* made all generated methods final; saves a few percent speed according to 1881 Egor Ushakov (Java only). 1882 1883* removed most assignments from each lexer rule and even the Lexer.emit() 1884 call! All done in nextToken now. Saves on code gen size and a wee bit of 1885 execution speed probably. Variables became fields: type, channel, line, 1886 etc... Now emit() needs no args even. Again, Egor helped on this. 1887 1888March 17, 2007 1889 1890* Jonathan DeKlotz updated C# templates to be 3.0b6 current 1891 1892March 14, 2007 1893 1894* Manually-specified (...)=> force backtracking eval of that predicate. 1895 backtracking=true mode does not however. Added unit test. 1896 1897March 14, 2007 1898 1899* Fixed bug in lexer where ~T didn't compute the set from rule T. 1900 1901* Added -Xnoinlinedfa make all DFA with tables; no inline prediction with IFs 1902 1903* Fixed http://www.antlr.org/browse/ANTLR-80. 1904 Sem pred states didn't define lookahead vars. 1905 1906* Fixed http://www.antlr.org/browse/ANTLR-91. 1907 When forcing some acyclic DFA to be state tables, they broke. 1908 Forcing all DFA to be state tables should give same results. 1909 1910March 12, 2007 1911 1912* setTokenSource in CommonTokenStream didn't clear tokens list. 1913 setCharStream calls reset in Lexer. 1914 1915* Altered -depend. No longer printing grammar files for multiple input 1916 files with -depend. Doesn't show T__.g temp file anymore. Added 1917 TLexer.tokens. Added .h files if defined. 1918 1919February 11, 2007 1920 1921* Added -depend command-line option that, instead of processing files, 1922 it shows you what files the input grammar(s) depend on and what files 1923 they generate. For combined grammar T.g: 1924 1925 $ java org.antlr.Tool -depend T.g 1926 1927 You get: 1928 1929 TParser.java : T.g 1930 T.tokens : T.g 1931 T__.g : T.g 1932 1933 Now, assuming U.g is a tree grammar ref'd T's tokens: 1934 1935 $ java org.antlr.Tool -depend T.g U.g 1936 1937 TParser.java : T.g 1938 T.tokens : T.g 1939 T__.g : T.g 1940 U.g: T.tokens 1941 U.java : U.g 1942 U.tokens : U.g 1943 1944 Handles spaces by escaping them. Pays attention to -o, -fo and -lib. 1945 Dir 'x y' is a valid dir in current dir. 1946 1947 $ java org.antlr.Tool -depend -lib /usr/local/lib -o 'x y' T.g U.g 1948 x\ y/TParser.java : T.g 1949 x\ y/T.tokens : T.g 1950 x\ y/T__.g : T.g 1951 U.g: /usr/local/lib/T.tokens 1952 x\ y/U.java : U.g 1953 x\ y/U.tokens : U.g 1954 1955 You have API access via org.antlr.tool.BuildDependencyGenerator class: 1956 getGeneratedFileList(), getDependenciesFileList(). You can also access 1957 the output template: getDependencies(). The file 1958 org/antlr/tool/templates/depend.stg contains the template. You can 1959 modify as you want. File objects go in so you can play with path etc... 1960 1961February 10, 2007 1962 1963* no more .gl files generated. All .g all the time. 1964 1965* changed @finally to be @after and added a finally clause to the 1966 exception stuff. I also removed the superfluous "exception" 1967 keyword. Here's what the new syntax looks like: 1968 1969 a 1970 @after { System.out.println("ick"); } 1971 : 'a' 1972 ; 1973 catch[RecognitionException e] { System.out.println("foo"); } 1974 catch[IOException e] { System.out.println("io"); } 1975 finally { System.out.println("foobar"); } 1976 1977 @after executes after bookkeeping to set $rule.stop, $rule.tree but 1978 before scopes pop and any memoization happens. Dynamic scopes and 1979 memoization are still in generated finally block because they must 1980 exec even if error in rule. The @after action and tree setting 1981 stuff can technically be skipped upon syntax error in rule. [Later 1982 we might add something to finally to stick an ERROR token in the 1983 tree and set the return value.] Sequence goes: set $stop, $tree (if 1984 any), @after (if any), pop scopes (if any), memoize (if needed), 1985 grammar finally clause. Last 3 are in generated code's finally 1986 clause. 1987 19883.0b6 - January 31, 2007 1989 1990January 30, 2007 1991 1992* Fixed bug in IntervalSet.and: it returned the same empty set all the time 1993 rather than new empty set. Code altered the same empty set. 1994 1995* Made analysis terminate faster upon a decision that takes too long; 1996 it seemed to keep doing work for a while. Refactored some names 1997 and updated comments. Also made it terminate when it realizes it's 1998 non-LL(*) due to recursion. just added terminate conditions to loop 1999 in convert(). 2000 2001* Sometimes fatal non-LL(*) messages didn't appear; instead you got 2002 "antlr couldn't analyze", which is actually untrue. I had the 2003 order of some prints wrong in the DecisionProbe. 2004 2005* The code generator incorrectly detected when it could use a fixed, 2006 acyclic inline DFA (i.e., using an IF). Upon non-LL(*) decisions 2007 with predicates, analysis made cyclic DFA. But this stops 2008 the computation detecting whether they are cyclic. I just added 2009 a protection in front of the acyclic DFA generator to avoid if 2010 non-LL(*). Updated comments. 2011 2012January 23, 2007 2013 2014* Made tree node streams use adaptor to create navigation nodes. 2015 Thanks to Emond Papegaaij. 2016 2017January 22, 2007 2018 2019* Added lexer rule properties: start, stop 2020 2021January 1, 2007 2022 2023* analysis failsafe is back on; if a decision takes too long, it bails out 2024 and uses k=1 2025 2026January 1, 2007 2027 2028* += labels for rules only work for output option; previously elements 2029 of list were the return value structs, but are now either the tree or 2030 StringTemplate return value. You can label different rules now 2031 x+=a x+=b. 2032 2033December 30, 2006 2034 2035* Allow \" to work correctly in "..." template. 2036 2037December 28, 2006 2038 2039* errors that are now warnings: missing AST label type in trees. 2040 Also "no start rule detected" is warning. 2041 2042* tree grammars also can do rewrite=true for output=template. 2043 Only works for alts with single node or tree as alt elements. 2044 If you are going to use $text in a tree grammar or do rewrite=true 2045 for templates, you must use in your main: 2046 2047 nodes.setTokenStream(tokens); 2048 2049* You get a warning for tree grammars that do rewrite=true and 2050 output=template and have -> for alts that are not simple nodes 2051 or simple trees. new unit tests in TestRewriteTemplates at end. 2052 2053December 27, 2006 2054 2055* Error message appears when you use -> in tree grammar with 2056 output=template and rewrite=true for alt that is not simple 2057 node or tree ref. 2058 2059* no more $stop attribute for tree parsers; meaningless/useless. 2060 Removed from TreeRuleReturnScope also. 2061 2062* rule text attribute in tree parser must pull from token buffer. 2063 Makes no sense otherwise. added getTokenStream to TreeNodeStream 2064 so rule $text attr works. CommonTreeNodeStream etc... now let 2065 you set the token stream so you can access later from tree parser. 2066 $text is not well-defined for rules like 2067 2068 slist : stat+ ; 2069 2070 because stat is not a single node nor rooted with a single node. 2071 $slist.text will get only first stat. I need to add a warning about 2072 this... 2073 2074* Fixed http://www.antlr.org/browse/ANTLR-76 for Java. 2075 Enhanced TokenRewriteStream so it accepts any object; converts 2076 to string at last second. Allows you to rewrite with StringTemplate 2077 templates now :) 2078 2079* added rewrite option that makes -> template rewrites do replace ops for 2080 TokenRewriteStream input stream. In output=template and rewrite=true mode 2081 same as before 'cept that the parser does 2082 2083 ((TokenRewriteStream)input).replace( 2084 ((Token)retval.start).getTokenIndex(), 2085 input.LT(-1).getTokenIndex(), 2086 retval.st); 2087 2088 after each rewrite so that the input stream is altered. Later refs to 2089 $text will have rewrites. Here's a sample test program for grammar Rew. 2090 2091 FileReader groupFileR = new FileReader("Rew.stg"); 2092 StringTemplateGroup templates = new StringTemplateGroup(groupFileR); 2093 ANTLRInputStream input = new ANTLRInputStream(System.in); 2094 RewLexer lexer = new RewLexer(input); 2095 TokenRewriteStream tokens = new TokenRewriteStream(lexer); 2096 RewParser parser = new RewParser(tokens); 2097 parser.setTemplateLib(templates); 2098 parser.program(); 2099 System.out.println(tokens.toString()); 2100 groupFileR.close(); 2101 2102December 26, 2006 2103 2104* BaseTree.dupTree didn't dup recursively. 2105 2106December 24, 2006 2107 2108* Cleaned up some comments and removed field treeNode 2109 from MismatchedTreeNodeException class. It is "node" in 2110 RecognitionException. 2111 2112* Changed type from Object to BitSet for expecting fields in 2113 MismatchedSetException and MismatchedNotSetException 2114 2115* Cleaned up error printing in lexers and the messages that it creates. 2116 2117* Added this to TreeAdaptor: 2118 /** Return the token object from which this node was created. 2119 * Currently used only for printing an error message. 2120 * The error display routine in BaseRecognizer needs to 2121 * display where the input the error occurred. If your 2122 * tree of limitation does not store information that can 2123 * lead you to the token, you can create a token filled with 2124 * the appropriate information and pass that back. See 2125 * BaseRecognizer.getErrorMessage(). 2126 */ 2127 public Token getToken(Object t); 2128 2129December 23, 2006 2130 2131* made BaseRecognizer.displayRecognitionError nonstatic so people can 2132 override it. Not sure why it was static before. 2133 2134* Removed state/decision message that comes out of no 2135 viable alternative exceptions, as that was too much. 2136 removed the decision number from the early exit exception 2137 also. During development, you can simply override 2138 displayRecognitionError from BaseRecognizer to add the stuff 2139 back in if you want. 2140 2141* made output go to an output method you can override: emitErrorMessage() 2142 2143* general cleanup of the error emitting code in BaseRecognizer. Lots 2144 more stuff you can override: getErrorHeader, getTokenErrorDisplay, 2145 emitErrorMessage, getErrorMessage. 2146 2147December 22, 2006 2148 2149* Altered Tree.Parser.matchAny() so that it skips entire trees if 2150 node has children otherwise skips one node. Now this works to 2151 skip entire body of function if single-rooted subtree: 2152 ^(FUNC name=ID arg=ID .) 2153 2154* Added "reverse index" from node to stream index. Override 2155 fillReverseIndex() in CommonTreeNodeStream if you want to change. 2156 Use getNodeIndex(node) to find stream index for a specific tree node. 2157 See getNodeIndex(), reverseIndex(Set tokenTypes), 2158 reverseIndex(int tokenType), fillReverseIndex(). The indexing 2159 costs time and memory to fill, but pulling stuff out will be lots 2160 faster as it can jump from a node ptr straight to a stream index. 2161 2162* Added TreeNodeStream.get(index) to make it easier for interpreters to 2163 jump around in tree node stream. 2164 2165* New CommonTreeNodeStream buffers all nodes in stream for fast jumping 2166 around. It now has push/pop methods to invoke other locations in 2167 the stream for building interpreters. 2168 2169* Moved CommonTreeNodeStream to UnBufferedTreeNodeStream and removed 2170 Iterator implementation. moved toNodesOnlyString() to TestTreeNodeStream 2171 2172* [BREAKS ANY TREE IMPLEMENTATION] 2173 made CommonTreeNodeStream work with any tree node type. TreeAdaptor 2174 now implements isNil so must add; trivial, but does break back 2175 compatibility. 2176 2177December 17, 2006 2178 2179* Added traceIn/Out methods to recognizers so that you can override them; 2180 previously they were in-line print statements. The message has also 2181 been slightly improved. 2182 2183* Factored BuildParseTree into debug package; cleaned stuff up. Fixed 2184 unit tests. 2185 2186December 15, 2006 2187 2188* [BREAKS ANY TREE IMPLEMENTATION] 2189 org.antlr.runtime.tree.Tree; needed to add get/set for token start/stop 2190 index so CommonTreeAdaptor can assume Tree interface not CommonTree 2191 implementation. Otherwise, no way to create your own nodes that satisfy 2192 Tree because CommonTreeAdaptor was doing 2193 2194 public int getTokenStartIndex(Object t) { 2195 return ((CommonTree)t).startIndex; 2196 } 2197 2198 Added to Tree: 2199 2200 /** What is the smallest token index (indexing from 0) for this node 2201 * and its children? 2202 */ 2203 int getTokenStartIndex(); 2204 2205 void setTokenStartIndex(int index); 2206 2207 /** What is the largest token index (indexing from 0) for this node 2208 * and its children? 2209 */ 2210 int getTokenStopIndex(); 2211 2212 void setTokenStopIndex(int index); 2213 2214December 13, 2006 2215 2216* Added org.antlr.runtime.tree.DOTTreeGenerator so you can generate DOT 2217 diagrams easily from trees. 2218 2219 CharStream input = new ANTLRInputStream(System.in); 2220 TLexer lex = new TLexer(input); 2221 CommonTokenStream tokens = new CommonTokenStream(lex); 2222 TParser parser = new TParser(tokens); 2223 TParser.e_return r = parser.e(); 2224 Tree t = (Tree)r.tree; 2225 System.out.println(t.toStringTree()); 2226 DOTTreeGenerator gen = new DOTTreeGenerator(); 2227 StringTemplate st = gen.toDOT(t); 2228 System.out.println(st); 2229 2230* Changed the way mark()/rewind() work in CommonTreeNode stream to mirror 2231 more flexible solution in ANTLRStringStream. Forgot to set lastMarker 2232 anyway. Now you can rewind to non-most-recent marker. 2233 2234December 12, 2006 2235 2236* Temp lexer now end in .gl (T__.gl, for example) 2237 2238* TreeParser suffix no longer generated for tree grammars 2239 2240* Defined reset for lexer, parser, tree parser; rewinds the input stream also 2241 2242December 10, 2006 2243 2244* Made Grammar.abortNFAToDFAConversion() abort in middle of a DFA. 2245 2246December 9, 2006 2247 2248* fixed bug in OrderedHashSet.add(). It didn't track elements correctly. 2249 2250December 6, 2006 2251 2252* updated build.xml for future Ant compatibility, thanks to Matt Benson. 2253 2254* various tests in TestRewriteTemplate and TestSyntacticPredicateEvaluation 2255 were using the old 'channel' vs. new '$channel' notation. 2256 TestInterpretedParsing didn't pick up an earlier change to CommonToken. 2257 Reported by Matt Benson. 2258 2259* fixed platform dependent test failures in TestTemplates, supplied by Matt 2260 Benson. 2261 2262November 29, 2006 2263 2264* optimized semantic predicate evaluation so that p||!p yields true. 2265 2266November 22, 2006 2267 2268* fixed bug that prevented var = $rule.some_retval from working in anything 2269 but the first alternative of a rule or subrule. 2270 2271* attribute names containing digits were not allowed, this is now fixed, 2272 allowing attributes like 'name1' but not '1name1'. 2273 2274November 19, 2006 2275 2276* Removed LeftRecursionMessage and apparatus because it seems that I check 2277 for left recursion upfront before analysis and everything gets specified as 2278 recursion cycles at this point. 2279 2280November 16, 2006 2281 2282* TokenRewriteStream.replace was not passing programName to next method. 2283 2284November 15, 2006 2285 2286* updated DOT files for DFA generation to make smaller circles. 2287 2288* made epsilon edges italics in the NFA diagrams. 2289 22903.0b5 - November 15, 2006 2291 2292The biggest thing is that your grammar file names must match the grammar name 2293inside (your generated class names will also be different) and we use 2294$channel=HIDDEN now instead of channel=99 inside lexer actions. 2295Should be compatible other than that. Please look at complete list of 2296changes. 2297 2298November 14, 2006 2299 2300* Force token index to be -1 for CommonIndex in case not set. 2301 2302November 11, 2006 2303 2304* getUniqueID for TreeAdaptor now uses identityHashCode instead of hashCode. 2305 2306November 10, 2006 2307 2308* No grammar nondeterminism warning now when wildcard '.' is final alt. 2309 Examples: 2310 2311 a : A | B | . ; 2312 2313 A : 'a' 2314 | . 2315 ; 2316 2317 SL_COMMENT 2318 : '//' (options {greedy=false;} : .)* '\r'? '\n' 2319 ; 2320 2321 SL_COMMENT2 2322 : '//' (options {greedy=false;} : 'x'|.)* '\r'? '\n' 2323 ; 2324 2325 2326November 8, 2006 2327 2328* Syntactic predicates did not get hoisting properly upon non-LL(*) decision. Other hoisting issues fixed. Cleaned up code. 2329 2330* Removed failsafe that check to see if I'm spending too much time on a single DFA; I don't think we need it anymore. 2331 2332November 3, 2006 2333 2334* $text, $line, etc... were not working in assignments. Fixed and added 2335 test case. 2336 2337* $label.text translated to label.getText in lexer even if label was on a char 2338 2339November 2, 2006 2340 2341* Added error if you don't specify what the AST type is; actions in tree 2342 grammar won't work without it. 2343 2344 $ cat x.g 2345 tree grammar x; 2346 a : ID {String s = $ID.text;} ; 2347 2348 ANTLR Parser Generator Early Access Version 3.0b5 (??, 2006) 1989-2006 2349 error: x.g:0:0: (152) tree grammar x has no ASTLabelType option 2350 2351November 1, 2006 2352 2353* $text, $line, etc... were not working properly within lexer rule. 2354 2355October 32, 2006 2356 2357* Finally actions now execute before dynamic scopes are popped it in the 2358 rule. Previously was not possible to access the rules scoped variables 2359 in a finally action. 2360 2361October 29, 2006 2362 2363* Altered ActionTranslator to emit errors on setting read-only attributes 2364 such as $start, $stop, $text in a rule. Also forbid setting any attributes 2365 in rules/tokens referenced by a label or name. 2366 Setting dynamic scopes's attributes and your own parameter attributes 2367 is legal. 2368 2369October 27, 2006 2370 2371* Altered how ANTLR figures out what decision is associated with which 2372 block of grammar. Makes ANTLRWorks correctly find DFA for a block. 2373 2374October 26, 2006 2375 2376* Fixed bug where EOT transitions led to no NFA configs in a DFA state, 2377 yielding an error in DFA table generation. 2378 2379* renamed action.g to ActionTranslator.g 2380 the ActionTranslator class is now called ActionTranslatorLexer, as ANTLR 2381 generates this classname now. Fixed rest of codebase accordingly. 2382 2383* added rules recognizing setting of scopes' attributes to ActionTranslator.g 2384 the Objective C target needed access to the right-hand side of the assignment 2385 in order to generate correct code 2386 2387* changed ANTLRCore.sti to reflect the new mandatory templates to support the above 2388 namely: scopeSetAttributeRef, returnSetAttributeRef and the ruleSetPropertyRef_* 2389 templates, with the exception of ruleSetPropertyRef_text. we cannot set this attribute 2390 2391October 19, 2006 2392 2393* Fixed 2 bugs in DFA conversion that caused exceptions. 2394 altered functionality of getMinElement so it ignores elements<0. 2395 2396October 18, 2006 2397 2398* moved resetStateNumbersToBeContiguous() to after issuing of warnings; 2399 an internal error in that routine should make more sense as issues 2400 with decision will appear first. 2401 2402* fixed cut/paste bug I introduced when fixed EOF in min/max 2403 bug. Prevented C grammar from working briefly. 2404 2405October 17, 2006 2406 2407* Removed a failsafe that seems to be unnecessary that ensure DFA didn't 2408 get too big. It was resulting in some failures in code generation that 2409 led me on quite a strange debugging trip. 2410 2411October 16, 2006 2412 2413* Use channel=HIDDEN not channel=99 to put tokens on hidden channel. 2414 2415October 12, 2006 2416 2417* ANTLR now has a customizable message format for errors and warnings, 2418 to make it easier to fulfill requirements by IDEs and such. 2419 The format to be used can be specified via the '-message-format name' 2420 command line switch. The default for name is 'antlr', also available 2421 at the moment is 'gnu'. This is done via StringTemplate, for details 2422 on the requirements look in org/antlr/tool/templates/messages/formats/ 2423 2424* line numbers for lexers in combined grammars are now reported correctly. 2425 2426September 29, 2006 2427 2428* ANTLRReaderStream improperly checked for end of input. 2429 2430September 28, 2006 2431 2432* For ANTLRStringStream, LA(-1) was off by one...gave you LA(-2). 2433 24343.0b4 - August 24, 2006 2435 2436* error when no rules in grammar. doesn't crash now. 2437 2438* Token is now an interface. 2439 2440* remove dependence on non runtime classes in runtime package. 2441 2442* filename and grammar name must be same Foo in Foo.g. Generates FooParser, 2443 FooLexer, ... Combined grammar Foo generates Foo$Lexer.g which generates 2444 FooLexer.java. tree grammars generate FooTreeParser.java 2445 2446August 24, 2006 2447 2448* added C# target to lib, codegen, templates 2449 2450August 11, 2006 2451 2452* added tree arg to navigation methods in treeadaptor 2453 2454August 07, 2006 2455 2456* fixed bug related to (a|)+ on end of lexer rules. crashed instead 2457 of warning. 2458 2459* added warning that interpreter doesn't do synpreds yet 2460 2461* allow different source of classloader: 2462ClassLoader cl = Thread.currentThread().getContextClassLoader(); 2463if ( cl==null ) { 2464 cl = this.getClass().getClassLoader(); 2465} 2466 2467 2468July 26, 2006 2469 2470* compressed DFA edge tables significantly. All edge tables are 2471 unique. The transition table can reuse arrays. Look like this now: 2472 2473 public static readonly DFA30_transition0 = 2474 new short[] { 46, 46, -1, 46, 46, -1, -1, -1, -1, -1, -1, -1,...}; 2475 public static readonly DFA30_transition1 = 2476 new short[] { 21 }; 2477 public static readonly short[][] DFA30_transition = { 2478 DFA30_transition0, 2479 DFA30_transition0, 2480 DFA30_transition1, 2481 ... 2482 }; 2483 2484* If you defined both a label like EQ and '=', sometimes the '=' was 2485 used instead of the EQ label. 2486 2487* made headerFile template have same arg list as outputFile for consistency 2488 2489* outputFile, lexer, genericParser, parser, treeParser templates 2490 reference cyclicDFAs attribute which was no longer used after I 2491 started the new table-based DFA. I made cyclicDFADescriptors 2492 argument to outputFile and headerFile (only). I think this is 2493 correct as only OO languages will want the DFA in the recognizer. 2494 At the top level, C and friends can use it. Changed name to use 2495 cyclicDFAs again as it's a better name probably. Removed parameter 2496 from the lexer, ... For example, my parser template says this now: 2497 2498 <cyclicDFAs:cyclicDFA()> <! dump tables for all DFA !> 2499 2500* made all token ref token types go thru code gen's 2501 getTokenTypeAsTargetLabel() 2502 2503* no more computing DFA transition tables for acyclic DFA. 2504 2505July 25, 2006 2506 2507* fixed a place where I was adding syn predicates into rewrite stuff. 2508 2509* turned off invalid token index warning in AW support; had a problem. 2510 2511* bad location event generated with -debug for synpreds in autobacktrack mode. 2512 2513July 24, 2006 2514 2515* changed runtime.DFA so that it treats all chars and token types as 2516 char (unsigned 16 bit int). -1 becomes '\uFFFF' then or 65535. 2517 2518* changed MAX_STATE_TRANSITIONS_FOR_TABLE to be 65534 by default 2519 now. This means that all states can use a table to do transitions. 2520 2521* was not making synpreds on (C)* type loops with backtrack=true 2522 2523* was copying tree stuff and actions into synpreds with backtrack=true 2524 2525* was making synpreds on even single alt rules / blocks with backtrack=true 2526 25273.0b3 - July 21, 2006 2528 2529* ANTLR fails to analyze complex decisions much less frequently. It 2530 turns out that the set of decisions for which ANTLR fails (times 2531 out) is the same set (so far) of non-LL(*) decisions. Morever, I'm 2532 able to detect this situation quickly and report rather than timing 2533 out. Errors look like: 2534 2535 java.g:468:23: [fatal] rule concreteDimensions has non-LL(*) 2536 decision due to recursive rule invocations in alts 1,2. Resolve 2537 by left-factoring or using syntactic predicates with fixed k 2538 lookahead or use backtrack=true option. 2539 2540 This message only appears when k=*. 2541 2542* Shortened no viable alt messages to not include decision 2543 description: 2544 2545[compilationUnit, declaration]: line 8:8 decision=<<67:1: declaration 2546: ( ( fieldDeclaration )=> fieldDeclaration | ( methodDeclaration )=> 2547methodDeclaration | ( constructorDeclaration )=> 2548constructorDeclaration | ( classDeclaration )=> classDeclaration | ( 2549interfaceDeclaration )=> interfaceDeclaration | ( blockDeclaration )=> 2550blockDeclaration | emptyDeclaration );>> state 3 (decision=14) no 2551viable alt; token=[@1,184:187='java',<122>,8:8] 2552 2553 too long and hard to read. 2554 2555July 19, 2006 2556 2557* Code gen bug: states with no emanating edges were ignored by ST. 2558 Now an empty list is used. 2559 2560* Added grammar parameter to recognizer templates so they can access 2561 properties like getName(), ... 2562 2563July 10, 2006 2564 2565* Fixed the gated pred merged state bug. Added unit test. 2566 2567* added new method to Target: getTokenTypeAsTargetLabel() 2568 2569July 7, 2006 2570 2571* I was doing an AND instead of OR in the gated predicate stuff. 2572 Thanks to Stephen Kou! 2573 2574* Reduce op for combining predicates was insanely slow sometimes and 2575 didn't actually work well. Now it's fast and works. 2576 2577* There is a bug in merging of DFA stop states related to gated 2578 preds...turned it off for now. 2579 25803.0b2 - July 5, 2006 2581 2582July 5, 2006 2583 2584* token emission not properly protected in lexer filter mode. 2585 2586* EOT, EOT DFA state transition tables should be init'd to -1 (only 2587 was doing this for compressed tables). Fixed. 2588 2589* in trace mode, exit method not shown for memoized rules 2590 2591* added -Xmaxdfaedges to allow you to increase number of edges allowed 2592 for a single DFA state before it becomes "special" and can't fit in 2593 a simple table. 2594 2595* Bug in tables. Short are signed so min/max tables for DFA are now 2596 char[]. Bizarre. 2597 2598July 3, 2006 2599 2600* Added a method to reset the tool error state for current thread. 2601 See ErrorManager.java 2602 2603* [Got this working properly today] backtrack mode that let's you type 2604 in any old crap and ANTLR will backtrack if it can't figure out what 2605 you meant. No errors are reported by antlr during analysis. It 2606 implicitly adds a syn pred in front of every production, using them 2607 only if static grammar LL(*) analysis fails. Syn pred code is not 2608 generated if the pred is not used in a decision. 2609 2610 This is essentially a rapid prototyping mode. 2611 2612* Added backtracking report to the -report option 2613 2614* Added NFA->DFA conversion early termination report to the -report option 2615 2616* Added grammar level k and backtrack options to -report 2617 2618* Added a dozen unit tests to test autobacktrack NFA construction. 2619 2620* If you are using filter mode, you must manually use option 2621 memoize=true now. 2622 2623July 2, 2006 2624 2625* Added k=* option so you can set k=2, for example, on whole grammar, 2626 but an individual decision can be LL(*). 2627 2628* memoize option for grammars, rules, blocks. Remove -nomemo cmd-line option 2629 2630* but in DOT generator for DFA; fixed. 2631 2632* runtime.DFA reported errors even when backtracking 2633 2634July 1, 2006 2635 2636* Added -X option list to help 2637 2638* Syn preds were being hoisted into other rules, causing lots of extra 2639 backtracking. 2640 2641June 29, 2006 2642 2643* unnecessary files removed during build. 2644 2645* Matt Benson updated build.xml 2646 2647* Detecting use of synpreds in analysis now instead of codegen. In 2648 this way, I can avoid analyzing decisions in synpreds for synpreds 2649 not used in a DFA for a real rule. This is used to optimize things 2650 for backtrack option. 2651 2652* Code gen must add _fragment or whatever to end of pred name in 2653 template synpredRule to avoid having ANTLR know anything about 2654 method names. 2655 2656* Added -IdbgST option to emit ST delimiters at start/stop of all 2657 templates spit out. 2658 2659June 28, 2006 2660 2661* Tweaked message when ANTLR cannot handle analysis. 2662 26633.0b1 - June 27, 2006 2664 2665June 24, 2006 2666 2667* syn preds no longer generate little static classes; they also don't 2668 generate a whole bunch of extra crap in the rules built to test syn 2669 preds. Removed GrammarFragmentPointer class from runtime. 2670 2671June 23-24, 2006 2672 2673* added output option to -report output. 2674 2675* added profiling info: 2676 Number of rule invocations in "guessing" mode 2677 number of rule memoization cache hits 2678 number of rule memoization cache misses 2679 2680* made DFA DOT diagrams go left to right not top to bottom 2681 2682* I try to recursive overflow states now by resolving these states 2683 with semantic/syntactic predicates if they exist. The DFA is then 2684 deterministic rather than simply resolving by choosing first 2685 nondeterministic alt. I used to generated errors: 2686 2687~/tmp $ java org.antlr.Tool -dfa t.g 2688ANTLR Parser Generator Early Access Version 3.0b2 (July 5, 2006) 1989-2006 2689t.g:2:5: Alternative 1: after matching input such as A A A A A decision cannot predict what comes next due to recursion overflow to b from b 2690t.g:2:5: Alternative 2: after matching input such as A A A A A decision cannot predict what comes next due to recursion overflow to b from b 2691 2692 Now, I uses predicates if available and emits no warnings. 2693 2694* made sem preds share accept states. Previously, multiple preds in a 2695decision forked new accepts each time for each nondet state. 2696 2697June 19, 2006 2698 2699* Need parens around the prediction expressions in templates. 2700 2701* Referencing $ID.text in an action forced bad code gen in lexer rule ID. 2702 2703* Fixed a bug in how predicates are collected. The definition of 2704 "last predicated alternative" was incorrect in the analysis. Further, 2705 gated predicates incorrectly missed a case where an edge should become 2706 true (a tautology). 2707 2708* Removed an unnecessary input.consume() reference in the runtime/DFA class. 2709 2710June 14, 2006 2711 2712* -> ($rulelabel)? didn't generate proper code for ASTs. 2713 2714* bug in code gen (did not compile) 2715a : ID -> ID 2716 | ID -> ID 2717 ; 2718Problem is repeated ref to ID from left side. Juergen pointed this out. 2719 2720* use of tokenVocab with missing file yielded exception 2721 2722* (A|B)=> foo yielded an exception as (A|B) is a set not a block. Fixed. 2723 2724* Didn't set ID1= and INT1= for this alt: 2725 | ^(ID INT+ {System.out.print(\"^(\"+$ID+\" \"+$INT+\")\");}) 2726 2727* Fixed so repeated dangling state errors only occur once like: 2728t.g:4:17: the decision cannot distinguish between alternative(s) 2,1 for at least one input sequence 2729 2730* tracking of rule elements was on (making list defs at start of 2731 method) with templates instead of just with ASTs. Turned off. 2732 2733* Doesn't crash when you give it a missing file now. 2734 2735* -report: add output info: how many LL(1) decisions. 2736 2737June 13, 2006 2738 2739* ^(ROOT ID?) Didn't work; nor did any other nullable child list such as 2740 ^(ROOT ID* INT?). Now, I check to see if child list is nullable using 2741 Grammar.LOOK() and, if so, I generate an "IF lookahead is DOWN" gate 2742 around the child list so the whole thing is optional. 2743 2744* Fixed a bug in LOOK that made it not look through nullable rules. 2745 2746* Using AST suffixes or -> rewrite syntax now gives an error w/o a grammar 2747 output option. Used to crash ;) 2748 2749* References to EOF ended up with improper -1 refs instead of EOF in output. 2750 2751* didn't warn of ambig ref to $expr in rewrite; fixed. 2752list 2753 : '[' expr 'for' type ID 'in' expr ']' 2754 -> comprehension(expr={$expr.st},type={},list={},i={}) 2755 ; 2756 2757June 12, 2006 2758 2759* EOF works in the parser as a token name. 2760 2761* Rule b:(A B?)*; didn't display properly in AW due to the way ANTLR 2762 generated NFA. 2763 2764* "scope x;" in a rule for unknown x gives no error. Fixed. Added unit test. 2765 2766* Label type for refs to start/stop in tree parser and other parsers were 2767 not used. Lots of casting. Ick. Fixed. 2768 2769* couldn't refer to $tokenlabel in isolation; but need so we can test if 2770 something was matched. Fixed. 2771 2772* Lots of little bugs fixed in $x.y, %... translation due to new 2773 action translator. 2774 2775* Improperly tracking block nesting level; result was that you couldn't 2776 see $ID in action of rule "a : A+ | ID {Token t = $ID;} | C ;" 2777 2778* a : ID ID {$ID.text;} ; did not get a warning about ambiguous $ID ref. 2779 2780* No error was found on $COMMENT.text: 2781 2782COMMENT 2783 : '/*' (options {greedy=false;} : . )* '*/' 2784 {System.out.println("found method "+$COMMENT.text);} 2785 ; 2786 2787 $enclosinglexerrule scope does not exist. Use text or setText() here. 2788 2789June 11, 2006 2790 2791* Single return values are initialized now to default or to your spec. 2792 2793* cleaned up input stream stuff. Added ANTLRReaderStream, ANTLRInputStream 2794 and refactored. You can specify encodings now on ANTLRFileStream (and 2795 ANTLRInputStream) now. 2796 2797* You can set text local var now in a lexer rule and token gets that text. 2798 start/stop indexes are still set for the token. 2799 2800* Changed lexer slightly. Calling a nonfragment rule from a 2801 nonfragment rule does not set the overall token. 2802 2803June 10, 2006 2804 2805* Fixed bug where unnecessary escapes yield char==0 like '\{'. 2806 2807* Fixed analysis bug. This grammar didn't report a recursion warning: 2808x : y X 2809 | y Y 2810 ; 2811y : L y R 2812 | B 2813 ; 2814 The DFAState.equals() method was messed up. 2815 2816* Added @synpredgate {...} action so you can tell ANTLR how to gate actions 2817 in/out during syntactic predicate evaluation. 2818 2819* Fuzzy parsing should be more efficient. It should backtrack over a rule 2820 and then rewind and do it again "with feeling" to exec actions. It was 2821 actually doing it 3x not 2x. 2822 2823June 9, 2006 2824 2825* Gutted and rebuilt the action translator for $x.y, $x::y, ... 2826 Uses ANTLR v3 now for the first time inside v3 source. :) 2827 ActionTranslator.java 2828 2829* Fixed a bug where referencing a return value on a rule didn't work 2830 because later a ref to that rule's predefined properties didn't 2831 properly force a return value struct to be built. Added unit test. 2832 2833June 6, 2006 2834 2835* New DFA mechanisms. Cyclic DFA are implemented as state tables, 2836 encoded via strings as java cannot handle large static arrays :( 2837 States with edges emanating that have predicates are specially 2838 treated. A method is generated to do these states. The DFA 2839 simulation routine uses the "special" array to figure out if the 2840 state is special. See March 25, 2006 entry for description: 2841 http://www.antlr.org/blog/antlr3/codegen.tml. analysis.DFA now has 2842 all the state tables generated for code gen. CyclicCodeGenerator.java 2843 disappeared as it's unneeded code. :) 2844 2845* Internal general clean up of the DFA.states vs uniqueStates thing. 2846 Fixed lookahead decisions no longer fill uniqueStates. Waste of 2847 time. Also noted that when adding sem pred edges, I didn't check 2848 for state reuse. Fixed. 2849 2850June 4, 2006 2851 2852* When resolving ambig DFA states predicates, I did not add the new states 2853 to the list of unique DFA states. No observable effect on output except 2854 that DFA state numbers were not always contiguous for predicated decisions. 2855 I needed this fix for new DFA tables. 2856 28573.0ea10 - June 2, 2006 2858 2859June 2, 2006 2860 2861* Improved grammar stats and added syntactic pred tracking. 2862 2863June 1, 2006 2864 2865* Due to a type mismatch, the DebugParser.recoverFromMismatchedToken() 2866 method was not called. Debug events for mismatched token error 2867 notification were not sent to ANTLRWorks probably 2868 2869* Added getBacktrackingLevel() for any recognizer; needed for profiler. 2870 2871* Only writes profiling data for antlr grammar analysis with -profile set 2872 2873* Major update and bug fix to (runtime) Profiler. 2874 2875May 27, 2006 2876 2877* Added Lexer.skip() to force lexer to ignore current token and look for 2878 another; no token is created for current rule and is not passed on to 2879 parser (or other consumer of the lexer). 2880 2881* Parsers are much faster now. I removed use of java.util.Stack for pushing 2882 follow sets and use a hardcoded array stack instead. Dropped from 2883 5900ms to 3900ms for parse+lex time parsing entire java 1.4.2 source. Lex 2884 time alone was about 1500ms. Just looking at parse time, we get about 2x 2885 speed improvement. :) 2886 2887May 26, 2006 2888 2889* Fixed NFA construction so it generates NFA for (A*)* such that ANTLRWorks 2890 can display it properly. 2891 2892May 25, 2006 2893 2894* added abort method to Grammar so AW can terminate the conversion if it's 2895 taking too long. 2896 2897May 24, 2006 2898 2899* added method to get left recursive rules from grammar without doing full 2900 grammar analysis. 2901 2902* analysis, code gen not attempted if serious error (like 2903 left-recursion or missing rule definition) occurred while reading 2904 the grammar in and defining symbols. 2905 2906* added amazing optimization; reduces analysis time by 90% for java 2907 grammar; simple IF statement addition! 2908 29093.0ea9 - May 20, 2006 2910 2911* added global k value for grammar to limit lookahead for all decisions unless 2912overridden in a particular decision. 2913 2914* added failsafe so that any decision taking longer than 2 seconds to create 2915the DFA will fall back on k=1. Use -ImaxtimeforDFA n (in ms) to set the time. 2916 2917* added an option (turned off for now) to use multiple threads to 2918perform grammar analysis. Not much help on a 2-CPU computer as 2919garbage collection seems to peg the 2nd CPU already. :( Gotta wait for 2920a 4 CPU box ;) 2921 2922* switched from #src to // $ANTLR src directive. 2923 2924* CommonTokenStream.getTokens() looked past end of buffer sometimes. fixed. 2925 2926* unicode literals didn't really work in DOT output and generated code. fixed. 2927 2928* fixed the unit test rig so it compiles nicely with Java 1.5 2929 2930* Added ant build.xml file (reads build.properties file) 2931 2932* predicates sometimes failed to compile/eval properly due to missing (...) 2933 in IF expressions. Forced (..) 2934 2935* (...)? with only one alt were not optimized. Was: 2936 2937 // t.g:4:7: ( B )? 2938 int alt1=2; 2939 int LA1_0 = input.LA(1); 2940 if ( LA1_0==B ) { 2941 alt1=1; 2942 } 2943 else if ( LA1_0==-1 ) { 2944 alt1=2; 2945 } 2946 else { 2947 NoViableAltException nvae = 2948 new NoViableAltException("4:7: ( B )?", 1, 0, input); 2949 throw nvae; 2950 } 2951 2952is now: 2953 2954 // t.g:4:7: ( B )? 2955 int alt1=2; 2956 int LA1_0 = input.LA(1); 2957 if ( LA1_0==B ) { 2958 alt1=1; 2959 } 2960 2961 Smaller, faster and more readable. 2962 2963* Allow manual init of return values now: 2964 functionHeader returns [int x=3*4, char (*f)()=null] : ... ; 2965 2966* Added optimization for DFAs that fixed a codegen bug with rules in lexer: 2967 EQ : '=' ; 2968 ASSIGNOP : '=' | '+=' ; 2969 EQ is a subset of other rule. It did not given an error which is 2970 correct, but generated bad code. 2971 2972* ANTLR was sending column not char position to ANTLRWorks. 2973 2974* Bug fix: location 0, 0 emitted for synpreds and empty alts. 2975 2976* debugging event handshake how sends grammar file name. Added getGrammarFileName() to recognizers. Java.stg generates it: 2977 2978 public String getGrammarFileName() { return "<fileName>"; } 2979 2980* tree parsers can do arbitrary lookahead now including backtracking. I 2981 updated CommonTreeNodeStream. 2982 2983* added events for debugging tree parsers: 2984 2985 /** Input for a tree parser is an AST, but we know nothing for sure 2986 * about a node except its type and text (obtained from the adaptor). 2987 * This is the analog of the consumeToken method. Again, the ID is 2988 * the hashCode usually of the node so it only works if hashCode is 2989 * not implemented. 2990 */ 2991 public void consumeNode(int ID, String text, int type); 2992 2993 /** The tree parser looked ahead */ 2994 public void LT(int i, int ID, String text, int type); 2995 2996 /** The tree parser has popped back up from the child list to the 2997 * root node. 2998 */ 2999 public void goUp(); 3000 3001 /** The tree parser has descended to the first child of a the current 3002 * root node. 3003 */ 3004 public void goDown(); 3005 3006* Added DebugTreeNodeStream and DebugTreeParser classes 3007 3008* Added ctor because the debug tree node stream will need to ask quesitons about nodes and since nodes are just Object, it needs an adaptor to decode the nodes and get text/type info for the debugger. 3009 3010public CommonTreeNodeStream(TreeAdaptor adaptor, Tree tree); 3011 3012* added getter to TreeNodeStream: 3013 public TreeAdaptor getTreeAdaptor(); 3014 3015* Implemented getText/getType in CommonTreeAdaptor. 3016 3017* Added TraceDebugEventListener that can dump all events to stdout. 3018 3019* I broke down and make Tree implement getText 3020 3021* tree rewrites now gen location debug events. 3022 3023* added AST debug events to listener; added blank listener for convenience 3024 3025* updated debug events to send begin/end backtrack events for debugging 3026 3027* with a : (b->b) ('+' b -> ^(PLUS $a b))* ; you get b[0] each time as 3028 there is no loop in rewrite rule itself. Need to know context that 3029 the -> is inside the rule and hence b means last value of b not all 3030 values. 3031 3032* Bug in TokenRewriteStream; ops at indexes < start index blocked proper op. 3033 3034* Actions in ST rewrites "-> ({$op})()" were not translated 3035 3036* Added new action name: 3037 3038@rulecatch { 3039catch (RecognitionException re) { 3040 reportError(re); 3041 recover(input,re); 3042} 3043catch (Throwable t) { 3044 System.err.println(t); 3045} 3046} 3047Overrides rule catch stuff. 3048 3049* Isolated $ refs caused exception 3050 30513.0ea8 - March 11, 2006 3052 3053* added @finally {...} action like @init for rules. Executes in 3054 finally block (java target) after all other stuff like rule memoization. 3055 No code changes needs; ST just refs a new action: 3056 <ruleDescriptor.actions.finally> 3057 3058* hideous bug fixed: PLUS='+' didn't result in '+' rule in lexer 3059 3060* TokenRewriteStream didn't do toString() right when no rewrites had been done. 3061 3062* lexer errors in interpreter were not printed properly 3063 3064* bitsets are dumped in hex not decimal now for FOLLOW sets 3065 3066* /* epsilon */ is not printed now when printing out grammars with empty alts 3067 3068* Fixed another bug in tree rewrite stuff where it was checking that elements 3069 had at least one element. Strange...commented out for now to see if I can remember what's up. 3070 3071* Tree rewrites had problems when you didn't have x+=FOO variables. Rules 3072 like this work now: 3073 3074 a : (x=ID)? y=ID -> ($x $y)?; 3075 3076* filter=true for lexers turns on k=1 and backtracking for every token 3077 alternative. Put the rules in priority order. 3078 3079* added getLine() etc... to Tree to support better error reporting for 3080 trees. Added MismatchedTreeNodeException. 3081 3082* $templates::foo() is gone. added % as special template symbol. 3083 %foo(a={},b={},...) ctor (even shorter than $templates::foo(...)) 3084 %({name-expr})(a={},...) indirect template ctor reference 3085 3086 The above are parsed by antlr.g and translated by codegen.g 3087 The following are parsed manually here: 3088 3089 %{string-expr} anonymous template from string expr 3090 %{expr}.y = z; template attribute y of StringTemplate-typed expr to z 3091 %x.y = z; set template attribute y of x (always set never get attr) 3092 to z [languages like python without ';' must still use the 3093 ';' which the code generator is free to remove during code gen] 3094 3095* -> ({expr})(a={},...) notation for indirect template rewrite. 3096 expr is the name of the template. 3097 3098* $x[i]::y and $x[-i]::y notation for accesssing absolute scope stack 3099 indexes and relative negative scopes. $x[-1]::y is the y attribute 3100 of the previous scope (stack top - 1). 3101 3102* filter=true mode for lexers; can do this now...upon mismatch, just 3103 consumes a char and tries again: 3104lexer grammar FuzzyJava; 3105options {filter=true;} 3106 3107FIELD 3108 : TYPE WS? name=ID WS? (';'|'=') 3109 {System.out.println("found var "+$name.text);} 3110 ; 3111 3112* refactored char streams so ANTLRFileStream is now a subclass of 3113 ANTLRStringStream. 3114 3115* char streams for lexer now allowed nested backtracking in lexer. 3116 3117* added TokenLabelType for lexer/parser for all token labels 3118 3119* line numbers for error messages were not updated properly in antlr.g 3120 for strings, char literals and <<...>> 3121 3122* init action in lexer rules was before the type,start,line,... decls. 3123 3124* Tree grammars can now specify output; I've only tested output=templat 3125 though. 3126 3127* You can reference EOF now in the parser and lexer. It's just token type 3128 or char value -1. 3129 3130* Bug fix: $ID refs in the *lexer* were all messed up. Cleaned up the 3131 set of properties available... 3132 3133* Bug fix: .st not found in rule ref when rule has scope: 3134field 3135scope { 3136 StringTemplate funcDef; 3137} 3138 : ... 3139 {$field::funcDef = $field.st;} 3140 ; 3141it gets field_stack.st instead 3142 3143* return in backtracking must return retval or null if return value. 3144 3145* $property within a rule now works like $text, $st, ... 3146 3147* AST/Template Rewrites were not gated by backtracking==0 so they 3148 executed even when guessing. Auto AST construction is now gated also. 3149 3150* CommonTokenStream was somehow returning tokens not text in toString() 3151 3152* added useful methods to runtime.BitSet and also to CommonToken so you can 3153 update the text. Added nice Token stream method: 3154 3155 /** Given a start and stop index, return a List of all tokens in 3156 * the token type BitSet. Return null if no tokens were found. This 3157 * method looks at both on and off channel tokens. 3158 */ 3159 public List getTokens(int start, int stop, BitSet types); 3160 3161* literals are now passed in the .tokens files so you can ref them in 3162 tree parses, for example. 3163 3164* added basic exception handling; no labels, just general catches: 3165 3166a : {;}A | B ; 3167 exception 3168 catch[RecognitionException re] { 3169 System.out.println("recog error"); 3170 } 3171 catch[Exception e] { 3172 System.out.println("error"); 3173 } 3174 3175* Added method to TokenStream: 3176 public String toString(Token start, Token stop); 3177 3178* antlr generates #src lines in lexer grammars generated from combined grammars 3179 so error messages refer to original file. 3180 3181* lexers generated from combined grammars now use originally formatting. 3182 3183* predicates have $x.y stuff translated now. Warning: predicates might be 3184 hoisted out of context. 3185 3186* return values in return val structs are now public. 3187 3188* output=template with return values on rules was broken. I assume return values with ASTs was broken too. Fixed. 3189 31903.0ea7 - December 14, 2005 3191 3192* Added -print option to print out grammar w/o actions 3193 3194* Renamed BaseParser to be BaseRecognizer and even made Lexer derive from 3195 this; nice as it now shares backtracking support code. 3196 3197* Added syntactic predicates (...)=>. See December 4, 2005 entry: 3198 3199 http://www.antlr.org/blog/antlr3/lookahead.tml 3200 3201 Note that we have a new option for turning off rule memoization during 3202 backtracking: 3203 3204 -nomemo when backtracking don't generate memoization code 3205 3206* Predicates are now tested in order that you specify the alts. If you 3207 leave the last alt "naked" (w/o pred), it will assume a true pred rather 3208 than union of other preds. 3209 3210* Added gated predicates "{p}?=>" that literally turn off a production whereas 3211disambiguating predicates are only hoisted into the predictor when syntax alone 3212is not sufficient to uniquely predict alternatives. 3213 3214A : {p}? => "a" ; 3215B : {!p}? => ("a"|"b")+ ; 3216 3217* bug fixed related to predicates in predictor 3218lexer grammar w; 3219A : {p}? "a" ; 3220B : {!p}? ("a"|"b")+ ; 3221DFA is correct. A state splits for input "a" on the pred. 3222Generated code though was hosed. No pred tests in prediction code! 3223I added testLexerPreds() and others in TestSemanticPredicateEvaluation.java 3224 3225* added execAction template in case we want to do something in front of 3226 each action execution or something. 3227 3228* left-recursive cycles from rules w/o decisions were not detected. 3229 3230* undefined lexer rules were not announced! fixed. 3231 3232* unreachable messages for Tokens rule now indicate rule name not alt. E.g., 3233 3234 Ruby.lexer.g:24:1: The following token definitions are unreachable: IVAR 3235 3236* nondeterminism warnings improved for Tokens rule: 3237 3238Ruby.lexer.g:10:1: Multiple token rules can match input such as ""0".."9"": INT, FLOAT 3239As a result, tokens(s) FLOAT were disabled for that input 3240 3241 3242* DOT diagrams didn't show escaped char properly. 3243 3244* Char/string literals are now all 'abc' not "abc". 3245 3246* action syntax changed "@scope::actionname {action}" where scope defaults 3247 to "parser" if parser grammar or combined grammar, "lexer" if lexer grammar, 3248 and "treeparser" if tree grammar. The code generation targets decide 3249 what scopes are available. Each "scope" yields a hashtable for use in 3250 the output templates. The scopes full of actions are sent to all output 3251 file templates (currently headerFile and outputFile) as attribute actions. 3252 Then you can reference <actions.scope> to get the map of actions associated 3253 with scope and <actions.parser.header> to get the parser's header action 3254 for example. This should be very flexible. The target should only have 3255 to define which scopes are valid, but the action names should be variable 3256 so we don't have to recompile ANTLR to add actions to code gen templates. 3257 3258 grammar T; 3259 options {language=Java;} 3260 @header { package foo; } 3261 @parser::stuff { int i; } // names within scope not checked; target dependent 3262 @members { int i; } 3263 @lexer::header {head} 3264 @lexer::members { int j; } 3265 @headerfile::blort {...} // error: this target doesn't have headerfile 3266 @treeparser::members {...} // error: this is not a tree parser 3267 a 3268 @init {int i;} 3269 : ID 3270 ; 3271 ID : 'a'..'z'; 3272 3273 For now, the Java target uses members and header as a valid name. Within a 3274 rule, the init action name is valid. 3275 3276* changed $dynamicscope.value to $dynamicscope::value even if value is defined 3277 in same rule such as $function::name where rule function defines name. 3278 3279* $dynamicscope gets you the stack 3280 3281* rule scopes go like this now: 3282 3283 rule 3284 scope {...} 3285 scope slist,Symbols; 3286 : ... 3287 ; 3288 3289* Created RuleReturnScope as a generic rule return value. Makes it easier 3290 to do this: 3291 RuleReturnScope r = parser.program(); 3292 System.out.println(r.getTemplate().toString()); 3293 3294* $template, $tree, $start, etc... 3295 3296* $r.x in current rule. $r is ignored as fully-qualified name. $r.start works too 3297 3298* added warning about $r referring to both return value of rule and dynamic scope of rule 3299 3300* integrated StringTemplate in a very simple manner 3301 3302Syntax: 3303-> template(arglist) "..." 3304-> template(arglist) <<...>> 3305-> namedTemplate(arglist) 3306-> {free expression} 3307-> // empty 3308 3309Predicate syntax: 3310a : A B -> {p1}? foo(a={$A.text}) 3311 -> {p2}? foo(a={$B.text}) 3312 -> // return nothing 3313 3314An arg list is just a list of template attribute assignments to actions in curlies. 3315 3316There is a setTemplateLib() method for you to use with named template rewrites. 3317 3318Use a new option: 3319 3320grammar t; 3321options {output=template;} 3322... 3323 3324This all should work for tree grammars too, but I'm still testing. 3325 3326* fixed bugs where strings were improperly escaped in exceptions, comments, etc.. For example, newlines came out as newlines not the escaped version 3327 33283.0ea6 - November 13, 2005 3329 3330* turned off -debug/-profile, which was on by default 3331 3332* completely refactored the output templates; added some missing templates. 3333 3334* dramatically improved infinite recursion error messages (actually 3335 left-recursion never even was printed out before). 3336 3337* wasn't printing dangling state messages when it reanalyzes with k=1. 3338 3339* fixed a nasty bug in the analysis engine dealing with infinite recursion. 3340 Spent all day thinking about it and cleaned up the code dramatically. 3341 Bug fixed and software is more powerful and I understand it better! :) 3342 3343* improved verbose DFA nodes; organized by alt 3344 3345* got much better random phrase generation. For example: 3346 3347 $ java org.antlr.tool.RandomPhrase simple.g program 3348 int Ktcdn ';' method wh '(' ')' '{' return 5 ';' '}' 3349 3350* empty rules like "a : ;" generated code that didn't compile due to 3351 try/catch for RecognitionException. Generated code couldn't possibly 3352 throw that exception. 3353 3354* when printing out a grammar, such as in comments in generated code, 3355 ANTLR didn't print ast suffix stuff back out for literals. 3356 3357* This never exited loop: 3358 DATA : (options {greedy=false;}: .* '\n' )* '\n' '.' ; 3359 and now it works due to new default nongreedy .* Also this works: 3360 DATA : (options {greedy=false;}: .* '\n' )* '.' ; 3361 3362* Dot star ".*" syntax didn't work; in lexer it is nongreedy by 3363 default. In parser it is on greedy but also k=1 by default. Added 3364 unit tests. Added blog entry to describe. 3365 3366* ~T where T is the only token yielded an empty set but no error 3367 3368* Used to generate unreachable message here: 3369 3370 parser grammar t; 3371 a : ID a 3372 | ID 3373 ; 3374 3375 z.g:3:11: The following alternatives are unreachable: 2 3376 3377 In fact it should really be an error; now it generates: 3378 3379 no start rule in grammar t (no rule can obviously be followed by EOF) 3380 3381 Per next change item, ANTLR cannot know that EOF follows rule 'a'. 3382 3383* added error message indicating that ANTLR can't figure out what your 3384 start rule is. Required to properly generate code in some cases. 3385 3386* validating semantic predicates now work (if they are false, they 3387 throw a new FailedPredicateException 3388 3389* two hideous bug fixes in the IntervalSet, which made analysis go wrong 3390 in a few cases. Thanks to Oliver Zeigermann for finding lots of bugs 3391 and making suggested fixes (including the next two items)! 3392 3393* cyclic DFAs are now nonstatic and hence can access instance variables 3394 3395* labels are now allowed on lexical elements (in the lexer) 3396 3397* added some internal debugging options 3398 3399* ~'a'* and ~('a')* were not working properly; refactored antlr.g grammar 3400 34013.0ea5 - July 5, 2005 3402 3403* Using '\n' in a parser grammar resulted in a nonescaped version of '\n' in the token names table making compilation fail. I fixed this by reorganizing/cleaning up portion of ANTLR that deals with literals. See comment org.antlr.codegen.Target. 3404 3405* Target.getMaxCharValue() did not use the appropriate max value constant. 3406 3407* ALLCHAR was a constant when it should use the Target max value def. set complement for wildcard also didn't use the Target def. Generally cleaned up the max char value stuff. 3408 3409* Code gen didn't deal with ASTLabelType properly...I think even the 3.0ea7 example tree parser was broken! :( 3410 3411* Added a few more unit tests dealing with escaped literals 3412 34133.0ea4 - June 29, 2005 3414 3415* tree parsers work; added CommonTreeNodeStream. See simplecTreeParser 3416 example in examples-v3 tarball. 3417 3418* added superClass and ASTLabelType options 3419 3420* refactored Parser to have a BaseParser and added TreeParser 3421 3422* bug fix: actions being dumped in description strings; compile errors 3423 resulted 3424 34253.0ea3 - June 23, 2005 3426 3427Enhancements 3428 3429* Automatic tree construction operators are in: ! ^ ^^ 3430 3431* Tree construction rewrite rules are in 3432 -> {pred1}? rewrite1 3433 -> {pred2}? rewrite2 3434 ... 3435 -> rewriteN 3436 3437 The rewrite rules may be elements like ID, expr, $label, {node expr} 3438 and trees ^( <root> <children> ). You have have (...)?, (...)*, (...)+ 3439 subrules as well. 3440 3441 You may have rewrites in subrules not just at outer level of rule, but 3442 any -> rewrite forces auto AST construction off for that alternative 3443 of that rule. 3444 3445 To avoid cycles, copy semantics are used: 3446 3447 r : INT -> INT INT ; 3448 3449 means make two new nodes from the same INT token. 3450 3451 Repeated references to a rule element implies a copy for at least one 3452 tree: 3453 3454 a : atom -> ^(atom atom) ; // NOT CYCLE! (dup atom tree) 3455 3456* $ruleLabel.tree refers to tree created by matching the labeled element. 3457 3458* A description of the blocks/alts is generated as a comment in output code 3459 3460* A timestamp / signature is put at top of each generated code file 3461 34623.0ea2 - June 12, 2005 3463 3464Bug fixes 3465 3466* Some error messages were missing the stackTrace parameter 3467 3468* Removed the file locking mechanism as it's not cross platform 3469 3470* Some absolute vs relative path name problems with writing output 3471 files. Rules are now more concrete. -o option takes precedence 3472 // -o /tmp /var/lib/t.g => /tmp/T.java 3473 // -o subdir/output /usr/lib/t.g => subdir/output/T.java 3474 // -o . /usr/lib/t.g => ./T.java 3475 // -o /tmp subdir/t.g => /tmp/subdir/t.g 3476 // If they didn't specify a -o dir so just write to location 3477 // where grammar is, absolute or relative 3478 3479* does error checking on unknown option names now 3480 3481* Using just language code not locale name for error message file. I.e., 3482 the default (and for any English speaking locale) is en.stg not en_US.stg 3483 anymore. 3484 3485* The error manager now asks the Tool to panic rather than simply doing 3486 a System.exit(). 3487 3488* Lots of refactoring concerning grammar, rule, subrule options. Now 3489 detects invalid options. 3490 34913.0ea1 - June 1, 2005 3492 3493Initial early access release 3494