1 \input texinfo @c -*-texinfo-*-
3 @setfilename ../../info/wisent
4 @set TITLE Wisent Parser Development
5 @set AUTHOR Eric M. Ludlam, David Ponce, and Richard Y. Kim
6 @settitle @value{TITLE}
8 @c *************************************************************************
10 @c *************************************************************************
12 @c Merge all indexes into a single index for now.
13 @c We can always separate them later into two or more as needed.
20 @c @footnotestyle separate
26 Copyright @copyright{} 1988-1993, 1995, 1998-2004, 2007, 2012
27 Free Software Foundation, Inc.
29 @c Since we are both GNU manuals, we do not need to ack each other here.
31 Some texts are borrowed or adapted from the manual of Bison version
32 1.35. The text in section entitled ``Understanding the automaton'' is
33 adapted from the section ``Understanding Your Parser'' in the manual
34 of Bison version 1.49.
38 Permission is granted to copy, distribute and/or modify this document
39 under the terms of the GNU Free Documentation License, Version 1.3 or
40 any later version published by the Free Software Foundation; with no
41 Invariant Sections, with the Front-Cover texts being ``A GNU Manual,''
42 and with the Back-Cover Texts as in (a) below. A copy of the license
43 is included in the section entitled ``GNU Free Documentation License''.
45 (a) The FSF's Back-Cover Text is: ``You have the freedom to copy and
46 modify this GNU manual. Buying copies from the FSF supports it in
47 developing GNU and promoting software freedom.''
51 @dircategory Emacs misc features
53 * Wisent: (wisent). Semantic Wisent parser development.
60 @c @setchapternewpage odd
61 @c @setchapternewpage off
66 @author by @value{AUTHOR}
68 @vskip 0pt plus 1 fill
77 @c *************************************************************************
79 @c *************************************************************************
85 Wisent (the European Bison ;-) is an Emacs Lisp implementation of the
86 GNU Compiler Compiler Bison.
88 This manual describes how to use Wisent to develop grammars for
89 programming languages, and how to use grammars to parse language
90 source in Emacs buffers.
92 It also describes how Wisent is used with the @semantic{} tool set
93 described in the @ref{Top, Semantic Manual, Semantic Manual, semantic}.
104 * GNU Free Documentation License::
108 @node Wisent Overview
109 @chapter Wisent Overview
111 @dfn{Wisent} (the European Bison) is an implementation in Emacs Lisp
112 of the GNU Compiler Compiler Bison. Its code is a port of the C code
113 of GNU Bison 1.28 & 1.31.
115 For more details on the basic concepts for understanding Wisent, it is
116 worthwhile to read the @ref{Top, Bison Manual, bison}.
118 @uref{http://www.gnu.org/manual/bison/html_node/index.html}.
121 Wisent can generate compilers compatible with the @semantic{} tool set.
122 See the @ref{Top, Semantic Manual, , semantic}.
124 It benefits from these Bison features:
128 It uses a fast but not so space-efficient encoding for the parse
129 tables, described in Corbett's PhD thesis from Berkeley:
131 @cite{Static Semantics in Compiler Error Recovery}@*
132 June 1985, Report No. UCB/CSD 85/251.
136 For generating the lookahead sets, Wisent uses the well-known
137 technique of F. DeRemer and A. Pennello described in:
139 @cite{Efficient Computation of LALR(1) Look-Ahead Sets}@*
140 October 1982, ACM TOPLAS Vol 4 No 4, 615--49,
141 @uref{http://dx.doi.org/10.1145/69622.357187}.
145 Wisent resolves shift/reduce conflicts using operator precedence and
149 Parser error recovery is accomplished using rules which match the
150 special token @code{error}.
153 Nevertheless there are some fundamental differences between Bison and
158 Wisent is intended to be used in Emacs. It reads and produces Emacs
159 Lisp data structures. All the additional code used in grammars is
163 Contrary to Bison, Wisent does not generate a parser which combines
164 Emacs Lisp code and grammar constructs. They exist separately.
165 Wisent reads the grammar from a Lisp data structure and then generates
166 grammar constructs as tables. Afterward, the derived tables can be
167 included and byte-compiled in separate Emacs Lisp files, and be used
168 at a later time by the Wisent's parser engine.
171 Wisent allows multiple start nonterminals and allows a call to the
172 parsing function to be made for a particular start nonterminal. For
173 example, this is particularly useful to parse a region of an Emacs
174 buffer. @semantic{} heavily depends on the availability of this feature.
178 @chapter Wisent Grammar
180 @cindex context-free grammar
182 In order for Wisent to parse a language, it must be described by a
183 @dfn{context-free grammar}. That is a grammar specified as rules that
184 can be applied regardless of context. For more information, see
185 @ref{Language and Grammar, , , bison}, in the Bison manual.
189 The formal grammar is formulated using @dfn{terminal} and
190 @dfn{nonterminal} items. Terminals can be Emacs Lisp symbols or
191 characters, and nonterminals are symbols only.
194 Terminals (also known as @dfn{tokens}) represent the lexical
195 elements of the language like numbers, strings, etc..
197 For example @samp{PLUS} can represent the operator @samp{+}.
199 Nonterminal symbols are described by rules:
203 RESULT @equiv{} COMPONENTS@dots{}
207 @samp{RESULT} is a nonterminal that this rule describes and
208 @samp{COMPONENTS} are various terminals and nonterminals that are put
209 together by this rule.
211 For example, this rule:
215 exp @equiv{} exp PLUS exp
219 Says that two groupings of type @samp{exp}, with a @samp{PLUS} token
220 in between, can be combined into a larger grouping of type @samp{exp}.
225 * Compiling a grammar::
229 @node Grammar format, Example, Wisent Grammar, Wisent Grammar
230 @comment node-name, next, previous, up
231 @section Grammar format
233 @cindex grammar format
234 To be acceptable by Wisent a context-free grammar must respect a
235 particular format. That is, must be represented as an Emacs Lisp list
238 @code{(@var{terminals} @var{assocs} . @var{non-terminals})}
242 Is the list of terminal symbols used in the grammar.
244 @cindex associativity
246 Specify the associativity of @var{terminals}. It is @code{nil} when
247 there is no associativity defined, or an alist of
248 @w{@code{(@var{assoc-type} . @var{assoc-value})}} elements.
250 @var{assoc-type} must be one of the @code{default-prec},
251 @code{nonassoc}, @code{left} or @code{right} symbols. When
252 @var{assoc-type} is @code{default-prec}, @var{assoc-value} must be
253 @code{nil} or @code{t} (the default). Otherwise it is a list of
254 tokens which must have been previously declared in @var{terminals}.
256 For details, see @ref{Contextual Precedence, , , bison}, in the
260 Is the list of nonterminal definitions. Each definition has the form:
262 @code{(@var{nonterm} . @var{rules})}
264 Where @var{nonterm} is the nonterminal symbol defined and
265 @var{rules} the list of rules that describe this nonterminal. Each
268 @code{(@var{components} [@var{precedence}] [@var{action}])}
274 Is a list of various terminals and nonterminals that are put together
281 (exp ((exp ?+ exp)) ;; exp: exp '+' exp
286 Says that two groupings of type @samp{exp}, with a @samp{+} token in
287 between, can be combined into a larger grouping of type @samp{exp}.
289 @cindex grammar coding conventions
290 By convention, a nonterminal symbol should be in lower case, such as
291 @samp{exp}, @samp{stmt} or @samp{declaration}. Terminal symbols
292 should be upper case to distinguish them from nonterminals: for
293 example, @samp{INTEGER}, @samp{IDENTIFIER}, @samp{IF} or
294 @samp{RETURN}. A terminal symbol that represents a particular keyword
295 in the language is conventionally the same as that keyword converted
296 to upper case. The terminal symbol @code{error} is reserved for error
299 @cindex middle-rule actions
300 Scattered among the components can be @dfn{middle-rule} actions.
301 Usually only @var{action} is provided (@pxref{action}).
303 If @var{components} in a rule is @code{nil}, it means that the rule
304 can match the empty string. For example, here is how to define a
305 comma-separated sequence of zero or more @samp{exp} groupings:
309 (expseq (nil) ;; expseq: ;; empty
310 ((expseq1)) ;; | expseq1
313 (expseq1 ((exp)) ;; expseq1: exp
314 ((expseq1 ?, exp)) ;; | expseq1 ',' exp
319 @cindex precedence level
321 Assign the rule the precedence of the given terminal item, overriding
322 the precedence that would be deduced for it, that is the one of the
323 last terminal in it. Notice that only terminals declared in
324 @var{assocs} have a precedence level. The altered rule precedence
325 then affects how conflicts involving that rule are resolved.
327 @var{precedence} is an optional vector of one terminal item.
329 Here is how @var{precedence} solves the problem of unary minus.
330 First, declare a precedence for a fictitious terminal symbol named
331 @code{UMINUS}. There are no tokens of this type, but the symbol
332 serves to stand for its precedence:
336 ((default-prec t) ;; This is the default
342 Now the precedence of @code{UMINUS} can be used in specific rules:
346 (exp @dots{} ;; exp: @dots{}
347 ((exp ?- exp)) ;; | exp '-' exp
349 ((?- exp) [UMINUS]) ;; | '-' exp %prec UMINUS
355 If you forget to append @code{[UMINUS]} to the rule for unary minus,
356 Wisent silently assumes that minus has its usual precedence. This
357 kind of problem can be tricky to debug, since one typically discovers
358 the mistake only by testing the code.
360 Using @code{(default-prec nil)} declaration makes it easier to
361 discover this kind of problem systematically. It causes rules that
362 lack a @var{precedence} modifier to have no precedence, even if the
363 last terminal symbol mentioned in their components has a declared
366 If @code{(default-prec nil)} is in effect, you must specify
367 @var{precedence} for all rules that participate in precedence conflict
368 resolution. Then you will see any shift/reduce conflict until you
369 tell Wisent how to resolve it, either by changing your grammar or by
370 adding an explicit precedence. This will probably add declarations to
371 the grammar, but it helps to protect against incorrect rule
374 The effect of @code{(default-prec nil)} can be reversed by giving
375 @code{(default-prec t)}, which is the default.
377 For more details, see @ref{Contextual Precedence, , , bison}, in the
380 It is important to understand that @var{assocs} declarations defines
381 associativity but also assign a precedence level to terminals. All
382 terminals declared in the same @code{left}, @code{right} or
383 @code{nonassoc} association get the same precedence level. The
384 precedence level is increased at each new association.
386 On the other hand, @var{precedence} explicitly assign the precedence
387 level of the given terminal to a rule.
389 @cindex semantic actions
390 @item @anchor{action}action
391 An action is an optional Emacs Lisp function call, like this:
395 The result of an action determines the semantic value of a rule.
397 From an implementation standpoint, the function call will be embedded
398 in a lambda expression, and several useful local variables will be
404 Where @var{n} is a positive integer. Like in Bison, the value of
405 @code{$@var{n}} is the semantic value of the @var{n}th element of
406 @var{components}, starting from 1. It can be of any Lisp data
409 @vindex $region@var{n}
411 Where @var{n} is a positive integer. For each @code{$@var{n}}
412 variable defined there is a corresponding @code{$region@var{n}}
413 variable. Its value is a pair @code{(@var{start-pos} .
414 @var{end-pos})} that represent the start and end positions (in the
415 lexical input stream) of the @code{$@var{n}} value. It can be
416 @code{nil} when the component positions are not available, like for an
417 empty string component for example.
421 Its value is the leftmost and rightmost positions of input data
422 matched by all @var{components} in the rule. This is a pair
423 @code{(@var{leftmost-pos} . @var{rightmost-pos})}. It can be
424 @code{nil} when components positions are not available.
428 This variable is initialized with the nonterminal symbol
429 (@var{nonterm}) the rule belongs to. It could be useful to improve
430 error reporting or debugging. It is also used to automatically
431 provide incremental re-parse entry points for @semantic{} tags
432 (@pxref{Wisent Semantic}).
436 The value of @code{$action} is the symbolic name of the current
437 semantic action (@pxref{Debugging actions}).
440 When an action is not specified a default value is supplied, it is
441 @code{(identity $1)}. This means that the default semantic value of a
442 rule is the value of its first component. Excepted for a rule
443 matching the empty string, for which the default action is to return
448 @node Example, Compiling a grammar, Grammar format, Wisent Grammar
449 @comment node-name, next, previous, up
452 @cindex grammar example
453 Here is an example to parse simple infix arithmetic expressions. See
454 @ref{Infix Calc, , , bison}, in the Bison manual for details.
462 ;; Terminal associativity & precedence
473 (format "%s %s" $1 $2))
487 (string-to-number $1))
509 In the bison-like @dfn{WY} format (@pxref{Wisent Semantic}) the
510 grammar looks like this:
516 %nonassoc '=' ;; comparison
519 %left NEG ;; negation--unary minus
520 %right '^' ;; exponentiation
527 (format "%s %s" $1 $2)
541 (string-to-number $1)
564 @node Compiling a grammar, Conflicts, Example, Wisent Grammar
565 @comment node-name, next, previous, up
566 @section Compiling a grammar
569 After providing a context-free grammar in a suitable format, it must
570 be translated into a set of tables (an @dfn{automaton}) that will be
571 used to derive the parser. Like Bison, Wisent translates grammars that
572 must be @dfn{LALR(1)}.
574 @cindex LALR(1) grammar
575 @cindex look-ahead token
576 A grammar is @acronym{LALR(1)} if it is possible to tell how to parse
577 any portion of an input string with just a single token of look-ahead:
578 the @dfn{look-ahead token}. See @ref{Language and Grammar, , ,
579 bison}, in the Bison manual for more information.
581 @cindex grammar compilation
582 Grammar translation (compilation) is achieved by the function:
584 @cindex compiling a grammar
585 @vindex wisent-single-start-flag
586 @findex wisent-compile-grammar
587 @defun wisent-compile-grammar grammar &optional start-list
588 Compile @var{grammar} and return an @acronym{LALR(1)} automaton.
590 Optional argument @var{start-list} is a list of start symbols
591 (nonterminals). If @code{nil} the first nonterminal defined in the
592 grammar is the default start symbol. If @var{start-list} contains
593 only one element, it defines the start symbol. If @var{start-list}
594 contains more than one element, all are defined as potential start
595 symbols, unless @code{wisent-single-start-flag} is non-@code{nil}. In
596 that case the first element of @var{start-list} defines the start
597 symbol and others are ignored.
599 The @acronym{LALR(1)} automaton is a vector of the form:
601 @code{[@var{actions gotos starts functions}]}
605 A state/token matrix telling the parser what to do at every state
606 based on the current look-ahead token. That is shift, reduce, accept
607 or error. See also @ref{Wisent Parsing}.
610 A state/nonterminal matrix telling the parser the next state to go to
611 after reducing with each rule.
614 An alist which maps the allowed start symbols (nonterminals) to
615 lexical tokens that will be first shifted into the parser stack.
618 An obarray of semantic action symbols. A semantic action is actually
619 an Emacs Lisp function (lambda expression).
623 @node Conflicts, , Compiling a grammar, Wisent Grammar
624 @comment node-name, next, previous, up
627 Normally, a grammar should produce an automaton where at each state
628 the parser has only one action to do (@pxref{Wisent Parsing}).
630 @cindex ambiguous grammar
631 In certain cases, a grammar can produce an automaton where, at some
632 states, there are more than one action possible. Such a grammar is
633 @dfn{ambiguous}, and generates @dfn{conflicts}.
635 @cindex deterministic automaton
636 The parser can't be driven by an automaton which isn't completely
637 @dfn{deterministic}, that is which contains conflicts. It is
638 necessary to resolve the conflicts to eliminate them. Wisent resolves
639 conflicts like Bison does.
641 @cindex grammar conflicts
642 @cindex conflicts resolution
643 There are two sorts of conflicts:
646 @cindex shift/reduce conflicts
647 @item shift/reduce conflicts
648 When either a shift or a reduction would be valid at the same state.
650 Such conflicts are resolved by choosing to shift, unless otherwise
651 directed by operator precedence declarations.
652 See @ref{Shift/Reduce , , , bison}, in the Bison manual for more
655 @cindex reduce/reduce conflicts
656 @item reduce/reduce conflicts
657 That occurs if there are two or more rules that apply to the same
658 sequence of input. This usually indicates a serious error in the
661 Such conflicts are resolved by choosing to use the rule that appears
662 first in the grammar, but it is very risky to rely on this. Every
663 reduce/reduce conflict must be studied and usually eliminated. See
664 @ref{Reduce/Reduce , , , bison}, in the Bison manual for more
669 * Grammar Debugging::
670 * Understanding the automaton::
673 @node Grammar Debugging
674 @subsection Grammar debugging
676 @cindex grammar debugging
677 @cindex grammar verbose description
678 To help writing a new grammar, @code{wisent-compile-grammar} can
679 produce a verbose report containing a detailed description of the
680 grammar and parser (equivalent to what Bison reports with the
681 @option{--verbose} option).
683 To enable the verbose report you can set to non-@code{nil} the
686 @vindex wisent-verbose-flag
687 @deffn Option wisent-verbose-flag
688 non-@code{nil} means to report verbose information on generated parser.
691 Or interactively use the command:
693 @findex wisent-toggle-verbose-flag
694 @deffn Command wisent-toggle-verbose-flag
695 Toggle whether to report verbose information on generated parser.
698 The verbose report is printed in the temporary buffer
699 @code{*wisent-log*} when running interactively, or in file
700 @file{wisent.output} when running in batch mode. Different
701 reports are separated from each other by a line like this:
705 *** Wisent @var{source-file} - 2002-06-27 17:33
709 where @var{source-file} is the name of the Emacs Lisp file from which
710 the grammar was read. See @ref{Understanding the automaton}, for
711 details on the verbose report.
715 To help debugging the grammar compiler itself, you can set this
716 variable to print the content of some internal data structures:
718 @vindex wisent-debug-flag
719 @defvar wisent-debug-flag
720 non-@code{nil} means enable some debug stuff.
724 @node Understanding the automaton
725 @subsection Understanding the automaton
727 @cindex understanding the automaton
728 This section (took from the manual of Bison 1.49) describes how to use
729 the verbose report printed by @code{wisent-compile-grammar} to
730 understand the generated automaton, to tune or fix a grammar.
732 We will use the following example:
736 (let ((wisent-verbose-flag t)) ;; Print a verbose report!
737 (wisent-compile-grammar
738 '((NUM STR) ; %token NUM STR
740 ((left ?+ ?-) ; %left '+' '-';
741 (left ?*)) ; %left '*'
744 ((exp ?+ exp)) ; exp '+' exp
745 ((exp ?- exp)) ; | exp '-' exp
746 ((exp ?* exp)) ; | exp '*' exp
747 ((exp ?/ exp)) ; | exp '/' exp
755 'nil) ; no %start declarations
760 When evaluating the above expression, grammar compilation first issues
761 the following two clear messages:
765 Grammar contains 1 useless nonterminals and 1 useless rules
766 Grammar contains 7 shift/reduce conflicts
770 The @samp{*wisent-log*} buffer details things!
772 The first section reports conflicts that were solved using precedence
773 and/or associativity:
777 Conflict in state 7 between rule 1 and token '+' resolved as reduce.
778 Conflict in state 7 between rule 1 and token '-' resolved as reduce.
779 Conflict in state 7 between rule 1 and token '*' resolved as shift.
780 Conflict in state 8 between rule 2 and token '+' resolved as reduce.
781 Conflict in state 8 between rule 2 and token '-' resolved as reduce.
782 Conflict in state 8 between rule 2 and token '*' resolved as shift.
783 Conflict in state 9 between rule 3 and token '+' resolved as reduce.
784 Conflict in state 9 between rule 3 and token '-' resolved as reduce.
785 Conflict in state 9 between rule 3 and token '*' resolved as reduce.
789 The next section reports useless tokens, nonterminal and rules (note
790 that useless tokens might be used by the scanner):
794 Useless nonterminals:
799 Terminals which are not used:
810 The next section lists states that still have conflicts:
814 State 7 contains 1 shift/reduce conflict.
815 State 8 contains 1 shift/reduce conflict.
816 State 9 contains 1 shift/reduce conflict.
817 State 10 contains 4 shift/reduce conflicts.
821 The next section reproduces the grammar used:
836 And reports the uses of the symbols:
840 Terminals, with rules where they appear
852 Nonterminals, with rules where they appear
855 on left: 1 2 3 4 5, on right: 1 2 3 4
859 The report then details the automaton itself, describing each state
860 with it set of @dfn{items}, also known as @dfn{pointed rules}. Each
861 item is a production rule together with a point (marked by @samp{.})
862 that the input cursor.
868 NUM shift, and go to state 1
874 State 0 corresponds to being at the very beginning of the parsing, in
875 the initial rule, right before the start symbol (@samp{exp}). When
876 the parser returns to this state right after having reduced a rule
877 that produced an @samp{exp}, it jumps to state 2. If there is no such
878 transition on a nonterminal symbol, and the lookahead is a @samp{NUM},
879 then this token is shifted on the parse stack, and the control flow
880 jumps to state 1. Any other lookahead triggers a parse error.
888 exp -> NUM . (rule 5)
890 $default reduce using rule 5 (exp)
894 the rule 5, @samp{exp: NUM;}, is completed. Whatever the lookahead
895 (@samp{$default}), the parser will reduce it. If it was coming from
896 state 0, then, after this reduction it will return to state 0, and
897 will jump to state 2 (@samp{exp: go to state 2}).
903 exp -> exp . '+' exp (rule 1)
904 exp -> exp . '-' exp (rule 2)
905 exp -> exp . '*' exp (rule 3)
906 exp -> exp . '/' exp (rule 4)
908 $EOI shift, and go to state 11
909 '+' shift, and go to state 3
910 '-' shift, and go to state 4
911 '*' shift, and go to state 5
912 '/' shift, and go to state 6
916 In state 2, the automaton can only shift a symbol. For instance,
917 because of the item @samp{exp -> exp . '+' exp}, if the lookahead if
918 @samp{+}, it will be shifted on the parse stack, and the automaton
919 control will jump to state 3, corresponding to the item
920 @samp{exp -> exp . '+' exp}:
926 exp -> exp '+' . exp (rule 1)
928 NUM shift, and go to state 1
934 Since there is no default action, any other token than those listed
935 above will trigger a parse error.
937 The interpretation of states 4 to 6 is straightforward:
943 exp -> exp '-' . exp (rule 2)
945 NUM shift, and go to state 1
953 exp -> exp '*' . exp (rule 3)
955 NUM shift, and go to state 1
963 exp -> exp '/' . exp (rule 4)
965 NUM shift, and go to state 1
971 As was announced in beginning of the report, @samp{State 7 contains 1
972 shift/reduce conflict.}:
978 exp -> exp . '+' exp (rule 1)
979 exp -> exp '+' exp . (rule 1)
980 exp -> exp . '-' exp (rule 2)
981 exp -> exp . '*' exp (rule 3)
982 exp -> exp . '/' exp (rule 4)
984 '*' shift, and go to state 5
985 '/' shift, and go to state 6
987 '/' [reduce using rule 1 (exp)]
988 $default reduce using rule 1 (exp)
992 Indeed, there are two actions associated to the lookahead @samp{/}:
993 either shifting (and going to state 6), or reducing rule 1. The
994 conflict means that either the grammar is ambiguous, or the parser
995 lacks information to make the right decision. Indeed the grammar is
996 ambiguous, as, since we did not specify the precedence of @samp{/},
997 the sentence @samp{NUM + NUM / NUM} can be parsed as @samp{NUM + (NUM
998 / NUM)}, which corresponds to shifting @samp{/}, or as @samp{(NUM +
999 NUM) / NUM}, which corresponds to reducing rule 1.
1001 Because in @acronym{LALR(1)} parsing a single decision can be made,
1002 Wisent arbitrarily chose to disable the reduction, see
1003 @ref{Conflicts}. Discarded actions are reported in between square
1006 Note that all the previous states had a single possible action: either
1007 shifting the next token and going to the corresponding state, or
1008 reducing a single rule. In the other cases, i.e., when shifting
1009 @emph{and} reducing is possible or when @emph{several} reductions are
1010 possible, the lookahead is required to select the action. State 7 is
1011 one such state: if the lookahead is @samp{*} or @samp{/} then the
1012 action is shifting, otherwise the action is reducing rule 1. In other
1013 words, the first two items, corresponding to rule 1, are not eligible
1014 when the lookahead is @samp{*}, since we specified that @samp{*} has
1015 higher precedence that @samp{+}. More generally, some items are
1016 eligible only with some set of possible lookaheads.
1018 States 8 to 10 are similar:
1024 exp -> exp . '+' exp (rule 1)
1025 exp -> exp . '-' exp (rule 2)
1026 exp -> exp '-' exp . (rule 2)
1027 exp -> exp . '*' exp (rule 3)
1028 exp -> exp . '/' exp (rule 4)
1030 '*' shift, and go to state 5
1031 '/' shift, and go to state 6
1033 '/' [reduce using rule 2 (exp)]
1034 $default reduce using rule 2 (exp)
1040 exp -> exp . '+' exp (rule 1)
1041 exp -> exp . '-' exp (rule 2)
1042 exp -> exp . '*' exp (rule 3)
1043 exp -> exp '*' exp . (rule 3)
1044 exp -> exp . '/' exp (rule 4)
1046 '/' shift, and go to state 6
1048 '/' [reduce using rule 3 (exp)]
1049 $default reduce using rule 3 (exp)
1055 exp -> exp . '+' exp (rule 1)
1056 exp -> exp . '-' exp (rule 2)
1057 exp -> exp . '*' exp (rule 3)
1058 exp -> exp . '/' exp (rule 4)
1059 exp -> exp '/' exp . (rule 4)
1061 '+' shift, and go to state 3
1062 '-' shift, and go to state 4
1063 '*' shift, and go to state 5
1064 '/' shift, and go to state 6
1066 '+' [reduce using rule 4 (exp)]
1067 '-' [reduce using rule 4 (exp)]
1068 '*' [reduce using rule 4 (exp)]
1069 '/' [reduce using rule 4 (exp)]
1070 $default reduce using rule 4 (exp)
1074 Observe that state 10 contains conflicts due to the lack of precedence
1075 of @samp{/} wrt @samp{+}, @samp{-}, and @samp{*}, but also because the
1076 associativity of @samp{/} is not specified.
1078 Finally, the state 11 (plus 12) is named the @dfn{final state}, or the
1079 @dfn{accepting state}:
1085 $EOI shift, and go to state 12
1095 The end of input is shifted @samp{$EOI shift,} and the parser exits
1096 successfully (@samp{go to state 12}, that terminates).
1098 @node Wisent Parsing
1099 @chapter Wisent Parsing
1101 @cindex bottom-up parser
1102 @cindex shift-reduce parser
1103 The Wisent's parser is what is called a @dfn{bottom-up} or
1104 @dfn{shift-reduce} parser which repeatedly:
1109 That is pushes the value of the last lexical token read (the
1110 look-ahead token) into a value stack, and reads a new one.
1114 That is replaces a nonterminal by its semantic value. The values of
1115 the components which form the right hand side of a rule are popped
1116 from the value stack and reduced by the semantic action of this rule.
1117 The result is pushed back on top of value stack.
1120 The parser will stop on:
1125 When all input has been successfully parsed. The semantic value of
1126 the start nonterminal is on top of the value stack.
1128 @cindex syntax error
1130 When a syntax error (an unexpected token in input) has been detected.
1131 At this point the parser issues an error message and either stops or
1132 calls a recovery routine to try to resume parsing.
1135 @cindex table-driven parser
1136 The above elementary actions are driven by the @acronym{LALR(1)}
1137 automaton built by @code{wisent-compile-grammar} from a context-free
1140 The Wisent's parser is entered by calling the function:
1142 @findex wisent-parse
1143 @defun wisent-parse automaton lexer &optional error start
1144 Parse input using the automaton specified in @var{automaton}.
1148 Is an @acronym{LALR(1)} automaton generated by
1149 @code{wisent-compile-grammar} (@pxref{Wisent Grammar}).
1152 Is a function with no argument called by the parser to obtain the next
1153 terminal (token) in input (@pxref{Writing a lexer}).
1156 Is an optional reporting function called when a parse error occurs.
1157 It receives a message string to report. It defaults to the function
1158 @code{wisent-message} (@pxref{Report errors}).
1161 Specify the start symbol (nonterminal) used by the parser as its goal.
1162 It defaults to the start symbol defined in the grammar
1163 (@pxref{Wisent Grammar}).
1167 The following two normal hooks permit to do some useful processing
1168 respectively before to start parsing, and after the parser terminated.
1170 @vindex wisent-pre-parse-hook
1171 @defvar wisent-pre-parse-hook
1172 Normal hook run just before entering the @var{LR} parser engine.
1175 @vindex wisent-post-parse-hook
1176 @defvar wisent-post-parse-hook
1177 Normal hook run just after the @var{LR} parser engine terminated.
1185 * Debugging actions::
1188 @node Writing a lexer
1189 @section What the parser must receive
1191 It is important to understand that the parser does not parse
1192 characters, but lexical tokens, and does not know anything about
1193 characters in text streams!
1195 @cindex lexical analysis
1198 Reading input data to produce lexical tokens is performed by a lexer
1199 (also called a scanner) in a lexical analysis step, before the syntax
1200 analysis step performed by the parser. The parser automatically calls
1201 the lexer when it needs the next token to parse.
1203 @cindex lexical tokens
1204 A Wisent's lexer is an Emacs Lisp function with no argument. It must
1205 return a valid lexical token of the form:
1207 @code{(@var{token-class value} [@var{start} . @var{end}])}
1211 Is a category of lexical token identifying a terminal as specified in
1212 the grammar (@pxref{Wisent Grammar}). It can be a symbol or a character
1216 Is the value of the lexical token. It can be of any valid Emacs Lisp
1221 Are the optional beginning and ending positions of @var{value} in the
1225 When there are no more tokens to read the lexer must return the token
1226 @code{(list wisent-eoi-term)} to each request.
1228 @vindex wisent-eoi-term
1229 @defvar wisent-eoi-term
1230 Predefined constant, End-Of-Input terminal symbol.
1233 @code{wisent-lex} is an example of a lexer that reads lexical tokens
1234 produced by a @semantic{} lexer, and translates them into lexical tokens
1235 suitable to the Wisent parser. See also @ref{Wisent Lex}.
1237 To call the lexer in a semantic action use the function
1238 @code{wisent-lexer}. See also @ref{Actions goodies}.
1240 @node Actions goodies
1241 @section Variables and macros useful in grammar actions.
1243 @vindex wisent-input
1244 @defvar wisent-input
1245 The last token read.
1246 This variable only has meaning in the scope of @code{wisent-parse}.
1249 @findex wisent-lexer
1251 Obtain the next terminal in input.
1254 @findex wisent-region
1255 @defun wisent-region &rest positions
1256 Return the start/end positions of the region including
1257 @var{positions}. Each element of @var{positions} is a pair
1258 @w{@code{(@var{start-pos} . @var{end-pos})}} or @code{nil}. The
1259 returned value is the pair @w{@code{(@var{min-start-pos} .
1260 @var{max-end-pos})}} or @code{nil} if no @var{positions} are
1265 @section The error reporting function
1267 @cindex error reporting
1268 When the parser encounters a syntax error it calls a user-defined
1269 function. It must be an Emacs Lisp function with one argument: a
1270 string containing the message to report.
1272 By default the parser uses this function to report error messages:
1274 @findex wisent-message
1275 @defun wisent-message string &rest args
1276 Print a one-line message if @code{wisent-parse-verbose-flag} is set.
1277 Pass @var{string} and @var{args} arguments to @dfn{message}.
1282 @code{wisent-message} uses the following function to print lexical
1285 @defun wisent-token-to-string token
1286 Return a printed representation of lexical token @var{token}.
1289 The general printed form of a lexical token is:
1291 @w{@code{@var{token}(@var{value})@@@var{location}}}
1294 To control the verbosity of the parser you can set to non-@code{nil}
1297 @vindex wisent-parse-verbose-flag
1298 @deffn Option wisent-parse-verbose-flag
1299 non-@code{nil} means to issue more messages while parsing.
1302 Or interactively use the command:
1304 @findex wisent-parse-toggle-verbose-flag
1305 @deffn Command wisent-parse-toggle-verbose-flag
1306 Toggle whether to issue more messages while parsing.
1309 When the error reporting function is entered the variable
1310 @code{wisent-input} contains the unexpected token as returned by the
1313 The error reporting function can be called from a semantic action too
1314 using the special macro @code{wisent-error}. When called from a
1315 semantic action entered by error recovery (@pxref{Error recovery}) the
1316 value of the variable @code{wisent-recovering} is non-@code{nil}.
1318 @node Error recovery
1319 @section Error recovery
1321 @cindex error recovery
1322 The error recovery mechanism of the Wisent's parser conforms to the
1323 one Bison uses. See @ref{Error Recovery, , , bison}, in the Bison
1327 To recover from a syntax error you must write rules to recognize the
1328 special token @code{error}. This is a terminal symbol that is
1329 automatically defined and reserved for error handling.
1331 When the parser encounters a syntax error, it pops the state stack
1332 until it finds a state that allows shifting the @code{error} token.
1333 After it has been shifted, if the old look-ahead token is not
1334 acceptable to be shifted next, the parser reads tokens and discards
1335 them until it finds a token which is acceptable.
1337 @cindex error recovery strategy
1338 Strategies for error recovery depend on the choice of error rules in
1339 the grammar. A simple and useful strategy is simply to skip the rest
1340 of the current statement if an error is detected:
1344 (statement (( error ?; )) ;; on error, skip until ';' is read
1349 It is also useful to recover to the matching close-delimiter of an
1350 opening-delimiter that has already been parsed:
1354 (primary (( ?@{ expr ?@} ))
1361 @cindex error recovery actions
1362 Note that error recovery rules may have actions, just as any other
1363 rules can. Here are some predefined hooks, variables, functions or
1364 macros, useful in such actions:
1366 @vindex wisent-nerrs
1367 @defvar wisent-nerrs
1368 The number of parse errors encountered so far.
1371 @vindex wisent-recovering
1372 @defvar wisent-recovering
1373 non-@code{nil} means that the parser is recovering.
1374 This variable only has meaning in the scope of @code{wisent-parse}.
1377 @findex wisent-error
1378 @defun wisent-error msg
1379 Call the user supplied error reporting function with message
1380 @var{msg} (@pxref{Report errors}).
1382 For an example of use, @xref{wisent-skip-token}.
1385 @findex wisent-errok
1387 Resume generating error messages immediately for subsequent syntax
1390 The parser suppress error message for syntax errors that happens
1391 shortly after the first, until three consecutive input tokens have
1392 been successfully shifted.
1394 Calling @code{wisent-errok} in an action, make error messages resume
1395 immediately. No error messages will be suppressed if you call it in
1396 an error rule's action.
1398 For an example of use, @xref{wisent-skip-token}.
1401 @findex wisent-clearin
1402 @defun wisent-clearin
1403 Discard the current lookahead token.
1404 This will cause a new lexical token to be read.
1406 In an error rule's action the previous lookahead token is reanalyzed
1407 immediately. @code{wisent-clearin} may be called to clear this token.
1409 For example, suppose that on a parse error, an error handling routine
1410 is called that advances the input stream to some point where parsing
1411 should once again commence. The next symbol returned by the lexical
1412 scanner is probably correct. The previous lookahead token ought to
1413 be discarded with @code{wisent-clearin}.
1415 For an example of use, @xref{wisent-skip-token}.
1418 @findex wisent-abort
1420 Abort parsing and save the lookahead token.
1423 @findex wisent-set-region
1424 @defun wisent-set-region start end
1425 Change the region of text matched by the current nonterminal.
1426 @var{start} and @var{end} are respectively the beginning and end
1427 positions of the region occupied by the group of components associated
1428 to this nonterminal. If @var{start} or @var{end} values are not a
1429 valid positions the region is set to @code{nil}.
1431 For an example of use, @xref{wisent-skip-token}.
1434 @vindex wisent-discarding-token-functions
1435 @defvar wisent-discarding-token-functions
1436 List of functions to be called when discarding a lexical token.
1437 These functions receive the lexical token discarded.
1438 When the parser encounters unexpected tokens, it can discards them,
1439 based on what directed by error recovery rules. Either when the
1440 parser reads tokens until one is found that can be shifted, or when an
1441 semantic action calls the function @code{wisent-skip-token} or
1442 @code{wisent-skip-block}.
1443 For language specific hooks, make sure you define this as a local
1446 For example, in @semantic{}, this hook is set to the function
1447 @code{wisent-collect-unmatched-syntax} to collect unmatched lexical
1448 tokens (@pxref{Useful functions}).
1451 @findex wisent-skip-token
1452 @defun wisent-skip-token
1453 @anchor{wisent-skip-token}
1454 Skip the lookahead token in order to resume parsing.
1456 Must be used in error recovery semantic actions.
1458 It typically looks like this:
1462 (wisent-message "%s: skip %s" $action
1463 (wisent-token-to-string wisent-input))
1465 'wisent-discarding-token-functions wisent-input)
1472 @findex wisent-skip-block
1473 @defun wisent-skip-block
1474 Safely skip a block in order to resume parsing.
1476 Must be used in error recovery semantic actions.
1478 A block is data between an open-delimiter (syntax class @code{(}) and
1479 a matching close-delimiter (syntax class @code{)}):
1483 (a parenthesized block)
1484 [a block between brackets]
1485 @{a block between braces@}
1489 The following example uses @code{wisent-skip-block} to safely skip a
1490 block delimited by @samp{LBRACE} (@code{@{}) and @samp{RBRACE}
1491 (@code{@}}) tokens, when a syntax error occurs in
1492 @samp{other-components}:
1496 (block ((LBRACE other-components RBRACE))
1499 (wisent-skip-block))
1505 @node Debugging actions
1506 @section Debugging semantic actions
1508 @cindex semantic action symbols
1509 Each semantic action is represented by a symbol interned in an
1510 @dfn{obarray} that is part of the @acronym{LALR(1)} automaton
1511 (@pxref{Compiling a grammar}). @code{symbol-function} on a semantic
1512 action symbol return the semantic action lambda expression.
1514 A semantic action symbol name has the form
1515 @code{@var{nonterminal}:@var{index}}, where @var{nonterminal} is the
1516 name of the nonterminal symbol the action belongs to, and @var{index}
1517 is an action sequence number within the scope of @var{nonterminal}.
1518 For example, this nonterminal definition:
1523 line [@code{input:0}]
1525 (format "%s %s" $1 $2) [@code{input:1}]
1530 Will produce two semantic actions, and associated symbols:
1534 A default action that returns @code{$1}.
1537 That returns @code{(format "%s %s" $1 $2)}.
1540 @cindex debugging semantic actions
1541 Debugging uses the Lisp debugger to investigate what is happening
1542 during execution of semantic actions.
1543 Three commands are available to debug semantic actions. They receive
1547 @item The automaton that contains the semantic action.
1549 @item The semantic action symbol.
1552 @findex wisent-debug-on-entry
1553 @deffn Command wisent-debug-on-entry automaton function
1554 Request @var{automaton}'s @var{function} to invoke debugger each time it is called.
1555 @var{function} must be a semantic action symbol that exists in @var{automaton}.
1558 @findex wisent-cancel-debug-on-entry
1559 @deffn Command wisent-cancel-debug-on-entry automaton function
1560 Undo effect of @code{wisent-debug-on-entry} on @var{automaton}'s @var{function}.
1561 @var{function} must be a semantic action symbol that exists in @var{automaton}.
1564 @findex wisent-debug-show-entry
1565 @deffn Command wisent-debug-show-entry automaton function
1566 Show the source of @var{automaton}'s semantic action @var{function}.
1567 @var{function} must be a semantic action symbol that exists in @var{automaton}.
1570 @node Wisent Semantic
1571 @chapter How to use Wisent with Semantic
1574 This section presents how the Wisent's parser can be used to produce
1575 @dfn{tags} for the @semantic{} tool set.
1577 @semantic{} tags form a hierarchy of Emacs Lisp data structures that
1578 describes a program in a way independent of programming languages.
1579 Tags map program declarations, like functions, methods, variables,
1580 data types, classes, includes, grammar rules, etc..
1582 @cindex WY grammar format
1583 To use the Wisent parser with @semantic{} you have to define
1584 your grammar in @dfn{WY} form, a grammar format very close
1585 to the one used by Bison.
1587 Please @inforef{top, Semantic Grammar Framework Manual, grammar-fw}
1588 for more information on @semantic{} grammars.
1595 @node Grammar styles
1596 @section Grammar styles
1598 @cindex grammar styles
1599 @semantic{} parsing heavily depends on how you wrote the grammar.
1600 There are mainly two styles to write a Wisent's grammar intended to be
1601 used with the @semantic{} tool set: the @dfn{Iterative style} and the
1602 @dfn{Bison style}. Each one has pros and cons, and in certain cases
1603 it can be worth a mix of the two styles!
1609 * Start nonterminals::
1610 * Useful functions::
1613 @node Iterative style, Bison style, Grammar styles, Grammar styles
1614 @subsection Iterative style
1616 @cindex grammar iterative style
1617 The @dfn{iterative style} is the preferred style to use with @semantic{}.
1618 It relies on an iterative parser back-end mechanism which parses start
1619 nonterminals one at a time and automagically skips unexpected lexical
1622 Compared to rule-based iterative functions (@pxref{Bison style}),
1623 iterative parsers are better in that they can handle obscure errors
1627 Each start nonterminal must produces a @dfn{raw tag} by calling a
1628 @code{TAG}-like grammar macro with appropriate parameters. See also
1629 @ref{Start nonterminals}.
1631 @cindex expanded tag
1632 Then, each parsing iteration automatically translates a raw tag into
1633 @dfn{expanded tags}, updating the raw tag structure with internal
1634 properties and buffer related data.
1636 After parsing completes, it results in a tree of expanded tags.
1638 The following example is a snippet of the iterative style Java grammar
1639 provided in the @semantic{} distribution in the file
1640 @file{semantic/wisent/java-tags.wy}.
1645 ;; Alternate entry points
1646 ;; - Needed by partial re-parse
1647 %start formal_parameter
1649 ;; - Needed by EXPANDFULL clauses
1650 %start formal_parameters
1653 formal_parameter_list
1655 (EXPANDFULL $1 formal_parameters)
1663 | formal_parameter COMMA
1664 | formal_parameter RPAREN
1668 : formal_parameter_modifier_opt type variable_declarator_id
1669 (VARIABLE-TAG $3 $2 nil :typemodifiers $1)
1675 It shows the use of the @code{EXPANDFULL} grammar macro to parse a
1676 @samp{PAREN_BLOCK} which contains a @samp{formal_parameter_list}.
1677 @code{EXPANDFULL} tells to recursively parse @samp{formal_parameters}
1678 inside @samp{PAREN_BLOCK}. The parser iterates until it digested all
1679 available input data inside the @samp{PAREN_BLOCK}, trying to match
1680 any of the @samp{formal_parameters} rules:
1687 @item @samp{formal_parameter COMMA}
1689 @item @samp{formal_parameter RPAREN}
1692 At each iteration it will return a @samp{formal_parameter} raw tag,
1693 or @code{nil} to skip unwanted (single @samp{LPAREN} or @samp{RPAREN}
1694 for example) or unexpected input data. Those raw tags will be
1695 automatically expanded by the iterative back-end parser.
1698 @subsection Bison style
1700 @cindex grammar bison style
1701 What we call the @dfn{Bison style} is the traditional style of Bison's
1702 grammars. Compared to iterative style, it is not straightforward to
1703 use grammars written in Bison style in @semantic{}. Mainly because such
1704 grammars are designed to parse the whole input data in one pass, and
1705 don't use the iterative parser back-end mechanism (@pxref{Iterative
1706 style}). With Bison style the parser is called once to parse the
1707 grammar start nonterminal.
1709 The following example is a snippet of the Bison style Java grammar
1710 provided in the @semantic{} distribution in the file
1711 @file{semantic/wisent/java.wy}.
1715 %start formal_parameter
1718 formal_parameter_list
1719 : formal_parameter_list COMMA formal_parameter
1726 : formal_parameter_modifier_opt type variable_declarator_id
1728 (VARIABLE-TAG $3 $2 :typemodifiers $1)
1734 The first consequence is that syntax errors are not automatically
1735 handled by @semantic{}. Thus, it is necessary to explicitly handle
1736 them at the grammar level, providing error recovery rules to skip
1737 unexpected input data.
1739 The second consequence is that the iterative parser can't do automatic
1740 tag expansion, except for the start nonterminal value. It is
1741 necessary to explicitly expand tags from concerned semantic actions by
1742 calling the grammar macro @code{EXPANDTAG} with a raw tag as
1743 parameter. See also @ref{Start nonterminals}, for incremental
1744 re-parse considerations.
1747 @subsection Mixed style
1749 @cindex grammar mixed style
1754 %start prologue epilogue declaration nonterminal rule
1769 SYMBOL COLON rules SEMI
1770 (TAG $1 'nonterminal :children $3)
1775 (apply 'nconc (nreverse $1))
1788 name type comps prec action elt)
1791 (TAG name 'rule :type type :value comps :prec prec :expr action)
1797 This example shows how iterative and Bison styles can be combined in
1798 the same grammar to obtain a good compromise between grammar
1799 complexity and an efficient parsing strategy in an interactive
1802 @samp{nonterminal} is parsed using iterative style via the main
1803 @samp{grammar} rule. The semantic action uses the @code{TAG} macro to
1804 produce a raw tag, automagically expanded by @semantic{}.
1806 But @samp{rules} part is parsed in Bison style! Why?
1808 Rule delimiters are the colon (@code{:}), that follows the nonterminal
1809 name, and a final semicolon (@code{;}). Unfortunately these
1810 delimiters are not @code{open-paren}/@code{close-paren} type, and the
1811 Emacs' syntactic analyzer can't easily isolate data between them to
1812 produce a @samp{RULES_PART} parenthesis-block-like lexical token.
1813 Consequently it is not possible to use @code{EXPANDFULL} to iterate in
1814 @samp{RULES_PART}, like this:
1819 SYMBOL COLON rules SEMI
1820 (TAG $1 'nonterminal :children $3)
1824 RULES_PART ;; @strong{Map a parenthesis-block-like lexical token}
1825 (EXPANDFULL $1 'rules)
1838 name type comps prec action elt)
1840 (TAG name 'rule :type type :value comps :prec prec :expr action)
1846 In such cases, when it is difficult for Emacs to obtain
1847 parenthesis-block-like lexical tokens, the best solution is to use the
1848 traditional Bison style with error recovery!
1850 In some extreme cases, it can also be convenient to extend the lexer,
1851 to deliver new lexical tokens, to simplify the grammar.
1853 @node Start nonterminals
1854 @subsection Start nonterminals
1856 @cindex start nonterminals
1857 @cindex @code{reparse-symbol} property
1858 When you write a grammar for @semantic{}, it is important to carefully
1859 indicate the start nonterminals. Each one defines an entry point in
1860 the grammar, and after parsing its semantic value is returned to the
1861 back-end iterative engine. Consequently:
1863 @strong{The semantic value of a start nonterminal must be a produced
1864 by a TAG like grammar macro}.
1866 Start nonterminals are declared by @code{%start} statements. When
1867 nothing is specified the first nonterminal that appears in the grammar
1868 is the start nonterminal.
1870 Generally, the following nonterminals must be declared as start
1874 @item The main grammar entry point
1879 @item nonterminals passed to @code{EXPAND}/@code{EXPANDFULL}
1881 These grammar macros recursively parse a part of input data, based on
1882 rules of the given nonterminal.
1884 For example, the following will parse @samp{PAREN_BLOCK} data using
1885 the @samp{formal_parameters} rules:
1889 formal_parameter_list
1891 (EXPANDFULL $1 formal_parameters)
1896 The semantic value of @samp{formal_parameters} becomes the value of
1897 the @code{EXPANDFULL} expression. It is a list of @semantic{} tags
1898 spliced in the tags tree.
1900 Because the automaton must know that @samp{formal_parameters} is a
1901 start symbol, you must declare it like this:
1905 %start formal_parameters
1911 @cindex incremental re-parse
1912 @cindex reparse-symbol
1913 The @code{EXPANDFULL} macro has a side effect it is important to know,
1914 related to the incremental re-parse mechanism of @semantic{}: the
1915 nonterminal symbol parameter passed to @code{EXPANDFULL} also becomes
1916 the @code{reparse-symbol} property of the tag returned by the
1917 @code{EXPANDFULL} expression.
1919 When buffer's data mapped by a tag is modified, @semantic{}
1920 schedules an incremental re-parse of that data, using the tag's
1921 @code{reparse-symbol} property as start nonterminal.
1923 @strong{The rules associated to such start symbols must be carefully
1924 reviewed to ensure that the incremental parser will work!}
1926 Things are a little bit different when the grammar is written in Bison
1929 @strong{The @code{reparse-symbol} property is set to the nonterminal
1930 symbol the rule that explicitly uses @code{EXPANDTAG} belongs to.}
1939 name type comps prec action elt)
1942 (TAG name 'rule :type type :value comps :prec prec :expr action)
1948 Set the @code{reparse-symbol} property of the expanded tag to
1949 @samp{rule}. A important consequence is that:
1951 @strong{Every nonterminal having any rule that calls @code{EXPANDTAG}
1952 in a semantic action, should be declared as a start symbol!}
1954 @node Useful functions
1955 @subsection Useful functions
1957 Here is a description of some predefined functions it might be useful
1958 to know when writing new code to use Wisent in @semantic{}:
1960 @findex wisent-collect-unmatched-syntax
1961 @defun wisent-collect-unmatched-syntax input
1962 Add @var{input} lexical token to the cache of unmatched tokens, in
1963 variable @code{semantic-unmatched-syntax-cache}.
1965 See implementation of the function @code{wisent-skip-token} in
1966 @ref{Error recovery}, for an example of use.
1970 @section The Wisent Lex lexer
1972 @findex semantic-lex
1973 The lexical analysis step of @semantic{} is performed by the general
1974 function @code{semantic-lex}. For more information, @inforef{Writing
1975 Lexers, ,semantic-langdev}.
1977 @code{semantic-lex} produces lexical tokens of the form:
1981 @code{(@var{token-class start} . @var{end})}
1987 Is a symbol that identifies a lexical token class, like @code{symbol},
1988 @code{string}, @code{number}, or @code{PAREN_BLOCK}.
1992 Are the start and end positions of mapped data in the input buffer.
1995 The Wisent's parser doesn't depend on the nature of analyzed input
1996 stream (buffer, string, etc.), and requires that lexical tokens have a
1997 different form (@pxref{Writing a lexer}):
2001 @code{(@var{token-class value} [@var{start} . @var{end}])}
2005 @cindex lexical token mapping
2006 @code{wisent-lex} is the default Wisent's lexer used in @semantic{}.
2008 @vindex wisent-lex-istream
2011 Return the next available lexical token in Wisent's form.
2013 The variable @code{wisent-lex-istream} contains the list of lexical
2014 tokens produced by @code{semantic-lex}. Pop the next token available
2015 and convert it to a form suitable for the Wisent's parser.
2018 Mapping of lexical tokens as produced by @code{semantic-lex} into
2019 equivalent Wisent lexical tokens is straightforward:
2023 (@var{token-class start} . @var{end})
2024 @result{} (@var{token-class value start} . @var{end})
2028 @var{value} is the input @code{buffer-substring} from @var{start} to
2031 @node GNU Free Documentation License
2032 @appendix GNU Free Documentation License
2034 @include doclicense.texi
2047 @c Following comments are for the benefit of ispell.
2049 @c LocalWords: Wisent automagically wisent Wisent's LALR obarray