2 <title>CUP User's Manual</title>
7 <img src="cup_logo.gif" alt="[CUP Logo Image]">
9 <h1>CUP User's Manual</h1>
10 <h3><a href="http://www.cc.gatech.edu/gvu/people/Faculty/Scott.E.Hudson.html">
11 Scott E. Hudson</a><br>
12 <a href="http://www.cc.gatech.edu/gvu/gvutop.html">
13 Graphics Visualization and Usability Center</a><br>
14 <a href="http://www.gatech.edu/TechHome.html">
15 Georgia Institute of Technology</a></h3>
16 Modified by <a href="http://www.princeton.edu/~frankf">Frank
17 Flannery</a>, <a href="http://www.pdos.lcs.mit.edu/~cananian/">C. Scott Ananian</a>,
18 <a href="http://www.cs.princeton.edu/~danwang">Dan Wang</a> with advice from
19 <a href="http://www.cs.princeton.edu/~appel">Andrew W. Appel</a><br>
20 Last updated July 1999 (v0.10j)
23 <h3>Table of Contents</h3>
25 <dt> i. <dd> <a href="#about">About CUP Version 0.10</a>
26 <dt> 1. <dd> <a href="#intro">Introduction and Example</a>
27 <dt> 2. <dd> <a href="#spec">Specification Syntax</a>
28 <dt> 3. <dd> <a href="#running">Running CUP</a>
29 <dt> 4. <dd> <a href="#parser">Customizing the Parser</a>
30 <dt> 5. <dd> <a href="#scanner">Scanner interface</a>
31 <dt> 6. <dd> <a href="#errors">Error Recovery</a>
32 <dt> 7. <dd> <a href="#conclusion">Conclusion</a>
33 <dt> <dd> <a href="#refs">References</a>
34 <dt> A. <dd> <a href="#appendixa">Grammar for CUP Specification Files</a>
35 <dt> B. <dd> <a href="#appendixb">A Very Simple Example Scanner</a>
36 <dt> C. <dd> <a href="#changes">Incompatibilites between CUP 0.9 and CUP 0.10</a>
37 <dt> D. <dd> <a href="#bugs">Bugs</a>
38 <dt> E. <dd> <a href="#version">Change log</a>
42 <h3>i. About CUP Version 0.10</h3>
44 0.10 of CUP adds many new changes and features over the previous releases
45 of version 0.9. These changes attempt to make CUP more like its
46 predecessor, YACC. As a result, the old 0.9 parser specifications for CUP are
47 not compatible and a reading of <a href="#changes">appendix C</a> of the new
48 manual will be necessary to write new specifications. The new version,
49 however, gives the user more power and options, making parser specifications
53 <h3>1. Introduction and Example</h3></a>
55 This manual describes the basic operation and use of the
56 Java<a href="#trademark">(tm)</a>
57 Based Constructor of Useful Parsers (CUP for short).
58 CUP is a system for generating LALR parsers from simple specifications.
59 It serves the same role as the widely used program YACC
60 <a href="#YACCref">[1]</a> and in fact offers most of the features of YACC.
61 However, CUP is written in Java, uses specifications including embedded
62 Java code, and produces parsers which are implemented in Java.<p>
64 Although this manual covers all aspects of the CUP system, it is relatively
65 brief, and assumes you have at least a little bit of knowledge of LR
66 parsing. A working knowledge of YACC is also very helpful in
67 understanding how CUP specifications work.
68 A number of compiler construction textbooks (such as
69 <a href="#dragonbook">[2</a>,<a href="#crafting">3]</a>) cover this material,
70 and discuss the YACC system (which is quite similar to this one) as a
71 specific example. In addition, Andrew Appel's <a
72 href="http://www.cs.princeton.edu/~appel/modern/java/">Modern Compiler
73 Implementation in Java</a> textbook <a href="#modernjava">[4]</a> uses
74 and describes CUP in the context of compiler construction.
77 Using CUP involves creating a simple specification based on the
78 grammar for which a parser is needed, along with construction of a
79 scanner capable of breaking characters up into meaningful tokens (such
80 as keywords, numbers, and special symbols).<p>
82 As a simple example, consider a
83 system for evaluating simple arithmetic expressions over integers.
84 This system would read expressions from standard input (each terminated
85 with a semicolon), evaluate them, and print the result on standard output.
86 A grammar for the input to such a system might look like: <pre>
87 expr_list ::= expr_list expr_part | expr_part
88 expr_part ::= expr ';'
89 expr ::= expr '+' expr | expr '-' expr | expr '*' expr
90 | expr '/' expr | expr '%' expr | '(' expr ')'
93 To specify a parser based on this grammar, our first step is to identify and
94 name the set of terminal symbols that will appear on input, and the set of
95 non-terminal symbols. In this case, the non-terminals are:
97 <pre><tt> expr_list, expr_part </tt> and <tt> expr </tt>.</pre>
99 For terminal names we might choose:
101 <pre><tt> SEMI, PLUS, MINUS, TIMES, DIVIDE, MOD, NUMBER, LPAREN,</tt>
102 and <tt>RPAREN</tt></pre>
104 The experienced user will note a problem with the above grammar. It is
105 ambiguous. An ambiguous grammar is a grammar which, given a certain
106 input, can reduce the parts of the input in two different ways such as
107 to give two different answers. Take the above grammar, for
108 example. given the following input: <br>
109 <tt>3 + 4 * 6</tt><br>
110 The grammar can either evaluate the <tt>3 + 4</tt> and then multiply
111 seven by six, or it can evaluate <tt>4 * 6</tt> and then add three.
112 Older versions of CUP forced the user to write unambiguous grammars, but
113 now there is a construct allowing the user to specify precedences and
114 associativities for terminals. This means that the above ambiguous
115 grammar can be used, after specifying precedences and associativities.
116 There is more explanation later.
118 Based on these namings we can construct a small CUP specification
121 <pre><tt>// CUP specification for a simple expression evaluator (no actions)
123 import java_cup.runtime.*;
125 /* Preliminaries to set up and use the scanner. */
126 init with {: scanner.init(); :};
127 scan with {: return scanner.next_token(); :};
129 /* Terminals (tokens returned by the scanner). */
130 terminal SEMI, PLUS, MINUS, TIMES, DIVIDE, MOD;
131 terminal UMINUS, LPAREN, RPAREN;
132 terminal Integer NUMBER;
135 non terminal expr_list, expr_part;
136 non terminal Integer expr, term, factor;
139 precedence left PLUS, MINUS;
140 precedence left TIMES, DIVIDE, MOD;
141 precedence left UMINUS;
144 expr_list ::= expr_list expr_part |
146 expr_part ::= expr SEMI;
147 expr ::= expr PLUS expr
152 | MINUS expr %prec UMINUS
158 We will consider each part of the specification syntax in detail later.
159 However, here we can quickly see that the specification contains four
160 main parts. The first part provides preliminary and miscellaneous declarations
161 to specify how the parser is to be generated, and supply parts of the
162 runtime code. In this case we indicate that the <tt>java_cup.runtime</tt>
163 classes should be imported, then supply a small bit of initialization code,
164 and some code for invoking the scanner to retrieve the next input token.
165 The second part of the specification declares terminals and non-terminals,
166 and associates object classes with each. In this case, the terminals
167 are declared as either with no type, or of type
168 <tt>Integer</tt>. The specified type of the
169 terminal or non-terminal is the type of the value of those terminals or
170 non-terminals. If no type is specified, the terminal or non-terminal
171 carries no value. Here, no type indicates that these
172 terminals and non-terminals hold no value.
173 The third part specifies the precedence and
174 associativity of terminals. The last precedence declaration give its
175 terminals the highest precedence. The final
176 part of the specification contains the grammar.<p>
179 To produce a parser from this specification we use the CUP generator.
180 If this specification were stored in a file <tt>parser.cup</tt>, then
181 (on a Unix system at least) we might invoke CUP using a command like:
182 <pre><tt> java java_cup.Main < parser.cup</tt> </pre>
183 or (starting with CUP 0.10k):
184 <pre><tt> java java_cup.Main parser.cup</tt> </pre>
185 The system will produce two Java source files containing
186 parts of the generated parser: <tt>sym.java</tt> and <tt>parser.java</tt>
187 (these names can be changed with command-line options; see
188 <A HREF="#running">below</a>).
189 As you might expect, these two files contain declarations for the classes
190 <tt>sym</tt> and <tt>parser</tt>. The <tt>sym</tt> class contains a series of
191 constant declarations, one for each terminal symbol. This is typically used
192 by the scanner to refer to symbols (e.g. with code such as
193 "<tt>return new Symbol(sym.SEMI);</tt>" ). The <tt>parser</tt> class
194 implements the parser itself.<p>
196 The specification above, while constructing a full parser, does not perform
197 any semantic actions &emdash; it will only indicate success or failure of a parse.
198 To calculate and print values of each expression, we must embed Java
199 code within the parser to carry out actions at various points. In CUP,
200 actions are contained in <i>code strings</i> which are surrounded by delimiters
201 of the form <tt>{:</tt> and <tt>:}</tt> (we can see examples of this in the
202 <tt>init with</tt> and <tt>scan with</tt> clauses above). In general, the
203 system records all characters within the delimiters, but does not try to check
204 that it contains valid Java code.<p>
206 A more complete CUP specification for our example system (with actions
207 embedded at various points in the grammar) is shown below:<br>
209 <pre><tt>// CUP specification for a simple expression evaluator (w/ actions)
211 import java_cup.runtime.*;
213 /* Preliminaries to set up and use the scanner. */
214 init with {: scanner.init(); :};
215 scan with {: return scanner.next_token(); :};
217 /* Terminals (tokens returned by the scanner). */
218 terminal SEMI, PLUS, MINUS, TIMES, DIVIDE, MOD;
219 terminal UMINUS, LPAREN, RPAREN;
220 terminal Integer NUMBER;
223 non terminal expr_list, expr_part;
224 non terminal Integer expr;
227 precedence left PLUS, MINUS;
228 precedence left TIMES, DIVIDE, MOD;
229 precedence left UMINUS;
232 expr_list ::= expr_list expr_part
237 {: System.out.println("= " + e); :}
241 expr ::= expr:e1 PLUS expr:e2
242 {: RESULT = new Integer(e1.intValue() + e2.intValue()); :}
244 expr:e1 MINUS expr:e2
245 {: RESULT = new Integer(e1.intValue() - e2.intValue()); :}
247 expr:e1 TIMES expr:e2
248 {: RESULT = new Integer(e1.intValue() * e2.intValue()); :}
250 expr:e1 DIVIDE expr:e2
251 {: RESULT = new Integer(e1.intValue() / e2.intValue()); :}
254 {: RESULT = new Integer(e1.intValue() % e2.intValue()); :}
260 {: RESULT = new Integer(0 - e.intValue()); :}
268 Here we can see several changes. Most importantly, code to be executed at
269 various points in the parse is included inside code strings delimited by
270 <tt>{:</tt> and <tt>:}</tt>. In addition, labels have been placed on various
271 symbols in the right hand side of productions. For example in:<br>
272 <pre> expr:e1 PLUS expr:e2
273 {: RESULT = new Integer(e1.intValue() + e2.intValue()); :}
276 the first non-terminal <tt>expr</tt> has been labeled with <tt>e1</tt>, and
277 the second with <tt>e2</tt>. The left hand side value
278 of each production is always implicitly labeled as <tt>RESULT</tt>.<p>
280 Each symbol appearing in a production is represented at runtime by an
281 object of type <tt>Symbol</tt> on the parse stack. The labels refer to
282 the instance variable <tt>value</tt> in those objects. In the
283 expression <tt>expr:e1 PLUS expr:e2</tt>, <tt>e1</tt> and <tt>e2</tt>
284 refer to objects of type Integer. These objects are in the value fields
285 of the objects of type <tt>Symbol</tt> representing those non-terminals
286 on the parse stack. <tt>RESULT</tt> is of type <tt>Integer</tt> as
287 well, since the resulting non-terminal <tt>expr</tt> was declared as of
288 type <tt>Integer</tt>. This object becomes the <tt>value</tt> instance
289 variable of a new <tt>Symbol</tt> object.<p>
291 For each label, two more variables accessible to the user are declared.
292 A left and right value labels are passed to the code string, so that the
293 user can find out where the left and right side of each terminal or
294 non-terminal is in the input stream. The name of these variables is the
295 label name, plus <tt>left</tt> or <tt>right</tt>. for example, given
296 the right hand side of a production <tt>expr:e1 PLUS expr:e2</tt> the
297 user could not only access variables <tt>e1</tt> and <tt>e2</tt>, but
298 also <tt>e1left, e1right, e2left</tt> and <tt>e2right</tt>. these
299 variables are of type <tt>int</tt>.<p>
303 The final step in creating a working parser is to create a <i>scanner</i> (also
304 known as a <i>lexical analyzer</i> or simply a <i>lexer</i>). This routine is
305 responsible for reading individual characters, removing things things like
306 white space and comments, recognizing which terminal symbols from the
307 grammar each group of characters represents, then returning Symbol objects
308 representing these symbols to the parser.
309 The terminals will be retrieved with a call to the
310 scanner function. In the example, the parser will call
311 <tt>scanner.next_token()</tt>. The scanner should return objects of
312 type <tt>java_cup.runtime.Symbol</tt>. This type is very different than
313 older versions of CUP's <tt>java_cup.runtime.symbol</tt>. These Symbol
314 objects contains the instance variable <tt>value</tt> of type Object,
316 set by the lexer. This variable refers to the value of that symbol, and
317 the type of object in value should be of the same type as declared in
318 the <tt>terminal</tt> and <tt>non terminal</tt> declarations. In the
319 above example, if the lexer wished to pass a NUMBER token, it should
320 create a <tt>Symbol</tt> with the <tt>value</tt> instance variable
321 filled with an object of type <tt>Integer</tt>. <code>Symbol</code>
322 objects corresponding to terminals and non-terminals with no value
323 have a null value field.<p>
325 The code contained in the <tt>init with</tt> clause of the specification
326 will be executed before any tokens are requested. Each token will be
327 requested using whatever code is found in the <tt>scan with</tt> clause.
328 Beyond this, the exact form the scanner takes is up to you; however
329 note that each call to the scanner function should return a new
330 instance of <code>java_cup.runtime.Symbol</code> (or a subclass).
331 These symbol objects are annotated with parser information and pushed
332 onto a stack; reusing objects will result in the parser annotations
333 being scrambled. As of CUP 0.10j, <code>Symbol</code> reuse should be
334 detected if it occurs; the parser will throw an <code>Error</code>
335 telling you to fix your scanner.<p>
337 In the <a href="#spec">next section</a> a more detailed and formal
338 explanation of all parts of a CUP specification will be given.
339 <a href="#running">Section 3</a> describes options for running the
340 CUP system. <a href="#parser">Section 4</a> discusses the details
341 of how to customize a CUP parser, while <a href="#scanner">section 5</a>
342 discusses the scanner interface added in CUP 0.10j. <a href="#errors">Section
343 6</a> considers error recovery. Finally, <a href="#conclusion">Section 7</a>
344 provides a conclusion.
347 <h3>2. Specification Syntax</h3></a>
348 Now that we have seen a small example, we present a complete description of all
349 parts of a CUP specification. A specification has four sections with
350 a total of eight specific parts (however, most of these are optional).
351 A specification consists of:
353 <li> <a href="#package_spec">package and import specifications</a>,
354 <li> <a href="#code_part">user code components</a>,
355 <li> <a href="#symbol_list">symbol (terminal and non-terminal) lists</a>,
356 <li> <a href="#precedence">precedence declarations</a>, and
357 <li> <a href="#production_list">the grammar</a>.
359 Each of these parts must appear in the order presented here. (A complete
360 grammar for the specification language is given in
361 <a href="#appendixa">Appendix A</a>.) The particulars of each part of
362 the specification are described in the subsections below.<p>
364 <h5><a name="package_spec">Package and Import Specifications</a></h5>
366 A specification begins with optional <tt>package</tt> and <tt>import</tt>
367 declarations. These have the same syntax, and play the same
368 role, as the package and import declarations found in a normal Java program.
369 A package declaration is of the form:
371 <pre><tt> package <i>name</i>;</tt></pre>
373 where name <tt><i>name</i></tt> is a Java package identifier, possibly in
374 several parts separated by ".". In general, CUP employs Java lexical
375 conventions. So for example, both styles of Java comments are supported,
376 and identifiers are constructed beginning with a letter, dollar
377 sign ($), or underscore (_), which can then be followed by zero or more
378 letters, numbers, dollar signs, and underscores.<p>
380 After an optional <tt>package</tt> declaration, there can be zero or more
381 <tt>import</tt> declarations. As in a Java program these have the form:
383 <pre><tt> import <i>package_name.class_name</i>;</tt>
386 <pre><tt> import <i>package_name</i>.*;</tt>
389 The package declaration indicates what package the <tt>sym</tt> and
390 <tt>parser</tt> classes that are generated by the system will be in.
391 Any import declarations that appear in the specification will also appear
392 in the source file for the <tt>parser</tt> class allowing various names from
393 that package to be used directly in user supplied action code.
395 <h5><a name="code_part">User Code Components</a></h5>
397 Following the optional <tt>package</tt> and <tt>import</tt> declarations
398 are a series of optional declarations that allow user code to be included
399 as part of the generated parser (see <a href="#parser">Section 4</a> for a
400 full description of how the parser uses this code). As a part of the parser
401 file, a separate non-public class to contain all embedded user actions is
402 produced. The first <tt>action code</tt> declaration section allows code to
403 be included in this class. Routines and variables for use by the code
404 embedded in the grammar would normally be placed in this section (a typical
405 example might be symbol table manipulation routines). This declaration takes
408 <pre><tt> action code {: ... :};</tt>
411 where <tt>{: ... :}</tt> is a code string whose contents will be placed
412 directly within the <tt>action class</tt> class declaration.<p>
414 After the <tt>action code</tt> declaration is an optional
415 <tt>parser code</tt> declaration. This declaration allows methods and
416 variable to be placed directly within the generated parser class.
417 Although this is less common, it can be helpful when customizing the
418 parser &emdash; it is possible for example, to include scanning methods inside
419 the parser and/or override the default error reporting routines. This
420 declaration is very similar to the <tt>action code</tt> declaration and
423 <pre><tt> parser code {: ... :};</tt>
426 Again, code from the code string is placed directly into the generated parser
429 Next in the specification is the optional <tt>init</tt> declaration
432 <pre><tt> init with {: ... :};</tt></pre>
434 This declaration provides code that will be executed by the parser
435 before it asks for the first token. Typically, this is used to initialize
436 the scanner as well as various tables and other data structures that might
437 be needed by semantic actions. In this case, the code given in the code
438 string forms the body of a <tt>void</tt> method inside the <tt>parser</tt>
441 The final (optional) user code section of the specification indicates how
442 the parser should ask for the next token from the scanner. This has the
445 <pre><tt> scan with {: ... :};</tt></pre>
447 As with the <tt>init</tt> clause, the contents of the code string forms
448 the body of a method in the generated parser. However, in this case
449 the method returns an object of type <tt>java_cup.runtime.Symbol</tt>.
450 Consequently the code found in the <tt>scan with</tt> clause should
451 return such a value. See <a href="#scanner">section 5</a> for
452 information on the default behavior if the <code>scan with</code>
453 section is omitted.<p>
455 As of CUP 0.10j the action code, parser code, init code, and scan with
456 sections may appear in any order. They must, however, precede the
459 <h5><a name="symbol_list">Symbol Lists</a></h5>
461 Following user supplied code comes the first required part of the
462 specification: the symbol lists. These declarations are responsible
463 for naming and supplying a type for each terminal and non-terminal
464 symbol that appears in the grammar. As indicated above, each terminal
465 and non-terminal symbol is represented at runtime with a <tt>Symbol</tt>
467 the case of terminals, these are returned by the scanner and placed on
468 the parse stack. The lexer should put the value of the terminal in the
469 <tt>value</tt> instance variable.
470 In the case of non-terminals these replace a series
471 of <tt>Symbol</tt> objects on the parse stack whenever the right hand side of
472 some production is recognized. In order to tell the parser which object
473 types should be used for which symbol, <tt>terminal</tt> and
474 <tt>non terminal</tt> declarations are used. These take the forms:
476 <pre><tt> terminal <i>classname</i> <i>name1, name2,</i> ...;</tt>
477 <tt> non terminal <i>classname</i> <i>name1, name2,</i> ...;</tt>
478 <tt> terminal <i>name1, name2,</i> ...;</tt>
483 <pre><tt> non terminal <i>name1, name2,</i> ...;</tt>
486 where <tt><i>classname</i></tt> can be a multiple part name separated with
488 <tt><i>classname</i></tt> specified represents the type of the value of
489 that terminal or non-terminal. When accessing these values through
490 labels, the users uses the type declared. the <tt><i>classname</i></tt>
491 can be of any type. If no <tt><i>classname</i></tt> is given, then the
492 terminal or non-terminal holds no value. a label referring to such a
493 symbol with have a null value. As of CUP 0.10j, you may specify
494 non-terminals the declaration "<code>nonterminal</code>" (note, no
495 space) as well as the original "<code>non terminal</code>" spelling.<p>
497 Names of terminals and non-terminals cannot be CUP reserved words;
498 these include "code", "action", "parser", "terminal", "non",
499 "nonterminal", "init", "scan", "with", "start", "precedence", "left",
500 "right", "nonassoc", "import", and "package".<p>
502 <h5><a name="precedence">Precedence and Associativity declarations</a></h5>
504 The third section, which is optional, specifies the precedences and
505 associativity of terminals. This is useful for parsing with ambiguous
506 grammars, as done in the example above. There are three type of
507 precedence/associativity declarations:
509 precedence left <i>terminal</i>[, <i>terminal</i>...];
510 precedence right <i>terminal</i>[, <i>terminal</i>...];
511 precedence nonassoc <i>terminal</i>[, <i>terminal</i>...];
514 The comma separated list indicates that those terminals should have the
515 associativity specified at that precedence level and the precedence of
516 that declaration. The order of precedence, from highest to lowest, is
517 bottom to top. Hence, this declares that multiplication and division have
518 higher precedence than addition and subtraction:
520 precedence left ADD, SUBTRACT;
521 precedence left TIMES, DIVIDE;
523 Precedence resolves shift reduce problems. For example, given the input
524 to the above example parser <tt>3 + 4 * 8</tt>, the parser doesn't know
525 whether to reduce <tt>3 + 4</tt> or shift the '*' onto the stack.
526 However, since '*' has a higher precedence than '+', it will be shifted
527 and the multiplication will be performed before the addition.<p>
529 CUP assigns each one of its terminals a precedence according to these
530 declarations. Any terminals not in this declaration have lowest
531 precedence. CUP also assigns each of its productions a precedence.
532 That precedence is equal to the precedence of the last terminal in that
533 production. If the production has no terminals, then it has lowest
534 precedence. For example, <tt>expr ::= expr TIMES expr</tt> would have
535 the same precedence as <tt>TIMES</tt>. When there is a shift/reduce
536 conflict, the parser determines whether the terminal to be shifted has a
537 higher precedence, or if the production to reduce by does. If the
538 terminal has higher precedence, it it shifted, if the production has
539 higher precedence, a reduce is performed. If they have equal
540 precedence, associativity of the terminal determine what happens.<p>
542 An associativity is assigned to each terminal used in the
543 precedence/associativity declarations. The three associativities are
544 <tt>left, right</tt> and <tt>nonassoc</tt> Associativities are also
545 used to resolve shift/reduce conflicts, but only in the case of equal
546 precedences. If the associativity of the terminal that can be shifted
547 is <tt>left</tt>, then a reduce is performed. This means, if the input
548 is a string of additions, like <tt>3 + 4 + 5 + 6 + 7</tt>, the parser
549 will <i>always</i> reduce them from left to right, in this case,
550 starting with <tt>3 + 4</tt>. If the associativity of the terminal is
551 <tt>right</tt>, it is shifted onto the stack. hence, the reductions
552 will take place from right to left. So, if PLUS were declared with
553 associativity of <tt>right</tt>, the <tt>6 + 7</tt> would be reduced
554 first in the above string. If a terminal is declared as
555 <tt>nonassoc</tt>, then two consecutive occurrences of equal precedence
556 non-associative terminals generates an error. This is useful for
557 comparison operations. For example, if the input string is
558 <tt>6 == 7 == 8 == 9</tt>, the parser should generate an error. If '=='
559 is declared as <tt>nonassoc</tt> then an error will be generated. <p>
561 All terminals not used in the precedence/associativity declarations are
562 treated as lowest precedence. If a shift/reduce error results,
563 involving two such terminals, it cannot be resolved, as the above
564 conflicts are, so it will be reported.<p>
566 <h5><a name="production_list">The Grammar</a></h5>
568 The final section of a CUP declaration provides the grammar. This
569 section optionally starts with a declaration of the form:
571 <pre><tt> start with <i>non-terminal</i>;</tt>
574 This indicates which non-terminal is the <i>start</i> or <i>goal</i>
575 non-terminal for parsing. If a start non-terminal is not explicitly
576 declared, then the non-terminal on the left hand side of the first
577 production will be used. At the end of a successful parse, CUP returns
578 an object of type <tt>java_cup.runtime.Symbol</tt>. This
579 <tt>Symbol</tt>'s value instance variable contains the final reduction
582 The grammar itself follows the optional <tt>start</tt> declaration. Each
583 production in the grammar has a left hand side non-terminal followed by
584 the symbol "<tt>::=</tt>", which is then followed by a series of zero or more
585 actions, terminal, or non-terminal
586 symbols, followed by an optional contextual precedence assignment,
587 and terminated with a semicolon (;).<p>
589 <a name="label_part">
591 Each symbol on the right hand side can optionally be labeled with a name.
592 Label names appear after the symbol name separated by a colon (:). Label
593 names must be unique within the production, and can be used within action
594 code to refer to the value of the symbol. Along with the label, two
595 more variables are created, which are the label plus <tt>left</tt> and
596 the label plus <tt>right</tt>. These are <tt>int</tt> values that
597 contain the right and left locations of what the terminal or
598 non-terminal covers in the input file. These values must be properly
599 initialized in the terminals by the lexer. The left and right values
600 then propagate to non-terminals to which productions reduce.<p>
602 If there are several productions for the same non-terminal they may be
603 declared together. In this case the productions start with the non-terminal
604 and "<tt>::=</tt>". This is followed by multiple right hand sides each
605 separated by a bar (|). The full set of productions is then terminated by a
608 Actions appear in the right hand side as code strings (e.g., Java code inside
609 <tt>{:</tt> ... <tt>:}</tt> delimiters). These are executed by the parser
610 at the point when the portion of the production to the left of the
611 action has been recognized. (Note that the scanner will have returned the
612 token one past the point of the action since the parser needs this extra
613 <i>lookahead</i> token for recognition.)<p>
617 Contextual precedence assignments follow all the symbols and actions of
618 the right hand side of the production whose precedence it is assigning.
619 Contextual precedence assignment allows a production to be assigned a
620 precedence not based on the last terminal in it. A good example is
621 shown in the above sample parser specification:
624 precedence left PLUS, MINUS;
625 precedence left TIMES, DIVIDE, MOD;
626 precedence left UMINUS;
628 expr ::= MINUS expr:e
629 {: RESULT = new Integer(0 - e.intValue()); :}
633 Here, there production is declared as having the precedence of UMINUS.
634 Hence, the parser can give the MINUS sign two different precedences,
635 depending on whether it is a unary minus or a subtraction operation.
638 <h3>3. Running CUP</h3></a>
640 As mentioned above, CUP is written in Java. To invoke it, one needs
641 to use the Java interpreter to invoke the static method
642 <tt>java_cup.Main()</tt>, passing an array of strings containing options.
643 Assuming a Unix machine, the simplest way to do this is typically to invoke it
644 directly from the command line with a command such as:
646 <pre><tt> java java_cup.Main <i>options</i> < <i>inputfile</i></tt></pre>
648 Once running, CUP expects to find a specification file on standard input
649 and produces two Java source files as output. Starting with CUP 0.10k,
650 the final command-line argument may be a filename, in which case the
651 specification will be read from that file instead of from standard input.<p>
653 In addition to the specification file, CUP's behavior can also be changed
654 by passing various options to it. Legal options are documented in
655 <code>Main.java</code> and include:
657 <dt><tt>-package</tt> <i>name</i>
658 <dd>Specify that the <tt>parser</tt> and <tt>sym</tt> classes are to be
659 placed in the named package. By default, no package specification
660 is put in the generated code (hence the classes default to the special
663 <dt><tt>-parser</tt> <i>name</i>
664 <dd>Output parser and action code into a file (and class) with the given
665 name instead of the default of "<tt>parser</tt>".
667 <dt><tt>-symbols</tt> <i>name</i>
668 <dd>Output the symbol constant code into a class with the given
669 name instead of the default of "<tt>sym</tt>".
671 <dt><tt>-interface</tt>
672 <dd>Outputs the symbol constant code as an <code>interface</code>
673 rather than as a <code>class</code>.
675 <dt><tt>-nonterms</tt>
676 <dd>Place constants for non-terminals into the symbol constant class.
677 The parser does not need these symbol constants, so they are not normally
678 output. However, it can be very helpful to refer to these constants
679 when debugging a generated parser.
681 <dt><tt>-expect</tt> <i>number</i>
682 <dd>During parser construction the system may detect that an ambiguous
683 situation would occur at runtime. This is called a <i>conflict</i>.
684 In general, the parser may be unable to decide whether to <i>shift</i>
685 (read another symbol) or <i>reduce</i> (replace the recognized right
686 hand side of a production with its left hand side). This is called a
687 <i>shift/reduce conflict</i>. Similarly, the parser may not be able
688 to decide between reduction with two different productions. This is
689 called a <i>reduce/reduce conflict</i>. Normally, if one or more of
690 these conflicts occur, parser generation is aborted. However, in
691 certain carefully considered cases it may be advantageous to
692 arbitrarily break such a conflict. In this case CUP uses YACC
693 convention and resolves shift/reduce conflicts by shifting, and
694 reduce/reduce conflicts using the "highest priority" production (the
695 one declared first in the specification). In order to enable automatic
696 breaking of conflicts the <tt>-expect</tt> option must be given
697 indicating exactly how many conflicts are expected. Conflicts
698 resolved by precedences and associativities are not reported.
700 <dt><tt>-compact_red</tt>
701 <dd>Including this option enables a table compaction optimization involving
702 reductions. In particular, it allows the most common reduce entry in
703 each row of the parse action table to be used as the default for that
704 row. This typically saves considerable room in the tables, which can
705 grow to be very large. This optimization has the effect of replacing
706 all error entries in a row with the default reduce entry. While this
707 may sound dangerous, if not down right incorrect, it turns out that this
708 does not affect the correctness of the parser. In particular, some
709 changes of this type are inherent in LALR parsers (when compared to
710 canonical LR parsers), and the resulting parsers will still never
711 read past the first token at which the error could be detected.
712 The parser can, however, make extra erroneous reduces before detecting
713 the error, so this can degrade the parser's ability to do
714 <a href="#errors">error recovery</a>.
715 (Refer to reference [2] pp. 244-247 or reference [3] pp. 190-194 for a
716 complete explanation of this compaction technique.) <br><br>
718 This option is typically used to work-around the java bytecode
719 limitations on table initialization code sizes. However, CUP
720 0.10h introduced a string-encoding for the parser tables which
721 is not subject to the standard method-size limitations.
722 Consequently, use of this option should no longer be required
726 <dd>This options causes all warning messages (as opposed to error messages)
727 produced by the system to be suppressed.
729 <dt><tt>-nosummary</tt>
730 <dd>Normally, the system prints a summary listing such things as the
731 number of terminals, non-terminals, parse states, etc. at the end of
732 its run. This option suppresses that summary.
734 <dt><tt>-progress</tt>
735 <dd>This option causes the system to print short messages indicating its
736 progress through various parts of the parser generation process.
738 <dt><tt>-dump_grammar</tt>
739 <dt><tt>-dump_states</tt>
740 <dt><tt>-dump_tables</tt>
742 <dd> These options cause the system to produce a human readable dump of
743 the grammar, the constructed parse states (often needed to resolve
744 parse conflicts), and the parse tables (rarely needed), respectively.
745 The <tt>-dump</tt> option can be used to produce all of these dumps.
748 <dd>This option adds detailed timing statistics to the normal summary of
749 results. This is normally of great interest only to maintainers of
753 <dd>This option produces voluminous internal debugging information about
754 the system as it runs. This is normally of interest only to maintainers
755 of the system itself.
757 <dt><tt>-nopositions</tt>
758 <dd>This option keeps CUP from generating code to propagate the left
759 and right hand values of terminals to non-terminals, and then from
760 non-terminals to other terminals. If the left and right values aren't
761 going to be used by the parser, then it will save some runtime
762 computation to not generate these position propagations. This option
763 also keeps the left and right label variables from being generated, so
764 any reference to these will cause an error.
766 <dt><tt>-noscanner</tt>
767 <dd>CUP 0.10j introduced <a href="#scanner">improved scanner
768 integration</a> and a new interface,
769 <code>java_cup.runtime.Scanner</code>. By default, the
770 generated parser refers to this interface, which means you cannot
771 use these parsers with CUP runtimes older than 0.10j. If your
772 parser does not use the new scanner integration features, then you
773 may specify the <code>-noscanner</code> option to suppress the
774 <code>java_cup.runtime.Scanner</code> references and allow
775 compatibility with old runtimes. Not many people should have reason
778 <dt><tt>-version</tt>
779 <dd>Invoking CUP with the <code>-version</code> flag will cause it
780 to print out the working version of CUP and halt. This allows
781 automated CUP version checking for Makefiles, install scripts and
782 other applications which may require it.
786 <h3>4. Customizing the Parser</h3></a>
788 Each generated parser consists of three generated classes. The
789 <tt>sym</tt> class (which can be renamed using the <tt>-symbols</tt>
790 option) simply contains a series of <tt>int</tt> constants,
791 one for each terminal. Non-terminals are also included if the <tt>-nonterms</tt>
792 option is given. The source file for the <tt>parser</tt> class (which can
793 be renamed using the <tt>-parser</tt> option) actually contains two
794 class definitions, the public <tt>parser</tt> class that implements the
795 actual parser, and another non-public class (called <tt>CUP$action</tt>) which
796 encapsulates all user actions contained in the grammar, as well as code from
797 the <tt>action code</tt> declaration. In addition to user supplied code, this
798 class contains one method: <tt>CUP$do_action</tt> which consists of a large
799 switch statement for selecting and executing various fragments of user
800 supplied action code. In general, all names beginning with the prefix of
801 <tt>CUP$</tt> are reserved for internal uses by CUP generated code. <p>
803 The <tt>parser</tt> class contains the actual generated parser. It is
804 a subclass of <tt>java_cup.runtime.lr_parser</tt> which implements a
805 general table driven framework for an LR parser. The generated <tt>parser</tt>
806 class provides a series of tables for use by the general framework.
807 Three tables are provided:
809 <dt>the production table
810 <dd>provides the symbol number of the left hand side non-terminal, along with
811 the length of the right hand side, for each production in the grammar,
813 <dd>indicates what action (shift, reduce, or error) is to be taken on each
814 lookahead symbol when encountered in each state, and
815 <dt>the reduce-goto table
816 <dd>indicates which state to shift to after reduces (under each non-terminal
819 (Note that the action and reduce-goto tables are not stored as simple arrays,
820 but use a compacted "list" structure to save a significant amount of space.
821 See comments the runtime system source code for details.)<p>
823 Beyond the parse tables, generated (or inherited) code provides a series
824 of methods that can be used to customize the generated parser. Some of these
825 methods are supplied by code found in part of the specification and can
826 be customized directly in that fashion. The others are provided by the
827 <tt>lr_parser</tt> base class and can be overridden with new versions (via
828 the <tt>parser code</tt> declaration) to customize the system. Methods
829 available for customization include:
831 <dt><tt>public void user_init()</tt>
832 <dd>This method is called by the parser prior to asking for the first token
833 from the scanner. The body of this method contains the code from the
834 <tt>init with</tt> clause of the the specification.
835 <dt><a name="scan_method"><tt>public java_cup.runtime.Symbol scan()</tt></a>
836 <dd>This method encapsulates the scanner and is called each time a new
837 terminal is needed by the parser. The body of this method is
838 supplied by the <tt>scan with</tt> clause of the specification, if
839 present; otherwise it returns <code>getScanner().next_token()</code>.
840 <dt><tt>public java_cup.runtime.Scanner getScanner()</tt>
841 <dd>Returns the default scanner. See <a href="#scanner">section 5</a>.
842 <dt><tt>public void setScanner(java_cup.runtime.Scanner s)</tt>
843 <dd>Sets the default scanner. See <a href="#scanner">section 5</a>.
844 <dt><tt> public void report_error(String message, Object info)</tt>
845 <dd>This method should be called whenever an error message is to be issued. In
846 the default implementation of this method, the first parameter provides
847 the text of a message which is printed on <tt>System.err</tt>
848 and the second parameter is simply ignored. It is very typical to
849 override this method in order to provide a more sophisticated error
851 <dt><tt>public void report_fatal_error(String message, Object info)</tt>
852 <dd>This method should be called whenever a non-recoverable error occurs. It
853 responds by calling <tt>report_error()</tt>, then aborts parsing
854 by calling the parser method <tt>done_parsing()</tt>, and finally
855 throws an exception. (In general <tt>done_parsing()</tt> should be called
856 at any point that parsing needs to be terminated early).
857 <dt><tt>public void syntax_error(Symbol cur_token)</tt>
858 <dd>This method is called by the parser as soon as a syntax error is detected
859 (but before error recovery is attempted). In the default implementation it
860 calls: <tt>report_error("Syntax error", null);</tt>.
861 <dt><tt>public void unrecovered_syntax_error(Symbol cur_token)</tt>
862 <dd>This method is called by the parser if it is unable to recover from a
863 syntax error. In the default implementation it calls:
864 <tt>report_fatal_error("Couldn't repair and continue parse", null);</tt>.
865 <dt><tt> protected int error_sync_size()</tt>
866 <dd>This method is called by the parser to determine how many tokens it must
867 successfully parse in order to consider an error recovery successful.
868 The default implementation returns 3. Values below 2 are not recommended.
869 See the section on <a href="#errors">error recovery</a> for details.
872 Parsing itself is performed by the method <tt>public Symbol parse()</tt>.
873 This method starts by getting references to each of the parse tables,
874 then initializes a <tt>CUP$action</tt> object (by calling
875 <tt>protected void init_actions()</tt>). Next it calls <tt>user_init()</tt>,
876 then fetches the first lookahead token with a call to <tt>scan()</tt>.
877 Finally, it begins parsing. Parsing continues until <tt>done_parsing()</tt>
878 is called (this is done automatically, for example, when the parser
879 accepts). It then returns a <tt>Symbol</tt> with the <tt>value</tt>
880 instance variable containing the RESULT of the start production, or
881 <tt>null</tt>, if there is no value.<p>
883 In addition to the normal parser, the runtime system also provides a debugging
884 version of the parser. This operates in exactly the same way as the normal
885 parser, but prints debugging messages (by calling
886 <tt>public void debug_message(String mess)</tt> whose default implementation
887 prints a message to <tt>System.err</tt>).<p>
889 Based on these routines, invocation of a CUP parser is typically done
892 /* create a parsing object */
893 parser parser_obj = new parser();
895 /* open input files, etc. here */
896 Symbol parse_tree = null;
900 parse_tree = parser_obj.debug_parse();
902 parse_tree = parser_obj.parse();
903 } catch (Exception e) {
904 /* do cleanup here - - possibly rethrow e */
906 /* do close out here */
911 <h3>5. Scanner Interface</h3></a>
913 In CUP 0.10j scanner integration was improved according to
914 suggestions made by <a href="http://www.smartsc.com">David MacMahon</a>.
915 The changes make it easier to incorporate JLex and other
916 automatically-generated scanners into CUP parsers.<p>
918 To use the new code, your scanner should implement the
919 <code>java_cup.runtime.Scanner</code> interface, defined as:
921 package java_cup.runtime;
923 public interface Scanner {
924 public Symbol next_token() throws java.lang.Exception;
928 In addition to the methods described in <a href="#parser">section
929 4</a>, the <code>java_cup.runtime.lr_parser</code> class has two new
930 accessor methods, <code>setScanner()</code> and <code>getScanner()</code>.
931 The default implementation of <a href="#scan_method"><code>scan()</code></a>
934 public Symbol scan() throws java.lang.Exception {
935 Symbol sym = getScanner().next_token();
936 return (sym!=null) ? sym : new Symbol(EOF_sym());
939 The generated parser also contains a constructor which takes a
940 <code>Scanner</code> and calls <code>setScanner()</code> with it. In
941 most cases, then, the <code>init with</code> and <code>scan
942 with</code> directives may be omitted. You can simply create the
943 parser with a reference to the desired scanner:
945 /* create a parsing object */
946 parser parser_obj = new parser(new my_scanner());
948 or set the scanner after the parser is created:
950 /* create a parsing object */
951 parser parser_obj = new parser();
952 /* set the default scanner */
953 parser_obj.setScanner(new my_scanner());
955 Note that because the parser uses look-ahead, resetting the scanner in
956 the middle of a parse is not recommended. If you attempt to use the
957 default implementation of <code>scan()</code> without first calling
958 <code>setScanner()</code>, a <code>NullPointerException</code> will be
960 As an example of scanner integration, the following three lines in the
961 lexer-generator input are all that is required to use a
962 <a href="http://www.cs.princeton.edu/~appel/modern/java/JLex/">JLex</a>
963 or <a href="http://www.jflex.de/">JFlex</A>
966 %implements java_cup.runtime.Scanner
968 %type java_cup.runtime.Symbol
970 The JLex directive <code>%cup</code>
971 abbreviates the above three directive in JLex versions 1.2.5 and above.
972 Invoking the parser with the JLex scanner is then simply:
974 parser parser_obj = new parser( new Yylex( some_InputStream_or_Reader));
977 Note CUP handles the JLex/JFlex convention of returning null on EOF
978 without a problem, so an <code>%eofval</code> directive is not
979 required in the JLex specification (this feature was added in CUP 0.10k).
981 The simple_calc example in the CUP distribution illustrates the use of
982 the scanner integration features with a hand-coded scanner.
983 The CUP website has a minimal CUP/JLex integration example for study.<p>
986 <h3>6. Error Recovery</h3></a>
988 A final important aspect of building parsers with CUP is
989 support for syntactic error recovery. CUP uses the same
990 error recovery mechanisms as YACC. In particular, it supports
991 a special error symbol (denoted simply as <tt>error</tt>).
992 This symbol plays the role of a special non-terminal which, instead of
993 being defined by productions, instead matches an erroneous input
996 The error symbol only comes into play if a syntax error is
997 detected. If a syntax error is detected then the parser tries to replace
998 some portion of the input token stream with <tt>error</tt> and then
999 continue parsing. For example, we might have productions such as:
1001 <pre><tt> stmt ::= expr SEMI | while_stmt SEMI | if_stmt SEMI | ... |
1005 This indicates that if none of the normal productions for <tt>stmt</tt> can
1006 be matched by the input, then a syntax error should be declared, and recovery
1007 should be made by skipping erroneous tokens (equivalent to matching and
1008 replacing them with <tt>error</tt>) up to a point at which the parse can
1009 be continued with a semicolon (and additional context that legally follows a
1010 statement). An error is considered to be recovered from if and only if a
1011 sufficient number of tokens past the <tt>error</tt> symbol can be successfully
1012 parsed. (The number of tokens required is determined by the
1013 <tt>error_sync_size()</tt> method of the parser and defaults to 3). <p>
1015 Specifically, the parser first looks for the closest state to the top
1016 of the parse stack that has an outgoing transition under
1017 <tt>error</tt>. This generally corresponds to working from
1018 productions that represent more detailed constructs (such as a specific
1019 kind of statement) up to productions that represent more general or
1020 enclosing constructs (such as the general production for all
1021 statements or a production representing a whole section of declarations)
1022 until we get to a place where an error recovery production
1023 has been provided for. Once the parser is placed into a configuration
1024 that has an immediate error recovery (by popping the stack to the first
1025 such state), the parser begins skipping tokens to find a point at
1026 which the parse can be continued. After discarding each token, the
1027 parser attempts to parse ahead in the input (without executing any
1028 embedded semantic actions). If the parser can successfully parse past
1029 the required number of tokens, then the input is backed up to the point
1030 of recovery and the parse is resumed normally (executing all actions).
1031 If the parse cannot be continued far enough, then another token is
1032 discarded and the parser again tries to parse ahead. If the end of
1033 input is reached without making a successful recovery (or there was no
1034 suitable error recovery state found on the parse stack to begin with)
1035 then error recovery fails.
1037 <a name="conclusion">
1038 <h3>7. Conclusion</h3></a>
1040 This manual has briefly described the CUP LALR parser generation system.
1041 CUP is designed to fill the same role as the well known YACC parser
1042 generator system, but is written in and operates entirely with Java code
1043 rather than C or C++. Additional details on the operation of the system can
1044 be found in the parser generator and runtime source code. See the CUP
1045 home page below for access to the API documentation for the system and its
1048 This document covers version 0.10j of the system. Check the CUP home
1050 <a href="http://www.cs.princeton.edu/~appel/modern/java/CUP/">
1051 http://www.cs.princeton.edu/~appel/modern/java/CUP/</a>
1052 for the latest release information, instructions for downloading the
1053 system, and additional news about CUP. Bug reports and other
1054 comments for the developers should be sent to the CUP maintainer,
1055 C. Scott Ananian, at
1056 <a href="mailto:cananian@alumni.princeton.edu">
1057 cananian@alumni.princeton.edu</a><p>
1059 CUP was originally written by
1060 <a href="http://www.cs.cmu.edu/~hudson/">
1061 Scott Hudson</a>, in August of 1995.<p>
1063 It was extended to support precedence by
1064 <a href="http://www.princeton.edu/~frankf">
1065 Frank Flannery</a>, in July of 1996.<p>
1067 On-going improvements have been done by
1068 <A HREF="http://www.pdos.lcs.mit.edu/~cananian">
1069 C. Scott Ananian</A>, the CUP maintainer, from December of 1997 to the
1073 <h3>References</h3></a>
1076 <dt><a name = "YACCref">[1]</a>
1078 "YACC &emdash; Yet Another Compiler Compiler",
1079 CS Technical Report #32,
1080 Bell Telephone Laboratories,
1084 <dt><a name = "dragonbook">[2]</a>
1085 <dd>A. Aho, R. Sethi, and J. Ullman,
1086 <i>Compilers: Principles, Techniques, and Tools</i>,
1087 Addison-Wesley Publishing,
1091 <dt><a name = "crafting">[3]</a>
1092 <dd>C. Fischer, and R. LeBlanc,
1093 <i>Crafting a Compiler with C</i>,
1094 Benjamin/Cummings Publishing,
1098 <dt><a name = "modernjava">[4]</a>
1099 <dd>Andrew W. Appel,
1100 <i>Modern Compiler Implementation in Java</i>,
1101 Cambridge University Press,
1107 <h3><a name="appendixa">
1108 Appendix A. Grammar for CUP Specification Files</a> (0.10j)</h3>
1110 <pre><tt>java_cup_spec ::= package_spec import_list code_parts
1111 symbol_list precedence_list start_spec
1113 package_spec ::= PACKAGE multipart_id SEMI | empty
1114 import_list ::= import_list import_spec | empty
1115 import_spec ::= IMPORT import_id SEMI
1116 code_part ::= action_code_part | parser_code_part |
1117 init_code | scan_code
1118 code_parts ::= code_parts code_part | empty
1119 action_code_part ::= ACTION CODE CODE_STRING opt_semi
1120 parser_code_part ::= PARSER CODE CODE_STRING opt_semi
1121 init_code ::= INIT WITH CODE_STRING opt_semi
1122 scan_code ::= SCAN WITH CODE_STRING opt_semi
1123 symbol_list ::= symbol_list symbol | symbol
1124 symbol ::= TERMINAL type_id declares_term |
1125 NON TERMINAL type_id declares_non_term |
1126 NONTERMINAL type_id declares_non_term |
1127 TERMINAL declares_term |
1128 NON TERMINAL declares_non_term |
1129 NONTERMIANL declared_non_term
1130 term_name_list ::= term_name_list COMMA new_term_id | new_term_id
1131 non_term_name_list ::= non_term_name_list COMMA new_non_term_id |
1133 declares_term ::= term_name_list SEMI
1134 declares_non_term ::= non_term_name_list SEMI
1135 precedence_list ::= precedence_l | empty
1136 precedence_l ::= precedence_l preced + preced;
1137 preced ::= PRECEDENCE LEFT terminal_list SEMI
1138 | PRECEDENCE RIGHT terminal_list SEMI
1139 | PRECEDENCE NONASSOC terminal_list SEMI
1140 terminal_list ::= terminal_list COMMA terminal_id | terminal_id
1141 start_spec ::= START WITH nt_id SEMI | empty
1142 production_list ::= production_list production | production
1143 production ::= nt_id COLON_COLON_EQUALS rhs_list SEMI
1144 rhs_list ::= rhs_list BAR rhs | rhs
1145 rhs ::= prod_part_list PERCENT_PREC term_id |
1147 prod_part_list ::= prod_part_list prod_part | empty
1148 prod_part ::= symbol_id opt_label | CODE_STRING
1149 opt_label ::= COLON label_id | empty
1150 multipart_id ::= multipart_id DOT ID | ID
1151 import_id ::= multipart_id DOT STAR | multipart_id
1152 type_id ::= multipart_id
1153 terminal_id ::= term_id
1154 term_id ::= symbol_id
1156 new_non_term_id ::= ID
1160 opt_semi ::= SEMI | empty
1165 <h3><a name = "appendixb">Appendix B. A Very Simple Example Scanner<a></h3>
1168 <tt>// Simple Example Scanner Class
1170 import java_cup.runtime.*;
1173 public class scanner {
1174 /* single lookahead character */
1175 protected static int next_char;
1177 /* advance input by one character */
1178 protected static void advance()
1179 throws java.io.IOException
1180 { next_char = System.in.read(); }
1182 /* initialize the scanner */
1183 public static void init()
1184 throws java.io.IOException
1187 /* recognize and return the next complete token */
1188 public static Symbol next_token()
1189 throws java.io.IOException
1194 case '0': case '1': case '2': case '3': case '4':
1195 case '5': case '6': case '7': case '8': case '9':
1196 /* parse a decimal integer */
1199 i_val = i_val * 10 + (next_char - '0');
1201 } while (next_char >= '0' && next_char <= '9');
1202 return new Symbol(sym.NUMBER, new Integer(i_val));
1204 case ';': advance(); return new Symbol(sym.SEMI);
1205 case '+': advance(); return new Symbol(sym.PLUS);
1206 case '-': advance(); return new Symbol(sym.MINUS);
1207 case '*': advance(); return new Symbol(sym.TIMES);
1208 case '/': advance(); return new Symbol(sym.DIVIDE);
1209 case '%': advance(); return new Symbol(sym.MOD);
1210 case '(': advance(); return new Symbol(sym.LPAREN);
1211 case ')': advance(); return new Symbol(sym.RPAREN);
1213 case -1: return new Symbol(sym.EOF);
1216 /* in this simple scanner we just ignore everything else */
1226 <h3>Appendix C: Incompatibilites between CUP 0.9 and CUP 0.10</a></h3>
1228 CUP version 0.10a is a major overhaul of CUP. The changes are severe,
1229 meaning no backwards compatibility to older versions.
1231 The changes consist of:
1233 <li> <a href="#lex_inter">A different lexical interface</a>,
1234 <li> <a href="#new_dec">New terminal/non-terminal declarations</a>,
1235 <li> <a href="#label_ref">Different label references</a>,
1236 <li> <a href="#RESULT_pass">A different way of passing RESULT</a>,
1237 <li> <a href="#pos_prop">New position values and propagation</a>,
1238 <li> <a href="#ret_val">Parser now returns a value</a>,
1239 <li> <a href="#prec_add">Terminal precedence declarations</a> and
1240 <li> <a href="#con_prec">Rule contextual precedence assignment</a>
1243 <h5><a name="lex_inter">Lexical Interface</a></h5>
1245 CUP now interfaces with the lexer in a completely different
1246 manner. In the previous releases, a new class was used for every
1247 distinct type of terminal. This release, however, uses only one class:
1248 The <tt>Symbol</tt> class. The <tt>Symbol</tt> class has three instance
1250 are significant to the parser when passing information from the lexer.
1251 The first is the <tt>value</tt> instance variable. This variable
1253 value of that terminal. It is of the type declared as the terminal type
1254 in the parser specification file. The second two are the instance
1255 variables <tt>left</tt> and <tt>right</tt>. They should be filled with
1256 the <tt>int</tt> value of
1257 where in the input file, character-wise, that terminal was found.<p>
1259 For more information, refer to the manual on <a href="#lex_part">scanners</a>.
1261 <h5><a name="new_dec">Terminal/Non-Terminal Declarations</a></h5>
1263 Terminal and non-terminal declarations now can be declared in two
1264 different ways to indicate the values of the terminals or
1265 non-terminals. The previous declarations of the form
1268 terminal <i>classname terminal</i> [, <i>terminal ...</i>];
1271 still works. The classname, however indicates the type of the value of
1272 the terminal or non-terminal, and does not indicate the type of object
1273 placed on the parse stack.
1275 A declaration, such as:
1278 terminal <i>terminal</i> [, <i>terminal ...</i>];
1281 indicates the terminals in the list hold no value.<p>
1283 For more information, refer to the manual on <a
1284 href="#symbol_list">declarations</a>.
1286 <h5><a name="label_ref">Label References</a></h5>
1288 Label references do not refer to the object on the parse stack, as in
1289 the old CUP, but rather to the value of the <tt>value</tt>
1290 instance variable of
1291 the <tt>Symbol</tt> that represents that terminal or non-terminal. Hence,
1292 references to terminal and non-terminal values is direct, as opposed to
1293 the old CUP, where the labels referred to objects containing the value
1294 of the terminal or non-terminal.<p>
1296 For more information, refer to the manual on <a href="#label_part">labels</a>.
1298 <h5><a name="RESULT_pass">RESULT Value</a></h5>
1300 The <tt>RESULT</tt> variable refers directly to the value of the
1302 to which a rule reduces, rather than to the object on the parse stack.
1303 Hence, <tt>RESULT</tt> is of the same type the non-terminal to which
1305 as declared in the non-terminal declaration. Again, the reference is
1306 direct, rather than to something that will contain the data.<p>
1308 For more information, refer to the manual on <a href="#RES_part">RESULT</a>.
1310 <h5><a name="pos_prop">Position Propagation</a></h5>
1312 For every label, two more variables are declared, which are the label
1313 plus <tt>left</tt> or the label plus <tt>right</tt>. These correspond
1315 right locations in the input stream to which that terminal or
1316 non-terminal came from. These values are propagated from the input
1317 terminals, so that the starting non-terminal should have a left value of
1318 0 and a right value of the location of the last character read.<p>
1320 For more information, refer to the manual on <a href="#label_part">positions</a>.
1322 <h5><a name="ret_val">Return Value</a></h5>
1324 A call to <tt>parse()</tt> or <tt>debug_parse()</tt> returns a
1325 Symbol. This Symbol is the start non-terminal, so the <tt>value</tt>
1326 instance variable contains the final <tt>RESULT</tt> assignment.
1328 <h5><a name="prec_add">Precedence</a></h5>
1330 CUP now has precedenced terminals. a new declaration section,
1331 occurring between the terminal and non-terminal declarations and the
1332 grammar specifies the precedence and associativity of rules. The
1333 declarations are of the form:
1336 precedence {left| right | nonassoc} <i>terminal</i>[, <i>terminal</i> ...];
1341 The terminals are assigned a precedence, where terminals on the same
1342 line have equal precedences, and the precedence declarations farther
1343 down the list of precedence declarations have higher precedence.
1344 <tt>left, right</tt> and <tt>nonassoc</tt> specify the associativity
1345 of these terminals. left
1346 associativity corresponds to a reduce on conflict, right to a shift on
1347 conflict, and nonassoc to an error on conflict. Hence, ambiguous
1348 grammars may now be used.<p>
1350 For more information, refer to the manual on <a
1351 href="#precedence">precedence</a>.
1353 <h5><a name="con_prec">Contextual Precedence</a></h5>
1355 Finally the new CUP adds contextual precedence. A production may be
1356 declare as followed:
1359 lhs ::= <i>{right hand side list of terminals, non-terminals and actions}</i>
1360 %prec <i>{terminal}</i>;
1363 this production would then have a precedence equal to the terminal
1364 specified after the <tt>%prec</tt>. Hence, shift/reduce conflicts can be
1365 contextually resolved. Note that the <tt>%prec</tt> <i>terminal</i>
1366 part comes after all actions strings. It does not come before the
1367 last action string.<p>
1369 For more information, refer to the manual on <a href="#cpp">contextual
1372 These changes implemented by:
1374 <a href="http://www.princeton.edu/~frankf">Frank Flannery</a><br>
1375 <a href="http://www.cs.princeton.edu">Department of Computer Science</a><br>
1376 <a href="http://www.princeton.edu">Princeton University</a><br>
1379 <h3>Appendix D: Bugs</a></h3>
1380 In this version of CUP it's difficult for the semantic action phrases (Java code attached
1381 to productions) to access the <tt>report_error</tt> method and other similar methods and
1382 objects defined in the <tt>parser code</tt> directive.
1384 This is because the parsing tables (and parsing engine) are in one object (belonging to
1385 class <tt>parser</tt> or whatever name is specified by the <strong>-parser</strong> directive),
1386 and the semantic actions are in another object (of class <tt>CUP$actions</tt>).
1388 However, there is a way to do it, though it's a bit inelegant.
1389 The action object has a <tt>private final</tt> field named
1390 <tt>parser</tt> that points to the parsing object. Thus,
1391 methods and instance variables of the parser can be accessed within semantic actions as:
1393 parser.report_error(message,info);
1397 Perhaps this will not be necessary in a future release, and that
1398 such methods and variables as <tt>report_error</tt> and
1399 <tt>mydata</tt> will be available
1400 directly from the semantic actions; we will achieve this by combining the
1401 "parser" object and the "actions" object together.
1404 For a list of any other currently known bugs in CUP, see
1405 <A HREF="http://www.cs.princeton.edu/~appel/modern/java/CUP/bugs.html">
1406 http://www.cs.princeton.edu/~appel/modern/java/CUP/bugs.html</A>.
1409 <h3>Appendix E: Change log</a></h3>
1412 <dt>0.9e<dd>March 1996, Scott Hudson's original version.
1413 <dt>0.10a<dd>August 1996, <a href="#about">several major changes</a> to
1415 <dt>0.10b<dd>November 1996, fixes a few minor bugs.
1416 <dt>0.10c<dd>July 1997, fixes a bug related to precedence declarations.
1417 <dt>0.10e<dd>September 1997, fixes a bug introduced in 0.10c relating
1418 to <tt>nonassoc</tt> precedence. Thanks to
1419 <a href="http://www.cs.purdue.edu/homes/hosking">Tony Hosking</a>
1420 for reporting the bug and providing the fix.
1421 Also recognizes carriage-return character as white space and fixes a
1422 number of other small bugs.
1423 <dt>0.10f<dd>December 1997, was a maintenance release. The CUP source
1424 was cleaned up for JDK 1.1.
1425 <dt>0.10g<dd>March 1998, adds new features and fixes old bugs.
1426 The behavior of RESULT assignments was normalized, and a problem
1427 with implicit start productions was fixed. The CUP grammar was
1428 extended to allow array types for terminals and non-terminals, and
1429 a command-line flag was added to allow the generation of a symbol
1430 <i>interface</i>, rather than class. Bugs associated with multiple
1431 invocations of a single parser object and multiple CUP classes in one
1432 package have been stomped on. Documentation was updated, as well.
1433 <dt>0.10h-0.10i<dd>February 1999, are maintenance releases.
1434 <dt>0.10j<dd>July 1999, broadened the CUP input grammar to allow more
1435 flexibility and improved scanner integration via the
1436 <code>java_cup.runtime.Scanner</code> interface.
1442 <a name="trademark">
1443 Java and HotJava are
1444 trademarks of <a href="http://www.sun.com/">Sun Microsystems, Inc.</a>,
1445 and refer to Sun's Java programming language and HotJava browser
1447 CUP is not sponsored by or affiliated with Sun Microsystems, Inc.