1 '\" te
2 .\" Copyright (c) 1992, X/Open Company Limited All Rights Reserved Portions Copyright (c) 1997, Sun Microsystems, Inc. All Rights Reserved
3 .\" Sun Microsystems, Inc. gratefully acknowledges The Open Group for permission to reproduce portions of its copyrighted documentation. Original documentation from The Open Group can be obtained online at
4 .\" http://www.opengroup.org/bookstore/.
5 .\" The Institute of Electrical and Electronics Engineers and The Open Group, have given us permission to reprint portions of their documentation. In the following statement, the phrase "this text" refers to portions of the system documentation. Portions of this text are reprinted and reproduced in electronic form in the Sun OS Reference Manual, from IEEE Std 1003.1, 2004 Edition, Standard for Information Technology -- Portable Operating System Interface (POSIX), The Open Group Base Specifications Issue 6, Copyright (C) 2001-2004 by the Institute of Electrical and Electronics Engineers, Inc and The Open Group. In the event of any discrepancy between these versions and the original IEEE and The Open Group Standard, the original IEEE and The Open Group Standard is the referee document. The original Standard can be obtained online at http://www.opengroup.org/unix/online.html.
6 .\" This notice shall appear on any product containing this material.
7 .\" The contents of this file are subject to the terms of the Common Development and Distribution License (the "License"). You may not use this file except in compliance with the License.
8 .\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE or http://www.opensolaris.org/os/licensing. See the License for the specific language governing permissions and limitations under the License.
9 .\" When distributing Covered Code, include this CDDL HEADER in each file and include the License file at usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your own identifying information: Portions Copyright [yyyy] [name of copyright owner]
10 .TH LEX 1 "Aug 22, 1997"
11 .SH NAME
12 lex \- generate programs for lexical tasks
13 .SH SYNOPSIS
14 .LP
15 .nf
16 \fBlex\fR [\fB-cntv\fR] [\fB-e\fR | \fB-w\fR] [\fB-V\fR \fB-Q\fR [y | n]] [\fIfile\fR]...
17 .fi
18
19 .SH DESCRIPTION
20 .sp
21 .LP
22 The \fBlex\fR utility generates C programs to be used in lexical processing of
23 character input, and that can be used as an interface to \fByacc\fR. The C
24 programs are generated from \fBlex\fR source code and conform to the ISO C
25 standard. Usually, the \fBlex\fR utility writes the program it generates to the
26 file \fBlex.yy.c\fR. The state of this file is unspecified if \fBlex\fR exits
27 with a non-zero exit status. See \fBEXTENDED DESCRIPTION\fR for a complete
28 description of the \fBlex\fR input language.
29 .SH OPTIONS
30 .sp
31 .LP
32 The following options are supported:
33 .sp
34 .ne 2
35 .na
36 \fB\fB-c\fR \fR
37 .ad
38 .RS 12n
39 Indicates C-language action (default option).
40 .RE
41
42 .sp
43 .ne 2
44 .na
45 \fB\fB-e\fR \fR
46 .ad
47 .RS 12n
48 Generates a program that can handle \fBEUC\fR characters (cannot be used with
49 the \fB-w\fR option). \fByytext[\|]\fR is of type \fBunsigned char[\|]\fR.
50 .RE
51
52 .sp
53 .ne 2
54 .na
55 \fB\fB-n\fR \fR
56 .ad
57 .RS 12n
58 Suppresses the summary of statistics usually written with the \fB-v\fR option.
59 If no table sizes are specified in the \fBlex\fR source code and the \fB-v\fR
60 option is not specified, then \fB-n\fR is implied.
61 .RE
62
63 .sp
64 .ne 2
65 .na
66 \fB\fB-t\fR \fR
67 .ad
68 .RS 12n
69 Writes the resulting program to standard output instead of \fBlex.yy.c\fR.
70 .RE
71
72 .sp
73 .ne 2
74 .na
75 \fB\fB-v\fR \fR
76 .ad
77 .RS 12n
78 Writes a summary of \fBlex\fR statistics to the standard error. (See the
79 discussion of \fBlex\fR table sizes under the heading \fBDefinitions in
80 lex\fR.) If table sizes are specified in the \fBlex\fR source code, and if the
81 \fB-n\fR option is not specified, the \fB-v\fR option may be enabled.
82 .RE
83
84 .sp
85 .ne 2
86 .na
87 \fB\fB-w\fR \fR
88 .ad
89 .RS 12n
90 Generates a program that can handle \fBEUC\fR characters (cannot be used with
91 the \fB-e\fR option). Unlike the \fB-e\fR option, \fByytext[\|]\fR is of type
92 \fBwchar_t[\|]\fR.
93 .RE
94
95 .sp
96 .ne 2
97 .na
98 \fB\fB-V\fR \fR
99 .ad
100 .RS 12n
101 Prints out version information on standard error.
102 .RE
103
104 .sp
105 .ne 2
106 .na
107 \fB\fB\fR\fB-Q\fR\fB[y|n]\fR \fR
108 .ad
109 .RS 12n
110 Prints out version information to output file \fBlex.yy.c\fR by using
111 \fB-Qy\fR. The \fB-Qn\fR option does not print out version information and is
112 the default.
113 .RE
114
115 .SH OPERANDS
116 .sp
117 .LP
118 The following operand is supported:
119 .sp
120 .ne 2
121 .na
122 \fB\fIfile\fR \fR
123 .ad
124 .RS 9n
125 A pathname of an input file. If more than one such \fIfile\fR is specified, all
126 files will be concatenated to produce a single \fBlex\fR program. If no
127 \fIfile\fR operands are specified, or if a \fIfile\fR operand is \fB\(mi\fR,
128 the standard input will be used.
129 .RE
130
131 .SH OUTPUT
132 .sp
133 .LP
134 The \fBlex\fR output files are described below.
135 .SS "Stdout"
136 .sp
137 .LP
138 If the \fB-t\fR option is specified, the text file of C source code output of
139 \fBlex\fR will be written to standard output.
140 .SS "Stderr"
141 .sp
142 .LP
143 If the \fB-t\fR option is specified informational, error and warning messages
144 concerning the contents of \fBlex\fR source code input will be written to the
145 standard error.
146 .sp
147 .LP
148 If the \fB-t\fR option is not specified:
149 .RS +4
150 .TP
151 1.
152 Informational error and warning messages concerning the contents of
153 \fBlex\fR source code input will be written to either the standard output or
154 standard error.
155 .RE
156 .RS +4
157 .TP
158 2.
159 If the \fB-v\fR option is specified and the \fB-n\fR option is not
160 specified, \fBlex\fR statistics will also be written to standard error. These
161 statistics may also be generated if table sizes are specified with a \fB%\fR
162 operator in the \fBDefinitions\fR \fBin\fR \fBlex\fR section (see \fBEXTENDED
163 DESCRIPTION\fR), as long as the \fB-n\fR option is not specified.
164 .RE
165 .SS "Output Files"
166 .sp
167 .LP
168 A text file containing C source code will be written to \fBlex.yy.c\fR, or to
169 the standard output if the \fB-t\fR option is present.
170 .SH EXTENDED DESCRIPTION
171 .sp
172 .LP
173 Each input file contains \fBlex\fR source code, which is a table of regular
174 expressions with corresponding actions in the form of C program fragments.
175 .sp
176 .LP
177 When \fBlex.yy.c\fR is compiled and linked with the \fBlex\fR library (using
178 the \fB\fR\fB-l\fR\fB l\fR operand with \fBc89\fR or \fBcc\fR), the resulting
179 program reads character input from the standard input and partitions it into
180 strings that match the given expressions.
181 .sp
182 .LP
183 When an expression is matched, these actions will occur:
184 .RS +4
185 .TP
186 .ie t \(bu
187 .el o
188 The input string that was matched is left in \fIyytext\fR as a null-terminated
189 string; \fIyytext\fR is either an external character array or a pointer to a
190 character string. As explained in \fBDefinitions in lex\fR, the type can be
191 explicitly selected using the \fB%array\fR or \fB%pointer\fR declarations, but
192 the default is \fB%array\fR.
193 .RE
194 .RS +4
195 .TP
196 .ie t \(bu
197 .el o
198 The external \fBint\fR \fIyyleng\fR is set to the length of the matching
199 string.
200 .RE
201 .RS +4
202 .TP
203 .ie t \(bu
204 .el o
205 The expression's corresponding program fragment, or action, is executed.
206 .RE
207 .sp
208 .LP
209 During pattern matching, \fBlex\fR searches the set of patterns for the single
210 longest possible match. Among rules that match the same number of characters,
211 the rule given first will be chosen.
212 .sp
213 .LP
214 The general format of \fBlex\fR source is:
215 .sp
216 .in +2
217 .nf
218 \fIDefinitions\fR
219 %%
220 \fIRules\fR
221 %%
222 \fIUser Subroutines\fR
223 .fi
224 .in -2
225
226 .sp
227 .LP
228 The first \fB%%\fR is required to mark the beginning of the rules (regular
229 expressions and actions); the second \fB%%\fR is required only if user
230 subroutines follow.
231 .sp
232 .LP
233 Any line in the \fBDefinitions\fR \fBin\fR \fBlex\fR section beginning with a
234 blank character will be assumed to be a C program fragment and will be copied
235 to the external definition area of the \fBlex.yy.c\fR file. Similarly, anything
236 in the \fBDefinitions\fR \fBin\fR \fBlex\fR section included between delimiter
237 lines containing only \fB%{\fR and \fB%}\fR will also be copied unchanged to
238 the external definition area of the \fBlex.yy.c\fR file.
239 .sp
240 .LP
241 Any such input (beginning with a blank character or within \fB%{\fR and
242 \fB%}\fR delimiter lines) appearing at the beginning of the \fIRules\fR section
243 before any rules are specified will be written to \fBlex.yy.c\fR after the
244 declarations of variables for the \fByylex\fR function and before the first
245 line of code in \fByylex\fR. Thus, user variables local to \fByylex\fR can be
246 declared here, as well as application code to execute upon entry to
247 \fByylex\fR.
248 .sp
249 .LP
250 The action taken by \fBlex\fR when encountering any input beginning with a
251 blank character or within \fB%{\fR and \fB%}\fR delimiter lines appearing in
252 the \fIRules\fR section but coming after one or more rules is undefined. The
253 presence of such input may result in an erroneous definition of the \fByylex\fR
254 function.
255 .SS "Definitions in lex"
256 .sp
257 .LP
258 \fBDefinitions\fR \fBin\fR \fBlex\fR appear before the first \fB%%\fR
259 delimiter. Any line in this section not contained between \fB%{\fR and \fB%}\fR
260 lines and not beginning with a blank character is assumed to define a \fBlex\fR
261 substitution string. The format of these lines is:
262 .sp
263 .in +2
264 .nf
265 \fIname substitute\fR
266 .fi
267 .in -2
268 .sp
269
270 .sp
271 .LP
272 If a \fIname\fR does not meet the requirements for identifiers in the ISO C
273 standard, the result is undefined. The string \fIsubstitute\fR will replace the
274 string \fI{\fR \fIname\fR \fI}\fR when it is used in a rule. The \fIname\fR
275 string is recognized in this context only when the braces are provided and when
276 it does not appear within a bracket expression or within double-quotes.
277 .sp
278 .LP
279 In the \fBDefinitions\fR \fBin\fR \fBlex\fR section, any line beginning with a
280 \fB%\fR (percent sign) character and followed by an alphanumeric word beginning
281 with either \fBs\fR or \fBS\fR defines a set of start conditions. Any line
282 beginning with a \fB%\fR followed by a word beginning with either \fBx\fR or
283 \fBX\fR defines a set of exclusive start conditions. When the generated scanner
284 is in a \fB%s\fR state, patterns with no state specified will be also active;
285 in a \fB%x\fR state, such patterns will not be active. The rest of the line,
286 after the first word, is considered to be one or more blank-character-separated
287 names of start conditions. Start condition names are constructed in the same
288 way as definition names. Start conditions can be used to restrict the matching
289 of regular expressions to one or more states as described in \fBRegular
290 expressions in lex\fR.
291 .sp
292 .LP
293 Implementations accept either of the following two mutually exclusive
294 declarations in the \fBDefinitions\fR \fBin\fR \fBlex\fR section:
295 .sp
296 .ne 2
297 .na
298 \fB\fB%array\fR \fR
299 .ad
300 .RS 13n
301 Declare the type of \fIyytext\fR to be a null-terminated character array.
302 .RE
303
304 .sp
305 .ne 2
306 .na
307 \fB\fB%pointer\fR \fR
308 .ad
309 .RS 13n
310 Declare the type of \fIyytext\fR to be a pointer to a null-terminated character
311 string.
312 .RE
313
314 .sp
315 .LP
316 \fBNote:\fR When using the \fB%pointer\fR option, you may not also use the
317 \fByyless\fR function to alter \fIyytext\fR.
318 .sp
319 .LP
320 \fB%array\fR is the default. If \fB%array\fR is specified (or neither
321 \fB%array\fR nor \fB%pointer\fR is specified), then the correct way to make an
322 external reference to \fIyyext\fR is with a declaration of the form:
323 .sp
324 .LP
325 \fBextern char\fR\fI yytext\fR\fB[\|]\fR
326 .sp
327 .LP
328 If \fB%pointer\fR is specified, then the correct external reference is of the
329 form:
330 .sp
331 .LP
332 \fBextern char *\fR\fIyytext\fR\fB;\fR
333 .sp
334 .LP
335 \fBlex\fR will accept declarations in the \fBDefinitions in lex\fR section for
336 setting certain internal table sizes. The declarations are shown in the
337 following table.
338 .sp
339 .LP
340 \fBTable\fR \fBSize\fR \fBDeclaration\fR \fBin\fR \fBlex\fR
341 .sp
342
343 .sp
344 .TS
345 box;
346 c c c
347 l l l .
348 \fBDeclaration\fR \fBDescription\fR \fBDefault\fR
349 _
350 \fB%p\fR\fIn\fR Number of positions 2500
351 \fB%n\fR\fIn\fR Number of states 500
352 \fB%a\fR\fI n\fR Number of transitions 2000
353 \fB%e\fR\fIn\fR Number of parse tree nodes 1000
354 \fB%k\fR\fIn\fR Number of packed character classes 10000
355 \fB%o\fR\fIn\fR Size of the output array 3000
356 .TE
357
358 .sp
359 .LP
360 Programs generated by \fBlex\fR need either the \fB-e\fR or \fB-w\fR option to
361 handle input that contains \fBEUC\fR characters from supplementary codesets. If
362 neither of these options is specified, \fByytext\fR is of the type
363 \fBchar[\|]\fR, and the generated program can handle only \fBASCII\fR
364 characters.
365 .sp
366 .LP
367 When the \fB-e\fR option is used, \fByytext\fR is of the type \fBunsigned\fR
368 \fBchar[\|]\fR and \fByyleng\fR gives the total number of \fIbytes\fR in the
369 matched string. With this option, the macros \fBinput()\fR,
370 \fBunput(\fIc\fR)\fR, and \fBoutput(\fIc\fR)\fR should do a byte-based
371 \fBI/O\fR in the same way as with the regular \fBASCII\fR \fBlex\fR. Two more
372 variables are available with the \fB-e\fR option, \fByywtext\fR and
373 \fByywleng\fR, which behave the same as \fByytext\fR and \fByyleng\fR would
374 under the \fB-w\fR option.
375 .sp
376 .LP
377 When the \fB-w\fR option is used, \fByytext\fR is of the type \fBwchar_t[\|]\fR
378 and \fByyleng\fR gives the total number of \fIcharacters\fR in the matched
379 string. If you supply your own \fBinput()\fR, \fBunput(\fIc\fR)\fR, or
380 \fBoutput(\fR\fIc\fR\fB)\fR macros with this option, they must return or accept
381 \fBEUC\fR characters in the form of wide character (\fBwchar_t\fR). This allows
382 a different interface between your program and the lex internals, to expedite
383 some programs.
384 .SS "Rules in lex"
385 .sp
386 .LP
387 The \fBRules\fR \fBin\fR \fBlex\fR source files are a table in which the left
388 column contains regular expressions and the right column contains actions (C
389 program fragments) to be executed when the expressions are recognized.
390 .sp
391 .in +2
392 .nf
393 \fIERE action\fR
394 \fIERE action\fR
395 \&...
396 .fi
397 .in -2
398
399 .sp
400 .LP
401 The extended regular expression (ERE) portion of a row will be separated from
402 \fIaction\fR by one or more blank characters. A regular expression containing
403 blank characters is recognized under one of the following conditions:
404 .RS +4
405 .TP
406 .ie t \(bu
407 .el o
408 The entire expression appears within double-quotes.
409 .RE
410 .RS +4
411 .TP
412 .ie t \(bu
413 .el o
414 The blank characters appear within double-quotes or square brackets.
415 .RE
416 .RS +4
417 .TP
418 .ie t \(bu
419 .el o
420 Each blank character is preceded by a backslash character.
421 .RE
422 .SS "User Subroutines in lex"
423 .sp
424 .LP
425 Anything in the user subroutines section will be copied to \fBlex.yy.c\fR
426 following \fByylex\fR.
427 .SS "Regular Expressions in lex"
428 .sp
429 .LP
430 The \fBlex\fR utility supports the set of Extended Regular Expressions (EREs)
431 described on \fBregex\fR(5) with the following additions and exceptions to the
432 syntax:
433 .sp
434 .ne 2
435 .na
436 \fB\fB\|.\|.\|.\fR \fR
437 .ad
438 .RS 14n
439 Any string enclosed in double-quotes will represent the characters within the
440 double-quotes as themselves, except that backslash escapes (which appear in the
441 following table) are recognized. Any backslash-escape sequence is terminated by
442 the closing quote. For example, "\|\e\|01""1" represents a single string: the
443 octal value 1 followed by the character 1.
444 .RE
445
446 .sp
447 .LP
448 \fI<\fR\fIstate\fR\fI>\fR\fIr\fR
449 .sp
450 .ne 2
451 .na
452 \fB<\fIstate1\fR, \fIstate2\fR, \|.\|.\|.\|>\fIr\fR\fR
453 .ad
454 .sp .6
455 .RS 4n
456 The regular expression \fIr\fR will be matched only when the program is in one
457 of the start conditions indicated by \fIstate\fR, \fIstate1\fR, and so forth.
458 For more information, see \fBActions in lex\fR. As an exception to the
459 typographical conventions of the rest of this document, in this case
460 <\fIstate\fR> does not represent a metavariable, but the literal angle-bracket
461 characters surrounding a symbol. The start condition is recognized as such only
462 at the beginning of a regular expression.
463 .RE
464
465 .sp
466 .ne 2
467 .na
468 \fB\fIr\fR/\fIx\fR \fR
469 .ad
470 .sp .6
471 .RS 4n
472 The regular expression \fIr\fR will be matched only if it is followed by an
473 occurrence of regular expression \fIx\fR. The token returned in \fIyytext\fR
474 will only match \fIr\fR. If the trailing portion of \fIr\fR matches the
475 beginning of \fIx\fR, the result is unspecified. The \fIr\fR expression cannot
476 include further trailing context or the \fB$\fR (match-end-of-line) operator;
477 \fIx\fR cannot include the \fB^\fR (match-beginning-of-line) operator, nor
478 trailing context, nor the \fB$\fR operator. That is, only one occurrence of
479 trailing context is allowed in a \fBlex\fR regular expression, and the \fB^\fR
480 operator only can be used at the beginning of such an expression. A further
481 restriction is that the trailing-context operator \fB/\fR (slash) cannot be
482 grouped within parentheses.
483 .RE
484
485 .sp
486 .ne 2
487 .na
488 \fB\fB{\fR\fIname\fR\fB}\fR \fR
489 .ad
490 .sp .6
491 .RS 4n
492 When \fIname\fR is one of the substitution symbols from the \fIDefinitions\fR
493 section, the string, including the enclosing braces, will be replaced by the
494 \fIsubstitute\fR value. The \fIsubstitute\fR value will be treated in the
495 extended regular expression as if it were enclosed in parentheses. No
496 substitution will occur if \fB{\fR\fIname\fR\fB}\fR occurs within a bracket
497 expression or within double-quotes.
498 .RE
499
500 .sp
501 .LP
502 Within an \fBERE,\fR a backslash character (\fB\|\e\e\fR, \fB\e\|a\fR,
503 \fB\e\|b\fR, \fB\e\|f\fR, \fB\e\|n\fR, \fB\e\|r\fR, \fB\e\|t\fR, \fB\e\|v\fR)
504 is considered to begin an escape sequence. In addition, the escape sequences in
505 the following table will be recognized.
506 .sp
507 .LP
508 A literal newline character cannot occur within an \fBERE;\fR the escape
509 sequence \fB\e\|n\fR can be used to represent a newline character. A newline
510 character cannot be matched by a period operator.
511 .sp
512 .LP
513 \fBEscape Sequences in lex\fR
514 .sp
515
516 .sp
517 .TS
518 box;
519 c c c
520 c c c .
521 Escape Sequences in lex
522 _
523 Escape Sequence Description Meaning
524 _
525 \e\fIdigits\fR T{
526 A backslash character followed by the longest sequence of one, two or three octal-digit characters (01234567). Ifall of the digits are 0, (that is, representation of the NUL character), the behavior is undefined.
527 T} T{
528 The character whose encoding is represented by the one-, two- or three-digit octal integer. Multi-byte characters require multiple, concatenated escape sequences of this type, including the leading \e for each byte.
529 T}
530 _
531 \e\fBx\fR\fIdigits\fR T{
532 A backslash character followed by the longest sequence of hexadecimal-digit characters (01234567abcdefABCDEF). If all of the digits are 0, (that is, representation of the NUL character), the behavior is undefined.
533 T} T{
534 The character whose encoding is represented by the hexadecimal integer.
535 T}
536 _
537 \e\fIc\fR T{
538 A backslash character followed by any character not described in this table. (\e\e, \ea, \eb, \ef, \een, \er, \et, \ev).
539 T} The character c, unchanged.
540 .TE
541
542 .sp
543 .LP
544 The order of precedence given to extended regular expressions for \fBlex\fR is
545 as shown in the following table, from high to low.
546 .sp
547 .ne 2
548 .na
549 \fB\fBNote\fR: \fR
550 .ad
551 .RS 10n
552 The escaped characters entry is not meant to imply that these are operators,
553 but they are included in the table to show their relationships to the true
554 operators. The start condition, trailing context and anchoring notations have
555 been omitted from the table because of the placement restrictions described in
556 this section; they can only appear at the beginning or ending of an \fBERE.\fR
557 .RE
558
559 .sp
560
561 .sp
562 .TS
563 box;
564 c c
565 l l .
566 ERE Precedence in lex
567 _
568 \fIcollation-related bracket symbols\fR \fB[= =] [: :] [. .]\fR
569 \fIescaped characters\fR \fB\e<\fR\fIspecial character\fR>
570 \fIbracket expression\fR \fB[ ]\fR
571 \fIquoting\fR \fB".\|.\|."\fR
572 \fIgrouping\fR \fB()\fR
573 \fIdefinition\fR \fB{\fR\fIname\fR}
574 \fIsingle-character RE duplication\fR \fB* + ?\fR
575 \fIconcatenation\fR
576 \fIinterval expression\fR \fB{\fR\fIm\fR,\fIn\fR}
577 \fIalternation\fR \fB|\fR
578 .TE
579
580 .sp
581 .LP
582 The \fBERE\fR anchoring operators (\fB\|^\fR and \fB$\fR\|) do not appear in
583 the table. With \fBlex\fR regular expressions, these operators are restricted
584 in their use: the \fB^\fR operator can only be used at the beginning of an
585 entire regular expression, and the \fB$\fR operator only at the end. The
586 operators apply to the entire regular expression. Thus, for example, the
587 pattern (\fB^abc)|(def$\fR) is undefined; it can instead be written as two
588 separate rules, one with the regular expression \fB^abc\fR and one with
589 \fBdef$\fR, which share a common action via the special \fB|\fR action (see
590 below). If the pattern were written \fB^abc|def$\fR, it would match either of
591 \fBabc\fR or \fBdef\fR on a line by itself.
592 .sp
593 .LP
594 Unlike the general \fBERE\fR rules, embedded anchoring is not allowed by most
595 historical \fBlex\fR implementations. An example of embedded anchoring would be
596 for patterns such as (^)foo($) to match \fBfoo\fR when it exists as a complete
597 word. This functionality can be obtained using existing \fBlex\fR features:
598 .sp
599 .in +2
600 .nf
601 ^foo/[ \e\|n]|
602 " foo"/[ \e\|n] /* found foo as a separate word */
603 .fi
604 .in -2
605
606 .sp
607 .LP
608 Notice also that \fB$\fR is a form of trailing context (it is equivalent to
609 \fB/\e\|n\fR and as such cannot be used with regular expressions containing
610 another instance of the operator (see the preceding discussion of trailing
611 context).
612 .sp
613 .LP
614 The additional regular expressions trailing-context operator \fB/\fR (slash)
615 can be used as an ordinary character if presented within double-quotes,
616 \fB"\|/\|"\fR; preceded by a backslash, \fB\e\|/\fR; or within a bracket
617 expression, \fB[\|/\|]\fR. The start-condition \fB<\fR and \fB>\fR operators
618 are special only in a start condition at the beginning of a regular expression;
619 elsewhere in the regular expression they are treated as ordinary characters.
620 .sp
621 .LP
622 The following examples clarify the differences between \fBlex\fR regular
623 expressions and regular expressions appearing elsewhere in this document. For
624 regular expressions of the form \fIr\fR/\fIx\fR, the string matching \fIr\fR is
625 always returned; confusion may arise when the beginning of \fIx\fR matches the
626 trailing portion of \fIr\fR. For example, given the regular expression a*b/cc
627 and the input \fBaaabcc\fR, \fIyytext\fR would contain the string \fBaaab\fR on
628 this match. But given the regular expression x*/xy and the input \fBxxxy\fR,
629 the token \fBxxx\fR, not \fBxx\fR, is returned by some implementations because
630 \fBxxx\fR matches x*.
631 .sp
632 .LP
633 In the rule ab*/bc, the b* at the end of \fIr\fR will extend \fIr\fR's match
634 into the beginning of the trailing context, so the result is unspecified. If
635 this rule were ab/bc, however, the rule matches the text \fBab\fR when it is
636 followed by the text \fBbc\fR. In this latter case, the matching of \fIr\fR
637 cannot extend into the beginning of \fIx\fR, so the result is specified.
638 .SS "Actions in lex"
639 .sp
640 .LP
641 The action to be taken when an \fBERE\fR is matched can be a C program fragment
642 or the special actions described below; the program fragment can contain one or
643 more C statements, and can also include special actions. The empty C statement
644 \fB;\fR is a valid action; any string in the \fBlex.yy.c\fR input that matches
645 the pattern portion of such a rule is effectively ignored or skipped. However,
646 the absence of an action is not valid, and the action \fBlex\fR takes in such a
647 condition is undefined.
648 .sp
649 .LP
650 The specification for an action, including C statements and special actions,
651 can extend across several lines if enclosed in braces:
652 .sp
653 .in +2
654 .nf
655 ERE <one or more blanks> { program statement
656 program statement }
657 .fi
658 .in -2
659 .sp
660
661 .sp
662 .LP
663 The default action when a string in the input to a \fBlex.yy.c\fR program is
664 not matched by any expression is to copy the string to the output. Because the
665 default behavior of a program generated by \fBlex\fR is to read the input and
666 copy it to the output, a minimal \fBlex\fR source program that has just
667 \fB%%\fR generates a C program that simply copies the input to the output
668 unchanged.
669 .sp
670 .LP
671 Four special actions are available:
672 .sp
673 .in +2
674 .nf
675 | ECHO; REJECT; BEGIN
676 .fi
677 .in -2
678 .sp
679
680 .sp
681 .ne 2
682 .na
683 \fB|\fR
684 .ad
685 .RS 12n
686 The action | means that the action for the next rule is the action for this
687 rule. Unlike the other three actions, | cannot be enclosed in braces or be
688 semicolon-terminated. It must be specified alone, with no other actions.
689 .RE
690
691 .sp
692 .ne 2
693 .na
694 \fB\fBECHO;\fR \fR
695 .ad
696 .RS 12n
697 Writes the contents of the string \fIyytext\fR on the output.
698 .RE
699
700 .sp
701 .ne 2
702 .na
703 \fB\fBREJECT;\fR \fR
704 .ad
705 .RS 12n
706 Usually only a single expression is matched by a given string in the input.
707 \fBREJECT\fR means "continue to the next expression that matches the current
708 input," and causes whatever rule was the second choice after the current rule
709 to be executed for the same input. Thus, multiple rules can be matched and
710 executed for one input string or overlapping input strings. For example, given
711 the regular expressions \fBxyz\fR and \fBxy\fR and the input \fBxyz\fR, usually
712 only the regular expression \fBxyz\fR would match. The next attempted match
713 would start after z. If the last action in the \fBxyz\fR rule is \fBREJECT\fR ,
714 both this rule and the \fBxy\fR rule would be executed. The \fBREJECT\fR action
715 may be implemented in such a fashion that flow of control does not continue
716 after it, as if it were equivalent to a \fBgoto\fR to another part of
717 \fByylex\fR. The use of \fBREJECT\fR may result in somewhat larger and slower
718 scanners.
719 .RE
720
721 .sp
722 .ne 2
723 .na
724 \fB\fBBEGIN\fR \fR
725 .ad
726 .RS 12n
727 The action:
728 .sp
729 \fBBEGIN\fR \fInewstate\fR\fB;\fR
730 .sp
731 switches the state (start condition) to \fInewstate\fR. If the string
732 \fInewstate\fR has not been declared previously as a start condition in the
733 \fBDefinitions\fR \fBin\fR \fBlex\fR section, the results are unspecified. The
734 initial state is indicated by the digit \fB0\fR or the token \fBINITIAL\fR.
735 .RE
736
737 .sp
738 .LP
739 The functions or macros described below are accessible to user code included in
740 the \fBlex\fR input. It is unspecified whether they appear in the C code output
741 of \fBlex\fR, or are accessible only through the \fB\fR\fB-l\fR\fB l\fR operand
742 to \fBc89\fR or \fBcc\fR (the \fBlex\fR library).
743 .sp
744 .ne 2
745 .na
746 \fB\fBint\fR \fByylex(void)\fR \fR
747 .ad
748 .RS 21n
749 Performs lexical analysis on the input; this is the primary function generated
750 by the \fBlex\fR utility. The function returns zero when the end of input is
751 reached; otherwise it returns non-zero values (tokens) determined by the
752 actions that are selected.
753 .RE
754
755 .sp
756 .ne 2
757 .na
758 \fB\fBint\fR \fByymore(void)\fR \fR
759 .ad
760 .RS 21n
761 When called, indicates that when the next input string is recognized, it is to
762 be appended to the current value of \fIyytext\fR rather than replacing it; the
763 value in \fIyyleng\fR is adjusted accordingly.
764 .RE
765
766 .sp
767 .ne 2
768 .na
769 \fB\fBint\fR\fIyyless(int\fR\fB n\fR\fI)\fR \fR
770 .ad
771 .RS 21n
772 Retains \fIn\fR initial characters in \fIyytext\fR, NUL-terminated, and treats
773 the remaining characters as if they had not been read; the value in
774 \fIyyleng\fR is adjusted accordingly.
775 .RE
776
777 .sp
778 .ne 2
779 .na
780 \fB\fBint\fR \fBinput(void)\fR \fR
781 .ad
782 .RS 21n
783 Returns the next character from the input, or zero on end-of-file. It obtains
784 input from the stream pointer \fIyyin\fR, although possibly via an intermediate
785 buffer. Thus, once scanning has begun, the effect of altering the value of
786 \fIyyin\fR is undefined. The character read is removed from the input stream of
787 the scanner without any processing by the scanner.
788 .RE
789
790 .sp
791 .ne 2
792 .na
793 \fB\fBint\fR \fBunput(int\fR \fB\fIc\fR\fR\fB)\fR \fR
794 .ad
795 .RS 21n
796 Returns the character \fIc\fR to the input; \fIyytext\fR and \fIyyleng\fR are
797 undefined until the next expression is matched. The result of using \fIunput\fR
798 for more characters than have been input is unspecified.
799 .RE
800
801 .sp
802 .LP
803 The following functions appear only in the \fBlex\fR library accessible through
804 the \fB\fR\fB-l\fR\fB l\fR operand; they can therefore be redefined by a
805 portable application:
806 .sp
807 .ne 2
808 .na
809 \fB\fBint\fR \fByywrap(void)\fR \fR
810 .ad
811 .sp .6
812 .RS 4n
813 Called by \fByylex\fR at end-of-file; the default \fByywrap\fR always will
814 return 1. If the application requires \fByylex\fR to continue processing with
815 another source of input, then the application can include a function
816 \fByywrap\fR, which associates another file with the external variable
817 \fBFILE\fR *\fIyyin\fR and will return a value of zero.
818 .RE
819
820 .sp
821 .ne 2
822 .na
823 \fB\fBint\fR \fBmain(int\fR \fB\fIargc\fR,\fR \fBchar\fR \fB*\fIargv\fR[\|])\fR
824 \fR
825 .ad
826 .sp .6
827 .RS 4n
828 Calls \fByylex\fR to perform lexical analysis, then exits. The user code can
829 contain \fBmain\fR to perform application-specific operations, calling
830 \fByylex\fR as applicable.
831 .RE
832
833 .sp
834 .LP
835 The reason for breaking these functions into two lists is that only those
836 functions in \fBlibl.a\fR can be reliably redefined by a portable application.
837 .sp
838 .LP
839 Except for \fBinput\fR, \fBunput\fR and \fBmain\fR, all external and static
840 names generated by \fBlex\fR begin with the prefix \fByy\fR or \fBYY\fR.
841 .SH USAGE
842 .sp
843 .LP
844 Portable applications are warned that in the \fBRules in lex\fR section, an
845 \fBERE\fR without an action is not acceptable, but need not be detected as
846 erroneous by \fBlex\fR. This may result in compilation or run-time errors.
847 .sp
848 .LP
849 The purpose of \fBinput\fR is to take characters off the input stream and
850 discard them as far as the lexical analysis is concerned. A common use is to
851 discard the body of a comment once the beginning of a comment is recognized.
852 .sp
853 .LP
854 The \fBlex\fR utility is not fully internationalized in its treatment of
855 regular expressions in the \fBlex\fR source code or generated lexical analyzer.
856 It would seem desirable to have the lexical analyzer interpret the regular
857 expressions given in the \fBlex\fR source according to the environment
858 specified when the lexical analyzer is executed, but this is not possible with
859 the current \fBlex\fR technology. Furthermore, the very nature of the lexical
860 analyzers produced by \fBlex\fR must be closely tied to the lexical
861 requirements of the input language being described, which will frequently be
862 locale-specific anyway. (For example, writing an analyzer that is used for
863 French text will not automatically be useful for processing other languages.)
864 .SH EXAMPLES
865 .LP
866 \fBExample 1 \fRUsing lex
867 .sp
868 .LP
869 The following is an example of a \fBlex\fR program that implements a
870 rudimentary scanner for a Pascal-like syntax:
871
872 .sp
873 .in +2
874 .nf
875 %{
876 /* need this for the call to atof() below */
877 #include <math.h>
878 /* need this for printf(), fopen() and stdin below */
879 #include <stdio.h>
880 %}
881
882 DIGIT [0-9]
883 ID [a-z][a-z0-9]*
884 %%
885
886 {DIGIT}+ {
887 printf("An integer: %s (%d)\en", yytext,
888 atoi(yytext));
889 }
890
891 {DIGIT}+"."{DIGIT}* {
892 printf("A float: %s (%g)\en", yytext,
893 atof(yytext));
894 }
895
896 if|then|begin|end|procedure|function {
897 printf("A keyword: %s\en", yytext);
898 }
899
900 {ID} printf("An identifier: %s\en", yytext);
901
902 "+"|"-"|"*"|"/" printf("An operator: %s\en", yytext);
903
904 "{"[^}\en]*"}" /* eat up one-line comments */
905
906 [ \et\en]+ /* eat up white space */
907
908 \&. printf("Unrecognized character: %s\en", yytext);
909
910 %%
911
912 int main(int argc, char *argv[\|])
913 {
914 ++argv, --argc; /* skip over program name */
915 if (argc > 0)
916 yyin = fopen(argv[0], "r");
917 else
918 yyin = stdin;
919
920 yylex();
921 }
922 .fi
923 .in -2
924 .sp
925
926 .SH ENVIRONMENT VARIABLES
927 .sp
928 .LP
929 See \fBenviron\fR(5) for descriptions of the following environment variables
930 that affect the execution of \fBlex\fR: \fBLANG\fR, \fBLC_ALL\fR,
931 \fBLC_COLLATE\fR, \fBLC_CTYPE\fR, \fBLC_MESSAGES\fR, and \fBNLSPATH\fR.
932 .SH EXIT STATUS
933 .sp
934 .LP
935 The following exit values are returned:
936 .sp
937 .ne 2
938 .na
939 \fB\fB0\fR \fR
940 .ad
941 .RS 7n
942 Successful completion.
943 .RE
944
945 .sp
946 .ne 2
947 .na
948 \fB\fB>0\fR \fR
949 .ad
950 .RS 7n
951 An error occurred.
952 .RE
953
954 .SH ATTRIBUTES
955 .sp
956 .LP
957 See \fBattributes\fR(5) for descriptions of the following attributes:
958 .sp
959
960 .sp
961 .TS
962 box;
963 c | c
964 l | l .
965 ATTRIBUTE TYPE ATTRIBUTE VALUE
966 _
967 Interface Stability Standard
968 .TE
969
970 .SH SEE ALSO
971 .sp
972 .LP
973 \fByacc\fR(1), \fBattributes\fR(5), \fBenviron\fR(5), \fBregex\fR(5),
974 \fBstandards\fR(5)
975 .SH NOTES
976 .sp
977 .LP
978 If routines such as \fByyback()\fR, \fByywrap()\fR, and \fByylock()\fR
979 in \fB\|.l\fR (ell) files are to be external C functions, the command line to
980 compile a C++ program must define the \fB__EXTERN_C__\fR macro. For example:
981 .sp
982 .in +2
983 .nf
984 example% \fBCC -D__EXTERN_C__ ... file\fR
985 .fi
986 .in -2
987 .sp
988