xref: /NextBSD/lib/libedit/tokenizer.c (revision 287e3b14e9552995def1802ec9c5034f4adf28ec)
1 /*	$NetBSD: tokenizer.c,v 1.21 2011/08/16 16:25:15 christos Exp $	*/
2 
3 /*-
4  * Copyright (c) 1992, 1993
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * This code is derived from software contributed to Berkeley by
8  * Christos Zoulas of Cornell University.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. Neither the name of the University nor the names of its contributors
19  *    may be used to endorse or promote products derived from this software
20  *    without specific prior written permission.
21  *
22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  */
34 
35 #include "config.h"
36 #if !defined(lint) && !defined(SCCSID)
37 #if 0
38 static char sccsid[] = "@(#)tokenizer.c	8.1 (Berkeley) 6/4/93";
39 #else
40 __RCSID("$NetBSD: tokenizer.c,v 1.21 2011/08/16 16:25:15 christos Exp $");
41 #endif
42 #endif /* not lint && not SCCSID */
43 #include <sys/cdefs.h>
44 __FBSDID("$FreeBSD$");
45 
46 /* We build this file twice, once as NARROW, once as WIDE. */
47 /*
48  * tokenize.c: Bourne shell like tokenizer
49  */
50 #include <string.h>
51 #include <stdlib.h>
52 #include "histedit.h"
53 #include "chartype.h"
54 
55 typedef enum {
56 	Q_none, Q_single, Q_double, Q_one, Q_doubleone
57 } quote_t;
58 
59 #define	TOK_KEEP	1
60 #define	TOK_EAT		2
61 
62 #define	WINCR		20
63 #define	AINCR		10
64 
65 #define	IFS		STR("\t \n")
66 
67 #define	tok_malloc(a)		malloc(a)
68 #define	tok_free(a)		free(a)
69 #define	tok_realloc(a, b)	realloc(a, b)
70 #define	tok_strdup(a)		Strdup(a)
71 
72 
TYPE(tokenizer)73 struct TYPE(tokenizer) {
74 	Char	*ifs;		/* In field separator			 */
75 	size_t	 argc, amax;	/* Current and maximum number of args	 */
76 	Char   **argv;		/* Argument list			 */
77 	Char	*wptr, *wmax;	/* Space and limit on the word buffer	 */
78 	Char	*wstart;	/* Beginning of next word		 */
79 	Char	*wspace;	/* Space of word buffer			 */
80 	quote_t	 quote;		/* Quoting state			 */
81 	int	 flags;		/* flags;				 */
82 };
83 
84 
85 private void FUN(tok,finish)(TYPE(Tokenizer) *);
86 
87 
88 /* FUN(tok,finish)():
89  *	Finish a word in the tokenizer.
90  */
91 private void
FUN(tok,finish)92 FUN(tok,finish)(TYPE(Tokenizer) *tok)
93 {
94 
95 	*tok->wptr = '\0';
96 	if ((tok->flags & TOK_KEEP) || tok->wptr != tok->wstart) {
97 		tok->argv[tok->argc++] = tok->wstart;
98 		tok->argv[tok->argc] = NULL;
99 		tok->wstart = ++tok->wptr;
100 	}
101 	tok->flags &= ~TOK_KEEP;
102 }
103 
104 
105 /* FUN(tok,init)():
106  *	Initialize the tokenizer
107  */
TYPE(Tokenizer)108 public TYPE(Tokenizer) *
109 FUN(tok,init)(const Char *ifs)
110 {
111 	TYPE(Tokenizer) *tok = tok_malloc(sizeof(*tok));
112 
113 	if (tok == NULL)
114 		return NULL;
115 	tok->ifs = tok_strdup(ifs ? ifs : IFS);
116 	if (tok->ifs == NULL) {
117 		tok_free(tok);
118 		return NULL;
119 	}
120 	tok->argc = 0;
121 	tok->amax = AINCR;
122 	tok->argv = tok_malloc(sizeof(*tok->argv) * tok->amax);
123 	if (tok->argv == NULL) {
124 		tok_free(tok->ifs);
125 		tok_free(tok);
126 		return NULL;
127 	}
128 	tok->argv[0] = NULL;
129 	tok->wspace = tok_malloc(WINCR * sizeof(*tok->wspace));
130 	if (tok->wspace == NULL) {
131 		tok_free(tok->argv);
132 		tok_free(tok->ifs);
133 		tok_free(tok);
134 		return NULL;
135 	}
136 	tok->wmax = tok->wspace + WINCR;
137 	tok->wstart = tok->wspace;
138 	tok->wptr = tok->wspace;
139 	tok->flags = 0;
140 	tok->quote = Q_none;
141 
142 	return tok;
143 }
144 
145 
146 /* FUN(tok,reset)():
147  *	Reset the tokenizer
148  */
149 public void
FUN(tok,reset)150 FUN(tok,reset)(TYPE(Tokenizer) *tok)
151 {
152 
153 	tok->argc = 0;
154 	tok->wstart = tok->wspace;
155 	tok->wptr = tok->wspace;
156 	tok->flags = 0;
157 	tok->quote = Q_none;
158 }
159 
160 
161 /* FUN(tok,end)():
162  *	Clean up
163  */
164 public void
FUN(tok,end)165 FUN(tok,end)(TYPE(Tokenizer) *tok)
166 {
167 
168 	tok_free(tok->ifs);
169 	tok_free(tok->wspace);
170 	tok_free(tok->argv);
171 	tok_free(tok);
172 }
173 
174 
175 
176 /* FUN(tok,line)():
177  *	Bourne shell (sh(1)) like tokenizing
178  *	Arguments:
179  *		tok	current tokenizer state (setup with FUN(tok,init)())
180  *		line	line to parse
181  *	Returns:
182  *		-1	Internal error
183  *		 3	Quoted return
184  *		 2	Unmatched double quote
185  *		 1	Unmatched single quote
186  *		 0	Ok
187  *	Modifies (if return value is 0):
188  *		argc	number of arguments
189  *		argv	argument array
190  *		cursorc	if !NULL, argv element containing cursor
191  *		cursorv	if !NULL, offset in argv[cursorc] of cursor
192  */
193 public int
FUN(tok,line)194 FUN(tok,line)(TYPE(Tokenizer) *tok, const TYPE(LineInfo) *line,
195     int *argc, const Char ***argv, int *cursorc, int *cursoro)
196 {
197 	const Char *ptr;
198 	int cc, co;
199 
200 	cc = co = -1;
201 	ptr = line->buffer;
202 	for (ptr = line->buffer; ;ptr++) {
203 		if (ptr >= line->lastchar)
204 			ptr = STR("");
205 		if (ptr == line->cursor) {
206 			cc = (int)tok->argc;
207 			co = (int)(tok->wptr - tok->wstart);
208 		}
209 		switch (*ptr) {
210 		case '\'':
211 			tok->flags |= TOK_KEEP;
212 			tok->flags &= ~TOK_EAT;
213 			switch (tok->quote) {
214 			case Q_none:
215 				tok->quote = Q_single;	/* Enter single quote
216 							 * mode */
217 				break;
218 
219 			case Q_single:	/* Exit single quote mode */
220 				tok->quote = Q_none;
221 				break;
222 
223 			case Q_one:	/* Quote this ' */
224 				tok->quote = Q_none;
225 				*tok->wptr++ = *ptr;
226 				break;
227 
228 			case Q_double:	/* Stay in double quote mode */
229 				*tok->wptr++ = *ptr;
230 				break;
231 
232 			case Q_doubleone:	/* Quote this ' */
233 				tok->quote = Q_double;
234 				*tok->wptr++ = *ptr;
235 				break;
236 
237 			default:
238 				return -1;
239 			}
240 			break;
241 
242 		case '"':
243 			tok->flags &= ~TOK_EAT;
244 			tok->flags |= TOK_KEEP;
245 			switch (tok->quote) {
246 			case Q_none:	/* Enter double quote mode */
247 				tok->quote = Q_double;
248 				break;
249 
250 			case Q_double:	/* Exit double quote mode */
251 				tok->quote = Q_none;
252 				break;
253 
254 			case Q_one:	/* Quote this " */
255 				tok->quote = Q_none;
256 				*tok->wptr++ = *ptr;
257 				break;
258 
259 			case Q_single:	/* Stay in single quote mode */
260 				*tok->wptr++ = *ptr;
261 				break;
262 
263 			case Q_doubleone:	/* Quote this " */
264 				tok->quote = Q_double;
265 				*tok->wptr++ = *ptr;
266 				break;
267 
268 			default:
269 				return -1;
270 			}
271 			break;
272 
273 		case '\\':
274 			tok->flags |= TOK_KEEP;
275 			tok->flags &= ~TOK_EAT;
276 			switch (tok->quote) {
277 			case Q_none:	/* Quote next character */
278 				tok->quote = Q_one;
279 				break;
280 
281 			case Q_double:	/* Quote next character */
282 				tok->quote = Q_doubleone;
283 				break;
284 
285 			case Q_one:	/* Quote this, restore state */
286 				*tok->wptr++ = *ptr;
287 				tok->quote = Q_none;
288 				break;
289 
290 			case Q_single:	/* Stay in single quote mode */
291 				*tok->wptr++ = *ptr;
292 				break;
293 
294 			case Q_doubleone:	/* Quote this \ */
295 				tok->quote = Q_double;
296 				*tok->wptr++ = *ptr;
297 				break;
298 
299 			default:
300 				return -1;
301 			}
302 			break;
303 
304 		case '\n':
305 			tok->flags &= ~TOK_EAT;
306 			switch (tok->quote) {
307 			case Q_none:
308 				goto tok_line_outok;
309 
310 			case Q_single:
311 			case Q_double:
312 				*tok->wptr++ = *ptr;	/* Add the return */
313 				break;
314 
315 			case Q_doubleone:   /* Back to double, eat the '\n' */
316 				tok->flags |= TOK_EAT;
317 				tok->quote = Q_double;
318 				break;
319 
320 			case Q_one:	/* No quote, more eat the '\n' */
321 				tok->flags |= TOK_EAT;
322 				tok->quote = Q_none;
323 				break;
324 
325 			default:
326 				return 0;
327 			}
328 			break;
329 
330 		case '\0':
331 			switch (tok->quote) {
332 			case Q_none:
333 				/* Finish word and return */
334 				if (tok->flags & TOK_EAT) {
335 					tok->flags &= ~TOK_EAT;
336 					return 3;
337 				}
338 				goto tok_line_outok;
339 
340 			case Q_single:
341 				return 1;
342 
343 			case Q_double:
344 				return 2;
345 
346 			case Q_doubleone:
347 				tok->quote = Q_double;
348 				*tok->wptr++ = *ptr;
349 				break;
350 
351 			case Q_one:
352 				tok->quote = Q_none;
353 				*tok->wptr++ = *ptr;
354 				break;
355 
356 			default:
357 				return -1;
358 			}
359 			break;
360 
361 		default:
362 			tok->flags &= ~TOK_EAT;
363 			switch (tok->quote) {
364 			case Q_none:
365 				if (Strchr(tok->ifs, *ptr) != NULL)
366 					FUN(tok,finish)(tok);
367 				else
368 					*tok->wptr++ = *ptr;
369 				break;
370 
371 			case Q_single:
372 			case Q_double:
373 				*tok->wptr++ = *ptr;
374 				break;
375 
376 
377 			case Q_doubleone:
378 				*tok->wptr++ = '\\';
379 				tok->quote = Q_double;
380 				*tok->wptr++ = *ptr;
381 				break;
382 
383 			case Q_one:
384 				tok->quote = Q_none;
385 				*tok->wptr++ = *ptr;
386 				break;
387 
388 			default:
389 				return -1;
390 
391 			}
392 			break;
393 		}
394 
395 		if (tok->wptr >= tok->wmax - 4) {
396 			size_t size = (size_t)(tok->wmax - tok->wspace + WINCR);
397 			Char *s = tok_realloc(tok->wspace,
398 			    size * sizeof(*s));
399 			if (s == NULL)
400 				return -1;
401 
402 			if (s != tok->wspace) {
403 				size_t i;
404 				for (i = 0; i < tok->argc; i++) {
405 				    tok->argv[i] =
406 					(tok->argv[i] - tok->wspace) + s;
407 				}
408 				tok->wptr = (tok->wptr - tok->wspace) + s;
409 				tok->wstart = (tok->wstart - tok->wspace) + s;
410 				tok->wspace = s;
411 			}
412 			tok->wmax = s + size;
413 		}
414 		if (tok->argc >= tok->amax - 4) {
415 			Char **p;
416 			tok->amax += AINCR;
417 			p = tok_realloc(tok->argv, tok->amax * sizeof(*p));
418 			if (p == NULL)
419 				return -1;
420 			tok->argv = p;
421 		}
422 	}
423  tok_line_outok:
424 	if (cc == -1 && co == -1) {
425 		cc = (int)tok->argc;
426 		co = (int)(tok->wptr - tok->wstart);
427 	}
428 	if (cursorc != NULL)
429 		*cursorc = cc;
430 	if (cursoro != NULL)
431 		*cursoro = co;
432 	FUN(tok,finish)(tok);
433 	*argv = (const Char **)tok->argv;
434 	*argc = (int)tok->argc;
435 	return 0;
436 }
437 
438 /* FUN(tok,str)():
439  *	Simpler version of tok_line, taking a NUL terminated line
440  *	and splitting into words, ignoring cursor state.
441  */
442 public int
FUN(tok,str)443 FUN(tok,str)(TYPE(Tokenizer) *tok, const Char *line, int *argc,
444     const Char ***argv)
445 {
446 	TYPE(LineInfo) li;
447 
448 	memset(&li, 0, sizeof(li));
449 	li.buffer = line;
450 	li.cursor = li.lastchar = Strchr(line, '\0');
451 	return FUN(tok,line(tok, &li, argc, argv, NULL, NULL));
452 }
453