xref: /freebsd/contrib/libedit/tokenizer.c (revision 43a5ec4eb41567cc92586503212743d89686d78f) 
1 /*	$NetBSD: tokenizer.c,v 1.28 2016/04/11 18:56:31 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 1.28 2016/04/11 18:56:31 christos Exp $");
41 #endif
42 #endif /* not lint && not SCCSID */
43 
44 /* We build this file twice, once as NARROW, once as WIDE. */
45 /*
46  * tokenize.c: Bourne shell like tokenizer
47  */
48 #include <stdlib.h>
49 #include <string.h>
50 
51 #include "histedit.h"
52 
53 typedef enum {
54 	Q_none, Q_single, Q_double, Q_one, Q_doubleone
55 } quote_t;
56 
57 #define	TOK_KEEP	1
58 #define	TOK_EAT		2
59 
60 #define	WINCR		20
61 #define	AINCR		10
62 
63 #define	IFS		STR("\t \n")
64 
65 #define	tok_malloc(a)		malloc(a)
66 #define	tok_free(a)		free(a)
67 #define	tok_realloc(a, b)	realloc(a, b)
68 
69 #ifdef NARROWCHAR
70 #define	Char			char
71 #define	FUN(prefix, rest)	prefix ## _ ## rest
72 #define	TYPE(type)		type
73 #define	STR(x)			x
74 #define	Strchr(s, c)		strchr(s, c)
75 #define	tok_strdup(s)		strdup(s)
76 #else
77 #define	Char			wchar_t
78 #define	FUN(prefix, rest)	prefix ## _w ## rest
79 #define	TYPE(type)		type ## W
80 #define	STR(x)			L ## x
81 #define	Strchr(s, c)		wcschr(s, c)
82 #define	tok_strdup(s)		wcsdup(s)
83 #endif
84 
85 struct TYPE(tokenizer) {
86 	Char	*ifs;		/* In field separator			 */
87 	size_t	 argc, amax;	/* Current and maximum number of args	 */
88 	Char   **argv;		/* Argument list			 */
89 	Char	*wptr, *wmax;	/* Space and limit on the word buffer	 */
90 	Char	*wstart;	/* Beginning of next word		 */
91 	Char	*wspace;	/* Space of word buffer			 */
92 	quote_t	 quote;		/* Quoting state			 */
93 	int	 flags;		/* flags;				 */
94 };
95 
96 
97 static void FUN(tok,finish)(TYPE(Tokenizer) *);
98 
99 
100 /* FUN(tok,finish)():
101  *	Finish a word in the tokenizer.
102  */
103 static void
104 FUN(tok,finish)(TYPE(Tokenizer) *tok)
105 {
106 
107 	*tok->wptr = '\0';
108 	if ((tok->flags & TOK_KEEP) || tok->wptr != tok->wstart) {
109 		tok->argv[tok->argc++] = tok->wstart;
110 		tok->argv[tok->argc] = NULL;
111 		tok->wstart = ++tok->wptr;
112 	}
113 	tok->flags &= ~TOK_KEEP;
114 }
115 
116 
117 /* FUN(tok,init)():
118  *	Initialize the tokenizer
119  */
120 TYPE(Tokenizer) *
121 FUN(tok,init)(const Char *ifs)
122 {
123 	TYPE(Tokenizer) *tok = tok_malloc(sizeof(*tok));
124 
125 	if (tok == NULL)
126 		return NULL;
127 	tok->ifs = tok_strdup(ifs ? ifs : IFS);
128 	if (tok->ifs == NULL) {
129 		tok_free(tok);
130 		return NULL;
131 	}
132 	tok->argc = 0;
133 	tok->amax = AINCR;
134 	tok->argv = tok_malloc(sizeof(*tok->argv) * tok->amax);
135 	if (tok->argv == NULL) {
136 		tok_free(tok->ifs);
137 		tok_free(tok);
138 		return NULL;
139 	}
140 	tok->argv[0] = NULL;
141 	tok->wspace = tok_malloc(WINCR * sizeof(*tok->wspace));
142 	if (tok->wspace == NULL) {
143 		tok_free(tok->argv);
144 		tok_free(tok->ifs);
145 		tok_free(tok);
146 		return NULL;
147 	}
148 	tok->wmax = tok->wspace + WINCR;
149 	tok->wstart = tok->wspace;
150 	tok->wptr = tok->wspace;
151 	tok->flags = 0;
152 	tok->quote = Q_none;
153 
154 	return tok;
155 }
156 
157 
158 /* FUN(tok,reset)():
159  *	Reset the tokenizer
160  */
161 void
162 FUN(tok,reset)(TYPE(Tokenizer) *tok)
163 {
164 
165 	tok->argc = 0;
166 	tok->wstart = tok->wspace;
167 	tok->wptr = tok->wspace;
168 	tok->flags = 0;
169 	tok->quote = Q_none;
170 }
171 
172 
173 /* FUN(tok,end)():
174  *	Clean up
175  */
176 void
177 FUN(tok,end)(TYPE(Tokenizer) *tok)
178 {
179 
180 	tok_free(tok->ifs);
181 	tok_free(tok->wspace);
182 	tok_free(tok->argv);
183 	tok_free(tok);
184 }
185 
186 
187 
188 /* FUN(tok,line)():
189  *	Bourne shell (sh(1)) like tokenizing
190  *	Arguments:
191  *		tok	current tokenizer state (setup with FUN(tok,init)())
192  *		line	line to parse
193  *	Returns:
194  *		-1	Internal error
195  *		 3	Quoted return
196  *		 2	Unmatched double quote
197  *		 1	Unmatched single quote
198  *		 0	Ok
199  *	Modifies (if return value is 0):
200  *		argc	number of arguments
201  *		argv	argument array
202  *		cursorc	if !NULL, argv element containing cursor
203  *		cursorv	if !NULL, offset in argv[cursorc] of cursor
204  */
205 int
206 FUN(tok,line)(TYPE(Tokenizer) *tok, const TYPE(LineInfo) *line,
207     int *argc, const Char ***argv, int *cursorc, int *cursoro)
208 {
209 	const Char *ptr;
210 	int cc, co;
211 
212 	cc = co = -1;
213 	ptr = line->buffer;
214 	for (ptr = line->buffer; ;ptr++) {
215 		if (ptr >= line->lastchar)
216 			ptr = STR("");
217 		if (ptr == line->cursor) {
218 			cc = (int)tok->argc;
219 			co = (int)(tok->wptr - tok->wstart);
220 		}
221 		switch (*ptr) {
222 		case '\'':
223 			tok->flags |= TOK_KEEP;
224 			tok->flags &= ~TOK_EAT;
225 			switch (tok->quote) {
226 			case Q_none:
227 				tok->quote = Q_single;	/* Enter single quote
228 							 * mode */
229 				break;
230 
231 			case Q_single:	/* Exit single quote mode */
232 				tok->quote = Q_none;
233 				break;
234 
235 			case Q_one:	/* Quote this ' */
236 				tok->quote = Q_none;
237 				*tok->wptr++ = *ptr;
238 				break;
239 
240 			case Q_double:	/* Stay in double quote mode */
241 				*tok->wptr++ = *ptr;
242 				break;
243 
244 			case Q_doubleone:	/* Quote this ' */
245 				tok->quote = Q_double;
246 				*tok->wptr++ = *ptr;
247 				break;
248 
249 			default:
250 				return -1;
251 			}
252 			break;
253 
254 		case '"':
255 			tok->flags &= ~TOK_EAT;
256 			tok->flags |= TOK_KEEP;
257 			switch (tok->quote) {
258 			case Q_none:	/* Enter double quote mode */
259 				tok->quote = Q_double;
260 				break;
261 
262 			case Q_double:	/* Exit double quote mode */
263 				tok->quote = Q_none;
264 				break;
265 
266 			case Q_one:	/* Quote this " */
267 				tok->quote = Q_none;
268 				*tok->wptr++ = *ptr;
269 				break;
270 
271 			case Q_single:	/* Stay in single quote mode */
272 				*tok->wptr++ = *ptr;
273 				break;
274 
275 			case Q_doubleone:	/* Quote this " */
276 				tok->quote = Q_double;
277 				*tok->wptr++ = *ptr;
278 				break;
279 
280 			default:
281 				return -1;
282 			}
283 			break;
284 
285 		case '\\':
286 			tok->flags |= TOK_KEEP;
287 			tok->flags &= ~TOK_EAT;
288 			switch (tok->quote) {
289 			case Q_none:	/* Quote next character */
290 				tok->quote = Q_one;
291 				break;
292 
293 			case Q_double:	/* Quote next character */
294 				tok->quote = Q_doubleone;
295 				break;
296 
297 			case Q_one:	/* Quote this, restore state */
298 				*tok->wptr++ = *ptr;
299 				tok->quote = Q_none;
300 				break;
301 
302 			case Q_single:	/* Stay in single quote mode */
303 				*tok->wptr++ = *ptr;
304 				break;
305 
306 			case Q_doubleone:	/* Quote this \ */
307 				tok->quote = Q_double;
308 				*tok->wptr++ = *ptr;
309 				break;
310 
311 			default:
312 				return -1;
313 			}
314 			break;
315 
316 		case '\n':
317 			tok->flags &= ~TOK_EAT;
318 			switch (tok->quote) {
319 			case Q_none:
320 				goto tok_line_outok;
321 
322 			case Q_single:
323 			case Q_double:
324 				*tok->wptr++ = *ptr;	/* Add the return */
325 				break;
326 
327 			case Q_doubleone:   /* Back to double, eat the '\n' */
328 				tok->flags |= TOK_EAT;
329 				tok->quote = Q_double;
330 				break;
331 
332 			case Q_one:	/* No quote, more eat the '\n' */
333 				tok->flags |= TOK_EAT;
334 				tok->quote = Q_none;
335 				break;
336 
337 			default:
338 				return 0;
339 			}
340 			break;
341 
342 		case '\0':
343 			switch (tok->quote) {
344 			case Q_none:
345 				/* Finish word and return */
346 				if (tok->flags & TOK_EAT) {
347 					tok->flags &= ~TOK_EAT;
348 					return 3;
349 				}
350 				goto tok_line_outok;
351 
352 			case Q_single:
353 				return 1;
354 
355 			case Q_double:
356 				return 2;
357 
358 			case Q_doubleone:
359 				tok->quote = Q_double;
360 				*tok->wptr++ = *ptr;
361 				break;
362 
363 			case Q_one:
364 				tok->quote = Q_none;
365 				*tok->wptr++ = *ptr;
366 				break;
367 
368 			default:
369 				return -1;
370 			}
371 			break;
372 
373 		default:
374 			tok->flags &= ~TOK_EAT;
375 			switch (tok->quote) {
376 			case Q_none:
377 				if (Strchr(tok->ifs, *ptr) != NULL)
378 					FUN(tok,finish)(tok);
379 				else
380 					*tok->wptr++ = *ptr;
381 				break;
382 
383 			case Q_single:
384 			case Q_double:
385 				*tok->wptr++ = *ptr;
386 				break;
387 
388 
389 			case Q_doubleone:
390 				*tok->wptr++ = '\\';
391 				tok->quote = Q_double;
392 				*tok->wptr++ = *ptr;
393 				break;
394 
395 			case Q_one:
396 				tok->quote = Q_none;
397 				*tok->wptr++ = *ptr;
398 				break;
399 
400 			default:
401 				return -1;
402 
403 			}
404 			break;
405 		}
406 
407 		if (tok->wptr >= tok->wmax - 4) {
408 			size_t size = (size_t)(tok->wmax - tok->wspace + WINCR);
409 			Char *s = tok_realloc(tok->wspace,
410 			    size * sizeof(*s));
411 			if (s == NULL)
412 				return -1;
413 
414 			if (s != tok->wspace) {
415 				size_t i;
416 				for (i = 0; i < tok->argc; i++) {
417 				    tok->argv[i] =
418 					(tok->argv[i] - tok->wspace) + s;
419 				}
420 				tok->wptr = (tok->wptr - tok->wspace) + s;
421 				tok->wstart = (tok->wstart - tok->wspace) + s;
422 				tok->wspace = s;
423 			}
424 			tok->wmax = s + size;
425 		}
426 		if (tok->argc >= tok->amax - 4) {
427 			Char **p;
428 			tok->amax += AINCR;
429 			p = tok_realloc(tok->argv, tok->amax * sizeof(*p));
430 			if (p == NULL) {
431 				tok->amax -= AINCR;
432 				return -1;
433 			}
434 			tok->argv = p;
435 		}
436 	}
437  tok_line_outok:
438 	if (cc == -1 && co == -1) {
439 		cc = (int)tok->argc;
440 		co = (int)(tok->wptr - tok->wstart);
441 	}
442 	if (cursorc != NULL)
443 		*cursorc = cc;
444 	if (cursoro != NULL)
445 		*cursoro = co;
446 	FUN(tok,finish)(tok);
447 	*argv = (const Char **)tok->argv;
448 	*argc = (int)tok->argc;
449 	return 0;
450 }
451 
452 /* FUN(tok,str)():
453  *	Simpler version of tok_line, taking a NUL terminated line
454  *	and splitting into words, ignoring cursor state.
455  */
456 int
457 FUN(tok,str)(TYPE(Tokenizer) *tok, const Char *line, int *argc,
458     const Char ***argv)
459 {
460 	TYPE(LineInfo) li;
461 
462 	memset(&li, 0, sizeof(li));
463 	li.buffer = line;
464 	li.cursor = li.lastchar = Strchr(line, '\0');
465 	return FUN(tok,line)(tok, &li, argc, argv, NULL, NULL);
466 }
467