compile.c revision 1.16 
1/* $NetBSD: compile.c,v 1.16 2020/03/27 17:39:53 christos Exp $ */
2
3/*
4 * Copyright (c) 2009, 2010, 2011, 2020 The NetBSD Foundation, Inc.
5 *
6 * This code is derived from software contributed to The NetBSD Foundation
7 * by Roy Marples.
8 *
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
11 * are met:
12 * 1. Redistributions of source code must retain the above copyright
13 *    notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 *    notice, this list of conditions and the following disclaimer in the
16 *    documentation and/or other materials provided with the distribution.
17 *
18 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
19 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
20 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
21 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
22 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
23 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
24 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
25 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
26 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
27 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28 */
29
30#if HAVE_NBTOOL_CONFIG_H
31#include "nbtool_config.h"
32#endif
33
34#include <sys/cdefs.h>
35__RCSID("$NetBSD: compile.c,v 1.16 2020/03/27 17:39:53 christos Exp $");
36
37#if !HAVE_NBTOOL_CONFIG_H || HAVE_SYS_ENDIAN_H
38#include <sys/endian.h>
39#endif
40
41#include <assert.h>
42#include <ctype.h>
43#include <err.h>
44#include <errno.h>
45#include <limits.h>
46#include <stdarg.h>
47#include <stdlib.h>
48#include <stdint.h>
49#include <stdio.h>
50#include <string.h>
51#include <term_private.h>
52#include <term.h>
53
54static void __printflike(2, 3)
55dowarn(int flags, const char *fmt, ...)
56{
57	va_list va;
58
59	errno = EINVAL;
60	if (flags & TIC_WARNING) {
61		va_start(va, fmt);
62		vwarnx(fmt, va);
63		va_end(va);
64	}
65}
66
67char *
68_ti_grow_tbuf(TBUF *tbuf, size_t len)
69{
70	char *buf;
71	size_t l;
72
73	_DIAGASSERT(tbuf != NULL);
74
75	l = tbuf->bufpos + len;
76	if (l > tbuf->buflen) {
77		if (tbuf->buflen == 0)
78			buf = malloc(l);
79		else
80			buf = realloc(tbuf->buf, l);
81		if (buf == NULL)
82			return NULL;
83		tbuf->buf = buf;
84		tbuf->buflen = l;
85	}
86	return tbuf->buf;
87}
88
89const char *
90_ti_find_cap(TIC *tic, TBUF *tbuf, char type, short ind)
91{
92	size_t n;
93	uint16_t num;
94	const char *cap;
95
96	_DIAGASSERT(tbuf != NULL);
97
98	cap = tbuf->buf;
99	for (n = tbuf->entries; n > 0; n--) {
100		num = _ti_decode_16(&cap);
101		if ((short)num == ind)
102			return cap;
103		switch (type) {
104		case 'f':
105			cap++;
106			break;
107		case 'n':
108			cap += _ti_numsize(tic);
109			break;
110		case 's':
111			num = _ti_decode_16(&cap);
112			cap += num;
113			break;
114		}
115	}
116
117	errno = ESRCH;
118	return NULL;
119}
120
121const char *
122_ti_find_extra(TIC *tic, TBUF *tbuf, const char *code)
123{
124	size_t n;
125	uint16_t num;
126	const char *cap;
127
128	_DIAGASSERT(tbuf != NULL);
129	_DIAGASSERT(code != NULL);
130
131	cap = tbuf->buf;
132	for (n = tbuf->entries; n > 0; n--) {
133		num = _ti_decode_16(&cap);
134		if (strcmp(cap, code) == 0)
135			return cap + num;
136		cap += num;
137		switch (*cap++) {
138		case 'f':
139			cap++;
140			break;
141		case 'n':
142			cap += _ti_numsize(tic);
143			break;
144		case 's':
145			num = _ti_decode_16(&cap);
146			cap += num;
147			break;
148		}
149	}
150
151	errno = ESRCH;
152	return NULL;
153}
154
155char *
156_ti_getname(int rtype, const char *orig)
157{
158	char *name;
159
160	if (rtype == TERMINFO_RTYPE) {
161		if (asprintf(&name, "%s@v3", orig) < 0)
162			name = NULL;
163	} else {
164		name = strdup(orig);
165	}
166	return name;
167}
168
169size_t
170_ti_store_extra(TIC *tic, int wrn, const char *id, char type, char flag,
171    int num, const char *str, size_t strl, int flags)
172{
173	size_t l;
174
175	_DIAGASSERT(tic != NULL);
176
177	if (strcmp(id, "use") != 0) {
178		if (_ti_find_extra(tic, &tic->extras, id) != NULL)
179			return 0;
180		if (!(flags & TIC_EXTRA)) {
181			if (wrn != 0)
182				dowarn(flags, "%s: %s: unknown capability",
183				    tic->name, id);
184			return 0;
185		}
186	}
187
188	l = strlen(id) + 1;
189	if (l > UINT16_T_MAX) {
190		dowarn(flags, "%s: %s: cap name is too long", tic->name, id);
191		return 0;
192	}
193
194	if (!_ti_grow_tbuf(&tic->extras,
195		l + strl + sizeof(uint16_t) + _ti_numsize(tic) + 1))
196		return 0;
197	_ti_encode_buf_count_str(&tic->extras, id, l);
198	tic->extras.buf[tic->extras.bufpos++] = type;
199	switch (type) {
200	case 'f':
201		tic->extras.buf[tic->extras.bufpos++] = flag;
202		break;
203	case 'n':
204		_ti_encode_buf_num(&tic->extras, num, tic->rtype);
205		break;
206	case 's':
207		_ti_encode_buf_count_str(&tic->extras, str, strl);
208		break;
209	}
210	tic->extras.entries++;
211	return 1;
212}
213
214static void
215_ti_encode_buf(char **cap, const TBUF *buf)
216{
217	if (buf->entries == 0) {
218		_ti_encode_16(cap, 0);
219	} else {
220		_ti_encode_16(cap, buf->bufpos + sizeof(uint16_t));
221		_ti_encode_16(cap, buf->entries);
222		_ti_encode_str(cap, buf->buf, buf->bufpos);
223	}
224}
225
226ssize_t
227_ti_flatten(uint8_t **buf, const TIC *tic)
228{
229	size_t buflen, len, alen, dlen;
230	char *cap;
231
232	_DIAGASSERT(buf != NULL);
233	_DIAGASSERT(tic != NULL);
234
235	len = strlen(tic->name) + 1;
236	if (tic->alias == NULL)
237		alen = 0;
238	else
239		alen = strlen(tic->alias) + 1;
240	if (tic->desc == NULL)
241		dlen = 0;
242	else
243		dlen = strlen(tic->desc) + 1;
244
245	buflen = sizeof(char) +
246	    sizeof(uint16_t) + len +
247	    sizeof(uint16_t) + alen +
248	    sizeof(uint16_t) + dlen +
249	    (sizeof(uint16_t) * 2) + tic->flags.bufpos +
250	    (sizeof(uint16_t) * 2) + tic->nums.bufpos +
251	    (sizeof(uint16_t) * 2) + tic->strs.bufpos +
252	    (sizeof(uint16_t) * 2) + tic->extras.bufpos;
253
254	*buf = malloc(buflen);
255	if (*buf == NULL)
256		return -1;
257
258	cap = (char *)*buf;
259	*cap++ = tic->rtype;
260
261	_ti_encode_count_str(&cap, tic->name, len);
262	_ti_encode_count_str(&cap, tic->alias, alen);
263	_ti_encode_count_str(&cap, tic->desc, dlen);
264
265	_ti_encode_buf(&cap, &tic->flags);
266
267	_ti_encode_buf(&cap, &tic->nums);
268	_ti_encode_buf(&cap, &tic->strs);
269	_ti_encode_buf(&cap, &tic->extras);
270
271	return (uint8_t *)cap - *buf;
272}
273
274static int
275encode_string(const char *term, const char *cap, TBUF *tbuf, const char *str,
276    int flags)
277{
278	int slash, i, num;
279	char ch, *p, *s, last;
280
281	if (_ti_grow_tbuf(tbuf, strlen(str) + 1) == NULL)
282		return -1;
283	p = s = tbuf->buf + tbuf->bufpos;
284	slash = 0;
285	last = '\0';
286	/* Convert escape codes */
287	while ((ch = *str++) != '\0') {
288		if (ch == '\n') {
289			/* Following a newline, strip leading whitespace from
290			 * capability strings. */
291			while (isspace((unsigned char)*str))
292				str++;
293			continue;
294		}
295		if (slash == 0 && ch == '\\') {
296			slash = 1;
297			continue;
298		}
299		if (slash == 0) {
300			if (last != '%' && ch == '^') {
301				ch = *str++;
302				if (((unsigned char)ch) >= 128)
303					dowarn(flags,
304					    "%s: %s: illegal ^ character",
305					    term, cap);
306				if (ch == '\0')
307					break;
308				if (ch == '?')
309					ch = '\177';
310				else if ((ch &= 037) == 0)
311					ch = (char)128;
312			} else if (!isprint((unsigned char)ch))
313				dowarn(flags,
314				    "%s: %s: unprintable character",
315				    term, cap);
316			*p++ = ch;
317			last = ch;
318			continue;
319		}
320		slash = 0;
321		if (ch >= '0' && ch <= '7') {
322			num = ch - '0';
323			for (i = 0; i < 2; i++) {
324				if (*str < '0' || *str > '7') {
325					if (isdigit((unsigned char)*str))
326						dowarn(flags,
327						    "%s: %s: non octal"
328						    " digit", term, cap);
329					else
330						break;
331				}
332				num = num * 8 + *str++ - '0';
333			}
334			if (num == 0)
335				num = 0200;
336			*p++ = (char)num;
337			continue;
338		}
339		switch (ch) {
340		case 'a':
341			*p++ = '\a';
342			break;
343		case 'b':
344			*p++ = '\b';
345			break;
346		case 'e': /* FALLTHROUGH */
347		case 'E':
348			*p++ = '\033';
349			break;
350		case 'f':
351			*p++ = '\014';
352			break;
353		case 'l': /* FALLTHROUGH */
354		case 'n':
355			*p++ = '\n';
356			break;
357		case 'r':
358			*p++ = '\r';
359			break;
360		case 's':
361			*p++ = ' ';
362			break;
363		case 't':
364			*p++ = '\t';
365			break;
366		default:
367			/* We should warn here */
368		case '^':
369		case ',':
370		case ':':
371		case '|':
372			*p++ = ch;
373			break;
374		}
375		last = ch;
376	}
377	*p++ = '\0';
378	tbuf->bufpos += (size_t)(p - s);
379	return 0;
380}
381
382char *
383_ti_get_token(char **cap, char sep)
384{
385	char esc, *token;
386
387	while (isspace((unsigned char)**cap))
388		(*cap)++;
389	if (**cap == '\0')
390		return NULL;
391
392	/* We can't use stresep(3) as ^ we need two escape chars */
393	esc = '\0';
394	for (token = *cap;
395	     **cap != '\0' && (esc != '\0' || **cap != sep);
396	     (*cap)++)
397	{
398		if (esc == '\0') {
399			if (**cap == '\\' || **cap == '^')
400				esc = **cap;
401		} else {
402			/* termcap /E/ is valid */
403			if (sep == ':' && esc == '\\' && **cap == 'E')
404				esc = 'x';
405			else
406				esc = '\0';
407		}
408	}
409
410	if (**cap != '\0')
411		*(*cap)++ = '\0';
412
413	return token;
414}
415
416static int
417_ti_find_rtype(const char *cap)
418{
419	for (const char *ptr = cap; (ptr = strchr(ptr, '#')) != NULL;) {
420		if (strtol(++ptr, NULL, 0) > SHRT_MAX) {
421			return TERMINFO_RTYPE;
422		}
423	}
424	return TERMINFO_RTYPE_O1;
425}
426
427int
428_ti_encode_buf_id_num(TBUF *tbuf, int ind, int num, size_t len)
429{
430	if (!_ti_grow_tbuf(tbuf, sizeof(uint16_t) + len))
431		return 0;
432	_ti_encode_buf_16(tbuf, ind);
433	if (len == sizeof(uint32_t))
434		_ti_encode_buf_32(tbuf, num);
435	else
436		_ti_encode_buf_16(tbuf, num);
437	tbuf->entries++;
438	return 1;
439}
440
441int
442_ti_encode_buf_id_count_str(TBUF *tbuf, int ind, const void *buf, size_t len)
443{
444	if (!_ti_grow_tbuf(tbuf, 2 * sizeof(uint16_t) + len))
445		return 0;
446	_ti_encode_buf_16(tbuf, ind);
447	_ti_encode_buf_count_str(tbuf, buf, len);
448	tbuf->entries++;
449	return 1;
450}
451
452int
453_ti_encode_buf_id_flags(TBUF *tbuf, int ind, int flag)
454{
455	if (!_ti_grow_tbuf(tbuf, sizeof(uint16_t) + 1))
456		return 0;
457	_ti_encode_buf_16(tbuf, ind);
458	tbuf->buf[tbuf->bufpos++] = flag;
459	tbuf->entries++;
460	return 1;
461}
462
463TIC *
464_ti_compile(char *cap, int flags)
465{
466	char *token, *p, *e, *name, *desc, *alias;
467	signed char flag;
468	long cnum;
469	short ind;
470	int num;
471	size_t len;
472	TBUF buf;
473	TIC *tic;
474
475	_DIAGASSERT(cap != NULL);
476
477	name = _ti_get_token(&cap, ',');
478	if (name == NULL) {
479		dowarn(flags, "no separator found: %s", cap);
480		return NULL;
481	}
482	desc = strrchr(name, '|');
483	if (desc != NULL)
484		*desc++ = '\0';
485	alias = strchr(name, '|');
486	if (alias != NULL)
487		*alias++ = '\0';
488
489	if (strlen(name) > UINT16_MAX - 1) {
490		dowarn(flags, "%s: name too long", name);
491		return NULL;
492	}
493	if (desc != NULL && strlen(desc) > UINT16_MAX - 1) {
494		dowarn(flags, "%s: description too long: %s", name, desc);
495		return NULL;
496	}
497	if (alias != NULL && strlen(alias) > UINT16_MAX - 1) {
498		dowarn(flags, "%s: alias too long: %s", name, alias);
499		return NULL;
500	}
501
502	tic = calloc(sizeof(*tic), 1);
503	if (tic == NULL)
504		return NULL;
505
506	tic->rtype = (flags & TIC_COMPAT_V1) ? TERMINFO_RTYPE_O1 :
507	    _ti_find_rtype(cap);
508	buf.buf = NULL;
509	buf.buflen = 0;
510
511	tic->name = _ti_getname(tic->rtype, name);
512	if (tic->name == NULL)
513		goto error;
514	if (alias != NULL && flags & TIC_ALIAS) {
515		tic->alias = _ti_getname(tic->rtype, alias);
516		if (tic->alias == NULL)
517			goto error;
518	}
519	if (desc != NULL && flags & TIC_DESCRIPTION) {
520		tic->desc = strdup(desc);
521		if (tic->desc == NULL)
522			goto error;
523	}
524
525	for (token = _ti_get_token(&cap, ',');
526	     token != NULL && *token != '\0';
527	     token = _ti_get_token(&cap, ','))
528	{
529		/* Skip commented caps */
530		if (!(flags & TIC_COMMENT) && token[0] == '.')
531			continue;
532
533		/* Obsolete entries */
534		if (token[0] == 'O' && token[1] == 'T') {
535			if (!(flags & TIC_EXTRA))
536				continue;
537			token += 2;
538		}
539
540		/* str cap */
541		p = strchr(token, '=');
542		if (p != NULL) {
543			*p++ = '\0';
544			/* Don't use the string if we already have it */
545			ind = (short)_ti_strindex(token);
546			if (ind != -1 &&
547			    _ti_find_cap(tic, &tic->strs, 's', ind) != NULL)
548				continue;
549
550			/* Encode the string to our scratch buffer */
551			buf.bufpos = 0;
552			if (encode_string(tic->name, token,
553				&buf, p, flags) == -1)
554				goto error;
555			if (buf.bufpos > UINT16_MAX - 1) {
556				dowarn(flags, "%s: %s: string is too long",
557				    tic->name, token);
558				continue;
559			}
560			if (!VALID_STRING(buf.buf)) {
561				dowarn(flags, "%s: %s: invalid string",
562				    tic->name, token);
563				continue;
564			}
565
566			if (ind == -1) {
567				if (!_ti_store_extra(tic, 1, token, 's', -1, -2,
568				    buf.buf, buf.bufpos, flags))
569					goto error;
570			} else {
571				if (!_ti_encode_buf_id_count_str(&tic->strs,
572				    ind, buf.buf, buf.bufpos))
573					goto error;
574			}
575			continue;
576		}
577
578		/* num cap */
579		p = strchr(token, '#');
580		if (p != NULL) {
581			*p++ = '\0';
582			/* Don't use the number if we already have it */
583			ind = (short)_ti_numindex(token);
584			if (ind != -1 &&
585			    _ti_find_cap(tic, &tic->nums, 'n', ind) != NULL)
586				continue;
587
588			cnum = strtol(p, &e, 0);
589			if (*e != '\0') {
590				dowarn(flags, "%s: %s: not a number",
591				    tic->name, token);
592				continue;
593			}
594			if (!VALID_NUMERIC(cnum) || cnum > INT32_MAX) {
595				dowarn(flags, "%s: %s: number %ld out of range",
596				    tic->name, token, cnum);
597				continue;
598			}
599
600			num = (int)cnum;
601			if (ind == -1) {
602				if (!_ti_store_extra(tic, 1, token, 'n', -1,
603				    num, NULL, 0, flags))
604					goto error;
605			} else {
606				if (!_ti_encode_buf_id_num(&tic->nums,
607				    ind, num, _ti_numsize(tic)))
608					    goto error;
609			}
610			continue;
611		}
612
613		flag = 1;
614		len = strlen(token) - 1;
615		if (token[len] == '@') {
616			flag = CANCELLED_BOOLEAN;
617			token[len] = '\0';
618		}
619		ind = (short)_ti_flagindex(token);
620		if (ind == -1 && flag == CANCELLED_BOOLEAN) {
621			if ((ind = (short)_ti_numindex(token)) != -1) {
622				if (_ti_find_cap(tic, &tic->nums, 'n', ind)
623				    != NULL)
624					continue;
625				if (!_ti_encode_buf_id_num(&tic->nums, ind,
626				    CANCELLED_NUMERIC, _ti_numsize(tic)))
627					goto error;
628				continue;
629			} else if ((ind = (short)_ti_strindex(token)) != -1) {
630				if (_ti_find_cap(tic, &tic->strs, 's', ind)
631				    != NULL)
632					continue;
633				if (!_ti_encode_buf_id_num(
634				    &tic->strs, ind, 0, sizeof(uint16_t)))
635					goto error;
636				continue;
637			}
638		}
639		if (ind == -1) {
640			if (!_ti_store_extra(tic, 1, token, 'f', flag, 0, NULL,
641			    0, flags))
642				goto error;
643		} else if (_ti_find_cap(tic, &tic->flags, 'f', ind) == NULL) {
644			if (!_ti_encode_buf_id_flags(&tic->flags, ind, flags))
645				goto error;
646		}
647	}
648
649	free(buf.buf);
650	return tic;
651
652error:
653	free(buf.buf);
654	_ti_freetic(tic);
655	return NULL;
656}
657
658void
659_ti_freetic(TIC *tic)
660{
661
662	if (tic != NULL) {
663		free(tic->name);
664		free(tic->alias);
665		free(tic->desc);
666		free(tic->extras.buf);
667		free(tic->flags.buf);
668		free(tic->nums.buf);
669		free(tic->strs.buf);
670		free(tic);
671	}
672}
673