1/*	$Id: tbl_data.c,v 1.59 2021/09/10 13:24:38 schwarze Exp $ */
2/*
3 * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011,2015,2017-2019,2021 Ingo Schwarze <schwarze@openbsd.org>
5 *
6 * Permission to use, copy, modify, and distribute this software for any
7 * purpose with or without fee is hereby granted, provided that the above
8 * copyright notice and this permission notice appear in all copies.
9 *
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17 */
18#include "config.h"
19
20#include <sys/types.h>
21
22#include <assert.h>
23#include <ctype.h>
24#include <stdint.h>
25#include <stdio.h>
26#include <stdlib.h>
27#include <string.h>
28#include <time.h>
29
30#include "mandoc_aux.h"
31#include "mandoc.h"
32#include "tbl.h"
33#include "libmandoc.h"
34#include "tbl_int.h"
35
36static	void		 getdata(struct tbl_node *, struct tbl_span *,
37				int, const char *, int *);
38static	struct tbl_span	*newspan(struct tbl_node *, int,
39				struct tbl_row *);
40
41
42static void
43getdata(struct tbl_node *tbl, struct tbl_span *dp,
44		int ln, const char *p, int *pos)
45{
46	struct tbl_dat	*dat, *pdat;
47	struct tbl_cell	*cp;
48	struct tbl_span	*pdp;
49	const char	*ccp;
50	int		 startpos, endpos;
51
52	/*
53	 * Determine the length of the string in the cell
54	 * and advance the parse point to the end of the cell.
55	 */
56
57	startpos = *pos;
58	ccp = p + startpos;
59	while (*ccp != '\0' && *ccp != tbl->opts.tab)
60		if (*ccp++ == '\\')
61			mandoc_escape(&ccp, NULL, NULL);
62	*pos = ccp - p;
63
64	/* Advance to the next layout cell, skipping spanners. */
65
66	cp = dp->last == NULL ? dp->layout->first : dp->last->layout->next;
67	while (cp != NULL && cp->pos == TBL_CELL_SPAN)
68		cp = cp->next;
69
70	/*
71	 * If the current layout row is out of cells, allocate
72	 * a new cell if another row of the table has at least
73	 * this number of columns, or discard the input if we
74	 * are beyond the last column of the table as a whole.
75	 */
76
77	if (cp == NULL) {
78		if (dp->layout->last->col + 1 < dp->opts->cols) {
79			cp = mandoc_calloc(1, sizeof(*cp));
80			cp->pos = TBL_CELL_LEFT;
81			cp->font = ESCAPE_FONTROMAN;
82			cp->spacing = SIZE_MAX;
83			dp->layout->last->next = cp;
84			cp->col = dp->layout->last->col + 1;
85			dp->layout->last = cp;
86		} else {
87			mandoc_msg(MANDOCERR_TBLDATA_EXTRA,
88			    ln, startpos, "%s", p + startpos);
89			while (p[*pos] != '\0')
90				(*pos)++;
91			return;
92		}
93	}
94
95	dat = mandoc_malloc(sizeof(*dat));
96	dat->layout = cp;
97	dat->next = NULL;
98	dat->string = NULL;
99	dat->hspans = 0;
100	dat->vspans = 0;
101	dat->block = 0;
102	dat->pos = TBL_DATA_NONE;
103
104	/*
105	 * Increment the number of vertical spans in a data cell above,
106	 * if this cell vertically extends one or more cells above.
107	 * The iteration must be done over data rows,
108	 * not over layout rows, because one layout row
109	 * can be reused for more than one data row.
110	 */
111
112	if (cp->pos == TBL_CELL_DOWN ||
113	    (*pos - startpos == 2 &&
114	     p[startpos] == '\\' && p[startpos + 1] == '^')) {
115		pdp = dp;
116		while ((pdp = pdp->prev) != NULL) {
117			pdat = pdp->first;
118			while (pdat != NULL &&
119			    pdat->layout->col < dat->layout->col)
120				pdat = pdat->next;
121			if (pdat == NULL)
122				break;
123			if (pdat->layout->pos != TBL_CELL_DOWN &&
124			    strcmp(pdat->string, "\\^") != 0) {
125				pdat->vspans++;
126				break;
127			}
128		}
129	}
130
131	/*
132	 * Count the number of horizontal spans to the right of this cell.
133	 * This is purely a matter of the layout, independent of the data.
134	 */
135
136	for (cp = cp->next; cp != NULL; cp = cp->next)
137		if (cp->pos == TBL_CELL_SPAN)
138			dat->hspans++;
139		else
140			break;
141
142	/* Append the new data cell to the data row. */
143
144	if (dp->last == NULL)
145		dp->first = dat;
146	else
147		dp->last->next = dat;
148	dp->last = dat;
149
150	/* Strip leading and trailing spaces, if requested. */
151
152	endpos = *pos;
153	if (dp->opts->opts & TBL_OPT_NOSPACE) {
154		while (p[startpos] == ' ')
155			startpos++;
156		while (endpos > startpos && p[endpos - 1] == ' ')
157			endpos--;
158	}
159
160	/*
161	 * Check for a continued-data scope opening.  This consists of a
162	 * trailing `T{' at the end of the line.  Subsequent lines,
163	 * until a standalone `T}', are included in our cell.
164	 */
165
166	if (endpos - startpos == 2 &&
167	    p[startpos] == 'T' && p[startpos + 1] == '{') {
168		tbl->part = TBL_PART_CDATA;
169		return;
170	}
171
172	dat->string = mandoc_strndup(p + startpos, endpos - startpos);
173
174	if (p[*pos] != '\0')
175		(*pos)++;
176
177	if ( ! strcmp(dat->string, "_"))
178		dat->pos = TBL_DATA_HORIZ;
179	else if ( ! strcmp(dat->string, "="))
180		dat->pos = TBL_DATA_DHORIZ;
181	else if ( ! strcmp(dat->string, "\\_"))
182		dat->pos = TBL_DATA_NHORIZ;
183	else if ( ! strcmp(dat->string, "\\="))
184		dat->pos = TBL_DATA_NDHORIZ;
185	else
186		dat->pos = TBL_DATA_DATA;
187
188	if ((dat->layout->pos == TBL_CELL_HORIZ ||
189	    dat->layout->pos == TBL_CELL_DHORIZ ||
190	    dat->layout->pos == TBL_CELL_DOWN) &&
191	    dat->pos == TBL_DATA_DATA && *dat->string != '\0')
192		mandoc_msg(MANDOCERR_TBLDATA_SPAN,
193		    ln, startpos, "%s", dat->string);
194}
195
196void
197tbl_cdata(struct tbl_node *tbl, int ln, const char *p, int pos)
198{
199	struct tbl_dat	*dat;
200	size_t		 sz;
201
202	dat = tbl->last_span->last;
203
204	if (p[pos] == 'T' && p[pos + 1] == '}') {
205		pos += 2;
206		if (tbl->opts.opts & TBL_OPT_NOSPACE)
207			while (p[pos] == ' ')
208				pos++;
209		if (p[pos] == tbl->opts.tab) {
210			tbl->part = TBL_PART_DATA;
211			pos++;
212			while (p[pos] != '\0')
213				getdata(tbl, tbl->last_span, ln, p, &pos);
214			return;
215		} else if (p[pos] == '\0') {
216			tbl->part = TBL_PART_DATA;
217			return;
218		}
219
220		/* Fallthrough: T} is part of a word. */
221	}
222
223	dat->pos = TBL_DATA_DATA;
224	dat->block = 1;
225
226	if (dat->string != NULL) {
227		sz = strlen(p + pos) + strlen(dat->string) + 2;
228		dat->string = mandoc_realloc(dat->string, sz);
229		(void)strlcat(dat->string, " ", sz);
230		(void)strlcat(dat->string, p + pos, sz);
231	} else
232		dat->string = mandoc_strdup(p + pos);
233
234	if (dat->layout->pos == TBL_CELL_DOWN)
235		mandoc_msg(MANDOCERR_TBLDATA_SPAN,
236		    ln, pos, "%s", dat->string);
237}
238
239static struct tbl_span *
240newspan(struct tbl_node *tbl, int line, struct tbl_row *rp)
241{
242	struct tbl_span	*dp;
243
244	dp = mandoc_calloc(1, sizeof(*dp));
245	dp->line = line;
246	dp->opts = &tbl->opts;
247	dp->layout = rp;
248	dp->prev = tbl->last_span;
249
250	if (dp->prev == NULL) {
251		tbl->first_span = dp;
252		tbl->current_span = NULL;
253	} else
254		dp->prev->next = dp;
255	tbl->last_span = dp;
256
257	return dp;
258}
259
260void
261tbl_data(struct tbl_node *tbl, int ln, const char *p, int pos)
262{
263	struct tbl_row	*rp;
264	struct tbl_cell	*cp;
265	struct tbl_span	*sp;
266
267	for (sp = tbl->last_span; sp != NULL; sp = sp->prev)
268		if (sp->pos == TBL_SPAN_DATA)
269			break;
270	rp = sp == NULL ? tbl->first_row :
271	    sp->layout->next == NULL ? sp->layout : sp->layout->next;
272	assert(rp != NULL);
273
274	if (p[1] == '\0') {
275		switch (p[0]) {
276		case '.':
277			/*
278			 * Empty request lines must be handled here
279			 * and cannot be discarded in roff_parseln()
280			 * because in the layout section, they
281			 * are significant and end the layout.
282			 */
283			return;
284		case '_':
285			sp = newspan(tbl, ln, rp);
286			sp->pos = TBL_SPAN_HORIZ;
287			return;
288		case '=':
289			sp = newspan(tbl, ln, rp);
290			sp->pos = TBL_SPAN_DHORIZ;
291			return;
292		default:
293			break;
294		}
295	}
296
297	/*
298	 * If the layout row contains nothing but horizontal lines,
299	 * allocate an empty span for it and assign the current span
300	 * to the next layout row accepting data.
301	 */
302
303	while (rp->next != NULL) {
304		if (rp->last->col + 1 < tbl->opts.cols)
305			break;
306		for (cp = rp->first; cp != NULL; cp = cp->next)
307			if (cp->pos != TBL_CELL_HORIZ &&
308			    cp->pos != TBL_CELL_DHORIZ)
309				break;
310		if (cp != NULL)
311			break;
312		sp = newspan(tbl, ln, rp);
313		sp->pos = TBL_SPAN_DATA;
314		rp = rp->next;
315	}
316
317	/* Process a real data row. */
318
319	sp = newspan(tbl, ln, rp);
320	sp->pos = TBL_SPAN_DATA;
321	while (p[pos] != '\0')
322		getdata(tbl, sp, ln, p, &pos);
323}
324