1/*	$OpenBSD: tbl_layout.c,v 1.37 2021/08/10 12:36:42 schwarze Exp $ */
2/*
3 * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2012, 2014, 2015, 2017, 2020, 2021
5 *               Ingo Schwarze <schwarze@openbsd.org>
6 *
7 * Permission to use, copy, modify, and distribute this software for any
8 * purpose with or without fee is hereby granted, provided that the above
9 * copyright notice and this permission notice appear in all copies.
10 *
11 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
12 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
13 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
14 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
15 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
16 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
17 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
18 */
19#include <sys/types.h>
20
21#include <ctype.h>
22#include <stdint.h>
23#include <stdio.h>
24#include <stdlib.h>
25#include <string.h>
26#include <time.h>
27
28#include "mandoc_aux.h"
29#include "mandoc.h"
30#include "tbl.h"
31#include "libmandoc.h"
32#include "tbl_int.h"
33
34struct	tbl_phrase {
35	char		 name;
36	enum tbl_cellt	 key;
37};
38
39static	const struct tbl_phrase keys[] = {
40	{ 'c',		 TBL_CELL_CENTRE },
41	{ 'r',		 TBL_CELL_RIGHT },
42	{ 'l',		 TBL_CELL_LEFT },
43	{ 'n',		 TBL_CELL_NUMBER },
44	{ 's',		 TBL_CELL_SPAN },
45	{ 'a',		 TBL_CELL_LONG },
46	{ '^',		 TBL_CELL_DOWN },
47	{ '-',		 TBL_CELL_HORIZ },
48	{ '_',		 TBL_CELL_HORIZ },
49	{ '=',		 TBL_CELL_DHORIZ }
50};
51
52#define KEYS_MAX ((int)(sizeof(keys)/sizeof(keys[0])))
53
54static	void		 mods(struct tbl_node *, struct tbl_cell *,
55				int, const char *, int *);
56static	void		 cell(struct tbl_node *, struct tbl_row *,
57				int, const char *, int *);
58static	struct tbl_cell *cell_alloc(struct tbl_node *, struct tbl_row *,
59				enum tbl_cellt);
60
61
62static void
63mods(struct tbl_node *tbl, struct tbl_cell *cp,
64		int ln, const char *p, int *pos)
65{
66	char		*endptr;
67	unsigned long	 spacing;
68	size_t		 sz;
69	int		 isz;
70	enum mandoc_esc	 fontesc;
71
72mod:
73	while (p[*pos] == ' ' || p[*pos] == '\t')
74		(*pos)++;
75
76	/* Row delimiters and cell specifiers end modifier lists. */
77
78	if (strchr(".,-=^_ACLNRSaclnrs", p[*pos]) != NULL)
79		return;
80
81	/* Throw away parenthesised expression. */
82
83	if ('(' == p[*pos]) {
84		(*pos)++;
85		while (p[*pos] && ')' != p[*pos])
86			(*pos)++;
87		if (')' == p[*pos]) {
88			(*pos)++;
89			goto mod;
90		}
91		mandoc_msg(MANDOCERR_TBLLAYOUT_PAR, ln, *pos, NULL);
92		return;
93	}
94
95	/* Parse numerical spacing from modifier string. */
96
97	if (isdigit((unsigned char)p[*pos])) {
98		if ((spacing = strtoul(p + *pos, &endptr, 10)) > 9)
99			mandoc_msg(MANDOCERR_TBLLAYOUT_SPC, ln, *pos,
100			    "%lu", spacing);
101		else
102			cp->spacing = spacing;
103		*pos = endptr - p;
104		goto mod;
105	}
106
107	switch (tolower((unsigned char)p[(*pos)++])) {
108	case 'b':
109		cp->font = ESCAPE_FONTBOLD;
110		goto mod;
111	case 'd':
112		cp->flags |= TBL_CELL_BALIGN;
113		goto mod;
114	case 'e':
115		cp->flags |= TBL_CELL_EQUAL;
116		goto mod;
117	case 'f':
118		break;
119	case 'i':
120		cp->font = ESCAPE_FONTITALIC;
121		goto mod;
122	case 'm':
123		mandoc_msg(MANDOCERR_TBLLAYOUT_MOD, ln, *pos, "m");
124		goto mod;
125	case 'p':
126	case 'v':
127		if (p[*pos] == '-' || p[*pos] == '+')
128			(*pos)++;
129		while (isdigit((unsigned char)p[*pos]))
130			(*pos)++;
131		goto mod;
132	case 't':
133		cp->flags |= TBL_CELL_TALIGN;
134		goto mod;
135	case 'u':
136		cp->flags |= TBL_CELL_UP;
137		goto mod;
138	case 'w':
139		sz = 0;
140		if (p[*pos] == '(') {
141			(*pos)++;
142			while (p[*pos + sz] != '\0' && p[*pos + sz] != ')')
143				sz++;
144		} else
145			while (isdigit((unsigned char)p[*pos + sz]))
146				sz++;
147		if (sz) {
148			free(cp->wstr);
149			cp->wstr = mandoc_strndup(p + *pos, sz);
150			*pos += sz;
151			if (p[*pos] == ')')
152				(*pos)++;
153		}
154		goto mod;
155	case 'x':
156		cp->flags |= TBL_CELL_WMAX;
157		goto mod;
158	case 'z':
159		cp->flags |= TBL_CELL_WIGN;
160		goto mod;
161	case '|':
162		if (cp->vert < 2)
163			cp->vert++;
164		else
165			mandoc_msg(MANDOCERR_TBLLAYOUT_VERT,
166			    ln, *pos - 1, NULL);
167		goto mod;
168	default:
169		mandoc_msg(MANDOCERR_TBLLAYOUT_CHAR,
170		    ln, *pos - 1, "%c", p[*pos - 1]);
171		goto mod;
172	}
173
174	while (p[*pos] == ' ' || p[*pos] == '\t')
175		(*pos)++;
176
177	/* Ignore parenthised font names for now. */
178
179	if (p[*pos] == '(')
180		goto mod;
181
182	isz = 0;
183	if (p[*pos] != '\0')
184		isz++;
185	if (strchr(" \t.", p[*pos + isz]) == NULL)
186		isz++;
187
188	fontesc = mandoc_font(p + *pos, isz);
189
190	switch (fontesc) {
191	case ESCAPE_FONTPREV:
192	case ESCAPE_ERROR:
193		mandoc_msg(MANDOCERR_FT_BAD,
194		    ln, *pos, "TS %s", p + *pos - 1);
195		break;
196	default:
197		cp->font = fontesc;
198		break;
199	}
200	*pos += isz;
201	goto mod;
202}
203
204static void
205cell(struct tbl_node *tbl, struct tbl_row *rp,
206		int ln, const char *p, int *pos)
207{
208	int		 i;
209	enum tbl_cellt	 c;
210
211	/* Handle leading vertical lines */
212
213	while (p[*pos] == ' ' || p[*pos] == '\t' || p[*pos] == '|') {
214		if (p[*pos] == '|') {
215			if (rp->vert < 2)
216				rp->vert++;
217			else
218				mandoc_msg(MANDOCERR_TBLLAYOUT_VERT,
219				    ln, *pos, NULL);
220		}
221		(*pos)++;
222	}
223
224again:
225	while (p[*pos] == ' ' || p[*pos] == '\t')
226		(*pos)++;
227
228	if (p[*pos] == '.' || p[*pos] == '\0')
229		return;
230
231	/* Parse the column position (`c', `l', `r', ...). */
232
233	for (i = 0; i < KEYS_MAX; i++)
234		if (tolower((unsigned char)p[*pos]) == keys[i].name)
235			break;
236
237	if (i == KEYS_MAX) {
238		mandoc_msg(MANDOCERR_TBLLAYOUT_CHAR,
239		    ln, *pos, "%c", p[*pos]);
240		(*pos)++;
241		goto again;
242	}
243	c = keys[i].key;
244
245	/* Special cases of spanners. */
246
247	if (c == TBL_CELL_SPAN) {
248		if (rp->last == NULL)
249			mandoc_msg(MANDOCERR_TBLLAYOUT_SPAN, ln, *pos, NULL);
250		else if (rp->last->pos == TBL_CELL_HORIZ ||
251		    rp->last->pos == TBL_CELL_DHORIZ)
252			c = rp->last->pos;
253	} else if (c == TBL_CELL_DOWN && rp == tbl->first_row)
254		mandoc_msg(MANDOCERR_TBLLAYOUT_DOWN, ln, *pos, NULL);
255
256	(*pos)++;
257
258	/* Allocate cell then parse its modifiers. */
259
260	mods(tbl, cell_alloc(tbl, rp, c), ln, p, pos);
261}
262
263void
264tbl_layout(struct tbl_node *tbl, int ln, const char *p, int pos)
265{
266	struct tbl_row	*rp;
267
268	rp = NULL;
269	for (;;) {
270		/* Skip whitespace before and after each cell. */
271
272		while (p[pos] == ' ' || p[pos] == '\t')
273			pos++;
274
275		switch (p[pos]) {
276		case ',':  /* Next row on this input line. */
277			pos++;
278			rp = NULL;
279			continue;
280		case '\0':  /* Next row on next input line. */
281			return;
282		case '.':  /* End of layout. */
283			pos++;
284			tbl->part = TBL_PART_DATA;
285
286			/*
287			 * When the layout is completely empty,
288			 * default to one left-justified column.
289			 */
290
291			if (tbl->first_row == NULL) {
292				tbl->first_row = tbl->last_row =
293				    mandoc_calloc(1, sizeof(*rp));
294			}
295			if (tbl->first_row->first == NULL) {
296				mandoc_msg(MANDOCERR_TBLLAYOUT_NONE,
297				    ln, pos, NULL);
298				cell_alloc(tbl, tbl->first_row,
299				    TBL_CELL_LEFT);
300				if (tbl->opts.lvert < tbl->first_row->vert)
301					tbl->opts.lvert = tbl->first_row->vert;
302				return;
303			}
304
305			/*
306			 * Search for the widest line
307			 * along the left and right margins.
308			 */
309
310			for (rp = tbl->first_row; rp; rp = rp->next) {
311				if (tbl->opts.lvert < rp->vert)
312					tbl->opts.lvert = rp->vert;
313				if (rp->last != NULL &&
314				    rp->last->col + 1 == tbl->opts.cols &&
315				    tbl->opts.rvert < rp->last->vert)
316					tbl->opts.rvert = rp->last->vert;
317
318				/* If the last line is empty, drop it. */
319
320				if (rp->next != NULL &&
321				    rp->next->first == NULL) {
322					free(rp->next);
323					rp->next = NULL;
324					tbl->last_row = rp;
325				}
326			}
327			return;
328		default:  /* Cell. */
329			break;
330		}
331
332		/*
333		 * If the last line had at least one cell,
334		 * start a new one; otherwise, continue it.
335		 */
336
337		if (rp == NULL) {
338			if (tbl->last_row == NULL ||
339			    tbl->last_row->first != NULL) {
340				rp = mandoc_calloc(1, sizeof(*rp));
341				if (tbl->last_row)
342					tbl->last_row->next = rp;
343				else
344					tbl->first_row = rp;
345				tbl->last_row = rp;
346			} else
347				rp = tbl->last_row;
348		}
349		cell(tbl, rp, ln, p, &pos);
350	}
351}
352
353static struct tbl_cell *
354cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos)
355{
356	struct tbl_cell	*p, *pp;
357
358	p = mandoc_calloc(1, sizeof(*p));
359	p->spacing = SIZE_MAX;
360	p->font = ESCAPE_FONTROMAN;
361	p->pos = pos;
362
363	if ((pp = rp->last) != NULL) {
364		pp->next = p;
365		p->col = pp->col + 1;
366	} else
367		rp->first = p;
368	rp->last = p;
369
370	if (tbl->opts.cols <= p->col)
371		tbl->opts.cols = p->col + 1;
372
373	return p;
374}
375