Annotation of src/usr.bin/mandoc/tbl_data.c, Revision 1.42
1.42 ! schwarze 1: /* $OpenBSD: tbl_data.c,v 1.41 2021/05/15 17:16:38 schwarze Exp $ */
1.1 schwarze 2: /*
1.4 schwarze 3: * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.42 ! schwarze 4: * Copyright (c) 2011,2015,2017-2019,2021 Ingo Schwarze <schwarze@openbsd.org>
1.1 schwarze 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: */
1.19 schwarze 18: #include <sys/types.h>
19:
1.1 schwarze 20: #include <assert.h>
21: #include <ctype.h>
1.40 schwarze 22: #include <stdint.h>
1.37 schwarze 23: #include <stdio.h>
1.1 schwarze 24: #include <stdlib.h>
25: #include <string.h>
1.4 schwarze 26: #include <time.h>
1.1 schwarze 27:
1.35 schwarze 28: #include "mandoc_aux.h"
1.4 schwarze 29: #include "mandoc.h"
1.35 schwarze 30: #include "tbl.h"
1.4 schwarze 31: #include "libmandoc.h"
1.36 schwarze 32: #include "tbl_int.h"
1.4 schwarze 33:
1.20 schwarze 34: static void getdata(struct tbl_node *, struct tbl_span *,
1.9 schwarze 35: int, const char *, int *);
1.17 schwarze 36: static struct tbl_span *newspan(struct tbl_node *, int,
1.9 schwarze 37: struct tbl_row *);
1.4 schwarze 38:
1.17 schwarze 39:
1.20 schwarze 40: static void
1.17 schwarze 41: getdata(struct tbl_node *tbl, struct tbl_span *dp,
1.4 schwarze 42: int ln, const char *p, int *pos)
1.1 schwarze 43: {
1.33 schwarze 44: struct tbl_dat *dat, *pdat;
1.4 schwarze 45: struct tbl_cell *cp;
1.33 schwarze 46: struct tbl_span *pdp;
1.42 ! schwarze 47: const char *ccp;
1.24 schwarze 48: int sv;
1.4 schwarze 49:
1.34 schwarze 50: /*
51: * Determine the length of the string in the cell
52: * and advance the parse point to the end of the cell.
53: */
54:
55: sv = *pos;
1.42 ! schwarze 56: ccp = p + sv;
! 57: while (*ccp != '\0' && *ccp != tbl->opts.tab)
! 58: if (*ccp++ == '\\')
! 59: mandoc_escape(&ccp, NULL, NULL);
! 60: *pos = ccp - p;
1.34 schwarze 61:
1.25 schwarze 62: /* Advance to the next layout cell, skipping spanners. */
63:
1.24 schwarze 64: cp = dp->last == NULL ? dp->layout->first : dp->last->layout->next;
65: while (cp != NULL && cp->pos == TBL_CELL_SPAN)
1.4 schwarze 66: cp = cp->next;
67:
1.6 schwarze 68: /*
1.30 schwarze 69: * If the current layout row is out of cells, allocate
70: * a new cell if another row of the table has at least
71: * this number of columns, or discard the input if we
72: * are beyond the last column of the table as a whole.
1.6 schwarze 73: */
74:
1.24 schwarze 75: if (cp == NULL) {
1.30 schwarze 76: if (dp->layout->last->col + 1 < dp->opts->cols) {
77: cp = mandoc_calloc(1, sizeof(*cp));
78: cp->pos = TBL_CELL_LEFT;
1.40 schwarze 79: cp->spacing = SIZE_MAX;
1.30 schwarze 80: dp->layout->last->next = cp;
81: cp->col = dp->layout->last->col + 1;
82: dp->layout->last = cp;
83: } else {
1.38 schwarze 84: mandoc_msg(MANDOCERR_TBLDATA_EXTRA,
85: ln, sv, "%s", p + sv);
1.34 schwarze 86: while (p[*pos] != '\0')
1.30 schwarze 87: (*pos)++;
88: return;
89: }
1.6 schwarze 90: }
91:
1.33 schwarze 92: dat = mandoc_malloc(sizeof(*dat));
1.4 schwarze 93: dat->layout = cp;
1.33 schwarze 94: dat->next = NULL;
95: dat->string = NULL;
96: dat->hspans = 0;
97: dat->vspans = 0;
98: dat->block = 0;
1.4 schwarze 99: dat->pos = TBL_DATA_NONE;
1.33 schwarze 100:
101: /*
102: * Increment the number of vertical spans in a data cell above,
103: * if this cell vertically extends one or more cells above.
104: * The iteration must be done over data rows,
105: * not over layout rows, because one layout row
106: * can be reused for more than one data row.
107: */
108:
1.34 schwarze 109: if (cp->pos == TBL_CELL_DOWN ||
110: (*pos - sv == 2 && p[sv] == '\\' && p[sv + 1] == '^')) {
1.33 schwarze 111: pdp = dp;
112: while ((pdp = pdp->prev) != NULL) {
113: pdat = pdp->first;
114: while (pdat != NULL &&
115: pdat->layout->col < dat->layout->col)
116: pdat = pdat->next;
117: if (pdat == NULL)
118: break;
1.34 schwarze 119: if (pdat->layout->pos != TBL_CELL_DOWN &&
120: strcmp(pdat->string, "\\^") != 0) {
1.33 schwarze 121: pdat->vspans++;
122: break;
123: }
124: }
125: }
126:
127: /*
128: * Count the number of horizontal spans to the right of this cell.
129: * This is purely a matter of the layout, independent of the data.
130: */
131:
1.24 schwarze 132: for (cp = cp->next; cp != NULL; cp = cp->next)
133: if (cp->pos == TBL_CELL_SPAN)
1.33 schwarze 134: dat->hspans++;
1.6 schwarze 135: else
136: break;
1.33 schwarze 137:
138: /* Append the new data cell to the data row. */
1.17 schwarze 139:
1.24 schwarze 140: if (dp->last == NULL)
141: dp->first = dat;
142: else
1.4 schwarze 143: dp->last->next = dat;
1.24 schwarze 144: dp->last = dat;
1.4 schwarze 145:
146: /*
147: * Check for a continued-data scope opening. This consists of a
148: * trailing `T{' at the end of the line. Subsequent lines,
149: * until a standalone `T}', are included in our cell.
150: */
1.1 schwarze 151:
1.24 schwarze 152: if (*pos - sv == 2 && p[sv] == 'T' && p[sv + 1] == '{') {
1.4 schwarze 153: tbl->part = TBL_PART_CDATA;
1.20 schwarze 154: return;
1.4 schwarze 155: }
1.1 schwarze 156:
1.24 schwarze 157: dat->string = mandoc_strndup(p + sv, *pos - sv);
1.1 schwarze 158:
1.34 schwarze 159: if (p[*pos] != '\0')
1.4 schwarze 160: (*pos)++;
1.1 schwarze 161:
162: if ( ! strcmp(dat->string, "_"))
1.4 schwarze 163: dat->pos = TBL_DATA_HORIZ;
1.1 schwarze 164: else if ( ! strcmp(dat->string, "="))
1.4 schwarze 165: dat->pos = TBL_DATA_DHORIZ;
1.1 schwarze 166: else if ( ! strcmp(dat->string, "\\_"))
1.4 schwarze 167: dat->pos = TBL_DATA_NHORIZ;
1.1 schwarze 168: else if ( ! strcmp(dat->string, "\\="))
1.4 schwarze 169: dat->pos = TBL_DATA_NDHORIZ;
1.1 schwarze 170: else
1.4 schwarze 171: dat->pos = TBL_DATA_DATA;
172:
1.24 schwarze 173: if ((dat->layout->pos == TBL_CELL_HORIZ ||
174: dat->layout->pos == TBL_CELL_DHORIZ ||
175: dat->layout->pos == TBL_CELL_DOWN) &&
176: dat->pos == TBL_DATA_DATA && *dat->string != '\0')
177: mandoc_msg(MANDOCERR_TBLDATA_SPAN,
1.38 schwarze 178: ln, sv, "%s", dat->string);
1.1 schwarze 179: }
180:
1.32 schwarze 181: void
1.22 schwarze 182: tbl_cdata(struct tbl_node *tbl, int ln, const char *p, int pos)
1.4 schwarze 183: {
184: struct tbl_dat *dat;
1.17 schwarze 185: size_t sz;
1.4 schwarze 186:
187: dat = tbl->last_span->last;
1.5 schwarze 188:
189: if (p[pos] == 'T' && p[pos + 1] == '}') {
190: pos += 2;
191: if (p[pos] == tbl->opts.tab) {
192: tbl->part = TBL_PART_DATA;
193: pos++;
1.27 schwarze 194: while (p[pos] != '\0')
195: getdata(tbl, tbl->last_span, ln, p, &pos);
1.32 schwarze 196: return;
1.24 schwarze 197: } else if (p[pos] == '\0') {
1.5 schwarze 198: tbl->part = TBL_PART_DATA;
1.32 schwarze 199: return;
1.5 schwarze 200: }
201:
202: /* Fallthrough: T} is part of a word. */
203: }
1.4 schwarze 204:
1.6 schwarze 205: dat->pos = TBL_DATA_DATA;
1.29 schwarze 206: dat->block = 1;
1.6 schwarze 207:
1.24 schwarze 208: if (dat->string != NULL) {
1.23 schwarze 209: sz = strlen(p + pos) + strlen(dat->string) + 2;
1.4 schwarze 210: dat->string = mandoc_realloc(dat->string, sz);
1.18 schwarze 211: (void)strlcat(dat->string, " ", sz);
1.23 schwarze 212: (void)strlcat(dat->string, p + pos, sz);
1.4 schwarze 213: } else
1.23 schwarze 214: dat->string = mandoc_strdup(p + pos);
1.4 schwarze 215:
1.24 schwarze 216: if (dat->layout->pos == TBL_CELL_DOWN)
1.38 schwarze 217: mandoc_msg(MANDOCERR_TBLDATA_SPAN,
218: ln, pos, "%s", dat->string);
1.4 schwarze 219: }
1.1 schwarze 220:
1.7 schwarze 221: static struct tbl_span *
1.9 schwarze 222: newspan(struct tbl_node *tbl, int line, struct tbl_row *rp)
1.7 schwarze 223: {
224: struct tbl_span *dp;
225:
1.24 schwarze 226: dp = mandoc_calloc(1, sizeof(*dp));
1.9 schwarze 227: dp->line = line;
1.13 schwarze 228: dp->opts = &tbl->opts;
1.7 schwarze 229: dp->layout = rp;
1.21 schwarze 230: dp->prev = tbl->last_span;
1.7 schwarze 231:
1.21 schwarze 232: if (dp->prev == NULL) {
233: tbl->first_span = dp;
1.8 schwarze 234: tbl->current_span = NULL;
1.21 schwarze 235: } else
236: dp->prev->next = dp;
237: tbl->last_span = dp;
1.7 schwarze 238:
1.28 schwarze 239: return dp;
1.7 schwarze 240: }
241:
1.20 schwarze 242: void
1.22 schwarze 243: tbl_data(struct tbl_node *tbl, int ln, const char *p, int pos)
1.1 schwarze 244: {
1.4 schwarze 245: struct tbl_row *rp;
1.30 schwarze 246: struct tbl_cell *cp;
1.31 schwarze 247: struct tbl_span *sp;
1.4 schwarze 248:
1.41 schwarze 249: for (sp = tbl->last_span; sp != NULL; sp = sp->prev)
250: if (sp->pos == TBL_SPAN_DATA)
251: break;
252: rp = sp == NULL ? tbl->first_row :
253: sp->layout->next == NULL ? sp->layout : sp->layout->next;
1.30 schwarze 254: assert(rp != NULL);
1.4 schwarze 255:
1.39 schwarze 256: if (p[1] == '\0') {
257: switch (p[0]) {
258: case '.':
259: /*
260: * Empty request lines must be handled here
261: * and cannot be discarded in roff_parseln()
262: * because in the layout section, they
263: * are significant and end the layout.
264: */
265: return;
266: case '_':
267: sp = newspan(tbl, ln, rp);
268: sp->pos = TBL_SPAN_HORIZ;
269: return;
270: case '=':
271: sp = newspan(tbl, ln, rp);
272: sp->pos = TBL_SPAN_DHORIZ;
273: return;
274: default:
275: break;
276: }
1.1 schwarze 277: }
1.30 schwarze 278:
279: /*
1.31 schwarze 280: * If the layout row contains nothing but horizontal lines,
281: * allocate an empty span for it and assign the current span
282: * to the next layout row accepting data.
1.30 schwarze 283: */
284:
1.31 schwarze 285: while (rp->next != NULL) {
286: if (rp->last->col + 1 < tbl->opts.cols)
287: break;
288: for (cp = rp->first; cp != NULL; cp = cp->next)
289: if (cp->pos != TBL_CELL_HORIZ &&
290: cp->pos != TBL_CELL_DHORIZ)
291: break;
292: if (cp != NULL)
293: break;
294: sp = newspan(tbl, ln, rp);
295: sp->pos = TBL_SPAN_DATA;
296: rp = rp->next;
1.30 schwarze 297: }
1.1 schwarze 298:
1.31 schwarze 299: /* Process a real data row. */
1.1 schwarze 300:
1.31 schwarze 301: sp = newspan(tbl, ln, rp);
302: sp->pos = TBL_SPAN_DATA;
303: while (p[pos] != '\0')
304: getdata(tbl, sp, ln, p, &pos);
1.1 schwarze 305: }