Annotation of src/usr.bin/mandoc/tbl_data.c, Revision 1.45
1.45 ! schwarze 1: /* $OpenBSD: tbl_data.c,v 1.44 2021/09/07 11:47:42 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;
1.44 schwarze 48: int startpos, endpos;
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:
1.44 schwarze 55: startpos = *pos;
56: ccp = p + startpos;
1.42 schwarze 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;
1.43 schwarze 79: cp->font = ESCAPE_FONTROMAN;
1.40 schwarze 80: cp->spacing = SIZE_MAX;
1.30 schwarze 81: dp->layout->last->next = cp;
82: cp->col = dp->layout->last->col + 1;
83: dp->layout->last = cp;
84: } else {
1.38 schwarze 85: mandoc_msg(MANDOCERR_TBLDATA_EXTRA,
1.44 schwarze 86: ln, startpos, "%s", p + startpos);
1.34 schwarze 87: while (p[*pos] != '\0')
1.30 schwarze 88: (*pos)++;
89: return;
90: }
1.6 schwarze 91: }
92:
1.33 schwarze 93: dat = mandoc_malloc(sizeof(*dat));
1.4 schwarze 94: dat->layout = cp;
1.33 schwarze 95: dat->next = NULL;
96: dat->string = NULL;
97: dat->hspans = 0;
98: dat->vspans = 0;
99: dat->block = 0;
1.4 schwarze 100: dat->pos = TBL_DATA_NONE;
1.33 schwarze 101:
102: /*
103: * Increment the number of vertical spans in a data cell above,
104: * if this cell vertically extends one or more cells above.
105: * The iteration must be done over data rows,
106: * not over layout rows, because one layout row
107: * can be reused for more than one data row.
108: */
109:
1.34 schwarze 110: if (cp->pos == TBL_CELL_DOWN ||
1.44 schwarze 111: (*pos - startpos == 2 &&
112: p[startpos] == '\\' && p[startpos + 1] == '^')) {
1.33 schwarze 113: pdp = dp;
114: while ((pdp = pdp->prev) != NULL) {
115: pdat = pdp->first;
116: while (pdat != NULL &&
117: pdat->layout->col < dat->layout->col)
118: pdat = pdat->next;
119: if (pdat == NULL)
120: break;
1.34 schwarze 121: if (pdat->layout->pos != TBL_CELL_DOWN &&
122: strcmp(pdat->string, "\\^") != 0) {
1.33 schwarze 123: pdat->vspans++;
124: break;
125: }
126: }
127: }
128:
129: /*
130: * Count the number of horizontal spans to the right of this cell.
131: * This is purely a matter of the layout, independent of the data.
132: */
133:
1.24 schwarze 134: for (cp = cp->next; cp != NULL; cp = cp->next)
135: if (cp->pos == TBL_CELL_SPAN)
1.33 schwarze 136: dat->hspans++;
1.6 schwarze 137: else
138: break;
1.33 schwarze 139:
140: /* Append the new data cell to the data row. */
1.17 schwarze 141:
1.24 schwarze 142: if (dp->last == NULL)
143: dp->first = dat;
144: else
1.4 schwarze 145: dp->last->next = dat;
1.24 schwarze 146: dp->last = dat;
1.4 schwarze 147:
148: /*
149: * Check for a continued-data scope opening. This consists of a
150: * trailing `T{' at the end of the line. Subsequent lines,
151: * until a standalone `T}', are included in our cell.
152: */
1.1 schwarze 153:
1.44 schwarze 154: if (*pos - startpos == 2 &&
155: p[startpos] == 'T' && p[startpos + 1] == '{') {
1.4 schwarze 156: tbl->part = TBL_PART_CDATA;
1.20 schwarze 157: return;
1.4 schwarze 158: }
1.1 schwarze 159:
1.44 schwarze 160: endpos = *pos;
161: if (dp->opts->opts & TBL_OPT_NOSPACE) {
162: while (p[startpos] == ' ')
163: startpos++;
164: while (endpos > startpos && p[endpos - 1] == ' ')
165: endpos--;
166: }
167: dat->string = mandoc_strndup(p + startpos, endpos - startpos);
1.1 schwarze 168:
1.34 schwarze 169: if (p[*pos] != '\0')
1.4 schwarze 170: (*pos)++;
1.1 schwarze 171:
172: if ( ! strcmp(dat->string, "_"))
1.4 schwarze 173: dat->pos = TBL_DATA_HORIZ;
1.1 schwarze 174: else if ( ! strcmp(dat->string, "="))
1.4 schwarze 175: dat->pos = TBL_DATA_DHORIZ;
1.1 schwarze 176: else if ( ! strcmp(dat->string, "\\_"))
1.4 schwarze 177: dat->pos = TBL_DATA_NHORIZ;
1.1 schwarze 178: else if ( ! strcmp(dat->string, "\\="))
1.4 schwarze 179: dat->pos = TBL_DATA_NDHORIZ;
1.1 schwarze 180: else
1.4 schwarze 181: dat->pos = TBL_DATA_DATA;
182:
1.24 schwarze 183: if ((dat->layout->pos == TBL_CELL_HORIZ ||
184: dat->layout->pos == TBL_CELL_DHORIZ ||
185: dat->layout->pos == TBL_CELL_DOWN) &&
186: dat->pos == TBL_DATA_DATA && *dat->string != '\0')
187: mandoc_msg(MANDOCERR_TBLDATA_SPAN,
1.44 schwarze 188: ln, startpos, "%s", dat->string);
1.1 schwarze 189: }
190:
1.32 schwarze 191: void
1.22 schwarze 192: tbl_cdata(struct tbl_node *tbl, int ln, const char *p, int pos)
1.4 schwarze 193: {
194: struct tbl_dat *dat;
1.17 schwarze 195: size_t sz;
1.4 schwarze 196:
197: dat = tbl->last_span->last;
1.5 schwarze 198:
199: if (p[pos] == 'T' && p[pos + 1] == '}') {
200: pos += 2;
1.45 ! schwarze 201: if (tbl->opts.opts & TBL_OPT_NOSPACE)
! 202: while (p[pos] == ' ')
! 203: pos++;
1.5 schwarze 204: if (p[pos] == tbl->opts.tab) {
205: tbl->part = TBL_PART_DATA;
206: pos++;
1.27 schwarze 207: while (p[pos] != '\0')
208: getdata(tbl, tbl->last_span, ln, p, &pos);
1.32 schwarze 209: return;
1.24 schwarze 210: } else if (p[pos] == '\0') {
1.5 schwarze 211: tbl->part = TBL_PART_DATA;
1.32 schwarze 212: return;
1.5 schwarze 213: }
214:
215: /* Fallthrough: T} is part of a word. */
216: }
1.4 schwarze 217:
1.6 schwarze 218: dat->pos = TBL_DATA_DATA;
1.29 schwarze 219: dat->block = 1;
1.6 schwarze 220:
1.24 schwarze 221: if (dat->string != NULL) {
1.23 schwarze 222: sz = strlen(p + pos) + strlen(dat->string) + 2;
1.4 schwarze 223: dat->string = mandoc_realloc(dat->string, sz);
1.18 schwarze 224: (void)strlcat(dat->string, " ", sz);
1.23 schwarze 225: (void)strlcat(dat->string, p + pos, sz);
1.4 schwarze 226: } else
1.23 schwarze 227: dat->string = mandoc_strdup(p + pos);
1.4 schwarze 228:
1.24 schwarze 229: if (dat->layout->pos == TBL_CELL_DOWN)
1.38 schwarze 230: mandoc_msg(MANDOCERR_TBLDATA_SPAN,
231: ln, pos, "%s", dat->string);
1.4 schwarze 232: }
1.1 schwarze 233:
1.7 schwarze 234: static struct tbl_span *
1.9 schwarze 235: newspan(struct tbl_node *tbl, int line, struct tbl_row *rp)
1.7 schwarze 236: {
237: struct tbl_span *dp;
238:
1.24 schwarze 239: dp = mandoc_calloc(1, sizeof(*dp));
1.9 schwarze 240: dp->line = line;
1.13 schwarze 241: dp->opts = &tbl->opts;
1.7 schwarze 242: dp->layout = rp;
1.21 schwarze 243: dp->prev = tbl->last_span;
1.7 schwarze 244:
1.21 schwarze 245: if (dp->prev == NULL) {
246: tbl->first_span = dp;
1.8 schwarze 247: tbl->current_span = NULL;
1.21 schwarze 248: } else
249: dp->prev->next = dp;
250: tbl->last_span = dp;
1.7 schwarze 251:
1.28 schwarze 252: return dp;
1.7 schwarze 253: }
254:
1.20 schwarze 255: void
1.22 schwarze 256: tbl_data(struct tbl_node *tbl, int ln, const char *p, int pos)
1.1 schwarze 257: {
1.4 schwarze 258: struct tbl_row *rp;
1.30 schwarze 259: struct tbl_cell *cp;
1.31 schwarze 260: struct tbl_span *sp;
1.4 schwarze 261:
1.41 schwarze 262: for (sp = tbl->last_span; sp != NULL; sp = sp->prev)
263: if (sp->pos == TBL_SPAN_DATA)
264: break;
265: rp = sp == NULL ? tbl->first_row :
266: sp->layout->next == NULL ? sp->layout : sp->layout->next;
1.30 schwarze 267: assert(rp != NULL);
1.4 schwarze 268:
1.39 schwarze 269: if (p[1] == '\0') {
270: switch (p[0]) {
271: case '.':
272: /*
273: * Empty request lines must be handled here
274: * and cannot be discarded in roff_parseln()
275: * because in the layout section, they
276: * are significant and end the layout.
277: */
278: return;
279: case '_':
280: sp = newspan(tbl, ln, rp);
281: sp->pos = TBL_SPAN_HORIZ;
282: return;
283: case '=':
284: sp = newspan(tbl, ln, rp);
285: sp->pos = TBL_SPAN_DHORIZ;
286: return;
287: default:
288: break;
289: }
1.1 schwarze 290: }
1.30 schwarze 291:
292: /*
1.31 schwarze 293: * If the layout row contains nothing but horizontal lines,
294: * allocate an empty span for it and assign the current span
295: * to the next layout row accepting data.
1.30 schwarze 296: */
297:
1.31 schwarze 298: while (rp->next != NULL) {
299: if (rp->last->col + 1 < tbl->opts.cols)
300: break;
301: for (cp = rp->first; cp != NULL; cp = cp->next)
302: if (cp->pos != TBL_CELL_HORIZ &&
303: cp->pos != TBL_CELL_DHORIZ)
304: break;
305: if (cp != NULL)
306: break;
307: sp = newspan(tbl, ln, rp);
308: sp->pos = TBL_SPAN_DATA;
309: rp = rp->next;
1.30 schwarze 310: }
1.1 schwarze 311:
1.31 schwarze 312: /* Process a real data row. */
1.1 schwarze 313:
1.31 schwarze 314: sp = newspan(tbl, ln, rp);
315: sp->pos = TBL_SPAN_DATA;
316: while (p[pos] != '\0')
317: getdata(tbl, sp, ln, p, &pos);
1.1 schwarze 318: }