Annotation of src/usr.bin/mandoc/tbl_data.c, Revision 1.38
1.38 ! schwarze 1: /* $OpenBSD: tbl_data.c,v 1.37 2018/12/14 01:17:46 schwarze Exp $ */
1.1 schwarze 2: /*
1.4 schwarze 3: * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.33 schwarze 4: * Copyright (c) 2011, 2015, 2017, 2018 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>
1.37 schwarze 22: #include <stdio.h>
1.1 schwarze 23: #include <stdlib.h>
24: #include <string.h>
1.4 schwarze 25: #include <time.h>
1.1 schwarze 26:
1.35 schwarze 27: #include "mandoc_aux.h"
1.4 schwarze 28: #include "mandoc.h"
1.35 schwarze 29: #include "tbl.h"
1.4 schwarze 30: #include "libmandoc.h"
1.36 schwarze 31: #include "tbl_int.h"
1.4 schwarze 32:
1.20 schwarze 33: static void getdata(struct tbl_node *, struct tbl_span *,
1.9 schwarze 34: int, const char *, int *);
1.17 schwarze 35: static struct tbl_span *newspan(struct tbl_node *, int,
1.9 schwarze 36: struct tbl_row *);
1.4 schwarze 37:
1.17 schwarze 38:
1.20 schwarze 39: static void
1.17 schwarze 40: getdata(struct tbl_node *tbl, struct tbl_span *dp,
1.4 schwarze 41: int ln, const char *p, int *pos)
1.1 schwarze 42: {
1.33 schwarze 43: struct tbl_dat *dat, *pdat;
1.4 schwarze 44: struct tbl_cell *cp;
1.33 schwarze 45: struct tbl_span *pdp;
1.24 schwarze 46: int sv;
1.4 schwarze 47:
1.34 schwarze 48: /*
49: * Determine the length of the string in the cell
50: * and advance the parse point to the end of the cell.
51: */
52:
53: sv = *pos;
54: while (p[*pos] != '\0' && p[*pos] != tbl->opts.tab)
55: (*pos)++;
56:
1.25 schwarze 57: /* Advance to the next layout cell, skipping spanners. */
58:
1.24 schwarze 59: cp = dp->last == NULL ? dp->layout->first : dp->last->layout->next;
60: while (cp != NULL && cp->pos == TBL_CELL_SPAN)
1.4 schwarze 61: cp = cp->next;
62:
1.6 schwarze 63: /*
1.30 schwarze 64: * If the current layout row is out of cells, allocate
65: * a new cell if another row of the table has at least
66: * this number of columns, or discard the input if we
67: * are beyond the last column of the table as a whole.
1.6 schwarze 68: */
69:
1.24 schwarze 70: if (cp == NULL) {
1.30 schwarze 71: if (dp->layout->last->col + 1 < dp->opts->cols) {
72: cp = mandoc_calloc(1, sizeof(*cp));
73: cp->pos = TBL_CELL_LEFT;
74: dp->layout->last->next = cp;
75: cp->col = dp->layout->last->col + 1;
76: dp->layout->last = cp;
77: } else {
1.38 ! schwarze 78: mandoc_msg(MANDOCERR_TBLDATA_EXTRA,
! 79: ln, sv, "%s", p + sv);
1.34 schwarze 80: while (p[*pos] != '\0')
1.30 schwarze 81: (*pos)++;
82: return;
83: }
1.6 schwarze 84: }
85:
1.33 schwarze 86: dat = mandoc_malloc(sizeof(*dat));
1.4 schwarze 87: dat->layout = cp;
1.33 schwarze 88: dat->next = NULL;
89: dat->string = NULL;
90: dat->hspans = 0;
91: dat->vspans = 0;
92: dat->block = 0;
1.4 schwarze 93: dat->pos = TBL_DATA_NONE;
1.33 schwarze 94:
95: /*
96: * Increment the number of vertical spans in a data cell above,
97: * if this cell vertically extends one or more cells above.
98: * The iteration must be done over data rows,
99: * not over layout rows, because one layout row
100: * can be reused for more than one data row.
101: */
102:
1.34 schwarze 103: if (cp->pos == TBL_CELL_DOWN ||
104: (*pos - sv == 2 && p[sv] == '\\' && p[sv + 1] == '^')) {
1.33 schwarze 105: pdp = dp;
106: while ((pdp = pdp->prev) != NULL) {
107: pdat = pdp->first;
108: while (pdat != NULL &&
109: pdat->layout->col < dat->layout->col)
110: pdat = pdat->next;
111: if (pdat == NULL)
112: break;
1.34 schwarze 113: if (pdat->layout->pos != TBL_CELL_DOWN &&
114: strcmp(pdat->string, "\\^") != 0) {
1.33 schwarze 115: pdat->vspans++;
116: break;
117: }
118: }
119: }
120:
121: /*
122: * Count the number of horizontal spans to the right of this cell.
123: * This is purely a matter of the layout, independent of the data.
124: */
125:
1.24 schwarze 126: for (cp = cp->next; cp != NULL; cp = cp->next)
127: if (cp->pos == TBL_CELL_SPAN)
1.33 schwarze 128: dat->hspans++;
1.6 schwarze 129: else
130: break;
1.33 schwarze 131:
132: /* Append the new data cell to the data row. */
1.17 schwarze 133:
1.24 schwarze 134: if (dp->last == NULL)
135: dp->first = dat;
136: else
1.4 schwarze 137: dp->last->next = dat;
1.24 schwarze 138: dp->last = dat;
1.4 schwarze 139:
140: /*
141: * Check for a continued-data scope opening. This consists of a
142: * trailing `T{' at the end of the line. Subsequent lines,
143: * until a standalone `T}', are included in our cell.
144: */
1.1 schwarze 145:
1.24 schwarze 146: if (*pos - sv == 2 && p[sv] == 'T' && p[sv + 1] == '{') {
1.4 schwarze 147: tbl->part = TBL_PART_CDATA;
1.20 schwarze 148: return;
1.4 schwarze 149: }
1.1 schwarze 150:
1.24 schwarze 151: dat->string = mandoc_strndup(p + sv, *pos - sv);
1.1 schwarze 152:
1.34 schwarze 153: if (p[*pos] != '\0')
1.4 schwarze 154: (*pos)++;
1.1 schwarze 155:
156: if ( ! strcmp(dat->string, "_"))
1.4 schwarze 157: dat->pos = TBL_DATA_HORIZ;
1.1 schwarze 158: else if ( ! strcmp(dat->string, "="))
1.4 schwarze 159: dat->pos = TBL_DATA_DHORIZ;
1.1 schwarze 160: else if ( ! strcmp(dat->string, "\\_"))
1.4 schwarze 161: dat->pos = TBL_DATA_NHORIZ;
1.1 schwarze 162: else if ( ! strcmp(dat->string, "\\="))
1.4 schwarze 163: dat->pos = TBL_DATA_NDHORIZ;
1.1 schwarze 164: else
1.4 schwarze 165: dat->pos = TBL_DATA_DATA;
166:
1.24 schwarze 167: if ((dat->layout->pos == TBL_CELL_HORIZ ||
168: dat->layout->pos == TBL_CELL_DHORIZ ||
169: dat->layout->pos == TBL_CELL_DOWN) &&
170: dat->pos == TBL_DATA_DATA && *dat->string != '\0')
171: mandoc_msg(MANDOCERR_TBLDATA_SPAN,
1.38 ! schwarze 172: ln, sv, "%s", dat->string);
1.1 schwarze 173: }
174:
1.32 schwarze 175: void
1.22 schwarze 176: tbl_cdata(struct tbl_node *tbl, int ln, const char *p, int pos)
1.4 schwarze 177: {
178: struct tbl_dat *dat;
1.17 schwarze 179: size_t sz;
1.4 schwarze 180:
181: dat = tbl->last_span->last;
1.5 schwarze 182:
183: if (p[pos] == 'T' && p[pos + 1] == '}') {
184: pos += 2;
185: if (p[pos] == tbl->opts.tab) {
186: tbl->part = TBL_PART_DATA;
187: pos++;
1.27 schwarze 188: while (p[pos] != '\0')
189: getdata(tbl, tbl->last_span, ln, p, &pos);
1.32 schwarze 190: return;
1.24 schwarze 191: } else if (p[pos] == '\0') {
1.5 schwarze 192: tbl->part = TBL_PART_DATA;
1.32 schwarze 193: return;
1.5 schwarze 194: }
195:
196: /* Fallthrough: T} is part of a word. */
197: }
1.4 schwarze 198:
1.6 schwarze 199: dat->pos = TBL_DATA_DATA;
1.29 schwarze 200: dat->block = 1;
1.6 schwarze 201:
1.24 schwarze 202: if (dat->string != NULL) {
1.23 schwarze 203: sz = strlen(p + pos) + strlen(dat->string) + 2;
1.4 schwarze 204: dat->string = mandoc_realloc(dat->string, sz);
1.18 schwarze 205: (void)strlcat(dat->string, " ", sz);
1.23 schwarze 206: (void)strlcat(dat->string, p + pos, sz);
1.4 schwarze 207: } else
1.23 schwarze 208: dat->string = mandoc_strdup(p + pos);
1.4 schwarze 209:
1.24 schwarze 210: if (dat->layout->pos == TBL_CELL_DOWN)
1.38 ! schwarze 211: mandoc_msg(MANDOCERR_TBLDATA_SPAN,
! 212: ln, pos, "%s", dat->string);
1.4 schwarze 213: }
1.1 schwarze 214:
1.7 schwarze 215: static struct tbl_span *
1.9 schwarze 216: newspan(struct tbl_node *tbl, int line, struct tbl_row *rp)
1.7 schwarze 217: {
218: struct tbl_span *dp;
219:
1.24 schwarze 220: dp = mandoc_calloc(1, sizeof(*dp));
1.9 schwarze 221: dp->line = line;
1.13 schwarze 222: dp->opts = &tbl->opts;
1.7 schwarze 223: dp->layout = rp;
1.21 schwarze 224: dp->prev = tbl->last_span;
1.7 schwarze 225:
1.21 schwarze 226: if (dp->prev == NULL) {
227: tbl->first_span = dp;
1.8 schwarze 228: tbl->current_span = NULL;
1.21 schwarze 229: } else
230: dp->prev->next = dp;
231: tbl->last_span = dp;
1.7 schwarze 232:
1.28 schwarze 233: return dp;
1.7 schwarze 234: }
235:
1.20 schwarze 236: void
1.22 schwarze 237: tbl_data(struct tbl_node *tbl, int ln, const char *p, int pos)
1.1 schwarze 238: {
1.4 schwarze 239: struct tbl_row *rp;
1.30 schwarze 240: struct tbl_cell *cp;
1.31 schwarze 241: struct tbl_span *sp;
1.4 schwarze 242:
1.30 schwarze 243: rp = (sp = tbl->last_span) == NULL ? tbl->first_row :
244: sp->pos == TBL_SPAN_DATA && sp->layout->next != NULL ?
245: sp->layout->next : sp->layout;
1.6 schwarze 246:
1.30 schwarze 247: assert(rp != NULL);
1.4 schwarze 248:
1.1 schwarze 249: if ( ! strcmp(p, "_")) {
1.31 schwarze 250: sp = newspan(tbl, ln, rp);
1.30 schwarze 251: sp->pos = TBL_SPAN_HORIZ;
1.20 schwarze 252: return;
1.1 schwarze 253: } else if ( ! strcmp(p, "=")) {
1.31 schwarze 254: sp = newspan(tbl, ln, rp);
1.30 schwarze 255: sp->pos = TBL_SPAN_DHORIZ;
1.20 schwarze 256: return;
1.1 schwarze 257: }
1.30 schwarze 258:
259: /*
1.31 schwarze 260: * If the layout row contains nothing but horizontal lines,
261: * allocate an empty span for it and assign the current span
262: * to the next layout row accepting data.
1.30 schwarze 263: */
264:
1.31 schwarze 265: while (rp->next != NULL) {
266: if (rp->last->col + 1 < tbl->opts.cols)
267: break;
268: for (cp = rp->first; cp != NULL; cp = cp->next)
269: if (cp->pos != TBL_CELL_HORIZ &&
270: cp->pos != TBL_CELL_DHORIZ)
271: break;
272: if (cp != NULL)
273: break;
274: sp = newspan(tbl, ln, rp);
275: sp->pos = TBL_SPAN_DATA;
276: rp = rp->next;
1.30 schwarze 277: }
1.1 schwarze 278:
1.31 schwarze 279: /* Process a real data row. */
1.1 schwarze 280:
1.31 schwarze 281: sp = newspan(tbl, ln, rp);
282: sp->pos = TBL_SPAN_DATA;
283: while (p[pos] != '\0')
284: getdata(tbl, sp, ln, p, &pos);
1.1 schwarze 285: }