Annotation of mandoc/man.c, Revision 1.161
1.161 ! schwarze 1: /* $Id: man.c,v 1.160 2015/04/19 14:25:41 schwarze Exp $ */
1.1 kristaps 2: /*
1.102 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.149 schwarze 4: * Copyright (c) 2013, 2014, 2015 Ingo Schwarze <schwarze@openbsd.org>
1.124 schwarze 5: * Copyright (c) 2011 Joerg Sonnenberger <joerg@netbsd.org>
1.1 kristaps 6: *
7: * Permission to use, copy, modify, and distribute this software for any
1.18 kristaps 8: * purpose with or without fee is hereby granted, provided that the above
9: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 10: *
1.150 schwarze 11: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.18 kristaps 12: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.150 schwarze 13: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.18 kristaps 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.
1.1 kristaps 18: */
1.47 kristaps 19: #include "config.h"
20:
1.41 kristaps 21: #include <sys/types.h>
22:
1.1 kristaps 23: #include <assert.h>
1.126 schwarze 24: #include <ctype.h>
1.1 kristaps 25: #include <stdarg.h>
26: #include <stdlib.h>
27: #include <stdio.h>
28: #include <string.h>
29:
1.150 schwarze 30: #include "mandoc_aux.h"
31: #include "mandoc.h"
32: #include "roff.h"
1.105 kristaps 33: #include "man.h"
1.150 schwarze 34: #include "libmandoc.h"
1.158 schwarze 35: #include "roff_int.h"
1.1 kristaps 36: #include "libman.h"
37:
1.129 schwarze 38: const char *const __man_macronames[MAN_MAX] = {
1.21 kristaps 39: "br", "TH", "SH", "SS",
1.129 schwarze 40: "TP", "LP", "PP", "P",
1.1 kristaps 41: "IP", "HP", "SM", "SB",
42: "BI", "IB", "BR", "RB",
1.11 kristaps 43: "R", "B", "I", "IR",
1.148 schwarze 44: "RI", "sp", "nf",
1.92 kristaps 45: "fi", "RE", "RS", "DT",
46: "UC", "PD", "AT", "in",
1.120 schwarze 47: "ft", "OP", "EX", "EE",
1.128 schwarze 48: "UR", "UE", "ll"
1.1 kristaps 49: };
50:
51: const char * const *man_macronames = __man_macronames;
52:
1.153 schwarze 53: static void man_descope(struct roff_man *, int, int);
54: static int man_ptext(struct roff_man *, int, char *, int);
55: static int man_pmacro(struct roff_man *, int, char *, int);
1.1 kristaps 56:
57:
1.45 kristaps 58: void
1.153 schwarze 59: man_endparse(struct roff_man *man)
1.1 kristaps 60: {
61:
1.144 schwarze 62: man_macroend(man);
1.1 kristaps 63: }
64:
65: int
1.153 schwarze 66: man_parseln(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 67: {
68:
1.150 schwarze 69: if (man->last->type != ROFFT_EQN || ln > man->last->line)
1.141 schwarze 70: man->flags |= MAN_NEWLINE;
1.97 kristaps 71:
1.119 schwarze 72: return (roff_getcontrol(man->roff, buf, &offs) ?
1.129 schwarze 73: man_pmacro(man, ln, buf, offs) :
74: man_ptext(man, ln, buf, offs));
1.1 kristaps 75: }
1.2 kristaps 76:
1.144 schwarze 77: static void
1.153 schwarze 78: man_descope(struct roff_man *man, int line, int offs)
1.94 kristaps 79: {
80: /*
81: * Co-ordinate what happens with having a next-line scope open:
82: * first close out the element scope (if applicable), then close
83: * out the block scope (also if applicable).
84: */
85:
1.144 schwarze 86: if (man->flags & MAN_ELINE) {
1.119 schwarze 87: man->flags &= ~MAN_ELINE;
1.144 schwarze 88: man_unscope(man, man->last->parent);
1.94 kristaps 89: }
1.144 schwarze 90: if ( ! (man->flags & MAN_BLINE))
91: return;
1.119 schwarze 92: man->flags &= ~MAN_BLINE;
1.144 schwarze 93: man_unscope(man, man->last->parent);
1.158 schwarze 94: roff_body_alloc(man, line, offs, man->last->tok);
1.94 kristaps 95: }
96:
1.1 kristaps 97: static int
1.153 schwarze 98: man_ptext(struct roff_man *man, int line, char *buf, int offs)
1.1 kristaps 99: {
1.61 kristaps 100: int i;
1.60 kristaps 101:
1.35 kristaps 102: /* Literal free-form text whitespace is preserved. */
103:
1.144 schwarze 104: if (man->flags & MAN_LITERAL) {
1.160 schwarze 105: roff_word_alloc(man, line, offs, buf + offs);
1.144 schwarze 106: man_descope(man, line, offs);
107: return(1);
1.35 kristaps 108: }
109:
1.144 schwarze 110: for (i = offs; buf[i] == ' '; i++)
1.31 kristaps 111: /* Skip leading whitespace. */ ;
1.48 kristaps 112:
1.121 schwarze 113: /*
114: * Blank lines are ignored right after headings
115: * but add a single vertical space elsewhere.
116: */
117:
1.144 schwarze 118: if (buf[i] == '\0') {
1.61 kristaps 119: /* Allocate a blank entry. */
1.144 schwarze 120: if (man->last->tok != MAN_SH &&
121: man->last->tok != MAN_SS) {
1.161 ! schwarze 122: roff_elem_alloc(man, line, offs, MAN_sp);
1.153 schwarze 123: man->next = ROFF_NEXT_SIBLING;
1.121 schwarze 124: }
1.118 schwarze 125: return(1);
1.31 kristaps 126: }
127:
1.129 schwarze 128: /*
1.63 kristaps 129: * Warn if the last un-escaped character is whitespace. Then
1.129 schwarze 130: * strip away the remaining spaces (tabs stay!).
1.63 kristaps 131: */
1.29 kristaps 132:
1.61 kristaps 133: i = (int)strlen(buf);
134: assert(i);
1.49 kristaps 135:
1.63 kristaps 136: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
1.64 kristaps 137: if (i > 1 && '\\' != buf[i - 2])
1.131 schwarze 138: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
139: line, i - 1, NULL);
1.63 kristaps 140:
141: for (--i; i && ' ' == buf[i]; i--)
142: /* Spin back to non-space. */ ;
143:
144: /* Jump ahead of escaped whitespace. */
145: i += '\\' == buf[i] ? 2 : 1;
146:
147: buf[i] = '\0';
148: }
1.160 schwarze 149: roff_word_alloc(man, line, offs, buf + offs);
1.65 kristaps 150:
151: /*
152: * End-of-sentence check. If the last character is an unescaped
153: * EOS character, then flag the node as being the end of a
154: * sentence. The front-end will know how to interpret this.
155: */
1.67 kristaps 156:
1.65 kristaps 157: assert(i);
1.122 schwarze 158: if (mandoc_eos(buf, (size_t)i))
1.119 schwarze 159: man->last->flags |= MAN_EOS;
1.31 kristaps 160:
1.144 schwarze 161: man_descope(man, line, offs);
162: return(1);
1.1 kristaps 163: }
164:
1.96 kristaps 165: static int
1.153 schwarze 166: man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 167: {
1.151 schwarze 168: struct roff_node *n;
1.143 schwarze 169: const char *cp;
1.151 schwarze 170: int tok;
1.134 schwarze 171: int i, ppos;
172: int bline;
1.143 schwarze 173: char mac[5];
1.1 kristaps 174:
1.107 kristaps 175: ppos = offs;
1.9 kristaps 176:
1.56 kristaps 177: /*
1.107 kristaps 178: * Copy the first word into a nil-terminated buffer.
1.143 schwarze 179: * Stop when a space, tab, escape, or eoln is encountered.
1.56 kristaps 180: */
1.62 kristaps 181:
1.107 kristaps 182: i = 0;
1.143 schwarze 183: while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
1.107 kristaps 184: mac[i++] = buf[offs++];
1.10 kristaps 185:
1.107 kristaps 186: mac[i] = '\0';
1.1 kristaps 187:
1.159 schwarze 188: tok = (i > 0 && i < 4) ? man_hash_find(mac) : TOKEN_NONE;
1.1 kristaps 189:
1.159 schwarze 190: if (tok == TOKEN_NONE) {
1.136 schwarze 191: mandoc_msg(MANDOCERR_MACRO, man->parse,
192: ln, ppos, buf + ppos - 1);
1.12 kristaps 193: return(1);
1.1 kristaps 194: }
195:
1.143 schwarze 196: /* Skip a leading escape sequence or tab. */
197:
198: switch (buf[offs]) {
199: case '\\':
200: cp = buf + offs + 1;
201: mandoc_escape(&cp, NULL, NULL);
202: offs = cp - buf;
203: break;
204: case '\t':
205: offs++;
206: break;
207: default:
208: break;
209: }
210:
211: /* Jump to the next non-whitespace word. */
1.1 kristaps 212:
1.144 schwarze 213: while (buf[offs] && buf[offs] == ' ')
1.107 kristaps 214: offs++;
1.1 kristaps 215:
1.129 schwarze 216: /*
1.62 kristaps 217: * Trailing whitespace. Note that tabs are allowed to be passed
218: * into the parser as "text", so we only warn about spaces here.
219: */
1.48 kristaps 220:
1.144 schwarze 221: if (buf[offs] == '\0' && buf[offs - 1] == ' ')
1.131 schwarze 222: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
223: ln, offs - 1, NULL);
1.48 kristaps 224:
1.129 schwarze 225: /*
1.149 schwarze 226: * Some macros break next-line scopes; otherwise, remember
227: * whether we are in next-line scope for a block head.
1.50 kristaps 228: */
1.34 kristaps 229:
1.149 schwarze 230: man_breakscope(man, tok);
1.134 schwarze 231: bline = man->flags & MAN_BLINE;
1.59 kristaps 232:
233: /* Call to handler... */
1.1 kristaps 234:
1.53 kristaps 235: assert(man_macros[tok].fp);
1.144 schwarze 236: (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf);
1.123 schwarze 237:
238: /* In quick mode (for mandocdb), abort after the NAME section. */
239:
1.144 schwarze 240: if (man->quick && tok == MAN_SH) {
1.130 schwarze 241: n = man->last;
1.150 schwarze 242: if (n->type == ROFFT_BODY &&
1.130 schwarze 243: strcmp(n->prev->child->string, "NAME"))
244: return(2);
245: }
1.1 kristaps 246:
1.129 schwarze 247: /*
1.135 schwarze 248: * If we are in a next-line scope for a block head,
249: * close it out now and switch to the body,
250: * unless the next-line scope is allowed to continue.
1.29 kristaps 251: */
252:
1.135 schwarze 253: if ( ! bline || man->flags & MAN_ELINE ||
254: man_macros[tok].flags & MAN_NSCOPED)
1.29 kristaps 255: return(1);
256:
1.144 schwarze 257: assert(man->flags & MAN_BLINE);
1.119 schwarze 258: man->flags &= ~MAN_BLINE;
1.11 kristaps 259:
1.144 schwarze 260: man_unscope(man, man->last->parent);
1.158 schwarze 261: roff_body_alloc(man, ln, ppos, man->last->tok);
1.144 schwarze 262: return(1);
1.149 schwarze 263: }
264:
265: void
1.153 schwarze 266: man_breakscope(struct roff_man *man, int tok)
1.149 schwarze 267: {
1.151 schwarze 268: struct roff_node *n;
1.149 schwarze 269:
270: /*
271: * An element next line scope is open,
272: * and the new macro is not allowed inside elements.
273: * Delete the element that is being broken.
274: */
275:
1.159 schwarze 276: if (man->flags & MAN_ELINE && (tok == TOKEN_NONE ||
1.149 schwarze 277: ! (man_macros[tok].flags & MAN_NSCOPED))) {
278: n = man->last;
1.150 schwarze 279: assert(n->type != ROFFT_TEXT);
1.149 schwarze 280: if (man_macros[n->tok].flags & MAN_NSCOPED)
281: n = n->parent;
282:
283: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
284: n->line, n->pos, "%s breaks %s",
1.159 schwarze 285: tok == TOKEN_NONE ? "TS" : man_macronames[tok],
1.149 schwarze 286: man_macronames[n->tok]);
287:
1.158 schwarze 288: roff_node_delete(man, n);
1.149 schwarze 289: man->flags &= ~MAN_ELINE;
290: }
291:
292: /*
293: * A block header next line scope is open,
294: * and the new macro is not allowed inside block headers.
295: * Delete the block that is being broken.
296: */
297:
1.159 schwarze 298: if (man->flags & MAN_BLINE && (tok == TOKEN_NONE ||
1.149 schwarze 299: man_macros[tok].flags & MAN_BSCOPE)) {
300: n = man->last;
1.150 schwarze 301: if (n->type == ROFFT_TEXT)
1.149 schwarze 302: n = n->parent;
303: if ( ! (man_macros[n->tok].flags & MAN_BSCOPE))
304: n = n->parent;
305:
1.150 schwarze 306: assert(n->type == ROFFT_HEAD);
1.149 schwarze 307: n = n->parent;
1.150 schwarze 308: assert(n->type == ROFFT_BLOCK);
1.149 schwarze 309: assert(man_macros[n->tok].flags & MAN_SCOPED);
310:
311: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
312: n->line, n->pos, "%s breaks %s",
1.159 schwarze 313: tok == TOKEN_NONE ? "TS" : man_macronames[tok],
1.149 schwarze 314: man_macronames[n->tok]);
315:
1.158 schwarze 316: roff_node_delete(man, n);
1.149 schwarze 317: man->flags &= ~MAN_BLINE;
318: }
1.112 kristaps 319: }
320:
321: const struct mparse *
1.153 schwarze 322: man_mparse(const struct roff_man *man)
1.112 kristaps 323: {
324:
1.119 schwarze 325: assert(man && man->parse);
326: return(man->parse);
1.126 schwarze 327: }
328:
329: void
1.151 schwarze 330: man_deroff(char **dest, const struct roff_node *n)
1.126 schwarze 331: {
332: char *cp;
333: size_t sz;
334:
1.150 schwarze 335: if (n->type != ROFFT_TEXT) {
1.126 schwarze 336: for (n = n->child; n; n = n->next)
337: man_deroff(dest, n);
338: return;
339: }
340:
1.127 schwarze 341: /* Skip leading whitespace and escape sequences. */
1.126 schwarze 342:
1.127 schwarze 343: cp = n->string;
344: while ('\0' != *cp) {
345: if ('\\' == *cp) {
346: cp++;
347: mandoc_escape((const char **)&cp, NULL, NULL);
348: } else if (isspace((unsigned char)*cp))
349: cp++;
350: else
1.126 schwarze 351: break;
1.127 schwarze 352: }
1.126 schwarze 353:
354: /* Skip trailing whitespace. */
355:
356: for (sz = strlen(cp); sz; sz--)
357: if (0 == isspace((unsigned char)cp[sz-1]))
358: break;
359:
360: /* Skip empty strings. */
361:
362: if (0 == sz)
363: return;
364:
365: if (NULL == *dest) {
366: *dest = mandoc_strndup(cp, sz);
367: return;
368: }
369:
370: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
371: free(*dest);
372: *dest = cp;
1.23 kristaps 373: }
CVSweb