Annotation of mandoc/man.c, Revision 1.180
1.180 ! schwarze 1: /* $Id: man.c,v 1.179 2018/08/25 12:33:03 schwarze Exp $ */
1.1 kristaps 2: /*
1.102 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.177 schwarze 4: * Copyright (c) 2013,2014,2015,2017,2018 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:
1.179 schwarze 38: static char *man_hasc(char *);
1.153 schwarze 39: static int man_ptext(struct roff_man *, int, char *, int);
40: static int man_pmacro(struct roff_man *, int, char *, int);
1.1 kristaps 41:
42:
43: int
1.153 schwarze 44: man_parseln(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 45: {
46:
1.150 schwarze 47: if (man->last->type != ROFFT_EQN || ln > man->last->line)
1.141 schwarze 48: man->flags |= MAN_NEWLINE;
1.97 kristaps 49:
1.165 schwarze 50: return roff_getcontrol(man->roff, buf, &offs) ?
1.129 schwarze 51: man_pmacro(man, ln, buf, offs) :
1.165 schwarze 52: man_ptext(man, ln, buf, offs);
1.1 kristaps 53: }
1.2 kristaps 54:
1.179 schwarze 55: /*
56: * If the string ends with \c, return a pointer to the backslash.
57: * Otherwise, return NULL.
58: */
59: static char *
60: man_hasc(char *start)
61: {
62: char *cp, *ep;
63:
64: ep = strchr(start, '\0') - 2;
65: if (ep < start || ep[0] != '\\' || ep[1] != 'c')
66: return NULL;
67: for (cp = ep; cp > start; cp--)
68: if (cp[-1] != '\\')
69: break;
70: return (ep - cp) % 2 ? NULL : ep;
71: }
72:
1.180 ! schwarze 73: void
1.179 schwarze 74: man_descope(struct roff_man *man, int line, int offs, char *start)
1.94 kristaps 75: {
1.179 schwarze 76: /* Trailing \c keeps next-line scope open. */
77:
1.180 ! schwarze 78: if (start != NULL && man_hasc(start) != NULL)
1.179 schwarze 79: return;
80:
1.94 kristaps 81: /*
82: * Co-ordinate what happens with having a next-line scope open:
1.180 ! schwarze 83: * first close out the element scopes (if applicable),
! 84: * then close out the block scope (also if applicable).
1.94 kristaps 85: */
86:
1.144 schwarze 87: if (man->flags & MAN_ELINE) {
1.180 ! schwarze 88: while (man->last->parent->type != ROFFT_ROOT &&
! 89: man_macro(man->last->parent->tok)->flags & MAN_ESCOPED)
! 90: man_unscope(man, man->last->parent);
1.119 schwarze 91: man->flags &= ~MAN_ELINE;
1.94 kristaps 92: }
1.144 schwarze 93: if ( ! (man->flags & MAN_BLINE))
94: return;
1.119 schwarze 95: man->flags &= ~MAN_BLINE;
1.144 schwarze 96: man_unscope(man, man->last->parent);
1.158 schwarze 97: roff_body_alloc(man, line, offs, man->last->tok);
1.94 kristaps 98: }
99:
1.1 kristaps 100: static int
1.153 schwarze 101: man_ptext(struct roff_man *man, int line, char *buf, int offs)
1.1 kristaps 102: {
1.61 kristaps 103: int i;
1.175 schwarze 104: char *ep;
1.60 kristaps 105:
1.35 kristaps 106: /* Literal free-form text whitespace is preserved. */
107:
1.144 schwarze 108: if (man->flags & MAN_LITERAL) {
1.160 schwarze 109: roff_word_alloc(man, line, offs, buf + offs);
1.179 schwarze 110: man_descope(man, line, offs, buf + offs);
1.165 schwarze 111: return 1;
1.35 kristaps 112: }
113:
1.144 schwarze 114: for (i = offs; buf[i] == ' '; i++)
1.31 kristaps 115: /* Skip leading whitespace. */ ;
1.48 kristaps 116:
1.121 schwarze 117: /*
1.175 schwarze 118: * Blank lines are ignored in next line scope
119: * and right after headings and cancel preceding \c,
120: * but add a single vertical space elsewhere.
1.121 schwarze 121: */
122:
1.144 schwarze 123: if (buf[i] == '\0') {
1.175 schwarze 124: if (man->flags & (MAN_ELINE | MAN_BLINE)) {
1.174 schwarze 125: mandoc_msg(MANDOCERR_BLK_BLANK, man->parse,
126: line, 0, NULL);
1.175 schwarze 127: return 1;
1.121 schwarze 128: }
1.175 schwarze 129: if (man->last->tok == MAN_SH || man->last->tok == MAN_SS)
130: return 1;
1.179 schwarze 131: if (man->last->type == ROFFT_TEXT &&
132: ((ep = man_hasc(man->last->string)) != NULL)) {
1.175 schwarze 133: *ep = '\0';
134: return 1;
135: }
136: roff_elem_alloc(man, line, offs, ROFF_sp);
137: man->next = ROFF_NEXT_SIBLING;
1.165 schwarze 138: return 1;
1.31 kristaps 139: }
140:
1.129 schwarze 141: /*
1.63 kristaps 142: * Warn if the last un-escaped character is whitespace. Then
1.129 schwarze 143: * strip away the remaining spaces (tabs stay!).
1.63 kristaps 144: */
1.29 kristaps 145:
1.61 kristaps 146: i = (int)strlen(buf);
147: assert(i);
1.49 kristaps 148:
1.63 kristaps 149: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
1.64 kristaps 150: if (i > 1 && '\\' != buf[i - 2])
1.131 schwarze 151: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
152: line, i - 1, NULL);
1.63 kristaps 153:
154: for (--i; i && ' ' == buf[i]; i--)
155: /* Spin back to non-space. */ ;
156:
157: /* Jump ahead of escaped whitespace. */
158: i += '\\' == buf[i] ? 2 : 1;
159:
160: buf[i] = '\0';
161: }
1.160 schwarze 162: roff_word_alloc(man, line, offs, buf + offs);
1.65 kristaps 163:
164: /*
165: * End-of-sentence check. If the last character is an unescaped
166: * EOS character, then flag the node as being the end of a
167: * sentence. The front-end will know how to interpret this.
168: */
1.67 kristaps 169:
1.65 kristaps 170: assert(i);
1.122 schwarze 171: if (mandoc_eos(buf, (size_t)i))
1.167 schwarze 172: man->last->flags |= NODE_EOS;
1.31 kristaps 173:
1.179 schwarze 174: man_descope(man, line, offs, buf + offs);
1.165 schwarze 175: return 1;
1.1 kristaps 176: }
177:
1.96 kristaps 178: static int
1.153 schwarze 179: man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 180: {
1.151 schwarze 181: struct roff_node *n;
1.143 schwarze 182: const char *cp;
1.170 schwarze 183: size_t sz;
184: enum roff_tok tok;
185: int ppos;
1.134 schwarze 186: int bline;
1.170 schwarze 187:
188: /* Determine the line macro. */
1.1 kristaps 189:
1.107 kristaps 190: ppos = offs;
1.170 schwarze 191: tok = TOKEN_NONE;
192: for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
193: offs++;
194: if (sz > 0 && sz < 4)
195: tok = roffhash_find(man->manmac, buf + ppos, sz);
1.159 schwarze 196: if (tok == TOKEN_NONE) {
1.136 schwarze 197: mandoc_msg(MANDOCERR_MACRO, man->parse,
198: ln, ppos, buf + ppos - 1);
1.165 schwarze 199: return 1;
1.1 kristaps 200: }
201:
1.143 schwarze 202: /* Skip a leading escape sequence or tab. */
203:
204: switch (buf[offs]) {
205: case '\\':
206: cp = buf + offs + 1;
207: mandoc_escape(&cp, NULL, NULL);
208: offs = cp - buf;
209: break;
210: case '\t':
211: offs++;
212: break;
213: default:
214: break;
215: }
216:
217: /* Jump to the next non-whitespace word. */
1.1 kristaps 218:
1.168 schwarze 219: while (buf[offs] == ' ')
1.107 kristaps 220: offs++;
1.1 kristaps 221:
1.129 schwarze 222: /*
1.62 kristaps 223: * Trailing whitespace. Note that tabs are allowed to be passed
224: * into the parser as "text", so we only warn about spaces here.
225: */
1.48 kristaps 226:
1.144 schwarze 227: if (buf[offs] == '\0' && buf[offs - 1] == ' ')
1.131 schwarze 228: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
229: ln, offs - 1, NULL);
1.48 kristaps 230:
1.129 schwarze 231: /*
1.149 schwarze 232: * Some macros break next-line scopes; otherwise, remember
233: * whether we are in next-line scope for a block head.
1.50 kristaps 234: */
1.34 kristaps 235:
1.149 schwarze 236: man_breakscope(man, tok);
1.134 schwarze 237: bline = man->flags & MAN_BLINE;
1.171 schwarze 238:
239: /*
240: * If the line in next-line scope ends with \c, keep the
241: * next-line scope open for the subsequent input line.
242: * That is not at all portable, only groff >= 1.22.4
243: * does it, but *if* this weird idiom occurs in a manual
244: * page, that's very likely what the author intended.
245: */
246:
1.180 ! schwarze 247: if (bline && man_hasc(buf + offs))
! 248: bline = 0;
1.59 kristaps 249:
250: /* Call to handler... */
1.1 kristaps 251:
1.177 schwarze 252: (*man_macro(tok)->fp)(man, tok, ln, ppos, &offs, buf);
1.123 schwarze 253:
254: /* In quick mode (for mandocdb), abort after the NAME section. */
255:
1.144 schwarze 256: if (man->quick && tok == MAN_SH) {
1.130 schwarze 257: n = man->last;
1.150 schwarze 258: if (n->type == ROFFT_BODY &&
1.130 schwarze 259: strcmp(n->prev->child->string, "NAME"))
1.165 schwarze 260: return 2;
1.130 schwarze 261: }
1.1 kristaps 262:
1.129 schwarze 263: /*
1.135 schwarze 264: * If we are in a next-line scope for a block head,
265: * close it out now and switch to the body,
266: * unless the next-line scope is allowed to continue.
1.29 kristaps 267: */
268:
1.180 ! schwarze 269: if (bline == 0 ||
! 270: (man->flags & MAN_BLINE) == 0 ||
! 271: man->flags & MAN_ELINE ||
1.177 schwarze 272: man_macro(tok)->flags & MAN_NSCOPED)
1.165 schwarze 273: return 1;
1.29 kristaps 274:
1.119 schwarze 275: man->flags &= ~MAN_BLINE;
1.144 schwarze 276: man_unscope(man, man->last->parent);
1.158 schwarze 277: roff_body_alloc(man, ln, ppos, man->last->tok);
1.165 schwarze 278: return 1;
1.149 schwarze 279: }
280:
281: void
1.153 schwarze 282: man_breakscope(struct roff_man *man, int tok)
1.149 schwarze 283: {
1.151 schwarze 284: struct roff_node *n;
1.149 schwarze 285:
286: /*
287: * An element next line scope is open,
288: * and the new macro is not allowed inside elements.
289: * Delete the element that is being broken.
290: */
291:
1.173 schwarze 292: if (man->flags & MAN_ELINE && (tok < MAN_TH ||
1.177 schwarze 293: (man_macro(tok)->flags & MAN_NSCOPED) == 0)) {
1.149 schwarze 294: n = man->last;
1.176 schwarze 295: if (n->type == ROFFT_TEXT)
296: n = n->parent;
297: if (n->tok < MAN_TH ||
1.180 ! schwarze 298: (man_macro(n->tok)->flags & (MAN_NSCOPED | MAN_ESCOPED))
! 299: == MAN_NSCOPED)
1.149 schwarze 300: n = n->parent;
301:
302: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
303: n->line, n->pos, "%s breaks %s",
1.173 schwarze 304: roff_name[tok], roff_name[n->tok]);
1.149 schwarze 305:
1.158 schwarze 306: roff_node_delete(man, n);
1.149 schwarze 307: man->flags &= ~MAN_ELINE;
1.164 schwarze 308: }
309:
310: /*
311: * Weird special case:
312: * Switching fill mode closes section headers.
313: */
314:
315: if (man->flags & MAN_BLINE &&
316: (tok == MAN_nf || tok == MAN_fi) &&
317: (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
318: n = man->last;
319: man_unscope(man, n);
320: roff_body_alloc(man, n->line, n->pos, n->tok);
321: man->flags &= ~MAN_BLINE;
1.149 schwarze 322: }
323:
324: /*
325: * A block header next line scope is open,
326: * and the new macro is not allowed inside block headers.
327: * Delete the block that is being broken.
328: */
329:
1.173 schwarze 330: if (man->flags & MAN_BLINE && (tok < MAN_TH ||
1.180 ! schwarze 331: man_macro(tok)->flags & MAN_XSCOPE)) {
1.149 schwarze 332: n = man->last;
1.150 schwarze 333: if (n->type == ROFFT_TEXT)
1.149 schwarze 334: n = n->parent;
1.176 schwarze 335: if (n->tok < MAN_TH ||
1.180 ! schwarze 336: (man_macro(n->tok)->flags & MAN_XSCOPE) == 0)
1.149 schwarze 337: n = n->parent;
338:
1.150 schwarze 339: assert(n->type == ROFFT_HEAD);
1.149 schwarze 340: n = n->parent;
1.150 schwarze 341: assert(n->type == ROFFT_BLOCK);
1.180 ! schwarze 342: assert(man_macro(n->tok)->flags & MAN_BSCOPED);
1.149 schwarze 343:
344: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
345: n->line, n->pos, "%s breaks %s",
1.173 schwarze 346: roff_name[tok], roff_name[n->tok]);
1.149 schwarze 347:
1.158 schwarze 348: roff_node_delete(man, n);
1.149 schwarze 349: man->flags &= ~MAN_BLINE;
350: }
1.166 schwarze 351: }
352:
353: void
354: man_state(struct roff_man *man, struct roff_node *n)
355: {
356:
357: switch(n->tok) {
358: case MAN_nf:
359: case MAN_EX:
1.167 schwarze 360: if (man->flags & MAN_LITERAL && ! (n->flags & NODE_VALID))
1.166 schwarze 361: mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
362: n->line, n->pos, "nf");
363: man->flags |= MAN_LITERAL;
364: break;
365: case MAN_fi:
366: case MAN_EE:
367: if ( ! (man->flags & MAN_LITERAL) &&
1.167 schwarze 368: ! (n->flags & NODE_VALID))
1.166 schwarze 369: mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
370: n->line, n->pos, "fi");
371: man->flags &= ~MAN_LITERAL;
372: break;
373: default:
374: break;
375: }
1.167 schwarze 376: man->last->flags |= NODE_VALID;
1.166 schwarze 377: }
378:
379: void
380: man_validate(struct roff_man *man)
381: {
382:
383: man->last = man->first;
384: man_node_validate(man);
385: man->flags &= ~MAN_LITERAL;
1.23 kristaps 386: }
CVSweb