Annotation of src/usr.bin/mandoc/man.c, Revision 1.134
1.134 ! schwarze 1: /* $OpenBSD: man.c,v 1.133 2018/12/31 08:38:17 schwarze Exp $ */
1.1 kristaps 2: /*
1.52 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.125 schwarze 4: * Copyright (c) 2013,2014,2015,2017,2018 Ingo Schwarze <schwarze@openbsd.org>
1.74 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
1.2 schwarze 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: *
1.98 schwarze 11: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.2 schwarze 12: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.98 schwarze 13: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.2 schwarze 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: */
1.14 schwarze 19: #include <sys/types.h>
20:
1.1 kristaps 21: #include <assert.h>
1.76 schwarze 22: #include <ctype.h>
1.1 kristaps 23: #include <stdarg.h>
24: #include <stdlib.h>
25: #include <stdio.h>
26: #include <string.h>
27:
1.98 schwarze 28: #include "mandoc_aux.h"
29: #include "mandoc.h"
30: #include "roff.h"
1.59 schwarze 31: #include "man.h"
1.98 schwarze 32: #include "libmandoc.h"
1.106 schwarze 33: #include "roff_int.h"
1.1 kristaps 34: #include "libman.h"
35:
1.127 schwarze 36: static char *man_hasc(char *);
1.101 schwarze 37: static int man_ptext(struct roff_man *, int, char *, int);
38: static int man_pmacro(struct roff_man *, int, char *, int);
1.1 kristaps 39:
40:
41: int
1.101 schwarze 42: man_parseln(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 43: {
1.25 schwarze 44:
1.98 schwarze 45: if (man->last->type != ROFFT_EQN || ln > man->last->line)
1.89 schwarze 46: man->flags |= MAN_NEWLINE;
1.54 schwarze 47:
1.113 schwarze 48: return roff_getcontrol(man->roff, buf, &offs) ?
1.79 schwarze 49: man_pmacro(man, ln, buf, offs) :
1.113 schwarze 50: man_ptext(man, ln, buf, offs);
1.1 kristaps 51: }
52:
1.127 schwarze 53: /*
54: * If the string ends with \c, return a pointer to the backslash.
55: * Otherwise, return NULL.
56: */
57: static char *
58: man_hasc(char *start)
59: {
60: char *cp, *ep;
61:
62: ep = strchr(start, '\0') - 2;
63: if (ep < start || ep[0] != '\\' || ep[1] != 'c')
64: return NULL;
65: for (cp = ep; cp > start; cp--)
66: if (cp[-1] != '\\')
67: break;
68: return (ep - cp) % 2 ? NULL : ep;
69: }
70:
1.128 schwarze 71: void
1.127 schwarze 72: man_descope(struct roff_man *man, int line, int offs, char *start)
1.52 schwarze 73: {
1.127 schwarze 74: /* Trailing \c keeps next-line scope open. */
75:
1.128 schwarze 76: if (start != NULL && man_hasc(start) != NULL)
1.127 schwarze 77: return;
78:
1.52 schwarze 79: /*
80: * Co-ordinate what happens with having a next-line scope open:
1.128 schwarze 81: * first close out the element scopes (if applicable),
82: * then close out the block scope (also if applicable).
1.52 schwarze 83: */
84:
1.92 schwarze 85: if (man->flags & MAN_ELINE) {
1.128 schwarze 86: while (man->last->parent->type != ROFFT_ROOT &&
87: man_macro(man->last->parent->tok)->flags & MAN_ESCOPED)
88: man_unscope(man, man->last->parent);
1.69 schwarze 89: man->flags &= ~MAN_ELINE;
1.52 schwarze 90: }
1.92 schwarze 91: if ( ! (man->flags & MAN_BLINE))
92: return;
1.69 schwarze 93: man->flags &= ~MAN_BLINE;
1.92 schwarze 94: man_unscope(man, man->last->parent);
1.106 schwarze 95: roff_body_alloc(man, line, offs, man->last->tok);
1.52 schwarze 96: }
97:
1.1 kristaps 98: static int
1.101 schwarze 99: man_ptext(struct roff_man *man, int line, char *buf, int offs)
1.1 kristaps 100: {
1.27 schwarze 101: int i;
1.123 schwarze 102: char *ep;
1.26 schwarze 103:
1.132 schwarze 104: /* In no-fill mode, whitespace is preserved on text lines. */
1.10 schwarze 105:
1.132 schwarze 106: if (man->flags & ROFF_NOFILL) {
1.108 schwarze 107: roff_word_alloc(man, line, offs, buf + offs);
1.127 schwarze 108: man_descope(man, line, offs, buf + offs);
1.113 schwarze 109: return 1;
1.10 schwarze 110: }
111:
1.92 schwarze 112: for (i = offs; buf[i] == ' '; i++)
1.10 schwarze 113: /* Skip leading whitespace. */ ;
1.18 schwarze 114:
1.71 schwarze 115: /*
1.123 schwarze 116: * Blank lines are ignored in next line scope
117: * and right after headings and cancel preceding \c,
118: * but add a single vertical space elsewhere.
1.71 schwarze 119: */
120:
1.92 schwarze 121: if (buf[i] == '\0') {
1.123 schwarze 122: if (man->flags & (MAN_ELINE | MAN_BLINE)) {
1.129 schwarze 123: mandoc_msg(MANDOCERR_BLK_BLANK, line, 0, NULL);
1.123 schwarze 124: return 1;
1.71 schwarze 125: }
1.123 schwarze 126: if (man->last->tok == MAN_SH || man->last->tok == MAN_SS)
127: return 1;
1.127 schwarze 128: if (man->last->type == ROFFT_TEXT &&
129: ((ep = man_hasc(man->last->string)) != NULL)) {
1.123 schwarze 130: *ep = '\0';
131: return 1;
132: }
133: roff_elem_alloc(man, line, offs, ROFF_sp);
134: man->next = ROFF_NEXT_SIBLING;
1.113 schwarze 135: return 1;
1.10 schwarze 136: }
1.1 kristaps 137:
1.79 schwarze 138: /*
1.27 schwarze 139: * Warn if the last un-escaped character is whitespace. Then
1.79 schwarze 140: * strip away the remaining spaces (tabs stay!).
1.27 schwarze 141: */
1.18 schwarze 142:
1.27 schwarze 143: i = (int)strlen(buf);
144: assert(i);
1.18 schwarze 145:
1.27 schwarze 146: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
147: if (i > 1 && '\\' != buf[i - 2])
1.129 schwarze 148: mandoc_msg(MANDOCERR_SPACE_EOL, line, i - 1, NULL);
1.18 schwarze 149:
1.27 schwarze 150: for (--i; i && ' ' == buf[i]; i--)
151: /* Spin back to non-space. */ ;
1.10 schwarze 152:
1.27 schwarze 153: /* Jump ahead of escaped whitespace. */
154: i += '\\' == buf[i] ? 2 : 1;
1.18 schwarze 155:
1.27 schwarze 156: buf[i] = '\0';
1.10 schwarze 157: }
1.108 schwarze 158: roff_word_alloc(man, line, offs, buf + offs);
1.28 schwarze 159:
160: /*
161: * End-of-sentence check. If the last character is an unescaped
162: * EOS character, then flag the node as being the end of a
163: * sentence. The front-end will know how to interpret this.
164: */
165:
166: assert(i);
1.72 schwarze 167: if (mandoc_eos(buf, (size_t)i))
1.115 schwarze 168: man->last->flags |= NODE_EOS;
1.10 schwarze 169:
1.127 schwarze 170: man_descope(man, line, offs, buf + offs);
1.113 schwarze 171: return 1;
1.1 kristaps 172: }
173:
1.53 schwarze 174: static int
1.101 schwarze 175: man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 176: {
1.99 schwarze 177: struct roff_node *n;
1.91 schwarze 178: const char *cp;
1.118 schwarze 179: size_t sz;
180: enum roff_tok tok;
181: int ppos;
1.84 schwarze 182: int bline;
1.118 schwarze 183:
184: /* Determine the line macro. */
1.1 kristaps 185:
1.59 schwarze 186: ppos = offs;
1.118 schwarze 187: tok = TOKEN_NONE;
188: for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
189: offs++;
190: if (sz > 0 && sz < 4)
191: tok = roffhash_find(man->manmac, buf + ppos, sz);
1.107 schwarze 192: if (tok == TOKEN_NONE) {
1.129 schwarze 193: mandoc_msg(MANDOCERR_MACRO, ln, ppos, "%s", buf + ppos - 1);
1.113 schwarze 194: return 1;
1.1 kristaps 195: }
196:
1.91 schwarze 197: /* Skip a leading escape sequence or tab. */
198:
199: switch (buf[offs]) {
200: case '\\':
201: cp = buf + offs + 1;
202: mandoc_escape(&cp, NULL, NULL);
203: offs = cp - buf;
204: break;
205: case '\t':
206: offs++;
207: break;
208: default:
209: break;
210: }
211:
212: /* Jump to the next non-whitespace word. */
1.1 kristaps 213:
1.116 schwarze 214: while (buf[offs] == ' ')
1.59 schwarze 215: offs++;
1.18 schwarze 216:
1.79 schwarze 217: /*
1.27 schwarze 218: * Trailing whitespace. Note that tabs are allowed to be passed
219: * into the parser as "text", so we only warn about spaces here.
220: */
1.18 schwarze 221:
1.92 schwarze 222: if (buf[offs] == '\0' && buf[offs - 1] == ' ')
1.129 schwarze 223: mandoc_msg(MANDOCERR_SPACE_EOL, ln, offs - 1, NULL);
1.1 kristaps 224:
1.79 schwarze 225: /*
1.97 schwarze 226: * Some macros break next-line scopes; otherwise, remember
227: * whether we are in next-line scope for a block head.
1.21 schwarze 228: */
229:
1.97 schwarze 230: man_breakscope(man, tok);
1.84 schwarze 231: bline = man->flags & MAN_BLINE;
1.119 schwarze 232:
233: /*
234: * If the line in next-line scope ends with \c, keep the
235: * next-line scope open for the subsequent input line.
236: * That is not at all portable, only groff >= 1.22.4
237: * does it, but *if* this weird idiom occurs in a manual
238: * page, that's very likely what the author intended.
239: */
240:
1.128 schwarze 241: if (bline && man_hasc(buf + offs))
242: bline = 0;
1.24 schwarze 243:
244: /* Call to handler... */
1.1 kristaps 245:
1.125 schwarze 246: (*man_macro(tok)->fp)(man, tok, ln, ppos, &offs, buf);
1.73 schwarze 247:
248: /* In quick mode (for mandocdb), abort after the NAME section. */
249:
1.92 schwarze 250: if (man->quick && tok == MAN_SH) {
1.80 schwarze 251: n = man->last;
1.98 schwarze 252: if (n->type == ROFFT_BODY &&
1.80 schwarze 253: strcmp(n->prev->child->string, "NAME"))
1.113 schwarze 254: return 2;
1.80 schwarze 255: }
1.1 kristaps 256:
1.79 schwarze 257: /*
1.84 schwarze 258: * If we are in a next-line scope for a block head,
259: * close it out now and switch to the body,
260: * unless the next-line scope is allowed to continue.
1.21 schwarze 261: */
262:
1.128 schwarze 263: if (bline == 0 ||
264: (man->flags & MAN_BLINE) == 0 ||
265: man->flags & MAN_ELINE ||
1.125 schwarze 266: man_macro(tok)->flags & MAN_NSCOPED)
1.113 schwarze 267: return 1;
1.1 kristaps 268:
1.69 schwarze 269: man->flags &= ~MAN_BLINE;
1.92 schwarze 270: man_unscope(man, man->last->parent);
1.106 schwarze 271: roff_body_alloc(man, ln, ppos, man->last->tok);
1.113 schwarze 272: return 1;
1.97 schwarze 273: }
274:
275: void
1.101 schwarze 276: man_breakscope(struct roff_man *man, int tok)
1.97 schwarze 277: {
1.99 schwarze 278: struct roff_node *n;
1.97 schwarze 279:
280: /*
281: * An element next line scope is open,
282: * and the new macro is not allowed inside elements.
283: * Delete the element that is being broken.
284: */
285:
1.121 schwarze 286: if (man->flags & MAN_ELINE && (tok < MAN_TH ||
1.125 schwarze 287: (man_macro(tok)->flags & MAN_NSCOPED) == 0)) {
1.97 schwarze 288: n = man->last;
1.124 schwarze 289: if (n->type == ROFFT_TEXT)
290: n = n->parent;
291: if (n->tok < MAN_TH ||
1.128 schwarze 292: (man_macro(n->tok)->flags & (MAN_NSCOPED | MAN_ESCOPED))
293: == MAN_NSCOPED)
1.97 schwarze 294: n = n->parent;
295:
1.129 schwarze 296: mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
297: "%s breaks %s", roff_name[tok], roff_name[n->tok]);
1.97 schwarze 298:
1.106 schwarze 299: roff_node_delete(man, n);
1.97 schwarze 300: man->flags &= ~MAN_ELINE;
1.112 schwarze 301: }
302:
303: /*
304: * Weird special case:
305: * Switching fill mode closes section headers.
306: */
307:
308: if (man->flags & MAN_BLINE &&
1.132 schwarze 309: (tok == ROFF_nf || tok == ROFF_fi) &&
1.112 schwarze 310: (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
311: n = man->last;
312: man_unscope(man, n);
313: roff_body_alloc(man, n->line, n->pos, n->tok);
314: man->flags &= ~MAN_BLINE;
1.97 schwarze 315: }
316:
317: /*
318: * A block header next line scope is open,
319: * and the new macro is not allowed inside block headers.
320: * Delete the block that is being broken.
321: */
322:
1.132 schwarze 323: if (man->flags & MAN_BLINE && tok != ROFF_nf && tok != ROFF_fi &&
324: (tok < MAN_TH || man_macro(tok)->flags & MAN_XSCOPE)) {
1.97 schwarze 325: n = man->last;
1.98 schwarze 326: if (n->type == ROFFT_TEXT)
1.97 schwarze 327: n = n->parent;
1.124 schwarze 328: if (n->tok < MAN_TH ||
1.128 schwarze 329: (man_macro(n->tok)->flags & MAN_XSCOPE) == 0)
1.97 schwarze 330: n = n->parent;
331:
1.98 schwarze 332: assert(n->type == ROFFT_HEAD);
1.97 schwarze 333: n = n->parent;
1.98 schwarze 334: assert(n->type == ROFFT_BLOCK);
1.128 schwarze 335: assert(man_macro(n->tok)->flags & MAN_BSCOPED);
1.97 schwarze 336:
1.129 schwarze 337: mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
338: "%s breaks %s", roff_name[tok], roff_name[n->tok]);
1.97 schwarze 339:
1.106 schwarze 340: roff_node_delete(man, n);
1.97 schwarze 341: man->flags &= ~MAN_BLINE;
342: }
1.4 schwarze 343: }