Annotation of src/usr.bin/mandoc/man.c, Revision 1.128
1.128 ! schwarze 1: /* $OpenBSD: man.c,v 1.127 2018/08/25 12:28:52 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:
1.10 schwarze 104: /* Literal free-form text whitespace is preserved. */
105:
1.92 schwarze 106: if (man->flags & MAN_LITERAL) {
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)) {
1.122 schwarze 123: mandoc_msg(MANDOCERR_BLK_BLANK, man->parse,
124: line, 0, NULL);
1.123 schwarze 125: return 1;
1.71 schwarze 126: }
1.123 schwarze 127: if (man->last->tok == MAN_SH || man->last->tok == MAN_SS)
128: return 1;
1.127 schwarze 129: if (man->last->type == ROFFT_TEXT &&
130: ((ep = man_hasc(man->last->string)) != NULL)) {
1.123 schwarze 131: *ep = '\0';
132: return 1;
133: }
134: roff_elem_alloc(man, line, offs, ROFF_sp);
135: man->next = ROFF_NEXT_SIBLING;
1.113 schwarze 136: return 1;
1.10 schwarze 137: }
1.1 kristaps 138:
1.79 schwarze 139: /*
1.27 schwarze 140: * Warn if the last un-escaped character is whitespace. Then
1.79 schwarze 141: * strip away the remaining spaces (tabs stay!).
1.27 schwarze 142: */
1.18 schwarze 143:
1.27 schwarze 144: i = (int)strlen(buf);
145: assert(i);
1.18 schwarze 146:
1.27 schwarze 147: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
148: if (i > 1 && '\\' != buf[i - 2])
1.81 schwarze 149: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
150: line, i - 1, NULL);
1.18 schwarze 151:
1.27 schwarze 152: for (--i; i && ' ' == buf[i]; i--)
153: /* Spin back to non-space. */ ;
1.10 schwarze 154:
1.27 schwarze 155: /* Jump ahead of escaped whitespace. */
156: i += '\\' == buf[i] ? 2 : 1;
1.18 schwarze 157:
1.27 schwarze 158: buf[i] = '\0';
1.10 schwarze 159: }
1.108 schwarze 160: roff_word_alloc(man, line, offs, buf + offs);
1.28 schwarze 161:
162: /*
163: * End-of-sentence check. If the last character is an unescaped
164: * EOS character, then flag the node as being the end of a
165: * sentence. The front-end will know how to interpret this.
166: */
167:
168: assert(i);
1.72 schwarze 169: if (mandoc_eos(buf, (size_t)i))
1.115 schwarze 170: man->last->flags |= NODE_EOS;
1.10 schwarze 171:
1.127 schwarze 172: man_descope(man, line, offs, buf + offs);
1.113 schwarze 173: return 1;
1.1 kristaps 174: }
175:
1.53 schwarze 176: static int
1.101 schwarze 177: man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 178: {
1.99 schwarze 179: struct roff_node *n;
1.91 schwarze 180: const char *cp;
1.118 schwarze 181: size_t sz;
182: enum roff_tok tok;
183: int ppos;
1.84 schwarze 184: int bline;
1.118 schwarze 185:
186: /* Determine the line macro. */
1.1 kristaps 187:
1.59 schwarze 188: ppos = offs;
1.118 schwarze 189: tok = TOKEN_NONE;
190: for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
191: offs++;
192: if (sz > 0 && sz < 4)
193: tok = roffhash_find(man->manmac, buf + ppos, sz);
1.107 schwarze 194: if (tok == TOKEN_NONE) {
1.85 schwarze 195: mandoc_msg(MANDOCERR_MACRO, man->parse,
196: ln, ppos, buf + ppos - 1);
1.113 schwarze 197: return 1;
1.1 kristaps 198: }
199:
1.91 schwarze 200: /* Skip a leading escape sequence or tab. */
201:
202: switch (buf[offs]) {
203: case '\\':
204: cp = buf + offs + 1;
205: mandoc_escape(&cp, NULL, NULL);
206: offs = cp - buf;
207: break;
208: case '\t':
209: offs++;
210: break;
211: default:
212: break;
213: }
214:
215: /* Jump to the next non-whitespace word. */
1.1 kristaps 216:
1.116 schwarze 217: while (buf[offs] == ' ')
1.59 schwarze 218: offs++;
1.18 schwarze 219:
1.79 schwarze 220: /*
1.27 schwarze 221: * Trailing whitespace. Note that tabs are allowed to be passed
222: * into the parser as "text", so we only warn about spaces here.
223: */
1.18 schwarze 224:
1.92 schwarze 225: if (buf[offs] == '\0' && buf[offs - 1] == ' ')
1.81 schwarze 226: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
227: ln, offs - 1, NULL);
1.1 kristaps 228:
1.79 schwarze 229: /*
1.97 schwarze 230: * Some macros break next-line scopes; otherwise, remember
231: * whether we are in next-line scope for a block head.
1.21 schwarze 232: */
233:
1.97 schwarze 234: man_breakscope(man, tok);
1.84 schwarze 235: bline = man->flags & MAN_BLINE;
1.119 schwarze 236:
237: /*
238: * If the line in next-line scope ends with \c, keep the
239: * next-line scope open for the subsequent input line.
240: * That is not at all portable, only groff >= 1.22.4
241: * does it, but *if* this weird idiom occurs in a manual
242: * page, that's very likely what the author intended.
243: */
244:
1.128 ! schwarze 245: if (bline && man_hasc(buf + offs))
! 246: bline = 0;
1.24 schwarze 247:
248: /* Call to handler... */
1.1 kristaps 249:
1.125 schwarze 250: (*man_macro(tok)->fp)(man, tok, ln, ppos, &offs, buf);
1.73 schwarze 251:
252: /* In quick mode (for mandocdb), abort after the NAME section. */
253:
1.92 schwarze 254: if (man->quick && tok == MAN_SH) {
1.80 schwarze 255: n = man->last;
1.98 schwarze 256: if (n->type == ROFFT_BODY &&
1.80 schwarze 257: strcmp(n->prev->child->string, "NAME"))
1.113 schwarze 258: return 2;
1.80 schwarze 259: }
1.1 kristaps 260:
1.79 schwarze 261: /*
1.84 schwarze 262: * If we are in a next-line scope for a block head,
263: * close it out now and switch to the body,
264: * unless the next-line scope is allowed to continue.
1.21 schwarze 265: */
266:
1.128 ! schwarze 267: if (bline == 0 ||
! 268: (man->flags & MAN_BLINE) == 0 ||
! 269: man->flags & MAN_ELINE ||
1.125 schwarze 270: man_macro(tok)->flags & MAN_NSCOPED)
1.113 schwarze 271: return 1;
1.1 kristaps 272:
1.69 schwarze 273: man->flags &= ~MAN_BLINE;
1.92 schwarze 274: man_unscope(man, man->last->parent);
1.106 schwarze 275: roff_body_alloc(man, ln, ppos, man->last->tok);
1.113 schwarze 276: return 1;
1.97 schwarze 277: }
278:
279: void
1.101 schwarze 280: man_breakscope(struct roff_man *man, int tok)
1.97 schwarze 281: {
1.99 schwarze 282: struct roff_node *n;
1.97 schwarze 283:
284: /*
285: * An element next line scope is open,
286: * and the new macro is not allowed inside elements.
287: * Delete the element that is being broken.
288: */
289:
1.121 schwarze 290: if (man->flags & MAN_ELINE && (tok < MAN_TH ||
1.125 schwarze 291: (man_macro(tok)->flags & MAN_NSCOPED) == 0)) {
1.97 schwarze 292: n = man->last;
1.124 schwarze 293: if (n->type == ROFFT_TEXT)
294: n = n->parent;
295: if (n->tok < MAN_TH ||
1.128 ! schwarze 296: (man_macro(n->tok)->flags & (MAN_NSCOPED | MAN_ESCOPED))
! 297: == MAN_NSCOPED)
1.97 schwarze 298: n = n->parent;
299:
300: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
301: n->line, n->pos, "%s breaks %s",
1.121 schwarze 302: roff_name[tok], roff_name[n->tok]);
1.97 schwarze 303:
1.106 schwarze 304: roff_node_delete(man, n);
1.97 schwarze 305: man->flags &= ~MAN_ELINE;
1.112 schwarze 306: }
307:
308: /*
309: * Weird special case:
310: * Switching fill mode closes section headers.
311: */
312:
313: if (man->flags & MAN_BLINE &&
314: (tok == MAN_nf || tok == MAN_fi) &&
315: (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
316: n = man->last;
317: man_unscope(man, n);
318: roff_body_alloc(man, n->line, n->pos, n->tok);
319: man->flags &= ~MAN_BLINE;
1.97 schwarze 320: }
321:
322: /*
323: * A block header next line scope is open,
324: * and the new macro is not allowed inside block headers.
325: * Delete the block that is being broken.
326: */
327:
1.121 schwarze 328: if (man->flags & MAN_BLINE && (tok < MAN_TH ||
1.128 ! schwarze 329: man_macro(tok)->flags & MAN_XSCOPE)) {
1.97 schwarze 330: n = man->last;
1.98 schwarze 331: if (n->type == ROFFT_TEXT)
1.97 schwarze 332: n = n->parent;
1.124 schwarze 333: if (n->tok < MAN_TH ||
1.128 ! schwarze 334: (man_macro(n->tok)->flags & MAN_XSCOPE) == 0)
1.97 schwarze 335: n = n->parent;
336:
1.98 schwarze 337: assert(n->type == ROFFT_HEAD);
1.97 schwarze 338: n = n->parent;
1.98 schwarze 339: assert(n->type == ROFFT_BLOCK);
1.128 ! schwarze 340: assert(man_macro(n->tok)->flags & MAN_BSCOPED);
1.97 schwarze 341:
342: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
343: n->line, n->pos, "%s breaks %s",
1.121 schwarze 344: roff_name[tok], roff_name[n->tok]);
1.97 schwarze 345:
1.106 schwarze 346: roff_node_delete(man, n);
1.97 schwarze 347: man->flags &= ~MAN_BLINE;
348: }
1.114 schwarze 349: }
350:
351: void
352: man_state(struct roff_man *man, struct roff_node *n)
353: {
354:
355: switch(n->tok) {
356: case MAN_nf:
357: case MAN_EX:
1.115 schwarze 358: if (man->flags & MAN_LITERAL && ! (n->flags & NODE_VALID))
1.114 schwarze 359: mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
360: n->line, n->pos, "nf");
361: man->flags |= MAN_LITERAL;
362: break;
363: case MAN_fi:
364: case MAN_EE:
365: if ( ! (man->flags & MAN_LITERAL) &&
1.115 schwarze 366: ! (n->flags & NODE_VALID))
1.114 schwarze 367: mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
368: n->line, n->pos, "fi");
369: man->flags &= ~MAN_LITERAL;
370: break;
371: default:
372: break;
373: }
1.115 schwarze 374: man->last->flags |= NODE_VALID;
1.114 schwarze 375: }
376:
377: void
378: man_validate(struct roff_man *man)
379: {
380:
381: man->last = man->first;
382: man_node_validate(man);
383: man->flags &= ~MAN_LITERAL;
1.4 schwarze 384: }