Annotation of src/usr.bin/mandoc/man.c, Revision 1.137
1.137 ! schwarze 1: /* $OpenBSD: man.c,v 1.136 2022/04/28 10:17:37 schwarze Exp $ */
1.1 kristaps 2: /*
1.136 schwarze 3: * Copyright (c) 2013-2015,2017-2019,2022 Ingo Schwarze <schwarze@openbsd.org>
1.52 schwarze 4: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
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:
1.136 schwarze 71: /*
72: * Rewind all open next-line scopes.
73: */
1.128 schwarze 74: void
1.127 schwarze 75: man_descope(struct roff_man *man, int line, int offs, char *start)
1.52 schwarze 76: {
1.137 ! schwarze 77: /* First close out all next-line element scopes, if any. */
1.52 schwarze 78:
1.92 schwarze 79: if (man->flags & MAN_ELINE) {
1.128 schwarze 80: while (man->last->parent->type != ROFFT_ROOT &&
81: man_macro(man->last->parent->tok)->flags & MAN_ESCOPED)
82: man_unscope(man, man->last->parent);
1.69 schwarze 83: man->flags &= ~MAN_ELINE;
1.52 schwarze 84: }
1.137 ! schwarze 85:
! 86: /* Trailing \c keeps next-line block scope open. */
! 87:
! 88: if (start != NULL && man_hasc(start) != NULL)
! 89: return;
! 90:
! 91: /* Close out the next-line block scope, if there is one. */
! 92:
1.92 schwarze 93: if ( ! (man->flags & MAN_BLINE))
94: return;
95: man_unscope(man, man->last->parent);
1.106 schwarze 96: roff_body_alloc(man, line, offs, man->last->tok);
1.135 schwarze 97: man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
1.52 schwarze 98: }
99:
1.1 kristaps 100: static int
1.101 schwarze 101: man_ptext(struct roff_man *man, int line, char *buf, int offs)
1.1 kristaps 102: {
1.27 schwarze 103: int i;
1.123 schwarze 104: char *ep;
1.26 schwarze 105:
1.132 schwarze 106: /* In no-fill mode, whitespace is preserved on text lines. */
1.10 schwarze 107:
1.132 schwarze 108: if (man->flags & ROFF_NOFILL) {
1.108 schwarze 109: roff_word_alloc(man, line, offs, buf + offs);
1.127 schwarze 110: man_descope(man, line, offs, buf + offs);
1.113 schwarze 111: return 1;
1.10 schwarze 112: }
113:
1.92 schwarze 114: for (i = offs; buf[i] == ' '; i++)
1.10 schwarze 115: /* Skip leading whitespace. */ ;
1.18 schwarze 116:
1.71 schwarze 117: /*
1.123 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.
1.71 schwarze 121: */
122:
1.92 schwarze 123: if (buf[i] == '\0') {
1.123 schwarze 124: if (man->flags & (MAN_ELINE | MAN_BLINE)) {
1.129 schwarze 125: mandoc_msg(MANDOCERR_BLK_BLANK, line, 0, NULL);
1.123 schwarze 126: return 1;
1.71 schwarze 127: }
1.123 schwarze 128: if (man->last->tok == MAN_SH || man->last->tok == MAN_SS)
129: return 1;
1.127 schwarze 130: if (man->last->type == ROFFT_TEXT &&
131: ((ep = man_hasc(man->last->string)) != NULL)) {
1.123 schwarze 132: *ep = '\0';
133: return 1;
134: }
135: roff_elem_alloc(man, line, offs, ROFF_sp);
136: man->next = ROFF_NEXT_SIBLING;
1.113 schwarze 137: return 1;
1.10 schwarze 138: }
1.1 kristaps 139:
1.79 schwarze 140: /*
1.27 schwarze 141: * Warn if the last un-escaped character is whitespace. Then
1.79 schwarze 142: * strip away the remaining spaces (tabs stay!).
1.27 schwarze 143: */
1.18 schwarze 144:
1.27 schwarze 145: i = (int)strlen(buf);
146: assert(i);
1.18 schwarze 147:
1.27 schwarze 148: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
149: if (i > 1 && '\\' != buf[i - 2])
1.129 schwarze 150: mandoc_msg(MANDOCERR_SPACE_EOL, 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) {
1.129 schwarze 195: mandoc_msg(MANDOCERR_MACRO, ln, ppos, "%s", buf + ppos - 1);
1.113 schwarze 196: return 1;
1.1 kristaps 197: }
198:
1.91 schwarze 199: /* Skip a leading escape sequence or tab. */
200:
201: switch (buf[offs]) {
202: case '\\':
203: cp = buf + offs + 1;
204: mandoc_escape(&cp, NULL, NULL);
205: offs = cp - buf;
206: break;
207: case '\t':
208: offs++;
209: break;
210: default:
211: break;
212: }
213:
214: /* Jump to the next non-whitespace word. */
1.1 kristaps 215:
1.116 schwarze 216: while (buf[offs] == ' ')
1.59 schwarze 217: offs++;
1.18 schwarze 218:
1.79 schwarze 219: /*
1.27 schwarze 220: * Trailing whitespace. Note that tabs are allowed to be passed
221: * into the parser as "text", so we only warn about spaces here.
222: */
1.18 schwarze 223:
1.92 schwarze 224: if (buf[offs] == '\0' && buf[offs - 1] == ' ')
1.129 schwarze 225: mandoc_msg(MANDOCERR_SPACE_EOL, ln, offs - 1, NULL);
1.1 kristaps 226:
1.79 schwarze 227: /*
1.97 schwarze 228: * Some macros break next-line scopes; otherwise, remember
229: * whether we are in next-line scope for a block head.
1.21 schwarze 230: */
231:
1.97 schwarze 232: man_breakscope(man, tok);
1.84 schwarze 233: bline = man->flags & MAN_BLINE;
1.119 schwarze 234:
235: /*
236: * If the line in next-line scope ends with \c, keep the
237: * next-line scope open for the subsequent input line.
238: * That is not at all portable, only groff >= 1.22.4
239: * does it, but *if* this weird idiom occurs in a manual
240: * page, that's very likely what the author intended.
241: */
242:
1.128 schwarze 243: if (bline && man_hasc(buf + offs))
244: bline = 0;
1.24 schwarze 245:
246: /* Call to handler... */
1.1 kristaps 247:
1.125 schwarze 248: (*man_macro(tok)->fp)(man, tok, ln, ppos, &offs, buf);
1.73 schwarze 249:
250: /* In quick mode (for mandocdb), abort after the NAME section. */
251:
1.92 schwarze 252: if (man->quick && tok == MAN_SH) {
1.80 schwarze 253: n = man->last;
1.98 schwarze 254: if (n->type == ROFFT_BODY &&
1.80 schwarze 255: strcmp(n->prev->child->string, "NAME"))
1.113 schwarze 256: return 2;
1.80 schwarze 257: }
1.1 kristaps 258:
1.79 schwarze 259: /*
1.84 schwarze 260: * If we are in a next-line scope for a block head,
261: * close it out now and switch to the body,
262: * unless the next-line scope is allowed to continue.
1.21 schwarze 263: */
264:
1.128 schwarze 265: if (bline == 0 ||
266: (man->flags & MAN_BLINE) == 0 ||
267: man->flags & MAN_ELINE ||
1.125 schwarze 268: man_macro(tok)->flags & MAN_NSCOPED)
1.113 schwarze 269: return 1;
1.1 kristaps 270:
1.92 schwarze 271: man_unscope(man, man->last->parent);
1.106 schwarze 272: roff_body_alloc(man, ln, ppos, man->last->tok);
1.135 schwarze 273: man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
1.113 schwarze 274: return 1;
1.97 schwarze 275: }
276:
1.136 schwarze 277: /*
278: * Rewind open next-line scopes
279: * unless the tok request or macro is allowed inside them.
280: */
1.97 schwarze 281: void
1.101 schwarze 282: man_breakscope(struct roff_man *man, int tok)
1.97 schwarze 283: {
1.99 schwarze 284: struct roff_node *n;
1.97 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:
1.121 schwarze 292: if (man->flags & MAN_ELINE && (tok < MAN_TH ||
1.125 schwarze 293: (man_macro(tok)->flags & MAN_NSCOPED) == 0)) {
1.97 schwarze 294: n = man->last;
1.124 schwarze 295: if (n->type == ROFFT_TEXT)
296: n = n->parent;
297: if (n->tok < MAN_TH ||
1.128 schwarze 298: (man_macro(n->tok)->flags & (MAN_NSCOPED | MAN_ESCOPED))
299: == MAN_NSCOPED)
1.97 schwarze 300: n = n->parent;
1.136 schwarze 301: for (;;) {
302: mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
303: "%s breaks %s", roff_name[tok], roff_name[n->tok]);
304: if (n->parent->type != ROFFT_ELEM ||
305: (man_macro(n->parent->tok)->flags &
306: MAN_ESCOPED) == 0)
307: break;
308: n = n->parent;
309: }
1.106 schwarze 310: roff_node_delete(man, n);
1.97 schwarze 311: man->flags &= ~MAN_ELINE;
1.112 schwarze 312: }
313:
314: /*
315: * Weird special case:
316: * Switching fill mode closes section headers.
317: */
318:
319: if (man->flags & MAN_BLINE &&
1.132 schwarze 320: (tok == ROFF_nf || tok == ROFF_fi) &&
1.112 schwarze 321: (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
322: n = man->last;
323: man_unscope(man, n);
324: roff_body_alloc(man, n->line, n->pos, n->tok);
1.135 schwarze 325: man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
1.97 schwarze 326: }
327:
328: /*
329: * A block header next line scope is open,
330: * and the new macro is not allowed inside block headers.
331: * Delete the block that is being broken.
332: */
333:
1.132 schwarze 334: if (man->flags & MAN_BLINE && tok != ROFF_nf && tok != ROFF_fi &&
335: (tok < MAN_TH || man_macro(tok)->flags & MAN_XSCOPE)) {
1.97 schwarze 336: n = man->last;
1.98 schwarze 337: if (n->type == ROFFT_TEXT)
1.97 schwarze 338: n = n->parent;
1.124 schwarze 339: if (n->tok < MAN_TH ||
1.128 schwarze 340: (man_macro(n->tok)->flags & MAN_XSCOPE) == 0)
1.97 schwarze 341: n = n->parent;
342:
1.98 schwarze 343: assert(n->type == ROFFT_HEAD);
1.97 schwarze 344: n = n->parent;
1.98 schwarze 345: assert(n->type == ROFFT_BLOCK);
1.128 schwarze 346: assert(man_macro(n->tok)->flags & MAN_BSCOPED);
1.97 schwarze 347:
1.129 schwarze 348: mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
349: "%s breaks %s", roff_name[tok], roff_name[n->tok]);
1.97 schwarze 350:
1.106 schwarze 351: roff_node_delete(man, n);
1.135 schwarze 352: man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
1.97 schwarze 353: }
1.4 schwarze 354: }