Annotation of mandoc/man.c, Revision 1.189
1.189 ! schwarze 1: /* $Id: man.c,v 1.188 2022/04/28 10:26:37 schwarze Exp $ */
1.1 kristaps 2: /*
1.188 schwarze 3: * Copyright (c) 2013-2015,2017-2019,2022 Ingo Schwarze <schwarze@openbsd.org>
1.102 schwarze 4: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
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:
1.188 schwarze 73: /*
74: * Rewind all open next-line scopes.
75: */
1.180 schwarze 76: void
1.179 schwarze 77: man_descope(struct roff_man *man, int line, int offs, char *start)
1.94 kristaps 78: {
1.189 ! schwarze 79: /* First close out all next-line element scopes, if any. */
1.94 kristaps 80:
1.144 schwarze 81: if (man->flags & MAN_ELINE) {
1.180 schwarze 82: while (man->last->parent->type != ROFFT_ROOT &&
83: man_macro(man->last->parent->tok)->flags & MAN_ESCOPED)
84: man_unscope(man, man->last->parent);
1.119 schwarze 85: man->flags &= ~MAN_ELINE;
1.94 kristaps 86: }
1.189 ! schwarze 87:
! 88: /* Trailing \c keeps next-line block scope open. */
! 89:
! 90: if (start != NULL && man_hasc(start) != NULL)
! 91: return;
! 92:
! 93: /* Close out the next-line block scope, if there is one. */
! 94:
1.144 schwarze 95: if ( ! (man->flags & MAN_BLINE))
96: return;
97: man_unscope(man, man->last->parent);
1.158 schwarze 98: roff_body_alloc(man, line, offs, man->last->tok);
1.187 schwarze 99: man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
1.94 kristaps 100: }
101:
1.1 kristaps 102: static int
1.153 schwarze 103: man_ptext(struct roff_man *man, int line, char *buf, int offs)
1.1 kristaps 104: {
1.61 kristaps 105: int i;
1.175 schwarze 106: char *ep;
1.60 kristaps 107:
1.184 schwarze 108: /* In no-fill mode, whitespace is preserved on text lines. */
1.35 kristaps 109:
1.184 schwarze 110: if (man->flags & ROFF_NOFILL) {
1.160 schwarze 111: roff_word_alloc(man, line, offs, buf + offs);
1.179 schwarze 112: man_descope(man, line, offs, buf + offs);
1.165 schwarze 113: return 1;
1.35 kristaps 114: }
115:
1.144 schwarze 116: for (i = offs; buf[i] == ' '; i++)
1.31 kristaps 117: /* Skip leading whitespace. */ ;
1.48 kristaps 118:
1.121 schwarze 119: /*
1.175 schwarze 120: * Blank lines are ignored in next line scope
121: * and right after headings and cancel preceding \c,
122: * but add a single vertical space elsewhere.
1.121 schwarze 123: */
124:
1.144 schwarze 125: if (buf[i] == '\0') {
1.175 schwarze 126: if (man->flags & (MAN_ELINE | MAN_BLINE)) {
1.181 schwarze 127: mandoc_msg(MANDOCERR_BLK_BLANK, line, 0, NULL);
1.175 schwarze 128: return 1;
1.121 schwarze 129: }
1.175 schwarze 130: if (man->last->tok == MAN_SH || man->last->tok == MAN_SS)
131: return 1;
1.179 schwarze 132: if (man->last->type == ROFFT_TEXT &&
133: ((ep = man_hasc(man->last->string)) != NULL)) {
1.175 schwarze 134: *ep = '\0';
135: return 1;
136: }
137: roff_elem_alloc(man, line, offs, ROFF_sp);
138: man->next = ROFF_NEXT_SIBLING;
1.165 schwarze 139: return 1;
1.31 kristaps 140: }
141:
1.129 schwarze 142: /*
1.63 kristaps 143: * Warn if the last un-escaped character is whitespace. Then
1.129 schwarze 144: * strip away the remaining spaces (tabs stay!).
1.63 kristaps 145: */
1.29 kristaps 146:
1.61 kristaps 147: i = (int)strlen(buf);
148: assert(i);
1.49 kristaps 149:
1.63 kristaps 150: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
1.64 kristaps 151: if (i > 1 && '\\' != buf[i - 2])
1.181 schwarze 152: mandoc_msg(MANDOCERR_SPACE_EOL, 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) {
1.181 schwarze 197: mandoc_msg(MANDOCERR_MACRO, ln, ppos, "%s", buf + ppos - 1);
1.165 schwarze 198: return 1;
1.1 kristaps 199: }
200:
1.143 schwarze 201: /* Skip a leading escape sequence or tab. */
202:
203: switch (buf[offs]) {
204: case '\\':
205: cp = buf + offs + 1;
206: mandoc_escape(&cp, NULL, NULL);
207: offs = cp - buf;
208: break;
209: case '\t':
210: offs++;
211: break;
212: default:
213: break;
214: }
215:
216: /* Jump to the next non-whitespace word. */
1.1 kristaps 217:
1.168 schwarze 218: while (buf[offs] == ' ')
1.107 kristaps 219: offs++;
1.1 kristaps 220:
1.129 schwarze 221: /*
1.62 kristaps 222: * Trailing whitespace. Note that tabs are allowed to be passed
223: * into the parser as "text", so we only warn about spaces here.
224: */
1.48 kristaps 225:
1.144 schwarze 226: if (buf[offs] == '\0' && buf[offs - 1] == ' ')
1.181 schwarze 227: mandoc_msg(MANDOCERR_SPACE_EOL, ln, offs - 1, NULL);
1.48 kristaps 228:
1.129 schwarze 229: /*
1.149 schwarze 230: * Some macros break next-line scopes; otherwise, remember
231: * whether we are in next-line scope for a block head.
1.50 kristaps 232: */
1.34 kristaps 233:
1.149 schwarze 234: man_breakscope(man, tok);
1.134 schwarze 235: bline = man->flags & MAN_BLINE;
1.171 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:
1.180 schwarze 245: if (bline && man_hasc(buf + offs))
246: bline = 0;
1.59 kristaps 247:
248: /* Call to handler... */
1.1 kristaps 249:
1.177 schwarze 250: (*man_macro(tok)->fp)(man, tok, ln, ppos, &offs, buf);
1.123 schwarze 251:
252: /* In quick mode (for mandocdb), abort after the NAME section. */
253:
1.144 schwarze 254: if (man->quick && tok == MAN_SH) {
1.130 schwarze 255: n = man->last;
1.150 schwarze 256: if (n->type == ROFFT_BODY &&
1.130 schwarze 257: strcmp(n->prev->child->string, "NAME"))
1.165 schwarze 258: return 2;
1.130 schwarze 259: }
1.1 kristaps 260:
1.129 schwarze 261: /*
1.135 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.
1.29 kristaps 265: */
266:
1.180 schwarze 267: if (bline == 0 ||
268: (man->flags & MAN_BLINE) == 0 ||
269: man->flags & MAN_ELINE ||
1.177 schwarze 270: man_macro(tok)->flags & MAN_NSCOPED)
1.165 schwarze 271: return 1;
1.29 kristaps 272:
1.144 schwarze 273: man_unscope(man, man->last->parent);
1.158 schwarze 274: roff_body_alloc(man, ln, ppos, man->last->tok);
1.187 schwarze 275: man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
1.165 schwarze 276: return 1;
1.149 schwarze 277: }
278:
1.188 schwarze 279: /*
280: * Rewind open next-line scopes
281: * unless the tok request or macro is allowed inside them.
282: */
1.149 schwarze 283: void
1.153 schwarze 284: man_breakscope(struct roff_man *man, int tok)
1.149 schwarze 285: {
1.151 schwarze 286: struct roff_node *n;
1.149 schwarze 287:
288: /*
289: * An element next line scope is open,
290: * and the new macro is not allowed inside elements.
291: * Delete the element that is being broken.
292: */
293:
1.173 schwarze 294: if (man->flags & MAN_ELINE && (tok < MAN_TH ||
1.177 schwarze 295: (man_macro(tok)->flags & MAN_NSCOPED) == 0)) {
1.149 schwarze 296: n = man->last;
1.176 schwarze 297: if (n->type == ROFFT_TEXT)
298: n = n->parent;
299: if (n->tok < MAN_TH ||
1.180 schwarze 300: (man_macro(n->tok)->flags & (MAN_NSCOPED | MAN_ESCOPED))
301: == MAN_NSCOPED)
1.149 schwarze 302: n = n->parent;
1.188 schwarze 303: for (;;) {
304: mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
305: "%s breaks %s", roff_name[tok], roff_name[n->tok]);
306: if (n->parent->type != ROFFT_ELEM ||
307: (man_macro(n->parent->tok)->flags &
308: MAN_ESCOPED) == 0)
309: break;
310: n = n->parent;
311: }
1.158 schwarze 312: roff_node_delete(man, n);
1.149 schwarze 313: man->flags &= ~MAN_ELINE;
1.164 schwarze 314: }
315:
316: /*
317: * Weird special case:
318: * Switching fill mode closes section headers.
319: */
320:
321: if (man->flags & MAN_BLINE &&
1.184 schwarze 322: (tok == ROFF_nf || tok == ROFF_fi) &&
1.164 schwarze 323: (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
324: n = man->last;
325: man_unscope(man, n);
326: roff_body_alloc(man, n->line, n->pos, n->tok);
1.187 schwarze 327: man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
1.149 schwarze 328: }
329:
330: /*
331: * A block header next line scope is open,
332: * and the new macro is not allowed inside block headers.
333: * Delete the block that is being broken.
334: */
335:
1.184 schwarze 336: if (man->flags & MAN_BLINE && tok != ROFF_nf && tok != ROFF_fi &&
337: (tok < MAN_TH || man_macro(tok)->flags & MAN_XSCOPE)) {
1.149 schwarze 338: n = man->last;
1.150 schwarze 339: if (n->type == ROFFT_TEXT)
1.149 schwarze 340: n = n->parent;
1.176 schwarze 341: if (n->tok < MAN_TH ||
1.180 schwarze 342: (man_macro(n->tok)->flags & MAN_XSCOPE) == 0)
1.149 schwarze 343: n = n->parent;
344:
1.150 schwarze 345: assert(n->type == ROFFT_HEAD);
1.149 schwarze 346: n = n->parent;
1.150 schwarze 347: assert(n->type == ROFFT_BLOCK);
1.180 schwarze 348: assert(man_macro(n->tok)->flags & MAN_BSCOPED);
1.149 schwarze 349:
1.181 schwarze 350: mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
351: "%s breaks %s", roff_name[tok], roff_name[n->tok]);
1.149 schwarze 352:
1.158 schwarze 353: roff_node_delete(man, n);
1.187 schwarze 354: man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
1.149 schwarze 355: }
1.23 kristaps 356: }
CVSweb