Annotation of src/usr.bin/mandoc/man.c, Revision 1.121
1.121 ! schwarze 1: /* $OpenBSD: man.c,v 1.120 2017/05/05 15:16:25 schwarze Exp $ */
1.1 kristaps 2: /*
1.52 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.119 schwarze 4: * Copyright (c) 2013, 2014, 2015, 2017 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:
1.101 schwarze 36: static void man_descope(struct roff_man *, int, int);
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:
1.92 schwarze 53: static void
1.101 schwarze 54: man_descope(struct roff_man *man, int line, int offs)
1.52 schwarze 55: {
56: /*
57: * Co-ordinate what happens with having a next-line scope open:
58: * first close out the element scope (if applicable), then close
59: * out the block scope (also if applicable).
60: */
61:
1.92 schwarze 62: if (man->flags & MAN_ELINE) {
1.69 schwarze 63: man->flags &= ~MAN_ELINE;
1.92 schwarze 64: man_unscope(man, man->last->parent);
1.52 schwarze 65: }
1.92 schwarze 66: if ( ! (man->flags & MAN_BLINE))
67: return;
1.69 schwarze 68: man->flags &= ~MAN_BLINE;
1.92 schwarze 69: man_unscope(man, man->last->parent);
1.106 schwarze 70: roff_body_alloc(man, line, offs, man->last->tok);
1.52 schwarze 71: }
72:
1.1 kristaps 73: static int
1.101 schwarze 74: man_ptext(struct roff_man *man, int line, char *buf, int offs)
1.1 kristaps 75: {
1.27 schwarze 76: int i;
1.26 schwarze 77:
1.10 schwarze 78: /* Literal free-form text whitespace is preserved. */
79:
1.92 schwarze 80: if (man->flags & MAN_LITERAL) {
1.108 schwarze 81: roff_word_alloc(man, line, offs, buf + offs);
1.92 schwarze 82: man_descope(man, line, offs);
1.113 schwarze 83: return 1;
1.10 schwarze 84: }
85:
1.92 schwarze 86: for (i = offs; buf[i] == ' '; i++)
1.10 schwarze 87: /* Skip leading whitespace. */ ;
1.18 schwarze 88:
1.71 schwarze 89: /*
90: * Blank lines are ignored right after headings
91: * but add a single vertical space elsewhere.
92: */
93:
1.92 schwarze 94: if (buf[i] == '\0') {
1.121 ! schwarze 95: man_breakscope(man, ROFF_sp);
1.27 schwarze 96: /* Allocate a blank entry. */
1.92 schwarze 97: if (man->last->tok != MAN_SH &&
98: man->last->tok != MAN_SS) {
1.120 schwarze 99: roff_elem_alloc(man, line, offs, ROFF_sp);
1.101 schwarze 100: man->next = ROFF_NEXT_SIBLING;
1.71 schwarze 101: }
1.113 schwarze 102: return 1;
1.10 schwarze 103: }
1.1 kristaps 104:
1.79 schwarze 105: /*
1.27 schwarze 106: * Warn if the last un-escaped character is whitespace. Then
1.79 schwarze 107: * strip away the remaining spaces (tabs stay!).
1.27 schwarze 108: */
1.18 schwarze 109:
1.27 schwarze 110: i = (int)strlen(buf);
111: assert(i);
1.18 schwarze 112:
1.27 schwarze 113: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
114: if (i > 1 && '\\' != buf[i - 2])
1.81 schwarze 115: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
116: line, i - 1, NULL);
1.18 schwarze 117:
1.27 schwarze 118: for (--i; i && ' ' == buf[i]; i--)
119: /* Spin back to non-space. */ ;
1.10 schwarze 120:
1.27 schwarze 121: /* Jump ahead of escaped whitespace. */
122: i += '\\' == buf[i] ? 2 : 1;
1.18 schwarze 123:
1.27 schwarze 124: buf[i] = '\0';
1.10 schwarze 125: }
1.108 schwarze 126: roff_word_alloc(man, line, offs, buf + offs);
1.28 schwarze 127:
128: /*
129: * End-of-sentence check. If the last character is an unescaped
130: * EOS character, then flag the node as being the end of a
131: * sentence. The front-end will know how to interpret this.
132: */
133:
134: assert(i);
1.72 schwarze 135: if (mandoc_eos(buf, (size_t)i))
1.115 schwarze 136: man->last->flags |= NODE_EOS;
1.10 schwarze 137:
1.92 schwarze 138: man_descope(man, line, offs);
1.113 schwarze 139: return 1;
1.1 kristaps 140: }
141:
1.53 schwarze 142: static int
1.101 schwarze 143: man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 144: {
1.99 schwarze 145: struct roff_node *n;
1.91 schwarze 146: const char *cp;
1.118 schwarze 147: size_t sz;
148: enum roff_tok tok;
149: int ppos;
1.84 schwarze 150: int bline;
1.118 schwarze 151:
152: /* Determine the line macro. */
1.1 kristaps 153:
1.59 schwarze 154: ppos = offs;
1.118 schwarze 155: tok = TOKEN_NONE;
156: for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
157: offs++;
158: if (sz > 0 && sz < 4)
159: tok = roffhash_find(man->manmac, buf + ppos, sz);
1.107 schwarze 160: if (tok == TOKEN_NONE) {
1.85 schwarze 161: mandoc_msg(MANDOCERR_MACRO, man->parse,
162: ln, ppos, buf + ppos - 1);
1.113 schwarze 163: return 1;
1.1 kristaps 164: }
165:
1.91 schwarze 166: /* Skip a leading escape sequence or tab. */
167:
168: switch (buf[offs]) {
169: case '\\':
170: cp = buf + offs + 1;
171: mandoc_escape(&cp, NULL, NULL);
172: offs = cp - buf;
173: break;
174: case '\t':
175: offs++;
176: break;
177: default:
178: break;
179: }
180:
181: /* Jump to the next non-whitespace word. */
1.1 kristaps 182:
1.116 schwarze 183: while (buf[offs] == ' ')
1.59 schwarze 184: offs++;
1.18 schwarze 185:
1.79 schwarze 186: /*
1.27 schwarze 187: * Trailing whitespace. Note that tabs are allowed to be passed
188: * into the parser as "text", so we only warn about spaces here.
189: */
1.18 schwarze 190:
1.92 schwarze 191: if (buf[offs] == '\0' && buf[offs - 1] == ' ')
1.81 schwarze 192: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
193: ln, offs - 1, NULL);
1.1 kristaps 194:
1.79 schwarze 195: /*
1.97 schwarze 196: * Some macros break next-line scopes; otherwise, remember
197: * whether we are in next-line scope for a block head.
1.21 schwarze 198: */
199:
1.97 schwarze 200: man_breakscope(man, tok);
1.84 schwarze 201: bline = man->flags & MAN_BLINE;
1.119 schwarze 202:
203: /*
204: * If the line in next-line scope ends with \c, keep the
205: * next-line scope open for the subsequent input line.
206: * That is not at all portable, only groff >= 1.22.4
207: * does it, but *if* this weird idiom occurs in a manual
208: * page, that's very likely what the author intended.
209: */
210:
211: if (bline) {
212: cp = strchr(buf + offs, '\0') - 2;
213: if (cp >= buf && cp[0] == '\\' && cp[1] == 'c')
214: bline = 0;
215: }
1.24 schwarze 216:
217: /* Call to handler... */
1.1 kristaps 218:
1.22 schwarze 219: assert(man_macros[tok].fp);
1.92 schwarze 220: (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf);
1.73 schwarze 221:
222: /* In quick mode (for mandocdb), abort after the NAME section. */
223:
1.92 schwarze 224: if (man->quick && tok == MAN_SH) {
1.80 schwarze 225: n = man->last;
1.98 schwarze 226: if (n->type == ROFFT_BODY &&
1.80 schwarze 227: strcmp(n->prev->child->string, "NAME"))
1.113 schwarze 228: return 2;
1.80 schwarze 229: }
1.1 kristaps 230:
1.79 schwarze 231: /*
1.84 schwarze 232: * If we are in a next-line scope for a block head,
233: * close it out now and switch to the body,
234: * unless the next-line scope is allowed to continue.
1.21 schwarze 235: */
236:
1.84 schwarze 237: if ( ! bline || man->flags & MAN_ELINE ||
238: man_macros[tok].flags & MAN_NSCOPED)
1.113 schwarze 239: return 1;
1.1 kristaps 240:
1.92 schwarze 241: assert(man->flags & MAN_BLINE);
1.69 schwarze 242: man->flags &= ~MAN_BLINE;
1.1 kristaps 243:
1.92 schwarze 244: man_unscope(man, man->last->parent);
1.106 schwarze 245: roff_body_alloc(man, ln, ppos, man->last->tok);
1.113 schwarze 246: return 1;
1.97 schwarze 247: }
248:
249: void
1.101 schwarze 250: man_breakscope(struct roff_man *man, int tok)
1.97 schwarze 251: {
1.99 schwarze 252: struct roff_node *n;
1.97 schwarze 253:
254: /*
255: * An element next line scope is open,
256: * and the new macro is not allowed inside elements.
257: * Delete the element that is being broken.
258: */
259:
1.121 ! schwarze 260: if (man->flags & MAN_ELINE && (tok < MAN_TH ||
1.97 schwarze 261: ! (man_macros[tok].flags & MAN_NSCOPED))) {
262: n = man->last;
1.98 schwarze 263: assert(n->type != ROFFT_TEXT);
1.97 schwarze 264: if (man_macros[n->tok].flags & MAN_NSCOPED)
265: n = n->parent;
266:
267: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
268: n->line, n->pos, "%s breaks %s",
1.121 ! schwarze 269: roff_name[tok], roff_name[n->tok]);
1.97 schwarze 270:
1.106 schwarze 271: roff_node_delete(man, n);
1.97 schwarze 272: man->flags &= ~MAN_ELINE;
1.112 schwarze 273: }
274:
275: /*
276: * Weird special case:
277: * Switching fill mode closes section headers.
278: */
279:
280: if (man->flags & MAN_BLINE &&
281: (tok == MAN_nf || tok == MAN_fi) &&
282: (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
283: n = man->last;
284: man_unscope(man, n);
285: roff_body_alloc(man, n->line, n->pos, n->tok);
286: man->flags &= ~MAN_BLINE;
1.97 schwarze 287: }
288:
289: /*
290: * A block header next line scope is open,
291: * and the new macro is not allowed inside block headers.
292: * Delete the block that is being broken.
293: */
294:
1.121 ! schwarze 295: if (man->flags & MAN_BLINE && (tok < MAN_TH ||
1.97 schwarze 296: man_macros[tok].flags & MAN_BSCOPE)) {
297: n = man->last;
1.98 schwarze 298: if (n->type == ROFFT_TEXT)
1.97 schwarze 299: n = n->parent;
300: if ( ! (man_macros[n->tok].flags & MAN_BSCOPE))
301: n = n->parent;
302:
1.98 schwarze 303: assert(n->type == ROFFT_HEAD);
1.97 schwarze 304: n = n->parent;
1.98 schwarze 305: assert(n->type == ROFFT_BLOCK);
1.97 schwarze 306: assert(man_macros[n->tok].flags & MAN_SCOPED);
307:
308: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
309: n->line, n->pos, "%s breaks %s",
1.121 ! schwarze 310: roff_name[tok], roff_name[n->tok]);
1.97 schwarze 311:
1.106 schwarze 312: roff_node_delete(man, n);
1.97 schwarze 313: man->flags &= ~MAN_BLINE;
314: }
1.62 schwarze 315: }
316:
317: const struct mparse *
1.101 schwarze 318: man_mparse(const struct roff_man *man)
1.62 schwarze 319: {
320:
1.69 schwarze 321: assert(man && man->parse);
1.113 schwarze 322: return man->parse;
1.114 schwarze 323: }
324:
325: void
326: man_state(struct roff_man *man, struct roff_node *n)
327: {
328:
329: switch(n->tok) {
330: case MAN_nf:
331: case MAN_EX:
1.115 schwarze 332: if (man->flags & MAN_LITERAL && ! (n->flags & NODE_VALID))
1.114 schwarze 333: mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
334: n->line, n->pos, "nf");
335: man->flags |= MAN_LITERAL;
336: break;
337: case MAN_fi:
338: case MAN_EE:
339: if ( ! (man->flags & MAN_LITERAL) &&
1.115 schwarze 340: ! (n->flags & NODE_VALID))
1.114 schwarze 341: mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
342: n->line, n->pos, "fi");
343: man->flags &= ~MAN_LITERAL;
344: break;
345: default:
346: break;
347: }
1.115 schwarze 348: man->last->flags |= NODE_VALID;
1.114 schwarze 349: }
350:
351: void
352: man_validate(struct roff_man *man)
353: {
354:
355: man->last = man->first;
356: man_node_validate(man);
357: man->flags &= ~MAN_LITERAL;
1.4 schwarze 358: }