Annotation of src/usr.bin/mandoc/man.c, Revision 1.122
1.122 ! schwarze 1: /* $OpenBSD: man.c,v 1.121 2017/05/08 20:33:40 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: /*
1.122 ! schwarze 90: * Blank lines are ignored in next line scope and right
! 91: * after headings but add a single vertical space elsewhere.
1.71 schwarze 92: */
93:
1.92 schwarze 94: if (buf[i] == '\0') {
1.122 ! schwarze 95: if (man->flags & (MAN_ELINE | MAN_BLINE))
! 96: mandoc_msg(MANDOCERR_BLK_BLANK, man->parse,
! 97: line, 0, NULL);
! 98: else if (man->last->tok != MAN_SH &&
1.92 schwarze 99: man->last->tok != MAN_SS) {
1.120 schwarze 100: roff_elem_alloc(man, line, offs, ROFF_sp);
1.101 schwarze 101: man->next = ROFF_NEXT_SIBLING;
1.71 schwarze 102: }
1.113 schwarze 103: return 1;
1.10 schwarze 104: }
1.1 kristaps 105:
1.79 schwarze 106: /*
1.27 schwarze 107: * Warn if the last un-escaped character is whitespace. Then
1.79 schwarze 108: * strip away the remaining spaces (tabs stay!).
1.27 schwarze 109: */
1.18 schwarze 110:
1.27 schwarze 111: i = (int)strlen(buf);
112: assert(i);
1.18 schwarze 113:
1.27 schwarze 114: if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
115: if (i > 1 && '\\' != buf[i - 2])
1.81 schwarze 116: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
117: line, i - 1, NULL);
1.18 schwarze 118:
1.27 schwarze 119: for (--i; i && ' ' == buf[i]; i--)
120: /* Spin back to non-space. */ ;
1.10 schwarze 121:
1.27 schwarze 122: /* Jump ahead of escaped whitespace. */
123: i += '\\' == buf[i] ? 2 : 1;
1.18 schwarze 124:
1.27 schwarze 125: buf[i] = '\0';
1.10 schwarze 126: }
1.108 schwarze 127: roff_word_alloc(man, line, offs, buf + offs);
1.28 schwarze 128:
129: /*
130: * End-of-sentence check. If the last character is an unescaped
131: * EOS character, then flag the node as being the end of a
132: * sentence. The front-end will know how to interpret this.
133: */
134:
135: assert(i);
1.72 schwarze 136: if (mandoc_eos(buf, (size_t)i))
1.115 schwarze 137: man->last->flags |= NODE_EOS;
1.10 schwarze 138:
1.92 schwarze 139: man_descope(man, line, offs);
1.113 schwarze 140: return 1;
1.1 kristaps 141: }
142:
1.53 schwarze 143: static int
1.101 schwarze 144: man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
1.1 kristaps 145: {
1.99 schwarze 146: struct roff_node *n;
1.91 schwarze 147: const char *cp;
1.118 schwarze 148: size_t sz;
149: enum roff_tok tok;
150: int ppos;
1.84 schwarze 151: int bline;
1.118 schwarze 152:
153: /* Determine the line macro. */
1.1 kristaps 154:
1.59 schwarze 155: ppos = offs;
1.118 schwarze 156: tok = TOKEN_NONE;
157: for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
158: offs++;
159: if (sz > 0 && sz < 4)
160: tok = roffhash_find(man->manmac, buf + ppos, sz);
1.107 schwarze 161: if (tok == TOKEN_NONE) {
1.85 schwarze 162: mandoc_msg(MANDOCERR_MACRO, man->parse,
163: ln, ppos, buf + ppos - 1);
1.113 schwarze 164: return 1;
1.1 kristaps 165: }
166:
1.91 schwarze 167: /* Skip a leading escape sequence or tab. */
168:
169: switch (buf[offs]) {
170: case '\\':
171: cp = buf + offs + 1;
172: mandoc_escape(&cp, NULL, NULL);
173: offs = cp - buf;
174: break;
175: case '\t':
176: offs++;
177: break;
178: default:
179: break;
180: }
181:
182: /* Jump to the next non-whitespace word. */
1.1 kristaps 183:
1.116 schwarze 184: while (buf[offs] == ' ')
1.59 schwarze 185: offs++;
1.18 schwarze 186:
1.79 schwarze 187: /*
1.27 schwarze 188: * Trailing whitespace. Note that tabs are allowed to be passed
189: * into the parser as "text", so we only warn about spaces here.
190: */
1.18 schwarze 191:
1.92 schwarze 192: if (buf[offs] == '\0' && buf[offs - 1] == ' ')
1.81 schwarze 193: mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
194: ln, offs - 1, NULL);
1.1 kristaps 195:
1.79 schwarze 196: /*
1.97 schwarze 197: * Some macros break next-line scopes; otherwise, remember
198: * whether we are in next-line scope for a block head.
1.21 schwarze 199: */
200:
1.97 schwarze 201: man_breakscope(man, tok);
1.84 schwarze 202: bline = man->flags & MAN_BLINE;
1.119 schwarze 203:
204: /*
205: * If the line in next-line scope ends with \c, keep the
206: * next-line scope open for the subsequent input line.
207: * That is not at all portable, only groff >= 1.22.4
208: * does it, but *if* this weird idiom occurs in a manual
209: * page, that's very likely what the author intended.
210: */
211:
212: if (bline) {
213: cp = strchr(buf + offs, '\0') - 2;
214: if (cp >= buf && cp[0] == '\\' && cp[1] == 'c')
215: bline = 0;
216: }
1.24 schwarze 217:
218: /* Call to handler... */
1.1 kristaps 219:
1.22 schwarze 220: assert(man_macros[tok].fp);
1.92 schwarze 221: (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf);
1.73 schwarze 222:
223: /* In quick mode (for mandocdb), abort after the NAME section. */
224:
1.92 schwarze 225: if (man->quick && tok == MAN_SH) {
1.80 schwarze 226: n = man->last;
1.98 schwarze 227: if (n->type == ROFFT_BODY &&
1.80 schwarze 228: strcmp(n->prev->child->string, "NAME"))
1.113 schwarze 229: return 2;
1.80 schwarze 230: }
1.1 kristaps 231:
1.79 schwarze 232: /*
1.84 schwarze 233: * If we are in a next-line scope for a block head,
234: * close it out now and switch to the body,
235: * unless the next-line scope is allowed to continue.
1.21 schwarze 236: */
237:
1.84 schwarze 238: if ( ! bline || man->flags & MAN_ELINE ||
239: man_macros[tok].flags & MAN_NSCOPED)
1.113 schwarze 240: return 1;
1.1 kristaps 241:
1.92 schwarze 242: assert(man->flags & MAN_BLINE);
1.69 schwarze 243: man->flags &= ~MAN_BLINE;
1.1 kristaps 244:
1.92 schwarze 245: man_unscope(man, man->last->parent);
1.106 schwarze 246: roff_body_alloc(man, ln, ppos, man->last->tok);
1.113 schwarze 247: return 1;
1.97 schwarze 248: }
249:
250: void
1.101 schwarze 251: man_breakscope(struct roff_man *man, int tok)
1.97 schwarze 252: {
1.99 schwarze 253: struct roff_node *n;
1.97 schwarze 254:
255: /*
256: * An element next line scope is open,
257: * and the new macro is not allowed inside elements.
258: * Delete the element that is being broken.
259: */
260:
1.121 schwarze 261: if (man->flags & MAN_ELINE && (tok < MAN_TH ||
1.97 schwarze 262: ! (man_macros[tok].flags & MAN_NSCOPED))) {
263: n = man->last;
1.98 schwarze 264: assert(n->type != ROFFT_TEXT);
1.97 schwarze 265: if (man_macros[n->tok].flags & MAN_NSCOPED)
266: n = n->parent;
267:
268: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
269: n->line, n->pos, "%s breaks %s",
1.121 schwarze 270: roff_name[tok], roff_name[n->tok]);
1.97 schwarze 271:
1.106 schwarze 272: roff_node_delete(man, n);
1.97 schwarze 273: man->flags &= ~MAN_ELINE;
1.112 schwarze 274: }
275:
276: /*
277: * Weird special case:
278: * Switching fill mode closes section headers.
279: */
280:
281: if (man->flags & MAN_BLINE &&
282: (tok == MAN_nf || tok == MAN_fi) &&
283: (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
284: n = man->last;
285: man_unscope(man, n);
286: roff_body_alloc(man, n->line, n->pos, n->tok);
287: man->flags &= ~MAN_BLINE;
1.97 schwarze 288: }
289:
290: /*
291: * A block header next line scope is open,
292: * and the new macro is not allowed inside block headers.
293: * Delete the block that is being broken.
294: */
295:
1.121 schwarze 296: if (man->flags & MAN_BLINE && (tok < MAN_TH ||
1.97 schwarze 297: man_macros[tok].flags & MAN_BSCOPE)) {
298: n = man->last;
1.98 schwarze 299: if (n->type == ROFFT_TEXT)
1.97 schwarze 300: n = n->parent;
301: if ( ! (man_macros[n->tok].flags & MAN_BSCOPE))
302: n = n->parent;
303:
1.98 schwarze 304: assert(n->type == ROFFT_HEAD);
1.97 schwarze 305: n = n->parent;
1.98 schwarze 306: assert(n->type == ROFFT_BLOCK);
1.97 schwarze 307: assert(man_macros[n->tok].flags & MAN_SCOPED);
308:
309: mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
310: n->line, n->pos, "%s breaks %s",
1.121 schwarze 311: roff_name[tok], roff_name[n->tok]);
1.97 schwarze 312:
1.106 schwarze 313: roff_node_delete(man, n);
1.97 schwarze 314: man->flags &= ~MAN_BLINE;
315: }
1.62 schwarze 316: }
317:
318: const struct mparse *
1.101 schwarze 319: man_mparse(const struct roff_man *man)
1.62 schwarze 320: {
321:
1.69 schwarze 322: assert(man && man->parse);
1.113 schwarze 323: return man->parse;
1.114 schwarze 324: }
325:
326: void
327: man_state(struct roff_man *man, struct roff_node *n)
328: {
329:
330: switch(n->tok) {
331: case MAN_nf:
332: case MAN_EX:
1.115 schwarze 333: if (man->flags & MAN_LITERAL && ! (n->flags & NODE_VALID))
1.114 schwarze 334: mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
335: n->line, n->pos, "nf");
336: man->flags |= MAN_LITERAL;
337: break;
338: case MAN_fi:
339: case MAN_EE:
340: if ( ! (man->flags & MAN_LITERAL) &&
1.115 schwarze 341: ! (n->flags & NODE_VALID))
1.114 schwarze 342: mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
343: n->line, n->pos, "fi");
344: man->flags &= ~MAN_LITERAL;
345: break;
346: default:
347: break;
348: }
1.115 schwarze 349: man->last->flags |= NODE_VALID;
1.114 schwarze 350: }
351:
352: void
353: man_validate(struct roff_man *man)
354: {
355:
356: man->last = man->first;
357: man_node_validate(man);
358: man->flags &= ~MAN_LITERAL;
1.4 schwarze 359: }