=================================================================== RCS file: /cvs/mandoc/mdoc_macro.c,v retrieving revision 1.103 retrieving revision 1.113 diff -u -p -r1.103 -r1.113 --- mandoc/mdoc_macro.c 2011/03/17 11:30:23 1.103 +++ mandoc/mdoc_macro.c 2011/12/03 23:59:14 1.113 @@ -1,6 +1,6 @@ -/* $Id: mdoc_macro.c,v 1.103 2011/03/17 11:30:23 kristaps Exp $ */ +/* $Id: mdoc_macro.c,v 1.113 2011/12/03 23:59:14 schwarze Exp $ */ /* - * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons + * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons * Copyright (c) 2010 Ingo Schwarze * * Permission to use, copy, modify, and distribute this software for any @@ -26,6 +26,7 @@ #include #include +#include "mdoc.h" #include "mandoc.h" #include "libmdoc.h" #include "libmandoc.h" @@ -50,6 +51,8 @@ static int in_line(MACRO_PROT_ARGS); static int obsolete(MACRO_PROT_ARGS); static int phrase_ta(MACRO_PROT_ARGS); +static int dword(struct mdoc *, int, int, + const char *, enum mdelim); static int append_delims(struct mdoc *, int, int *, char *); static enum mdoct lookup(enum mdoct, const char *); @@ -71,8 +74,8 @@ const struct mdoc_macro __mdoc_macros[MDOC_MAX] = { { in_line_eoln, MDOC_PROLOGUE }, /* Dd */ { in_line_eoln, MDOC_PROLOGUE }, /* Dt */ { in_line_eoln, MDOC_PROLOGUE }, /* Os */ - { blk_full, 0 }, /* Sh */ - { blk_full, 0 }, /* Ss */ + { blk_full, MDOC_PARSED }, /* Sh */ + { blk_full, MDOC_PARSED }, /* Ss */ { in_line_eoln, 0 }, /* Pp */ { blk_part_imp, MDOC_PARSED }, /* D1 */ { blk_part_imp, MDOC_PARSED }, /* Dl */ @@ -513,9 +516,9 @@ make_pending(struct mdoc_node *broken, enum mdoct tok, taker->pending = broken->pending; } broken->pending = breaker; - mdoc_vmsg(m, MANDOCERR_SCOPENEST, line, ppos, - "%s breaks %s", mdoc_macronames[tok], - mdoc_macronames[broken->tok]); + mandoc_vmsg(MANDOCERR_SCOPENEST, m->parse, line, ppos, + "%s breaks %s", mdoc_macronames[tok], + mdoc_macronames[broken->tok]); return(1); } @@ -541,9 +544,10 @@ rew_sub(enum mdoc_type t, struct mdoc *m, case (REWIND_THIS): break; case (REWIND_FORCE): - mdoc_vmsg(m, MANDOCERR_SCOPEBROKEN, line, ppos, - "%s breaks %s", mdoc_macronames[tok], - mdoc_macronames[n->tok]); + mandoc_vmsg(MANDOCERR_SCOPEBROKEN, m->parse, + line, ppos, "%s breaks %s", + mdoc_macronames[tok], + mdoc_macronames[n->tok]); /* FALLTHROUGH */ case (REWIND_MORE): n = n->parent; @@ -579,7 +583,41 @@ rew_sub(enum mdoc_type t, struct mdoc *m, return(1); } +/* + * Allocate a word and check whether it's punctuation or not. + * Punctuation consists of those tokens found in mdoc_isdelim(). + */ +static int +dword(struct mdoc *m, int line, + int col, const char *p, enum mdelim d) +{ + + if (DELIM_MAX == d) + d = mdoc_isdelim(p); + if ( ! mdoc_word_alloc(m, line, col, p)) + return(0); + + if (DELIM_OPEN == d) + m->last->flags |= MDOC_DELIMO; + + /* + * Closing delimiters only suppress the preceding space + * when they follow something, not when they start a new + * block or element, and not when they follow `No'. + * + * XXX Explicitly special-casing MDOC_No here feels + * like a layering violation. Find a better way + * and solve this in the code related to `No'! + */ + + else if (DELIM_CLOSE == d && m->last->prev && + m->last->prev->tok != MDOC_No) + m->last->flags |= MDOC_DELIMC; + + return(1); +} + static int append_delims(struct mdoc *m, int line, int *pos, char *buf) { @@ -592,27 +630,25 @@ append_delims(struct mdoc *m, int line, int *pos, char for (;;) { la = *pos; - ac = mdoc_zargs(m, line, pos, buf, ARGS_NOWARN, &p); + ac = mdoc_zargs(m, line, pos, buf, &p); if (ARGS_ERROR == ac) return(0); else if (ARGS_EOLN == ac) break; - assert(DELIM_NONE != mandoc_isdelim(p)); - if ( ! mdoc_word_alloc(m, line, la, p)) - return(0); + dword(m, line, la, p, DELIM_MAX); /* * If we encounter end-of-sentence symbols, then trigger * the double-space. * - * XXX: it's easy to allow this to propogate outward to + * XXX: it's easy to allow this to propagate outward to * the last symbol, such that `. )' will cause the * correct double-spacing. However, (1) groff isn't * smart enough to do this and (2) it would require * knowing which symbols break this behaviour, for - * example, `. ;' shouldn't propogate the double-space. + * example, `. ;' shouldn't propagate the double-space. */ if (mandoc_eos(p, strlen(p), 0)) m->last->flags |= MDOC_EOS; @@ -745,7 +781,7 @@ blk_exp_close(MACRO_PROT_ARGS) ntok = ARGS_QWORD == ac ? MDOC_MAX : lookup(tok, p); if (MDOC_MAX == ntok) { - if ( ! mdoc_word_alloc(m, line, lastarg, p)) + if ( ! dword(m, line, lastarg, p, DELIM_MAX)) return(0); continue; } @@ -869,7 +905,7 @@ in_line(MACRO_PROT_ARGS) * the word. */ - d = ARGS_QWORD == ac ? DELIM_NONE : mandoc_isdelim(p); + d = ARGS_QWORD == ac ? DELIM_NONE : mdoc_isdelim(p); if (DELIM_NONE != d) { /* @@ -904,7 +940,8 @@ in_line(MACRO_PROT_ARGS) if (DELIM_NONE == d) cnt++; - if ( ! mdoc_word_alloc(m, line, la, p)) + + if ( ! dword(m, line, la, p, d)) return(0); /* @@ -970,7 +1007,7 @@ blk_full(MACRO_PROT_ARGS) } /* - * This routine accomodates implicitly- and explicitly-scoped + * This routine accommodates implicitly- and explicitly-scoped * macro openings. Implicit ones first close out prior scope * (seen above). Delay opening the head until necessary to * allow leading punctuation to print. Special consideration @@ -1059,8 +1096,8 @@ blk_full(MACRO_PROT_ARGS) ARGS_PHRASE != ac && ARGS_PPHRASE != ac && ARGS_QWORD != ac && - DELIM_OPEN == mandoc_isdelim(p)) { - if ( ! mdoc_word_alloc(m, line, la, p)) + DELIM_OPEN == mdoc_isdelim(p)) { + if ( ! dword(m, line, la, p, DELIM_OPEN)) return(0); continue; } @@ -1112,7 +1149,7 @@ blk_full(MACRO_PROT_ARGS) ntok = ARGS_QWORD == ac ? MDOC_MAX : lookup(tok, p); if (MDOC_MAX == ntok) { - if ( ! mdoc_word_alloc(m, line, la, p)) + if ( ! dword(m, line, la, p, DELIM_MAX)) return(0); continue; } @@ -1222,8 +1259,8 @@ blk_part_imp(MACRO_PROT_ARGS) break; if (NULL == body && ARGS_QWORD != ac && - DELIM_OPEN == mandoc_isdelim(p)) { - if ( ! mdoc_word_alloc(m, line, la, p)) + DELIM_OPEN == mdoc_isdelim(p)) { + if ( ! dword(m, line, la, p, DELIM_OPEN)) return(0); continue; } @@ -1237,7 +1274,7 @@ blk_part_imp(MACRO_PROT_ARGS) ntok = ARGS_QWORD == ac ? MDOC_MAX : lookup(tok, p); if (MDOC_MAX == ntok) { - if ( ! mdoc_word_alloc(m, line, la, p)) + if ( ! dword(m, line, la, p, DELIM_MAX)) return(0); continue; } @@ -1267,7 +1304,7 @@ blk_part_imp(MACRO_PROT_ARGS) if (mandoc_eos(n->string, strlen(n->string), 1)) n->flags |= MDOC_EOS; - /* Up-propogate the end-of-space flag. */ + /* Up-propagate the end-of-space flag. */ if (n && (MDOC_EOS & n->flags)) { body->flags |= MDOC_EOS; @@ -1298,7 +1335,7 @@ blk_part_imp(MACRO_PROT_ARGS) * crufty use of `Op' breakage. */ if (n != body) - mdoc_vmsg(m, MANDOCERR_SCOPENEST, line, ppos, + mandoc_vmsg(MANDOCERR_SCOPENEST, m->parse, line, ppos, "%s broken", mdoc_macronames[tok]); if (n && ! rew_sub(MDOC_BODY, m, tok, line, ppos)) @@ -1353,9 +1390,9 @@ blk_part_exp(MACRO_PROT_ARGS) /* Flush out leading punctuation. */ if (NULL == head && ARGS_QWORD != ac && - DELIM_OPEN == mandoc_isdelim(p)) { + DELIM_OPEN == mdoc_isdelim(p)) { assert(NULL == body); - if ( ! mdoc_word_alloc(m, line, la, p)) + if ( ! dword(m, line, la, p, DELIM_OPEN)) return(0); continue; } @@ -1376,7 +1413,7 @@ blk_part_exp(MACRO_PROT_ARGS) assert(head); /* No check whether it's a macro! */ if (MDOC_Eo == tok) - if ( ! mdoc_word_alloc(m, line, la, p)) + if ( ! dword(m, line, la, p, DELIM_MAX)) return(0); if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos)) @@ -1394,7 +1431,7 @@ blk_part_exp(MACRO_PROT_ARGS) ntok = ARGS_QWORD == ac ? MDOC_MAX : lookup(tok, p); if (MDOC_MAX == ntok) { - if ( ! mdoc_word_alloc(m, line, la, p)) + if ( ! dword(m, line, la, p, DELIM_MAX)) return(0); continue; } @@ -1406,18 +1443,15 @@ blk_part_exp(MACRO_PROT_ARGS) /* Clean-up to leave in a consistent state. */ - if (NULL == head) { + if (NULL == head) if ( ! mdoc_head_alloc(m, line, ppos, tok)) return(0); - head = m->last; - } if (NULL == body) { if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos)) return(0); if ( ! mdoc_body_alloc(m, line, ppos, tok)) return(0); - body = m->last; } /* Standard appending of delimiters. */ @@ -1499,9 +1533,9 @@ in_line_argn(MACRO_PROT_ARGS) break; if ( ! (MDOC_IGNDELIM & mdoc_macros[tok].flags) && - ARGS_QWORD != ac && - 0 == j && DELIM_OPEN == mandoc_isdelim(p)) { - if ( ! mdoc_word_alloc(m, line, la, p)) + ARGS_QWORD != ac && 0 == j && + DELIM_OPEN == mdoc_isdelim(p)) { + if ( ! dword(m, line, la, p, DELIM_OPEN)) return(0); continue; } else if (0 == j) @@ -1529,29 +1563,13 @@ in_line_argn(MACRO_PROT_ARGS) if ( ! (MDOC_IGNDELIM & mdoc_macros[tok].flags) && ARGS_QWORD != ac && ! flushed && - DELIM_NONE != mandoc_isdelim(p)) { + DELIM_NONE != mdoc_isdelim(p)) { if ( ! rew_elem(m, tok)) return(0); flushed = 1; } - /* - * XXX: this is a hack to work around groff's ugliness - * as regards `Xr' and extraneous arguments. It should - * ideally be deprecated behaviour, but because this is - * code is no here, it's unlikely to be removed. - */ - -#ifdef __OpenBSD__ - if (MDOC_Xr == tok && j == maxargs) { - if ( ! mdoc_elem_alloc(m, line, la, MDOC_Ns, NULL)) - return(0); - if ( ! rew_elem(m, MDOC_Ns)) - return(0); - } -#endif - - if ( ! mdoc_word_alloc(m, line, la, p)) + if ( ! dword(m, line, la, p, DELIM_MAX)) return(0); j++; } @@ -1622,7 +1640,7 @@ in_line_eoln(MACRO_PROT_ARGS) ntok = ARGS_QWORD == ac ? MDOC_MAX : lookup(tok, p); if (MDOC_MAX == ntok) { - if ( ! mdoc_word_alloc(m, line, la, p)) + if ( ! dword(m, line, la, p, DELIM_MAX)) return(0); continue; } @@ -1692,7 +1710,7 @@ phrase(struct mdoc *m, int line, int ppos, char *buf) for (pos = ppos; ; ) { la = pos; - ac = mdoc_zargs(m, line, &pos, buf, 0, &p); + ac = mdoc_zargs(m, line, &pos, buf, &p); if (ARGS_ERROR == ac) return(0); @@ -1702,7 +1720,7 @@ phrase(struct mdoc *m, int line, int ppos, char *buf) ntok = ARGS_QWORD == ac ? MDOC_MAX : lookup_raw(p); if (MDOC_MAX == ntok) { - if ( ! mdoc_word_alloc(m, line, la, p)) + if ( ! dword(m, line, la, p, DELIM_MAX)) return(0); continue; } @@ -1737,7 +1755,7 @@ phrase_ta(MACRO_PROT_ARGS) for (;;) { la = *pos; - ac = mdoc_zargs(m, line, pos, buf, 0, &p); + ac = mdoc_zargs(m, line, pos, buf, &p); if (ARGS_ERROR == ac) return(0); @@ -1747,7 +1765,7 @@ phrase_ta(MACRO_PROT_ARGS) ntok = ARGS_QWORD == ac ? MDOC_MAX : lookup_raw(p); if (MDOC_MAX == ntok) { - if ( ! mdoc_word_alloc(m, line, la, p)) + if ( ! dword(m, line, la, p, DELIM_MAX)) return(0); continue; }