=================================================================== RCS file: /cvs/mandoc/mdoc_macro.c,v retrieving revision 1.62 retrieving revision 1.67 diff -u -p -r1.62 -r1.67 --- mandoc/mdoc_macro.c 2010/05/14 12:55:22 1.62 +++ mandoc/mdoc_macro.c 2010/05/15 08:54:04 1.67 @@ -1,4 +1,4 @@ -/* $Id: mdoc_macro.c,v 1.62 2010/05/14 12:55:22 kristaps Exp $ */ +/* $Id: mdoc_macro.c,v 1.67 2010/05/15 08:54:04 schwarze Exp $ */ /* * Copyright (c) 2008, 2009 Kristaps Dzonsons * @@ -26,6 +26,7 @@ #include #include "libmdoc.h" +#include "libmandoc.h" enum rew { REWIND_REWIND, @@ -621,26 +622,41 @@ rew_sub(enum mdoc_type t, struct mdoc *m, static int -append_delims(struct mdoc *mdoc, int line, int *pos, char *buf) +append_delims(struct mdoc *m, int line, int *pos, char *buf) { - int lastarg; + int la; enum margserr ac; char *p; - if (0 == buf[*pos]) + if ('\0' == buf[*pos]) return(1); for (;;) { - lastarg = *pos; - ac = mdoc_zargs(mdoc, line, pos, buf, ARGS_NOWARN, &p); + la = *pos; + ac = mdoc_zargs(m, line, pos, buf, ARGS_NOWARN, &p); if (ARGS_ERROR == ac) return(0); else if (ARGS_EOLN == ac) break; - assert(mdoc_isdelim(p)); - if ( ! mdoc_word_alloc(mdoc, line, lastarg, p)) + + assert(DELIM_NONE != mdoc_isdelim(p)); + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); + + /* + * If we encounter end-of-sentence symbols, then trigger + * the double-space. + * + * XXX: it's easy to allow this to propogate 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. + */ + if (mandoc_eos(p, strlen(p))) + m->last->flags |= MDOC_EOS; } return(1); @@ -734,10 +750,11 @@ blk_exp_close(MACRO_PROT_ARGS) static int in_line(MACRO_PROT_ARGS) { - int la, lastpunct, cnt, d, nc, nl; + int la, lastpunct, cnt, nc, nl; enum margverr av; enum mdoct ntok; enum margserr ac; + enum mdelim d; struct mdoc_arg *arg; char *p; @@ -830,9 +847,9 @@ in_line(MACRO_PROT_ARGS) * the word. */ - d = ARGS_QWORD == ac ? 0 : mdoc_isdelim(p); + d = ARGS_QWORD == ac ? DELIM_NONE : mdoc_isdelim(p); - if (ARGS_QWORD != ac && d) { + if (ARGS_QWORD != ac && DELIM_NONE != d) { if (0 == lastpunct && ! rew_elem(m, tok)) return(0); lastpunct = 1; @@ -842,7 +859,7 @@ in_line(MACRO_PROT_ARGS) lastpunct = 0; } - if ( ! d) + if (DELIM_NONE == d) cnt++; if ( ! mdoc_word_alloc(m, line, la, p)) return(0); @@ -888,7 +905,7 @@ in_line(MACRO_PROT_ARGS) static int blk_full(MACRO_PROT_ARGS) { - int la; + int la, nl; struct mdoc_arg *arg; struct mdoc_node *head; /* save of head macro */ struct mdoc_node *body; /* save of body macro */ @@ -900,6 +917,8 @@ blk_full(MACRO_PROT_ARGS) enum margverr av; char *p; + nl = MDOC_NEWLINE & m->flags; + /* Close out prior implicit scope. */ if ( ! (MDOC_EXPLICIT & mdoc_macros[tok].flags)) { @@ -969,14 +988,20 @@ blk_full(MACRO_PROT_ARGS) if (ARGS_EOLN == ac) break; + if (ARGS_PEND == ac) { + if (ARGS_PPHRASE == lac) + ac = ARGS_PPHRASE; + else + ac = ARGS_PHRASE; + } + /* Don't emit leading punct. for phrases. */ if (NULL == head && ARGS_PHRASE != ac && ARGS_PPHRASE != ac && - ARGS_PEND != ac && ARGS_QWORD != ac && - 1 == mdoc_isdelim(p)) { + DELIM_OPEN == mdoc_isdelim(p)) { if ( ! mdoc_word_alloc(m, line, la, p)) return(0); continue; @@ -986,29 +1011,18 @@ blk_full(MACRO_PROT_ARGS) if (NULL == head || ARGS_PHRASE == ac || - ARGS_PEND == ac || ARGS_PPHRASE == ac) { if ( ! mdoc_head_alloc(m, line, ppos, tok)) return(0); head = m->last; } - if (ARGS_PHRASE == ac || - ARGS_PEND == ac || - ARGS_PPHRASE == ac) { - /* - * Special treatment for the last phrase. A - * prior ARGS_PHRASE gets is handled as a - * regular ARGS_PHRASE, but a prior ARGS_PPHRASE - * has special handling. - */ - if (ARGS_PEND == ac && ARGS_ERROR == lac) - ac = ARGS_PHRASE; - else if (ARGS_PEND == ac && ARGS_PHRASE == lac) - ac = ARGS_PHRASE; - + if (ARGS_PHRASE == ac || ARGS_PPHRASE == ac) { + if (ARGS_PPHRASE == ac) + m->flags |= MDOC_PPHRASE; if ( ! phrase(m, line, la, buf, ac)) return(0); + m->flags &= ~MDOC_PPHRASE; if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos)) return(0); continue; @@ -1033,7 +1047,7 @@ blk_full(MACRO_PROT_ARGS) head = m->last; } - if (1 == ppos && ! append_delims(m, line, pos, buf)) + if (nl && ! append_delims(m, line, pos, buf)) return(0); /* If we've already opened our body, exit now. */ @@ -1073,7 +1087,7 @@ blk_full(MACRO_PROT_ARGS) static int blk_part_imp(MACRO_PROT_ARGS) { - int la; + int la, nl; enum mdoct ntok; enum margserr ac; char *p; @@ -1081,6 +1095,8 @@ blk_part_imp(MACRO_PROT_ARGS) struct mdoc_node *body; /* saved body context */ struct mdoc_node *n; + nl = MDOC_NEWLINE & m->flags; + /* * A macro that spans to the end of the line. This is generally * (but not necessarily) called as the first macro. The block @@ -1118,7 +1134,7 @@ blk_part_imp(MACRO_PROT_ARGS) break; if (NULL == body && ARGS_QWORD != ac && - 1 == mdoc_isdelim(p)) { + DELIM_OPEN == mdoc_isdelim(p)) { if ( ! mdoc_word_alloc(m, line, la, p)) return(0); continue; @@ -1151,11 +1167,32 @@ blk_part_imp(MACRO_PROT_ARGS) body = m->last; } + for (n = body->child; n && n->next; n = n->next) + /* Do nothing. */ ; + /* + * End of sentence spacing: if the last node is a text node and + * has a trailing period, then mark it as being end-of-sentence. + */ + + if (n && MDOC_TEXT == n->type && n->string) + if (mandoc_eos(n->string, strlen(n->string))) + n->flags |= MDOC_EOS; + + /* Up-propogate the end-of-space flag. */ + + if (n && (MDOC_EOS & n->flags)) { + body->flags |= MDOC_EOS; + body->parent->flags |= MDOC_EOS; + } + + /* * If we can't rewind to our body, then our scope has already * been closed by another macro (like `Oc' closing `Op'). This * is ugly behaviour nodding its head to OpenBSD's overwhelming * crufty use of `Op' breakage. + * + * FIXME - this should be ifdef'd OpenBSD? */ for (n = m->last; n; n = n->parent) if (body == n) @@ -1169,7 +1206,7 @@ blk_part_imp(MACRO_PROT_ARGS) /* Standard appending of delimiters. */ - if (1 == ppos && ! append_delims(m, line, pos, buf)) + if (nl && ! append_delims(m, line, pos, buf)) return(0); /* Rewind scope, if applicable. */ @@ -1216,7 +1253,7 @@ blk_part_exp(MACRO_PROT_ARGS) /* Flush out leading punctuation. */ if (NULL == head && ARGS_QWORD != ac && - 1 == mdoc_isdelim(p)) { + DELIM_OPEN == mdoc_isdelim(p)) { assert(NULL == body); if ( ! mdoc_word_alloc(m, line, la, p)) return(0); @@ -1361,7 +1398,7 @@ in_line_argn(MACRO_PROT_ARGS) if ( ! (MDOC_IGNDELIM & mdoc_macros[tok].flags) && ARGS_QWORD != ac && - 0 == j && 1 == mdoc_isdelim(p)) { + 0 == j && DELIM_OPEN == mdoc_isdelim(p)) { if ( ! mdoc_word_alloc(m, line, la, p)) return(0); continue; @@ -1389,7 +1426,8 @@ in_line_argn(MACRO_PROT_ARGS) if ( ! (MDOC_IGNDELIM & mdoc_macros[tok].flags) && ARGS_QWORD != ac && - ! flushed && mdoc_isdelim(p)) { + ! flushed && + DELIM_NONE != mdoc_isdelim(p)) { if ( ! rew_elem(m, tok)) return(0); flushed = 1; @@ -1543,22 +1581,19 @@ phrase(struct mdoc *m, int line, int ppos, char *buf, enum mdoct ntok; char *p; - assert(ARGS_PHRASE == ac || - ARGS_PEND == ac || - ARGS_PPHRASE == ac); + assert(ARGS_PHRASE == ac || ARGS_PPHRASE == ac); for (pos = ppos; ; ) { la = pos; - aac = mdoc_zargs(m, line, &pos, buf, ARGS_PPHRASED, &p); + aac = mdoc_zargs(m, line, &pos, buf, 0, &p); if (ARGS_ERROR == aac) return(0); if (ARGS_EOLN == aac) break; - ntok = ARGS_QWORD == aac || ARGS_PEND == ac ? - MDOC_MAX : lookup_raw(p); + ntok = ARGS_QWORD == aac ? MDOC_MAX : lookup_raw(p); if (MDOC_MAX == ntok) { if ( ! mdoc_word_alloc(m, line, la, p))