=================================================================== RCS file: /cvs/mandoc/mdoc_macro.c,v retrieving revision 1.30 retrieving revision 1.49 diff -u -p -r1.30 -r1.49 --- mandoc/mdoc_macro.c 2009/08/13 11:43:24 1.30 +++ mandoc/mdoc_macro.c 2010/03/31 08:04:57 1.49 @@ -1,4 +1,4 @@ -/* $Id: mdoc_macro.c,v 1.30 2009/08/13 11:43:24 kristaps Exp $ */ +/* $Id: mdoc_macro.c,v 1.49 2010/03/31 08:04:57 kristaps Exp $ */ /* * Copyright (c) 2008, 2009 Kristaps Dzonsons * @@ -14,11 +14,16 @@ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + #include #include #include #include #include +#include #include "libmdoc.h" @@ -26,6 +31,7 @@ #define REWIND_NOHALT (1 << 1) #define REWIND_HALT (1 << 2) +static int ctx_synopsis(MACRO_PROT_ARGS); static int obsolete(MACRO_PROT_ARGS); static int blk_part_exp(MACRO_PROT_ARGS); static int in_line_eoln(MACRO_PROT_ARGS); @@ -36,19 +42,18 @@ static int blk_exp_close(MACRO_PROT_ARGS); static int blk_part_imp(MACRO_PROT_ARGS); static int phrase(struct mdoc *, int, int, char *); -static int rew_dohalt(int, enum mdoc_type, +static int rew_dohalt(enum mdoct, enum mdoc_type, const struct mdoc_node *); -static int rew_alt(int); -static int rew_dobreak(int, const struct mdoc_node *); -static int rew_elem(struct mdoc *, int); -static int rew_impblock(struct mdoc *, int, int, int); -static int rew_expblock(struct mdoc *, int, int, int); -static int rew_subblock(enum mdoc_type, - struct mdoc *, int, int, int); -static int rew_last(struct mdoc *, struct mdoc_node *); /* FIXME: make const */ +static enum mdoct rew_alt(enum mdoct); +static int rew_dobreak(enum mdoct, const struct mdoc_node *); +static int rew_elem(struct mdoc *, enum mdoct); +static int rew_sub(enum mdoc_type, struct mdoc *, + enum mdoct, int, int); +static int rew_last(struct mdoc *, + const struct mdoc_node *); static int append_delims(struct mdoc *, int, int *, char *); -static int lookup(struct mdoc *, int, const char *); -static int lookup_raw(struct mdoc *, const char *); +static int lookup(int, const char *); +static int lookup_raw(const char *); static int swarn(struct mdoc *, enum mdoc_type, int, int, const struct mdoc_node *); @@ -94,8 +99,8 @@ const struct mdoc_macro __mdoc_macros[MDOC_MAX] = { { in_line_eoln, 0 }, /* Rv */ { in_line_argn, MDOC_CALLABLE | MDOC_PARSED }, /* St */ { in_line, MDOC_CALLABLE | MDOC_PARSED }, /* Va */ - { in_line, MDOC_CALLABLE | MDOC_PARSED }, /* Vt */ - { in_line, MDOC_CALLABLE | MDOC_PARSED }, /* Xr */ + { ctx_synopsis, MDOC_CALLABLE | MDOC_PARSED }, /* Vt */ + { in_line_argn, MDOC_CALLABLE | MDOC_PARSED }, /* Xr */ { in_line_eoln, 0 }, /* %A */ { in_line_eoln, 0 }, /* %B */ { in_line_eoln, 0 }, /* %D */ @@ -175,6 +180,7 @@ const struct mdoc_macro __mdoc_macros[MDOC_MAX] = { { in_line_eoln, 0 }, /* %Q */ { in_line_eoln, 0 }, /* br */ { in_line_eoln, 0 }, /* sp */ + { in_line_eoln, 0 }, /* %U */ }; const struct mdoc_macro * const mdoc_macros = __mdoc_macros; @@ -233,42 +239,51 @@ swarn(struct mdoc *mdoc, enum mdoc_type type, * are errors. */ int -mdoc_macroend(struct mdoc *mdoc) +mdoc_macroend(struct mdoc *m) { struct mdoc_node *n; /* Scan for open explicit scopes. */ - n = MDOC_VALID & mdoc->last->flags ? - mdoc->last->parent : mdoc->last; + n = MDOC_VALID & m->last->flags ? m->last->parent : m->last; for ( ; n; n = n->parent) { if (MDOC_BLOCK != n->type) continue; if ( ! (MDOC_EXPLICIT & mdoc_macros[n->tok].flags)) continue; - return(mdoc_nerr(mdoc, n, EOPEN)); + return(mdoc_nerr(m, n, EOPEN)); } - return(rew_last(mdoc, mdoc->first)); + /* Rewind to the first. */ + + return(rew_last(m, m->first)); } + +/* + * Look up a macro from within a subsequent context. + */ static int -lookup(struct mdoc *mdoc, int from, const char *p) +lookup(int from, const char *p) { + /* FIXME: make -diag lists be un-PARSED. */ if ( ! (MDOC_PARSED & mdoc_macros[from].flags)) return(MDOC_MAX); - return(lookup_raw(mdoc, p)); + return(lookup_raw(p)); } +/* + * Lookup a macro following the initial line macro. + */ static int -lookup_raw(struct mdoc *mdoc, const char *p) +lookup_raw(const char *p) { int res; - if (MDOC_MAX == (res = mdoc_hash_find(mdoc->htab, p))) + if (MDOC_MAX == (res = mdoc_hash_find(p))) return(MDOC_MAX); if (MDOC_CALLABLE & mdoc_macros[res].flags) return(res); @@ -277,7 +292,7 @@ lookup_raw(struct mdoc *mdoc, const char *p) static int -rew_last(struct mdoc *mdoc, struct mdoc_node *to) +rew_last(struct mdoc *mdoc, const struct mdoc_node *to) { assert(to); @@ -299,8 +314,12 @@ rew_last(struct mdoc *mdoc, struct mdoc_node *to) } -static int -rew_alt(int tok) +/* + * Return the opening macro of a closing one, e.g., `Ec' has `Eo' as its + * matching pair. + */ +static enum mdoct +rew_alt(enum mdoct tok) { switch (tok) { case (MDOC_Ac): @@ -350,7 +369,8 @@ rew_alt(int tok) * The scope-closing and so on occurs in the various rew_* routines. */ static int -rew_dohalt(int tok, enum mdoc_type type, const struct mdoc_node *p) +rew_dohalt(enum mdoct tok, enum mdoc_type type, + const struct mdoc_node *p) { if (MDOC_ROOT == p->type) @@ -380,7 +400,8 @@ rew_dohalt(int tok, enum mdoc_type type, const struct case (MDOC_Qq): /* FALLTHROUGH */ case (MDOC_Sq): - assert(MDOC_HEAD != type); + /* FALLTHROUGH */ + case (MDOC_Vt): assert(MDOC_TAIL != type); if (type == p->type && tok == p->tok) return(REWIND_REWIND); @@ -439,7 +460,6 @@ rew_dohalt(int tok, enum mdoc_type type, const struct if (type == p->type && tok == p->tok) return(REWIND_REWIND); break; - /* Multi-line explicit scope close. */ case (MDOC_Ac): /* FALLTHROUGH */ @@ -489,7 +509,7 @@ rew_dohalt(int tok, enum mdoc_type type, const struct * REWIND_NOHALT). */ static int -rew_dobreak(int tok, const struct mdoc_node *p) +rew_dobreak(enum mdoct tok, const struct mdoc_node *p) { assert(MDOC_ROOT != p->type); @@ -518,7 +538,6 @@ rew_dobreak(int tok, const struct mdoc_node *p) return(1); break; case (MDOC_Oc): - /* XXX - experimental! */ if (MDOC_Op == p->tok) return(1); break; @@ -536,7 +555,7 @@ rew_dobreak(int tok, const struct mdoc_node *p) static int -rew_elem(struct mdoc *mdoc, int tok) +rew_elem(struct mdoc *mdoc, enum mdoct tok) { struct mdoc_node *n; @@ -551,76 +570,50 @@ rew_elem(struct mdoc *mdoc, int tok) static int -rew_subblock(enum mdoc_type type, struct mdoc *mdoc, - int tok, int line, int ppos) +rew_sub(enum mdoc_type t, struct mdoc *m, + enum mdoct tok, int line, int ppos) { struct mdoc_node *n; int c; /* LINTED */ - for (n = mdoc->last; n; n = n->parent) { - c = rew_dohalt(tok, type, n); - if (REWIND_HALT == c) - return(1); + for (n = m->last; n; n = n->parent) { + c = rew_dohalt(tok, t, n); + if (REWIND_HALT == c) { + if (MDOC_BLOCK != t) + return(1); + if ( ! (MDOC_EXPLICIT & mdoc_macros[tok].flags)) + return(1); + return(mdoc_perr(m, line, ppos, ENOCTX)); + } if (REWIND_REWIND == c) break; else if (rew_dobreak(tok, n)) continue; - if ( ! swarn(mdoc, type, line, ppos, n)) + if ( ! swarn(m, t, line, ppos, n)) return(0); } assert(n); - return(rew_last(mdoc, n)); -} + if ( ! rew_last(m, n)) + return(0); - -static int -rew_expblock(struct mdoc *mdoc, int tok, int line, int ppos) -{ - struct mdoc_node *n; - int c; - - /* LINTED */ - for (n = mdoc->last; n; n = n->parent) { - c = rew_dohalt(tok, MDOC_BLOCK, n); - if (REWIND_HALT == c) - return(mdoc_perr(mdoc, line, ppos, ENOCTX)); - if (REWIND_REWIND == c) - break; - else if (rew_dobreak(tok, n)) - continue; - if ( ! swarn(mdoc, MDOC_BLOCK, line, ppos, n)) +#ifdef UGLY + /* + * The current block extends an enclosing block beyond a line + * break. Now that the current block ends, close the enclosing + * block, too. + */ + if (NULL != (n = n->pending)) { + assert(MDOC_HEAD == n->type); + if ( ! rew_last(m, n)) return(0); - } - - assert(n); - return(rew_last(mdoc, n)); -} - - -/* FIXME: can this be merged with subblock? */ -static int -rew_impblock(struct mdoc *mdoc, int tok, int line, int ppos) -{ - struct mdoc_node *n; - int c; - - /* LINTED */ - for (n = mdoc->last; n; n = n->parent) { - c = rew_dohalt(tok, MDOC_BLOCK, n); - if (REWIND_HALT == c) - return(1); - else if (REWIND_REWIND == c) - break; - else if (rew_dobreak(tok, n)) - continue; - if ( ! swarn(mdoc, MDOC_BLOCK, line, ppos, n)) + if ( ! mdoc_body_alloc(m, n->line, n->pos, n->tok)) return(0); } +#endif - assert(n); - return(rew_last(mdoc, n)); + return(1); } @@ -635,7 +628,7 @@ append_delims(struct mdoc *mdoc, int line, int *pos, c for (;;) { lastarg = *pos; - c = mdoc_args(mdoc, line, pos, buf, 0, &p); + c = mdoc_zargs(mdoc, line, pos, buf, ARGS_NOWARN, &p); assert(ARGS_PHRASE != c); if (ARGS_ERROR == c) @@ -645,7 +638,6 @@ append_delims(struct mdoc *mdoc, int line, int *pos, c assert(mdoc_isdelim(p)); if ( ! mdoc_word_alloc(mdoc, line, lastarg, p)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; } return(1); @@ -672,34 +664,31 @@ blk_exp_close(MACRO_PROT_ARGS) if ( ! (MDOC_CALLABLE & mdoc_macros[tok].flags)) { if (buf[*pos]) - if ( ! mdoc_pwarn(mdoc, line, ppos, ENOLINE)) + if ( ! mdoc_pwarn(m, line, ppos, ENOLINE)) return(0); - if ( ! rew_subblock(MDOC_BODY, mdoc, tok, line, ppos)) + if ( ! rew_sub(MDOC_BODY, m, tok, line, ppos)) return(0); - return(rew_expblock(mdoc, tok, line, ppos)); + return(rew_sub(MDOC_BLOCK, m, tok, line, ppos)); } - if ( ! rew_subblock(MDOC_BODY, mdoc, tok, line, ppos)) + if ( ! rew_sub(MDOC_BODY, m, tok, line, ppos)) return(0); - if (maxargs > 0) { - if ( ! mdoc_tail_alloc(mdoc, line, - ppos, rew_alt(tok))) + if (maxargs > 0) + if ( ! mdoc_tail_alloc(m, line, ppos, rew_alt(tok))) return(0); - mdoc->next = MDOC_NEXT_CHILD; - } for (flushed = j = 0; ; j++) { lastarg = *pos; if (j == maxargs && ! flushed) { - if ( ! rew_expblock(mdoc, tok, line, ppos)) + if ( ! rew_sub(MDOC_BLOCK, m, tok, line, ppos)) return(0); flushed = 1; } - c = mdoc_args(mdoc, line, pos, buf, tok, &p); + c = mdoc_args(m, line, pos, buf, tok, &p); if (ARGS_ERROR == c) return(0); @@ -708,35 +697,30 @@ blk_exp_close(MACRO_PROT_ARGS) if (ARGS_EOLN == c) break; - if (MDOC_MAX != (c = lookup(mdoc, tok, p))) { + if (MDOC_MAX != (c = lookup(tok, p))) { if ( ! flushed) { - if ( ! rew_expblock(mdoc, tok, - line, ppos)) + if ( ! rew_sub(MDOC_BLOCK, m, tok, line, ppos)) return(0); flushed = 1; } - if ( ! mdoc_macro(mdoc, c, line, lastarg, pos, buf)) + if ( ! mdoc_macro(m, c, line, lastarg, pos, buf)) return(0); break; } - if ( ! mdoc_word_alloc(mdoc, line, lastarg, p)) + if ( ! mdoc_word_alloc(m, line, lastarg, p)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; } - if ( ! flushed && ! rew_expblock(mdoc, tok, line, ppos)) + if ( ! flushed && ! rew_sub(MDOC_BLOCK, m, tok, line, ppos)) return(0); if (ppos > 1) return(1); - return(append_delims(mdoc, line, pos, buf)); + return(append_delims(m, line, pos, buf)); } -/* - * In-line macros where reserved words cause scope close-reopen. - */ static int in_line(MACRO_PROT_ARGS) { @@ -748,6 +732,7 @@ in_line(MACRO_PROT_ARGS) * Whether we allow ignored elements (those without content, * usually because of reserved words) to squeak by. */ + switch (tok) { case (MDOC_An): /* FALLTHROUGH */ @@ -769,7 +754,7 @@ in_line(MACRO_PROT_ARGS) for (arg = NULL;; ) { la = *pos; - c = mdoc_argv(mdoc, line, tok, &arg, pos, buf); + c = mdoc_argv(m, line, tok, &arg, pos, buf); if (ARGV_WORD == c) { *pos = la; @@ -786,7 +771,7 @@ in_line(MACRO_PROT_ARGS) for (cnt = 0, lastpunct = 1;; ) { la = *pos; - w = mdoc_args(mdoc, line, pos, buf, tok, &p); + w = mdoc_args(m, line, pos, buf, tok, &p); if (ARGS_ERROR == w) return(0); @@ -797,7 +782,7 @@ in_line(MACRO_PROT_ARGS) /* Quoted words shouldn't be looked-up. */ - c = ARGS_QWORD == w ? MDOC_MAX : lookup(mdoc, tok, p); + c = ARGS_QWORD == w ? MDOC_MAX : lookup(tok, p); /* * In this case, we've located a submacro and must @@ -807,25 +792,24 @@ in_line(MACRO_PROT_ARGS) */ if (MDOC_MAX != c) { - if (0 == lastpunct && ! rew_elem(mdoc, tok)) + if (0 == lastpunct && ! rew_elem(m, tok)) return(0); if (nc && 0 == cnt) { - if ( ! mdoc_elem_alloc(mdoc, line, ppos, - tok, arg)) + if ( ! mdoc_elem_alloc(m, line, ppos, tok, arg)) return(0); - if ( ! rew_last(mdoc, mdoc->last)) + if ( ! rew_last(m, m->last)) return(0); } else if ( ! nc && 0 == cnt) { mdoc_argv_free(arg); - if ( ! mdoc_pwarn(mdoc, line, ppos, EIGNE)) + if ( ! mdoc_pwarn(m, line, ppos, EIGNE)) return(0); } - c = mdoc_macro(mdoc, c, line, la, pos, buf); + c = mdoc_macro(m, c, line, la, pos, buf); if (0 == c) return(0); if (ppos > 1) return(1); - return(append_delims(mdoc, line, pos, buf)); + return(append_delims(m, line, pos, buf)); } /* @@ -837,91 +821,94 @@ in_line(MACRO_PROT_ARGS) d = mdoc_isdelim(p); if (ARGS_QWORD != w && d) { - if (0 == lastpunct && ! rew_elem(mdoc, tok)) + if (0 == lastpunct && ! rew_elem(m, tok)) return(0); lastpunct = 1; } else if (lastpunct) { - c = mdoc_elem_alloc(mdoc, line, ppos, tok, arg); - if (0 == c) + if ( ! mdoc_elem_alloc(m, line, ppos, tok, arg)) return(0); - mdoc->next = MDOC_NEXT_CHILD; lastpunct = 0; } if ( ! d) cnt++; - if ( ! mdoc_word_alloc(mdoc, line, la, p)) + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; + + /* + * `Fl' macros have their scope re-opened with each new + * word so that the `-' can be added to each one without + * having to parse out spaces. + */ + if (0 == lastpunct && MDOC_Fl == tok) { + if ( ! rew_elem(m, tok)) + return(0); + lastpunct = 1; + } } - if (0 == lastpunct && ! rew_elem(mdoc, tok)) + if (0 == lastpunct && ! rew_elem(m, tok)) return(0); /* * If no elements have been collected and we're allowed to have * empties (nc), open a scope and close it out. Otherwise, * raise a warning. - * */ + if (nc && 0 == cnt) { - c = mdoc_elem_alloc(mdoc, line, ppos, tok, arg); - if (0 == c) + if ( ! mdoc_elem_alloc(m, line, ppos, tok, arg)) return(0); - if ( ! rew_last(mdoc, mdoc->last)) + if ( ! rew_last(m, m->last)) return(0); } else if ( ! nc && 0 == cnt) { mdoc_argv_free(arg); - if ( ! mdoc_pwarn(mdoc, line, ppos, EIGNE)) + if ( ! mdoc_pwarn(m, line, ppos, EIGNE)) return(0); } if (ppos > 1) return(1); - return(append_delims(mdoc, line, pos, buf)); + return(append_delims(m, line, pos, buf)); } -/* - * Block full-explicit and full-implicit. - */ static int blk_full(MACRO_PROT_ARGS) { - int c, lastarg, reopen, dohead; + int c, la; struct mdoc_arg *arg; + struct mdoc_node *head; /* save of head macro */ + struct mdoc_node *body; /* save of body macro */ +#ifdef UGLY + struct mdoc_node *n; +#endif char *p; - /* - * Whether to process a block-head section. If this is - * non-zero, then a head will be opened for all line arguments. - * If not, then the head will always be empty and only a body - * will be opened, which will stay open at the eoln. - */ + /* Close out prior implicit scope. */ - switch (tok) { - case (MDOC_Nd): - dohead = 0; - break; - default: - dohead = 1; - break; - } - if ( ! (MDOC_EXPLICIT & mdoc_macros[tok].flags)) { - if ( ! rew_subblock(MDOC_BODY, mdoc, - tok, line, ppos)) + if ( ! rew_sub(MDOC_BODY, m, tok, line, ppos)) return(0); - if ( ! rew_impblock(mdoc, tok, line, ppos)) + if ( ! rew_sub(MDOC_BLOCK, m, tok, line, ppos)) return(0); } + /* + * This routine accomodates 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 + * for `It -column', which has phrase-part syntax instead of + * regular child nodes. + */ + for (arg = NULL;; ) { - lastarg = *pos; - c = mdoc_argv(mdoc, line, tok, &arg, pos, buf); + la = *pos; + c = mdoc_argv(m, line, tok, &arg, pos, buf); if (ARGV_WORD == c) { - *pos = lastarg; + *pos = la; break; } @@ -934,227 +921,243 @@ blk_full(MACRO_PROT_ARGS) return(0); } - if ( ! mdoc_block_alloc(mdoc, line, ppos, tok, arg)) + if ( ! mdoc_block_alloc(m, line, ppos, tok, arg)) return(0); - mdoc->next = MDOC_NEXT_CHILD; - if (0 == buf[*pos]) { - if ( ! mdoc_head_alloc(mdoc, line, ppos, tok)) - return(0); - if ( ! rew_subblock(MDOC_HEAD, mdoc, - tok, line, ppos)) - return(0); - if ( ! mdoc_body_alloc(mdoc, line, ppos, tok)) - return(0); - mdoc->next = MDOC_NEXT_CHILD; - return(1); - } + head = body = NULL; - if ( ! mdoc_head_alloc(mdoc, line, ppos, tok)) - return(0); + /* + * The `Nd' macro has all arguments in its body: it's a hybrid + * of block partial-explicit and full-implicit. Stupid. + */ - /* Immediately close out head and enter body, if applicable. */ - - if (0 == dohead) { - if ( ! rew_subblock(MDOC_HEAD, mdoc, tok, line, ppos)) + if (MDOC_Nd == tok) { + if ( ! mdoc_head_alloc(m, line, ppos, tok)) return(0); - if ( ! mdoc_body_alloc(mdoc, line, ppos, tok)) + head = m->last; + if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos)) return(0); + if ( ! mdoc_body_alloc(m, line, ppos, tok)) + return(0); + body = m->last; } - mdoc->next = MDOC_NEXT_CHILD; + for (;;) { + la = *pos; + c = mdoc_args(m, line, pos, buf, tok, &p); - for (reopen = 0;; ) { - lastarg = *pos; - c = mdoc_args(mdoc, line, pos, buf, tok, &p); - if (ARGS_ERROR == c) return(0); if (ARGS_EOLN == c) break; - if (ARGS_PHRASE == c) { - assert(dohead); - if (reopen && ! mdoc_head_alloc - (mdoc, line, ppos, tok)) + + /* Don't emit leading punct. for phrases. */ + + if (NULL == head && ARGS_PHRASE != c && + 1 == mdoc_isdelim(p)) { + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); - mdoc->next = MDOC_NEXT_CHILD; - /* - * Phrases are self-contained macro phrases used - * in the columnar output of a macro. They need - * special handling. - */ - if ( ! phrase(mdoc, line, lastarg, buf)) + continue; + } + + /* Always re-open head for phrases. */ + + if (NULL == head || ARGS_PHRASE == c) { + if ( ! mdoc_head_alloc(m, line, ppos, tok)) return(0); - if ( ! rew_subblock(MDOC_HEAD, mdoc, - tok, line, ppos)) - return(0); + head = m->last; + } - reopen = 1; + if (ARGS_PHRASE == c) { + if ( ! phrase(m, line, la, buf)) + return(0); + if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos)) + return(0); continue; } - if (MDOC_MAX == (c = lookup(mdoc, tok, p))) { - if ( ! mdoc_word_alloc(mdoc, line, lastarg, p)) + c = lookup(tok, p); + if (MDOC_MAX != c) { + if ( ! mdoc_macro(m, c, line, la, pos, buf)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; - continue; - } + break; + } + if ( ! mdoc_word_alloc(m, line, la, p)) + return(0); - if ( ! mdoc_macro(mdoc, c, line, lastarg, pos, buf)) + } + + if (NULL == head) { + if ( ! mdoc_head_alloc(m, line, ppos, tok)) return(0); - break; + head = m->last; } - if (1 == ppos && ! append_delims(mdoc, line, pos, buf)) + if (1 == ppos && ! append_delims(m, line, pos, buf)) return(0); - /* If the body's already open, then just return. */ - if (0 == dohead) + /* If we've already opened our body, exit now. */ + + if (NULL != body) return(1); - if ( ! rew_subblock(MDOC_HEAD, mdoc, tok, line, ppos)) +#ifdef UGLY + /* + * If there is an open (i.e., unvalidated) sub-block requiring + * explicit close-out, postpone switching the current block from + * head to body until the rew_sub() call closing out that + * sub-block. + */ + for (n = m->last; n && n != head; n = n->parent) { + if (MDOC_BLOCK == n->type && + MDOC_EXPLICIT & mdoc_macros[n->tok].flags && + ! (MDOC_VALID & n->flags)) { + assert( ! (MDOC_ACTED & n->flags)); + n->pending = head; + return(1); + } + } +#endif + + /* Close out scopes to remain in a consistent state. */ + + if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos)) return(0); - if ( ! mdoc_body_alloc(mdoc, line, ppos, tok)) + if ( ! mdoc_body_alloc(m, line, ppos, tok)) return(0); - mdoc->next = MDOC_NEXT_CHILD; return(1); } -/* - * Block partial-imnplicit scope. - */ static int blk_part_imp(MACRO_PROT_ARGS) { - int lastarg, c; + int la, c; char *p; - struct mdoc_node *blk, *body, *n; + struct mdoc_node *blk; /* saved block context */ + struct mdoc_node *body; /* saved body context */ + struct mdoc_node *n; - if ( ! mdoc_block_alloc(mdoc, line, ppos, tok, NULL)) - return(0); - mdoc->next = MDOC_NEXT_CHILD; - blk = mdoc->last; + /* + * A macro that spans to the end of the line. This is generally + * (but not necessarily) called as the first macro. The block + * has a head as the immediate child, which is always empty, + * followed by zero or more opening punctuation nodes, then the + * body (which may be empty, depending on the macro), then zero + * or more closing punctuation nodes. + */ - if ( ! mdoc_head_alloc(mdoc, line, ppos, tok)) + if ( ! mdoc_block_alloc(m, line, ppos, tok, NULL)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; - if ( ! mdoc_body_alloc(mdoc, line, ppos, tok)) + blk = m->last; + + if ( ! mdoc_head_alloc(m, line, ppos, tok)) return(0); - mdoc->next = MDOC_NEXT_CHILD; - body = mdoc->last; + if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos)) + return(0); - /* XXX - no known argument macros. */ + /* + * Open the body scope "on-demand", that is, after we've + * processed all our the leading delimiters (open parenthesis, + * etc.). + */ - for (;;) { - lastarg = *pos; - c = mdoc_args(mdoc, line, pos, buf, tok, &p); + for (body = NULL; ; ) { + la = *pos; + c = mdoc_args(m, line, pos, buf, tok, &p); + assert(ARGS_PHRASE != c); if (ARGS_ERROR == c) return(0); - if (ARGS_PUNCT == c) - break; if (ARGS_EOLN == c) break; + if (ARGS_PUNCT == c) + break; - if (MDOC_MAX == (c = lookup(mdoc, tok, p))) { - if ( ! mdoc_word_alloc(mdoc, line, lastarg, p)) + if (NULL == body && 1 == mdoc_isdelim(p)) { + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; continue; } - if ( ! mdoc_macro(mdoc, c, line, lastarg, pos, buf)) + if (NULL == body) { + if ( ! mdoc_body_alloc(m, line, ppos, tok)) + return(0); + body = m->last; + } + + if (MDOC_MAX != (c = lookup(tok, p))) { + if ( ! mdoc_macro(m, c, line, la, pos, buf)) + return(0); + break; + } + + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); - break; } - /* - * Since we know what our context is, we can rewind directly to - * it. This allows us to accomodate for our scope being - * violated by another token. - */ + /* Clean-ups to leave in a consistent state. */ - for (n = mdoc->last; n; n = n->parent) + if (NULL == body) { + if ( ! mdoc_body_alloc(m, line, ppos, tok)) + return(0); + body = m->last; + } + + /* + * 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. + */ + for (n = m->last; n; n = n->parent) if (body == n) break; - if (NULL == n && ! mdoc_nwarn(mdoc, body, EIMPBRK)) - return(0); + if (NULL == n && ! mdoc_nwarn(m, body, EIMPBRK)) + return(0); - if (n && ! rew_last(mdoc, body)) + if (n && ! rew_last(m, body)) return(0); - if (1 == ppos && ! append_delims(mdoc, line, pos, buf)) + /* Standard appending of delimiters. */ + + if (1 == ppos && ! append_delims(m, line, pos, buf)) return(0); - if (n && ! rew_last(mdoc, blk)) + /* Rewind scope, if applicable. */ + + if (n && ! rew_last(m, blk)) return(0); return(1); } -/* - * Block partial-explicit macros. - */ static int blk_part_exp(MACRO_PROT_ARGS) { - int lastarg, flushed, j, c, maxargs; + int la, c; + struct mdoc_node *head; /* keep track of head */ + struct mdoc_node *body; /* keep track of body */ char *p; - flushed = 0; - /* - * Number of arguments (head arguments). Only `Eo' has these, + * The opening of an explicit macro having zero or more leading + * punctuation nodes; a head with optional single element (the + * case of `Eo'); and a body that may be empty. */ - switch (tok) { - case (MDOC_Eo): - maxargs = 1; - break; - default: - maxargs = 0; - break; - } - - if ( ! mdoc_block_alloc(mdoc, line, ppos, tok, NULL)) + if ( ! mdoc_block_alloc(m, line, ppos, tok, NULL)) return(0); - mdoc->next = MDOC_NEXT_CHILD; - if (0 == maxargs) { - if ( ! mdoc_head_alloc(mdoc, line, ppos, tok)) - return(0); - if ( ! rew_subblock(MDOC_HEAD, mdoc, - tok, line, ppos)) - return(0); - if ( ! mdoc_body_alloc(mdoc, line, ppos, tok)) - return(0); - flushed = 1; - } else if ( ! mdoc_head_alloc(mdoc, line, ppos, tok)) - return(0); + for (head = body = NULL; ; ) { + la = *pos; + c = mdoc_args(m, line, pos, buf, tok, &p); - mdoc->next = MDOC_NEXT_CHILD; - - for (j = 0; ; j++) { - lastarg = *pos; - if (j == maxargs && ! flushed) { - if ( ! rew_subblock(MDOC_HEAD, mdoc, - tok, line, ppos)) - return(0); - flushed = 1; - if ( ! mdoc_body_alloc(mdoc, line, ppos, tok)) - return(0); - mdoc->next = MDOC_NEXT_CHILD; - } - - c = mdoc_args(mdoc, line, pos, buf, tok, &p); - assert(ARGS_PHRASE != c); - if (ARGS_ERROR == c) return(0); if (ARGS_PUNCT == c) @@ -1162,67 +1165,96 @@ blk_part_exp(MACRO_PROT_ARGS) if (ARGS_EOLN == c) break; - if (MDOC_MAX != (c = lookup(mdoc, tok, p))) { - if ( ! flushed) { - if ( ! rew_subblock(MDOC_HEAD, mdoc, - tok, line, ppos)) - return(0); - flushed = 1; - if ( ! mdoc_body_alloc(mdoc, line, - ppos, tok)) - return(0); - mdoc->next = MDOC_NEXT_CHILD; - } - if ( ! mdoc_macro(mdoc, c, line, lastarg, - pos, buf)) + assert(ARGS_PHRASE != c); + + /* Flush out leading punctuation. */ + + if (NULL == head && 1 == mdoc_isdelim(p)) { + assert(NULL == body); + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); - break; + continue; + } + + if (NULL == head) { + assert(NULL == body); + if ( ! mdoc_head_alloc(m, line, ppos, tok)) + return(0); + head = m->last; } - if ( ! flushed && mdoc_isdelim(p)) { - if ( ! rew_subblock(MDOC_HEAD, mdoc, - tok, line, ppos)) + /* + * `Eo' gobbles any data into the head, but most other + * macros just immediately close out and begin the body. + */ + + if (NULL == body) { + assert(head); + /* No check whether it's a macro! */ + if (MDOC_Eo == tok) + if ( ! mdoc_word_alloc(m, line, la, p)) + return(0); + + if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos)) return(0); - flushed = 1; - if ( ! mdoc_body_alloc(mdoc, line, ppos, tok)) + if ( ! mdoc_body_alloc(m, line, ppos, tok)) return(0); - mdoc->next = MDOC_NEXT_CHILD; + body = m->last; + + if (MDOC_Eo == tok) + continue; } - - if ( ! mdoc_word_alloc(mdoc, line, lastarg, p)) + + assert(NULL != head && NULL != body); + + if (MDOC_MAX != (c = lookup(tok, p))) { + if ( ! mdoc_macro(m, c, line, la, pos, buf)) + return(0); + break; + } + + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; } - if ( ! flushed) { - if ( ! rew_subblock(MDOC_HEAD, mdoc, tok, line, ppos)) + /* Clean-up to leave in a consistent state. */ + + if (NULL == head) { + if ( ! mdoc_head_alloc(m, line, ppos, tok)) return(0); - if ( ! mdoc_body_alloc(mdoc, line, ppos, tok)) + head = m->last; + } + + if (NULL == body) { + if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos)) return(0); - mdoc->next = MDOC_NEXT_CHILD; + if ( ! mdoc_body_alloc(m, line, ppos, tok)) + return(0); + body = m->last; } + /* Standard appending of delimiters. */ + if (ppos > 1) return(1); - return(append_delims(mdoc, line, pos, buf)); + + return(append_delims(m, line, pos, buf)); } -/* - * In-line macros where reserved words signal closure of the macro. - * Macros also have a fixed number of arguments. - */ static int in_line_argn(MACRO_PROT_ARGS) { - int lastarg, flushed, j, c, maxargs; + int la, flushed, j, c, maxargs; struct mdoc_arg *arg; char *p; - - /* - * Fixed maximum arguments per macro. Some of these have none - * and close as soon as the invocation is parsed. + /* + * A line macro that has a fixed number of arguments (maxargs). + * Only open the scope once the first non-leading-punctuation is + * found (unless MDOC_IGNDELIM is noted, like in `Pf'), then + * keep it open until the maximum number of arguments are + * exhausted. */ switch (tok) { @@ -1235,17 +1267,20 @@ in_line_argn(MACRO_PROT_ARGS) case (MDOC_Ux): maxargs = 0; break; + case (MDOC_Xr): + maxargs = 2; + break; default: maxargs = 1; break; } - for (arg = NULL;; ) { - lastarg = *pos; - c = mdoc_argv(mdoc, line, tok, &arg, pos, buf); + for (arg = NULL; ; ) { + la = *pos; + c = mdoc_argv(m, line, tok, &arg, pos, buf); if (ARGV_WORD == c) { - *pos = lastarg; + *pos = la; break; } @@ -1258,21 +1293,10 @@ in_line_argn(MACRO_PROT_ARGS) return(0); } - if ( ! mdoc_elem_alloc(mdoc, line, ppos, tok, arg)) - return(0); - mdoc->next = MDOC_NEXT_CHILD; + for (flushed = j = 0; ; ) { + la = *pos; + c = mdoc_args(m, line, pos, buf, tok, &p); - for (flushed = j = 0; ; j++) { - lastarg = *pos; - - if (j == maxargs && ! flushed) { - if ( ! rew_elem(mdoc, tok)) - return(0); - flushed = 1; - } - - c = mdoc_args(mdoc, line, pos, buf, tok, &p); - if (ARGS_ERROR == c) return(0); if (ARGS_PUNCT == c) @@ -1280,40 +1304,73 @@ in_line_argn(MACRO_PROT_ARGS) if (ARGS_EOLN == c) break; - if (MDOC_MAX != (c = lookup(mdoc, tok, p))) { - if ( ! flushed && ! rew_elem(mdoc, tok)) + if ( ! (MDOC_IGNDELIM & mdoc_macros[tok].flags) && + 0 == j && 1 == mdoc_isdelim(p)) { + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); + continue; + } else if (0 == j) + if ( ! mdoc_elem_alloc(m, line, la, tok, arg)) + return(0); + + if (j == maxargs && ! flushed) { + if ( ! rew_elem(m, tok)) + return(0); flushed = 1; - if ( ! mdoc_macro(mdoc, c, line, lastarg, pos, buf)) + } + + if (MDOC_MAX != (c = lookup(tok, p))) { + if ( ! flushed && ! rew_elem(m, tok)) return(0); + flushed = 1; + if ( ! mdoc_macro(m, c, line, la, pos, buf)) + return(0); + j++; break; } if ( ! (MDOC_IGNDELIM & mdoc_macros[tok].flags) && ! flushed && mdoc_isdelim(p)) { - if ( ! rew_elem(mdoc, tok)) + if ( ! rew_elem(m, tok)) return(0); flushed = 1; } - - if ( ! mdoc_word_alloc(mdoc, line, lastarg, p)) + + /* + * 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)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; + j++; } - if ( ! flushed && ! rew_elem(mdoc, tok)) + if (0 == j && ! mdoc_elem_alloc(m, line, la, tok, arg)) + return(0); + + /* Close out in a consistent state. */ + + if ( ! flushed && ! rew_elem(m, tok)) return(0); if (ppos > 1) return(1); - return(append_delims(mdoc, line, pos, buf)); + return(append_delims(m, line, pos, buf)); } -/* - * In-line macro that spans an entire line. May be callable, but has no - * subsequent parsed arguments. - */ static int in_line_eoln(MACRO_PROT_ARGS) { @@ -1323,11 +1380,11 @@ in_line_eoln(MACRO_PROT_ARGS) assert( ! (MDOC_PARSED & mdoc_macros[tok].flags)); - arg = NULL; + /* Parse macro arguments. */ - for (;;) { + for (arg = NULL; ; ) { la = *pos; - c = mdoc_argv(mdoc, line, tok, &arg, pos, buf); + c = mdoc_argv(m, line, tok, &arg, pos, buf); if (ARGV_WORD == c) { *pos = la; @@ -1342,43 +1399,69 @@ in_line_eoln(MACRO_PROT_ARGS) return(0); } - if ( ! mdoc_elem_alloc(mdoc, line, ppos, tok, arg)) + /* Open element scope. */ + + if ( ! mdoc_elem_alloc(m, line, ppos, tok, arg)) return(0); - mdoc->next = MDOC_NEXT_CHILD; + /* Parse argument terms. */ for (;;) { la = *pos; - w = mdoc_args(mdoc, line, pos, buf, tok, &p); + w = mdoc_args(m, line, pos, buf, tok, &p); if (ARGS_ERROR == w) return(0); if (ARGS_EOLN == w) break; - c = ARGS_QWORD == w ? MDOC_MAX : lookup(mdoc, tok, p); + c = ARGS_QWORD == w ? MDOC_MAX : lookup(tok, p); if (MDOC_MAX != c) { - if ( ! rew_elem(mdoc, tok)) + if ( ! rew_elem(m, tok)) return(0); - return(mdoc_macro(mdoc, c, line, la, pos, buf)); + return(mdoc_macro(m, c, line, la, pos, buf)); } - if ( ! mdoc_word_alloc(mdoc, line, la, p)) + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; } - return(rew_elem(mdoc, tok)); + /* Close out (no delimiters). */ + + return(rew_elem(m, tok)); } /* ARGSUSED */ static int +ctx_synopsis(MACRO_PROT_ARGS) +{ + + /* If we're not in the SYNOPSIS, go straight to in-line. */ + if (SEC_SYNOPSIS != m->lastsec) + return(in_line(m, tok, line, ppos, pos, buf)); + + /* If we're a nested call, same place. */ + if (ppos > 1) + return(in_line(m, tok, line, ppos, pos, buf)); + + /* + * XXX: this will open a block scope; however, if later we end + * up formatting the block scope, then child nodes will inherit + * the formatting. Be careful. + */ + + return(blk_part_imp(m, tok, line, ppos, pos, buf)); +} + + +/* ARGSUSED */ +static int obsolete(MACRO_PROT_ARGS) { - return(mdoc_pwarn(mdoc, line, ppos, EOBS)); + return(mdoc_pwarn(m, line, ppos, EOBS)); } @@ -1388,7 +1471,7 @@ obsolete(MACRO_PROT_ARGS) * macro is encountered. */ static int -phrase(struct mdoc *mdoc, int line, int ppos, char *buf) +phrase(struct mdoc *m, int line, int ppos, char *buf) { int c, w, la, pos; char *p; @@ -1397,24 +1480,23 @@ phrase(struct mdoc *mdoc, int line, int ppos, char *bu la = pos; /* Note: no calling context! */ - w = mdoc_zargs(mdoc, line, &pos, buf, &p); + w = mdoc_zargs(m, line, &pos, buf, 0, &p); if (ARGS_ERROR == w) return(0); if (ARGS_EOLN == w) break; - c = ARGS_QWORD == w ? MDOC_MAX : lookup_raw(mdoc, p); + c = ARGS_QWORD == w ? MDOC_MAX : lookup_raw(p); if (MDOC_MAX != c) { - if ( ! mdoc_macro(mdoc, c, line, la, &pos, buf)) + if ( ! mdoc_macro(m, c, line, la, &pos, buf)) return(0); - return(append_delims(mdoc, line, &pos, buf)); + return(append_delims(m, line, &pos, buf)); } - if ( ! mdoc_word_alloc(mdoc, line, la, p)) + if ( ! mdoc_word_alloc(m, line, la, p)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; } return(1);