=================================================================== RCS file: /cvs/mandoc/mdoc.c,v retrieving revision 1.13 retrieving revision 1.245 diff -u -p -r1.13 -r1.245 --- mandoc/mdoc.c 2009/01/01 20:40:16 1.13 +++ mandoc/mdoc.c 2015/04/18 17:28:36 1.245 @@ -1,33 +1,41 @@ -/* $Id: mdoc.c,v 1.13 2009/01/01 20:40:16 kristaps Exp $ */ +/* $Id: mdoc.c,v 1.245 2015/04/18 17:28:36 schwarze Exp $ */ /* - * Copyright (c) 2008 Kristaps Dzonsons + * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons + * Copyright (c) 2010, 2012-2015 Ingo Schwarze * * Permission to use, copy, modify, and distribute this software for any - * purpose with or without fee is hereby granted, provided that the - * above copyright notice and this permission notice appear in all - * copies. + * purpose with or without fee is hereby granted, provided that the above + * copyright notice and this permission notice appear in all copies. * - * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL - * WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED - * WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE - * AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL - * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR - * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER - * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR - * PERFORMANCE OF THIS SOFTWARE. + * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES + * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF + * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR + * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN + * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF + * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ +#include "config.h" + +#include + #include #include -#include #include -#include #include +#include #include +#include -#include "private.h" +#include "mandoc_aux.h" +#include "mandoc.h" +#include "roff.h" +#include "mdoc.h" +#include "libmandoc.h" +#include "libmdoc.h" -const char *const __mdoc_macronames[MDOC_MAX] = { - "\\\"", "Dd", "Dt", "Os", +const char *const __mdoc_macronames[MDOC_MAX + 1] = { + "Ap", "Dd", "Dt", "Os", "Sh", "Ss", "Pp", "D1", "Dl", "Bd", "Ed", "Bl", "El", "It", "Ad", "An", @@ -37,12 +45,9 @@ const char *const __mdoc_macronames[MDOC_MAX] = { "Ic", "In", "Li", "Nd", "Nm", "Op", "Ot", "Pa", "Rv", "St", "Va", "Vt", - /* LINTED */ - "Xr", "\%A", "\%B", "\%D", - /* LINTED */ - "\%I", "\%J", "\%N", "\%O", - /* LINTED */ - "\%P", "\%R", "\%T", "\%V", + "Xr", "%A", "%B", "%D", + "%I", "%J", "%N", "%O", + "%P", "%R", "%T", "%V", "Ac", "Ao", "Aq", "At", "Bc", "Bf", "Bo", "Bq", "Bsx", "Bx", "Db", "Dc", @@ -56,619 +61,756 @@ const char *const __mdoc_macronames[MDOC_MAX] = { "Tn", "Ux", "Xc", "Xo", "Fo", "Fc", "Oo", "Oc", "Bk", "Ek", "Bt", "Hf", - "Fr", "Ud", + "Fr", "Ud", "Lb", "Lp", + "Lk", "Mt", "Brq", "Bro", + "Brc", "%C", "Es", "En", + "Dx", "%Q", "br", "sp", + "%U", "Ta", "ll", "text", }; -const char *const __mdoc_argnames[MDOC_ARG_MAX] = { +const char *const __mdoc_argnames[MDOC_ARG_MAX] = { "split", "nosplit", "ragged", - "unfilled", "literal", "file", - "offset", "bullet", "dash", - "hyphen", "item", "enum", - "tag", "diag", "hang", - "ohang", "inset", "column", - "width", "compact", "std", - "p1003.1-88", "p1003.1-90", "p1003.1-96", - "p1003.1-2001", "p1003.1-2004", "p1003.1", - "p1003.1b", "p1003.1b-93", "p1003.1c-95", - "p1003.1g-2000", "p1003.2-92", "p1387.2-95", - "p1003.2", "p1387.2", "isoC-90", - "isoC-amd1", "isoC-tcor1", "isoC-tcor2", - "isoC-99", "ansiC", "ansiC-89", - "ansiC-99", "ieee754", "iso8802-3", - "xpg3", "xpg4", "xpg4.2", - "xpg4.3", "xbd5", "xcu5", - "xsh5", "xns5", "xns5.2d2.0", - "xcurses4.2", "susv2", "susv3", - "svid4", "filled", "words", + "unfilled", "literal", "file", + "offset", "bullet", "dash", + "hyphen", "item", "enum", + "tag", "diag", "hang", + "ohang", "inset", "column", + "width", "compact", "std", + "filled", "words", "emphasis", + "symbolic", "nested", "centered" }; -const struct mdoc_macro __mdoc_macros[MDOC_MAX] = { - { NULL, 0 }, /* \" */ - { NULL, MDOC_PROLOGUE }, /* Dd */ - { NULL, MDOC_PROLOGUE }, /* Dt */ - { NULL, MDOC_PROLOGUE }, /* Os */ - { macro_scoped, 0 }, /* Sh */ - { macro_scoped, 0 }, /* Ss */ - { macro_text, 0 }, /* Pp */ - { macro_scoped_line, 0 }, /* D1 */ - { macro_scoped_line, 0 }, /* Dl */ - { macro_scoped, MDOC_EXPLICIT }, /* Bd */ - { macro_close_explicit, 0 }, /* Ed */ - { macro_scoped, MDOC_EXPLICIT }, /* Bl */ - { macro_close_explicit, 0 }, /* El */ - { NULL, 0 }, /* It */ - { macro_text, MDOC_CALLABLE }, /* Ad */ - { NULL, 0 }, /* An */ - { macro_text, MDOC_CALLABLE }, /* Ar */ - { macro_constant, MDOC_QUOTABLE }, /* Cd */ - { macro_text, MDOC_CALLABLE }, /* Cm */ - { macro_text, MDOC_CALLABLE }, /* Dv */ - { macro_text, MDOC_CALLABLE }, /* Er */ - { macro_text, MDOC_CALLABLE }, /* Ev */ - { macro_constant, 0 }, /* Ex */ - { macro_text, MDOC_CALLABLE }, /* Fa */ - { macro_constant, 0 }, /* Fd */ - { macro_text, MDOC_CALLABLE }, /* Fl */ - { macro_text, MDOC_CALLABLE | MDOC_QUOTABLE }, /* Fn */ - { macro_text, 0 }, /* Ft */ - { macro_text, MDOC_CALLABLE }, /* Ic */ - { macro_constant, 0 }, /* In */ - { macro_text, MDOC_CALLABLE }, /* Li */ - { macro_constant, 0 }, /* Nd */ - { NULL, 0 }, /* Nm */ - { macro_scoped_line, MDOC_CALLABLE }, /* Op */ - { macro_obsolete, 0 }, /* Ot */ - { macro_text, MDOC_CALLABLE }, /* Pa */ - { macro_constant, 0 }, /* Rv */ - { macro_constant, 0 }, /* St */ - { macro_text, MDOC_CALLABLE }, /* Va */ - { macro_text, MDOC_CALLABLE }, /* Vt */ - { macro_text, MDOC_CALLABLE }, /* Xr */ - { macro_constant, MDOC_QUOTABLE }, /* %A */ - { macro_constant, MDOC_QUOTABLE }, /* %B */ - { macro_constant, MDOC_QUOTABLE }, /* %D */ - { macro_constant, MDOC_QUOTABLE }, /* %I */ - { macro_constant, MDOC_QUOTABLE }, /* %J */ - { macro_constant, MDOC_QUOTABLE }, /* %N */ - { macro_constant, MDOC_QUOTABLE }, /* %O */ - { macro_constant, MDOC_QUOTABLE }, /* %P */ - { macro_constant, MDOC_QUOTABLE }, /* %R */ - { macro_constant, MDOC_QUOTABLE }, /* %T */ - { macro_constant, MDOC_QUOTABLE }, /* %V */ - { NULL, 0 }, /* Ac */ - { NULL, 0 }, /* Ao */ - { macro_scoped_line, MDOC_CALLABLE }, /* Aq */ - { macro_constant, 0 }, /* At */ - { NULL, 0 }, /* Bc */ - { NULL, 0 }, /* Bf */ - { NULL, 0 }, /* Bo */ - { macro_scoped_line, MDOC_CALLABLE }, /* Bq */ - { macro_constant_delimited, 0 }, /* Bsx */ - { macro_constant_delimited, 0 }, /* Bx */ - { macro_constant, 0 }, /* Db */ - { NULL, 0 }, /* Dc */ - { NULL, 0 }, /* Do */ - { macro_scoped_line, MDOC_CALLABLE }, /* Dq */ - { NULL, 0 }, /* Ec */ - { NULL, 0 }, /* Ef */ - { macro_text, MDOC_CALLABLE }, /* Em */ - { NULL, 0 }, /* Eo */ - { macro_constant_delimited, 0 }, /* Fx */ - { macro_text, 0 }, /* Ms */ - { macro_constant_delimited, MDOC_CALLABLE }, /* No */ - { macro_constant_delimited, MDOC_CALLABLE }, /* Ns */ - { macro_constant_delimited, 0 }, /* Nx */ - { macro_constant_delimited, 0 }, /* Ox */ - { NULL, 0 }, /* Pc */ - { NULL, 0 }, /* Pf */ - { NULL, 0 }, /* Po */ - { macro_scoped_line, MDOC_CALLABLE }, /* Pq */ - { NULL, 0 }, /* Qc */ - { macro_scoped_line, MDOC_CALLABLE }, /* Ql */ - { NULL, 0 }, /* Qo */ - { macro_scoped_line, MDOC_CALLABLE }, /* Qq */ - { macro_scoped, MDOC_EXPLICIT }, /* Re */ - { macro_scoped, MDOC_EXPLICIT }, /* Rs */ - { NULL, 0 }, /* Sc */ - { NULL, 0 }, /* So */ - { macro_scoped_line, MDOC_CALLABLE }, /* Sq */ - { macro_constant, 0 }, /* Sm */ - { macro_text, MDOC_CALLABLE }, /* Sx */ - { macro_text, MDOC_CALLABLE }, /* Sy */ - { macro_text, MDOC_CALLABLE }, /* Tn */ - { macro_constant_delimited, 0 }, /* Ux */ - { NULL, 0 }, /* Xc */ - { NULL, 0 }, /* Xo */ - { NULL, 0 }, /* Fo */ - { NULL, 0 }, /* Fc */ - { NULL, 0 }, /* Oo */ - { NULL, 0 }, /* Oc */ - { NULL, 0 }, /* Bk */ - { NULL, 0 }, /* Ek */ - { macro_constant, 0 }, /* Bt */ - { macro_constant, 0 }, /* Hf */ - { macro_obsolete, 0 }, /* Fr */ - { macro_constant, 0 }, /* Ud */ -}; - const char * const *mdoc_macronames = __mdoc_macronames; const char * const *mdoc_argnames = __mdoc_argnames; -const struct mdoc_macro * const mdoc_macros = __mdoc_macros; +static void mdoc_node_free(struct roff_node *); +static void mdoc_node_unlink(struct roff_man *, + struct roff_node *); +static struct roff_node *node_alloc(struct roff_man *, int, int, + int, enum roff_type); +static void node_append(struct roff_man *, struct roff_node *); +static int mdoc_ptext(struct roff_man *, int, char *, int); +static int mdoc_pmacro(struct roff_man *, int, char *, int); -static struct mdoc_arg *argdup(size_t, const struct mdoc_arg *); -static void argfree(size_t, struct mdoc_arg *); -static void argcpy(struct mdoc_arg *, - const struct mdoc_arg *); -static char **paramdup(size_t, const char **); -static void paramfree(size_t, char **); -static void mdoc_node_freelist(struct mdoc_node *); -static void mdoc_node_append(struct mdoc *, int, - struct mdoc_node *); -static void mdoc_elem_free(struct mdoc_elem *); -static void mdoc_text_free(struct mdoc_text *); - - -const struct mdoc_node * -mdoc_result(struct mdoc *mdoc) +const struct roff_node * +mdoc_node(const struct roff_man *mdoc) { return(mdoc->first); } - -void -mdoc_free(struct mdoc *mdoc) +const struct roff_meta * +mdoc_meta(const struct roff_man *mdoc) { - if (mdoc->first) - mdoc_node_freelist(mdoc->first); - if (mdoc->htab) - mdoc_tokhash_free(mdoc->htab); - - free(mdoc); + return(&mdoc->meta); } - -struct mdoc * -mdoc_alloc(void *data, const struct mdoc_cb *cb) +void +mdoc_endparse(struct roff_man *mdoc) { - struct mdoc *p; - p = xcalloc(1, sizeof(struct mdoc)); - - p->data = data; - (void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb)); - - p->htab = mdoc_tokhash_alloc(); - return(p); + mdoc_macroend(mdoc); } - -int -mdoc_parseln(struct mdoc *mdoc, char *buf) +void +mdoc_addeqn(struct roff_man *mdoc, const struct eqn *ep) { - int c, i; - char tmp[5]; + struct roff_node *n; - if ('.' != *buf) { - mdoc_word_alloc(mdoc, 0, buf); - mdoc->next = MDOC_NEXT_SIBLING; - return(1); - } - - if (buf[1] && '\\' == buf[1]) - if (buf[2] && '\"' == buf[2]) - return(1); - - i = 1; - while (buf[i] && ! isspace(buf[i]) && i < (int)sizeof(tmp)) - i++; - - if (i == (int)sizeof(tmp)) - return(mdoc_err(mdoc, -1, 1, ERR_MACRO_NOTSUP)); - else if (i <= 2) - return(mdoc_err(mdoc, -1, 1, ERR_MACRO_NOTSUP)); - - i--; - - (void)memcpy(tmp, buf + 1, (size_t)i); - tmp[i++] = 0; - - if (MDOC_MAX == (c = mdoc_find(mdoc, tmp))) - return(mdoc_err(mdoc, c, 1, ERR_MACRO_NOTSUP)); - - while (buf[i] && isspace(buf[i])) - i++; - - return(mdoc_macro(mdoc, c, 1, &i, buf)); + n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, ROFFT_EQN); + n->eqn = ep; + if (ep->ln > mdoc->last->line) + n->flags |= MDOC_LINE; + node_append(mdoc, n); + mdoc->next = ROFF_NEXT_SIBLING; } - void -mdoc_msg(struct mdoc *mdoc, int pos, const char *fmt, ...) +mdoc_addspan(struct roff_man *mdoc, const struct tbl_span *sp) { - va_list ap; - char buf[256]; + struct roff_node *n; - if (NULL == mdoc->cb.mdoc_msg) - return; - - va_start(ap, fmt); - (void)vsnprintf(buf, sizeof(buf), fmt, ap); - va_end(ap); - - (*mdoc->cb.mdoc_msg)(mdoc->data, pos, buf); + n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, ROFFT_TBL); + n->span = sp; + node_append(mdoc, n); + mdoc->next = ROFF_NEXT_SIBLING; } - +/* + * Main parse routine. Parses a single line -- really just hands off to + * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()). + */ int -mdoc_err(struct mdoc *mdoc, int tok, int pos, enum mdoc_err type) +mdoc_parseln(struct roff_man *mdoc, int ln, char *buf, int offs) { - if (NULL == mdoc->cb.mdoc_err) - return(0); - return((*mdoc->cb.mdoc_err)(mdoc->data, tok, pos, type)); -} + if (mdoc->last->type != ROFFT_EQN || ln > mdoc->last->line) + mdoc->flags |= MDOC_NEWLINE; + /* + * Let the roff nS register switch SYNOPSIS mode early, + * such that the parser knows at all times + * whether this mode is on or off. + * Note that this mode is also switched by the Sh macro. + */ + if (roff_getreg(mdoc->roff, "nS")) + mdoc->flags |= MDOC_SYNOPSIS; + else + mdoc->flags &= ~MDOC_SYNOPSIS; -int -mdoc_warn(struct mdoc *mdoc, int tok, int pos, enum mdoc_warn type) -{ - - if (NULL == mdoc->cb.mdoc_warn) - return(0); - return((*mdoc->cb.mdoc_warn)(mdoc->data, tok, pos, type)); + return(roff_getcontrol(mdoc->roff, buf, &offs) ? + mdoc_pmacro(mdoc, ln, buf, offs) : + mdoc_ptext(mdoc, ln, buf, offs)); } - -int -mdoc_macro(struct mdoc *mdoc, int tok, int ppos, int *pos, char *buf) +void +mdoc_macro(MACRO_PROT_ARGS) { + assert(tok < MDOC_MAX); - if (MDOC_PROLOGUE & mdoc_macros[tok].flags && - SEC_PROLOGUE == mdoc->sec_lastn) - return(macro_prologue(mdoc, tok, ppos, pos, buf)); - - if (SEC_PROLOGUE == mdoc->sec_lastn) - return(mdoc_err(mdoc, tok, ppos, ERR_SEC_PROLOGUE)); - - if (NULL == (mdoc_macros[tok].fp)) { - (void)mdoc_err(mdoc, tok, ppos, ERR_MACRO_NOTSUP); - return(0); + if (mdoc->flags & MDOC_PBODY) { + if (tok == MDOC_Dt) { + mandoc_vmsg(MANDOCERR_DT_LATE, + mdoc->parse, line, ppos, + "Dt %s", buf + *pos); + return; + } + } else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) { + if (mdoc->meta.title == NULL) { + mandoc_vmsg(MANDOCERR_DT_NOTITLE, + mdoc->parse, line, ppos, "%s %s", + mdoc_macronames[tok], buf + *pos); + mdoc->meta.title = mandoc_strdup("UNTITLED"); + } + if (NULL == mdoc->meta.vol) + mdoc->meta.vol = mandoc_strdup("LOCAL"); + mdoc->flags |= MDOC_PBODY; } - - if (1 != ppos && ! (MDOC_CALLABLE & mdoc_macros[tok].flags)) { - (void)mdoc_err(mdoc, tok, ppos, ERR_MACRO_NOTCALL); - return(0); - } - - return((*mdoc_macros[tok].fp)(mdoc, tok, ppos, pos, buf)); + (*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf); } static void -mdoc_node_append(struct mdoc *mdoc, int pos, struct mdoc_node *p) +node_append(struct roff_man *mdoc, struct roff_node *p) { - const char *nn, *on, *nt, *ot, *act; - switch (p->type) { - case (MDOC_TEXT): - nn = p->data.text.string; - nt = "text"; + assert(mdoc->last); + assert(mdoc->first); + assert(p->type != ROFFT_ROOT); + + switch (mdoc->next) { + case ROFF_NEXT_SIBLING: + mdoc->last->next = p; + p->prev = mdoc->last; + p->parent = mdoc->last->parent; break; - case (MDOC_BODY): - nn = mdoc_macronames[p->data.body.tok]; - nt = "body"; + case ROFF_NEXT_CHILD: + mdoc->last->child = p; + p->parent = mdoc->last; break; - case (MDOC_ELEM): - nn = mdoc_macronames[p->data.elem.tok]; - nt = "elem"; - break; - case (MDOC_HEAD): - nn = mdoc_macronames[p->data.head.tok]; - nt = "head"; - break; - case (MDOC_BLOCK): - nn = mdoc_macronames[p->data.block.tok]; - nt = "block"; - break; default: abort(); /* NOTREACHED */ } - if (NULL == mdoc->first) { - assert(NULL == mdoc->last); - mdoc->first = p; - mdoc->last = p; - mdoc_msg(mdoc, pos, "parse: root %s `%s'", nt, nn); - return; - } + p->parent->nchild++; - switch (mdoc->last->type) { - case (MDOC_TEXT): - on = ""; - ot = "text"; + /* + * Copy over the normalised-data pointer of our parent. Not + * everybody has one, but copying a null pointer is fine. + */ + + switch (p->type) { + case ROFFT_BODY: + if (ENDBODY_NOT != p->end) + break; + /* FALLTHROUGH */ + case ROFFT_TAIL: + /* FALLTHROUGH */ + case ROFFT_HEAD: + p->norm = p->parent->norm; break; - case (MDOC_BODY): - on = mdoc_macronames[mdoc->last->data.body.tok]; - ot = "body"; + default: break; - case (MDOC_ELEM): - on = mdoc_macronames[mdoc->last->data.elem.tok]; - ot = "elem"; + } + + mdoc_valid_pre(mdoc, p); + + switch (p->type) { + case ROFFT_HEAD: + assert(p->parent->type == ROFFT_BLOCK); + p->parent->head = p; break; - case (MDOC_HEAD): - on = mdoc_macronames[mdoc->last->data.head.tok]; - ot = "head"; + case ROFFT_TAIL: + assert(p->parent->type == ROFFT_BLOCK); + p->parent->tail = p; break; - case (MDOC_BLOCK): - on = mdoc_macronames[mdoc->last->data.block.tok]; - ot = "block"; + case ROFFT_BODY: + if (p->end) + break; + assert(p->parent->type == ROFFT_BLOCK); + p->parent->body = p; break; default: - abort(); - /* NOTREACHED */ + break; } - switch (mdoc->next) { - case (MDOC_NEXT_SIBLING): - mdoc->last->next = p; - p->prev = mdoc->last; - p->parent = mdoc->last->parent; - act = "sibling"; + mdoc->last = p; + + switch (p->type) { + case ROFFT_TBL: + /* FALLTHROUGH */ + case ROFFT_TEXT: + mdoc_valid_post(mdoc); break; - case (MDOC_NEXT_CHILD): - mdoc->last->child = p; - p->parent = mdoc->last; - act = "child"; - break; default: - abort(); - /* NOTREACHED */ + break; } +} - mdoc_msg(mdoc, pos, "parse: %s `%s' %s of %s `%s'", - nt, nn, act, ot, on); +static struct roff_node * +node_alloc(struct roff_man *mdoc, int line, int pos, + int tok, enum roff_type type) +{ + struct roff_node *p; - mdoc->last = p; -} + p = mandoc_calloc(1, sizeof(*p)); + p->sec = mdoc->lastsec; + p->line = line; + p->pos = pos; + p->tok = tok; + p->type = type; + /* Flag analysis. */ + if (MDOC_SYNOPSIS & mdoc->flags) + p->flags |= MDOC_SYNPRETTY; + else + p->flags &= ~MDOC_SYNPRETTY; + if (MDOC_NEWLINE & mdoc->flags) + p->flags |= MDOC_LINE; + mdoc->flags &= ~MDOC_NEWLINE; + + return(p); +} + void -mdoc_head_alloc(struct mdoc *mdoc, int pos, int tok) +mdoc_tail_alloc(struct roff_man *mdoc, int line, int pos, int tok) { - struct mdoc_node *p; + struct roff_node *p; + p = node_alloc(mdoc, line, pos, tok, ROFFT_TAIL); + node_append(mdoc, p); + mdoc->next = ROFF_NEXT_CHILD; +} + +struct roff_node * +mdoc_head_alloc(struct roff_man *mdoc, int line, int pos, int tok) +{ + struct roff_node *p; + assert(mdoc->first); assert(mdoc->last); + p = node_alloc(mdoc, line, pos, tok, ROFFT_HEAD); + node_append(mdoc, p); + mdoc->next = ROFF_NEXT_CHILD; + return(p); +} - p = xcalloc(1, sizeof(struct mdoc_node)); +struct roff_node * +mdoc_body_alloc(struct roff_man *mdoc, int line, int pos, int tok) +{ + struct roff_node *p; - p->type = MDOC_HEAD; - p->data.head.tok = tok; + p = node_alloc(mdoc, line, pos, tok, ROFFT_BODY); + node_append(mdoc, p); + mdoc->next = ROFF_NEXT_CHILD; + return(p); +} - mdoc_node_append(mdoc, pos, p); +struct roff_node * +mdoc_endbody_alloc(struct roff_man *mdoc, int line, int pos, int tok, + struct roff_node *body, enum mdoc_endbody end) +{ + struct roff_node *p; + + body->flags |= MDOC_ENDED; + body->parent->flags |= MDOC_ENDED; + p = node_alloc(mdoc, line, pos, tok, ROFFT_BODY); + p->body = body; + p->norm = body->norm; + p->end = end; + node_append(mdoc, p); + mdoc->next = ROFF_NEXT_SIBLING; + return(p); } +struct roff_node * +mdoc_block_alloc(struct roff_man *mdoc, int line, int pos, + int tok, struct mdoc_arg *args) +{ + struct roff_node *p; + p = node_alloc(mdoc, line, pos, tok, ROFFT_BLOCK); + p->args = args; + if (p->args) + (args->refcnt)++; + + switch (tok) { + case MDOC_Bd: + /* FALLTHROUGH */ + case MDOC_Bf: + /* FALLTHROUGH */ + case MDOC_Bl: + /* FALLTHROUGH */ + case MDOC_En: + /* FALLTHROUGH */ + case MDOC_Rs: + p->norm = mandoc_calloc(1, sizeof(union mdoc_data)); + break; + default: + break; + } + node_append(mdoc, p); + mdoc->next = ROFF_NEXT_CHILD; + return(p); +} + void -mdoc_body_alloc(struct mdoc *mdoc, int pos, int tok) +mdoc_elem_alloc(struct roff_man *mdoc, int line, int pos, + int tok, struct mdoc_arg *args) { - struct mdoc_node *p; + struct roff_node *p; - assert(mdoc->first); - assert(mdoc->last); + p = node_alloc(mdoc, line, pos, tok, ROFFT_ELEM); + p->args = args; + if (p->args) + (args->refcnt)++; - p = xcalloc(1, sizeof(struct mdoc_node)); + switch (tok) { + case MDOC_An: + p->norm = mandoc_calloc(1, sizeof(union mdoc_data)); + break; + default: + break; + } + node_append(mdoc, p); + mdoc->next = ROFF_NEXT_CHILD; +} - p->type = MDOC_BODY; - p->data.body.tok = tok; +void +mdoc_word_alloc(struct roff_man *mdoc, int line, int pos, const char *p) +{ + struct roff_node *n; - mdoc_node_append(mdoc, pos, p); + n = node_alloc(mdoc, line, pos, MDOC_MAX, ROFFT_TEXT); + n->string = roff_strdup(mdoc->roff, p); + node_append(mdoc, n); + mdoc->next = ROFF_NEXT_SIBLING; } - void -mdoc_block_alloc(struct mdoc *mdoc, int pos, int tok, - size_t argsz, const struct mdoc_arg *args) +mdoc_word_append(struct roff_man *mdoc, const char *p) { - struct mdoc_node *p; + struct roff_node *n; + char *addstr, *newstr; - p = xcalloc(1, sizeof(struct mdoc_node)); + n = mdoc->last; + addstr = roff_strdup(mdoc->roff, p); + mandoc_asprintf(&newstr, "%s %s", n->string, addstr); + free(addstr); + free(n->string); + n->string = newstr; + mdoc->next = ROFF_NEXT_SIBLING; +} - p->type = MDOC_BLOCK; - p->data.block.tok = tok; - p->data.block.argc = argsz; - p->data.block.argv = argdup(argsz, args); +static void +mdoc_node_free(struct roff_node *p) +{ - mdoc_node_append(mdoc, pos, p); + if (p->type == ROFFT_BLOCK || p->type == ROFFT_ELEM) + free(p->norm); + if (p->string) + free(p->string); + if (p->args) + mdoc_argv_free(p->args); + free(p); } - -void -mdoc_elem_alloc(struct mdoc *mdoc, int pos, int tok, - size_t argsz, const struct mdoc_arg *args, - size_t paramsz, const char **params) +static void +mdoc_node_unlink(struct roff_man *mdoc, struct roff_node *n) { - struct mdoc_node *p; - p = xcalloc(1, sizeof(struct mdoc_node)); - p->type = MDOC_ELEM; - p->data.elem.tok = tok; - p->data.elem.sz = paramsz; - p->data.elem.args = paramdup(paramsz, params); - p->data.elem.argc = argsz; - p->data.elem.argv = argdup(argsz, args); + /* Adjust siblings. */ - mdoc_node_append(mdoc, pos, p); -} + if (n->prev) + n->prev->next = n->next; + if (n->next) + n->next->prev = n->prev; + /* Adjust parent. */ + if (n->parent) { + n->parent->nchild--; + if (n->parent->child == n) + n->parent->child = n->prev ? n->prev : n->next; + if (n->parent->last == n) + n->parent->last = n->prev ? n->prev : NULL; + } + + /* Adjust parse point, if applicable. */ + + if (mdoc && mdoc->last == n) { + if (n->prev) { + mdoc->last = n->prev; + mdoc->next = ROFF_NEXT_SIBLING; + } else { + mdoc->last = n->parent; + mdoc->next = ROFF_NEXT_CHILD; + } + } + + if (mdoc && mdoc->first == n) + mdoc->first = NULL; +} + void -mdoc_word_alloc(struct mdoc *mdoc, int pos, const char *word) +mdoc_node_delete(struct roff_man *mdoc, struct roff_node *p) { - struct mdoc_node *p; - p = xcalloc(1, sizeof(struct mdoc_node)); - p->type = MDOC_TEXT; - p->data.text.string = xstrdup(word); + while (p->child) { + assert(p->nchild); + mdoc_node_delete(mdoc, p->child); + } + assert(0 == p->nchild); - mdoc_node_append(mdoc, pos, p); + mdoc_node_unlink(mdoc, p); + mdoc_node_free(p); } +void +mdoc_node_relink(struct roff_man *mdoc, struct roff_node *p) +{ -static void -argfree(size_t sz, struct mdoc_arg *p) + mdoc_node_unlink(mdoc, p); + node_append(mdoc, p); +} + +/* + * Parse free-form text, that is, a line that does not begin with the + * control character. + */ +static int +mdoc_ptext(struct roff_man *mdoc, int line, char *buf, int offs) { - int i, j; + struct roff_node *n; + char *c, *ws, *end; - if (0 == sz) - return; + assert(mdoc->last); + n = mdoc->last; - assert(p); - /* LINTED */ - for (i = 0; i < (int)sz; i++) - if (p[i].sz > 0) { - assert(p[i].value); - /* LINTED */ - for (j = 0; j < (int)p[i].sz; j++) - free(p[i].value[j]); + /* + * Divert directly to list processing if we're encountering a + * columnar ROFFT_BLOCK with or without a prior ROFFT_BLOCK entry + * (a ROFFT_BODY means it's already open, in which case we should + * process within its context in the normal way). + */ + + if (n->tok == MDOC_Bl && n->type == ROFFT_BODY && + n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) { + /* `Bl' is open without any children. */ + mdoc->flags |= MDOC_FREECOL; + mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf); + return(1); + } + + if (n->tok == MDOC_It && n->type == ROFFT_BLOCK && + NULL != n->parent && + MDOC_Bl == n->parent->tok && + LIST_column == n->parent->norm->Bl.type) { + /* `Bl' has block-level `It' children. */ + mdoc->flags |= MDOC_FREECOL; + mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf); + return(1); + } + + /* + * Search for the beginning of unescaped trailing whitespace (ws) + * and for the first character not to be output (end). + */ + + /* FIXME: replace with strcspn(). */ + ws = NULL; + for (c = end = buf + offs; *c; c++) { + switch (*c) { + case ' ': + if (NULL == ws) + ws = c; + continue; + case '\t': + /* + * Always warn about trailing tabs, + * even outside literal context, + * where they should be put on the next line. + */ + if (NULL == ws) + ws = c; + /* + * Strip trailing tabs in literal context only; + * outside, they affect the next line. + */ + if (MDOC_LITERAL & mdoc->flags) + continue; + break; + case '\\': + /* Skip the escaped character, too, if any. */ + if (c[1]) + c++; + /* FALLTHROUGH */ + default: + ws = NULL; + break; } - free(p); -} + end = c + 1; + } + *end = '\0'; + if (ws) + mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse, + line, (int)(ws-buf), NULL); -static void -mdoc_elem_free(struct mdoc_elem *p) -{ + if (buf[offs] == '\0' && ! (mdoc->flags & MDOC_LITERAL)) { + mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse, + line, (int)(c - buf), NULL); - paramfree(p->sz, p->args); - argfree(p->argc, p->argv); -} + /* + * Insert a `sp' in the case of a blank line. Technically, + * blank lines aren't allowed, but enough manuals assume this + * behaviour that we want to work around it. + */ + mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL); + mdoc->next = ROFF_NEXT_SIBLING; + mdoc_valid_post(mdoc); + return(1); + } + mdoc_word_alloc(mdoc, line, offs, buf+offs); -static void -mdoc_block_free(struct mdoc_block *p) -{ + if (mdoc->flags & MDOC_LITERAL) + return(1); - argfree(p->argc, p->argv); -} + /* + * End-of-sentence check. If the last character is an unescaped + * EOS character, then flag the node as being the end of a + * sentence. The front-end will know how to interpret this. + */ + assert(buf < end); -static void -mdoc_text_free(struct mdoc_text *p) + if (mandoc_eos(buf+offs, (size_t)(end-buf-offs))) + mdoc->last->flags |= MDOC_EOS; + return(1); +} + +/* + * Parse a macro line, that is, a line beginning with the control + * character. + */ +static int +mdoc_pmacro(struct roff_man *mdoc, int ln, char *buf, int offs) { + struct roff_node *n; + const char *cp; + int tok; + int i, sv; + char mac[5]; - if (p->string) - free(p->string); -} + sv = offs; + /* + * Copy the first word into a nil-terminated buffer. + * Stop when a space, tab, escape, or eoln is encountered. + */ -void -mdoc_node_free(struct mdoc_node *p) -{ + i = 0; + while (i < 4 && strchr(" \t\\", buf[offs]) == NULL) + mac[i++] = buf[offs++]; - switch (p->type) { - case (MDOC_TEXT): - mdoc_text_free(&p->data.text); + mac[i] = '\0'; + + tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX; + + if (tok == MDOC_MAX) { + mandoc_msg(MANDOCERR_MACRO, mdoc->parse, + ln, sv, buf + sv - 1); + return(1); + } + + /* Skip a leading escape sequence or tab. */ + + switch (buf[offs]) { + case '\\': + cp = buf + offs + 1; + mandoc_escape(&cp, NULL, NULL); + offs = cp - buf; break; - case (MDOC_ELEM): - mdoc_elem_free(&p->data.elem); + case '\t': + offs++; break; - case (MDOC_BLOCK): - mdoc_block_free(&p->data.block); - break; default: break; } - free(p); -} + /* Jump to the next non-whitespace word. */ + while (buf[offs] && ' ' == buf[offs]) + offs++; -static void -mdoc_node_freelist(struct mdoc_node *p) -{ + /* + * Trailing whitespace. Note that tabs are allowed to be passed + * into the parser as "text", so we only warn about spaces here. + */ - if (p->child) - mdoc_node_freelist(p->child); - if (p->next) - mdoc_node_freelist(p->next); + if ('\0' == buf[offs] && ' ' == buf[offs - 1]) + mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse, + ln, offs - 1, NULL); - mdoc_node_free(p); -} + /* + * If an initial macro or a list invocation, divert directly + * into macro processing. + */ + if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) { + mdoc_macro(mdoc, tok, ln, sv, &offs, buf); + return(1); + } -int -mdoc_find(const struct mdoc *mdoc, const char *key) -{ + n = mdoc->last; + assert(mdoc->last); - return(mdoc_tokhash_find(mdoc->htab, key)); -} + /* + * If the first macro of a `Bl -column', open an `It' block + * context around the parsed macro. + */ + if (n->tok == MDOC_Bl && n->type == ROFFT_BODY && + n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) { + mdoc->flags |= MDOC_FREECOL; + mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf); + return(1); + } -static void -argcpy(struct mdoc_arg *dst, const struct mdoc_arg *src) -{ - int i; + /* + * If we're following a block-level `It' within a `Bl -column' + * context (perhaps opened in the above block or in ptext()), + * then open an `It' block context around the parsed macro. + */ - dst->arg = src->arg; - if (0 == (dst->sz = src->sz)) - return; - dst->value = xcalloc(dst->sz, sizeof(char *)); - for (i = 0; i < (int)dst->sz; i++) - dst->value[i] = xstrdup(src->value[i]); -} + if (n->tok == MDOC_It && n->type == ROFFT_BLOCK && + NULL != n->parent && + MDOC_Bl == n->parent->tok && + LIST_column == n->parent->norm->Bl.type) { + mdoc->flags |= MDOC_FREECOL; + mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf); + return(1); + } + /* Normal processing of a macro. */ -static struct mdoc_arg * -argdup(size_t argsz, const struct mdoc_arg *args) -{ - struct mdoc_arg *pp; - int i; + mdoc_macro(mdoc, tok, ln, sv, &offs, buf); - if (0 == argsz) - return(NULL); + /* In quick mode (for mandocdb), abort after the NAME section. */ - pp = xcalloc((size_t)argsz, sizeof(struct mdoc_arg)); - for (i = 0; i < (int)argsz; i++) - argcpy(&pp[i], &args[i]); + if (mdoc->quick && MDOC_Sh == tok && + SEC_NAME != mdoc->last->sec) + return(2); - return(pp); + return(1); } +enum mdelim +mdoc_isdelim(const char *p) +{ -static void -paramfree(size_t sz, char **p) + if ('\0' == p[0]) + return(DELIM_NONE); + + if ('\0' == p[1]) + switch (p[0]) { + case '(': + /* FALLTHROUGH */ + case '[': + return(DELIM_OPEN); + case '|': + return(DELIM_MIDDLE); + case '.': + /* FALLTHROUGH */ + case ',': + /* FALLTHROUGH */ + case ';': + /* FALLTHROUGH */ + case ':': + /* FALLTHROUGH */ + case '?': + /* FALLTHROUGH */ + case '!': + /* FALLTHROUGH */ + case ')': + /* FALLTHROUGH */ + case ']': + return(DELIM_CLOSE); + default: + return(DELIM_NONE); + } + + if ('\\' != p[0]) + return(DELIM_NONE); + + if (0 == strcmp(p + 1, ".")) + return(DELIM_CLOSE); + if (0 == strcmp(p + 1, "fR|\\fP")) + return(DELIM_MIDDLE); + + return(DELIM_NONE); +} + +void +mdoc_deroff(char **dest, const struct roff_node *n) { - int i; + char *cp; + size_t sz; - if (0 == sz) + if (n->type != ROFFT_TEXT) { + for (n = n->child; n; n = n->next) + mdoc_deroff(dest, n); return; + } - assert(p); - /* LINTED */ - for (i = 0; i < (int)sz; i++) - free(p[i]); - free(p); -} + /* Skip leading whitespace. */ + for (cp = n->string; '\0' != *cp; cp++) + if (0 == isspace((unsigned char)*cp)) + break; -static char ** -paramdup(size_t sz, const char **p) -{ - char **pp; - int i; + /* Skip trailing whitespace. */ + for (sz = strlen(cp); sz; sz--) + if (0 == isspace((unsigned char)cp[sz-1])) + break; + + /* Skip empty strings. */ + if (0 == sz) - return(NULL); + return; - pp = xcalloc(sz, sizeof(char *)); - for (i = 0; i < (int)sz; i++) - pp[i] = xstrdup(p[i]); + if (NULL == *dest) { + *dest = mandoc_strndup(cp, sz); + return; + } - return(pp); + mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp); + free(*dest); + *dest = cp; }