=================================================================== RCS file: /cvs/mandoc/mdoc.c,v retrieving revision 1.58 retrieving revision 1.194 diff -u -p -r1.58 -r1.194 --- mandoc/mdoc.c 2009/03/09 13:04:01 1.58 +++ mandoc/mdoc.c 2011/07/27 12:43:02 1.194 @@ -1,48 +1,40 @@ -/* $Id: mdoc.c,v 1.58 2009/03/09 13:04:01 kristaps Exp $ */ +/* $Id: mdoc.c,v 1.194 2011/07/27 12:43:02 kristaps Exp $ */ /* - * Copyright (c) 2008 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 - * 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 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. */ +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include + #include -#include -#include #include -#include #include +#include #include +#include -#include "private.h" +#include "mdoc.h" +#include "mandoc.h" +#include "libmdoc.h" +#include "libmandoc.h" -/* - * Main caller in the libmdoc library. This begins the parsing routine, - * handles allocation of data, and so forth. Most of the "work" is done - * in macro.c and validate.c. - */ - -static struct mdoc_node *mdoc_node_alloc(const struct mdoc *); -static int mdoc_node_append(struct mdoc *, - struct mdoc_node *); - -static int parsetext(struct mdoc *, int, char *); -static int parsemacro(struct mdoc *, int, char *); -static int macrowarn(struct mdoc *, int, const char *); - - const char *const __mdoc_macronames[MDOC_MAX] = { - "\\\"", "Dd", "Dt", "Os", + "Ap", "Dd", "Dt", "Os", "Sh", "Ss", "Pp", "D1", "Dl", "Bd", "Ed", "Bl", "El", "It", "Ad", "An", @@ -53,11 +45,11 @@ const char *const __mdoc_macronames[MDOC_MAX] = { "Nm", "Op", "Ot", "Pa", "Rv", "St", "Va", "Vt", /* LINTED */ - "Xr", "\%A", "\%B", "\%D", + "Xr", "%A", "%B", "%D", /* LINTED */ - "\%I", "\%J", "\%N", "\%O", + "%I", "%J", "%N", "%O", /* LINTED */ - "\%P", "\%R", "\%T", "\%V", + "%P", "%R", "%T", "%V", "Ac", "Ao", "Aq", "At", "Bc", "Bf", "Bo", "Bq", "Bsx", "Bx", "Db", "Dc", @@ -71,8 +63,14 @@ const char *const __mdoc_macronames[MDOC_MAX] = { "Tn", "Ux", "Xc", "Xo", "Fo", "Fc", "Oo", "Oc", "Bk", "Ek", "Bt", "Hf", - "Fr", "Ud", "Lb", "Ap", - "Lp" + "Fr", "Ud", "Lb", "Lp", + "Lk", "Mt", "Brq", "Bro", + /* LINTED */ + "Brc", "%C", "Es", "En", + /* LINTED */ + "Dx", "%Q", "br", "sp", + /* LINTED */ + "%U", "Ta" }; const char *const __mdoc_argnames[MDOC_ARG_MAX] = { @@ -84,37 +82,54 @@ const char *const __mdoc_argnames[MDOC_ARG_MAX] = { "ohang", "inset", "column", "width", "compact", "std", "filled", "words", "emphasis", - "symbolic" + "symbolic", "nested", "centered" }; const char * const *mdoc_macronames = __mdoc_macronames; const char * const *mdoc_argnames = __mdoc_argnames; +static void mdoc_node_free(struct mdoc_node *); +static void mdoc_node_unlink(struct mdoc *, + struct mdoc_node *); +static void mdoc_free1(struct mdoc *); +static void mdoc_alloc1(struct mdoc *); +static struct mdoc_node *node_alloc(struct mdoc *, int, int, + enum mdoct, enum mdoc_type); +static int node_append(struct mdoc *, + struct mdoc_node *); +#if 0 +static int mdoc_preptext(struct mdoc *, int, char *, int); +#endif +static int mdoc_ptext(struct mdoc *, int, char *, int); +static int mdoc_pmacro(struct mdoc *, int, char *, int); const struct mdoc_node * -mdoc_node(const struct mdoc *mdoc) +mdoc_node(const struct mdoc *m) { - return(mdoc->first); + assert( ! (MDOC_HALT & m->flags)); + return(m->first); } const struct mdoc_meta * -mdoc_meta(const struct mdoc *mdoc) +mdoc_meta(const struct mdoc *m) { - return(&mdoc->meta); + assert( ! (MDOC_HALT & m->flags)); + return(&m->meta); } -void -mdoc_free(struct mdoc *mdoc) +/* + * Frees volatile resources (parse tree, meta-data, fields). + */ +static void +mdoc_free1(struct mdoc *mdoc) { if (mdoc->first) - mdoc_node_freelist(mdoc->first); - if (mdoc->htab) - mdoc_tokhash_free(mdoc->htab); + mdoc_node_delete(mdoc, mdoc->first); if (mdoc->meta.title) free(mdoc->meta.title); if (mdoc->meta.os) @@ -125,145 +140,209 @@ mdoc_free(struct mdoc *mdoc) free(mdoc->meta.arch); if (mdoc->meta.vol) free(mdoc->meta.vol); + if (mdoc->meta.msec) + free(mdoc->meta.msec); + if (mdoc->meta.date) + free(mdoc->meta.date); +} + +/* + * Allocate all volatile resources (parse tree, meta-data, fields). + */ +static void +mdoc_alloc1(struct mdoc *mdoc) +{ + + memset(&mdoc->meta, 0, sizeof(struct mdoc_meta)); + mdoc->flags = 0; + mdoc->lastnamed = mdoc->lastsec = SEC_NONE; + mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node)); + mdoc->first = mdoc->last; + mdoc->last->type = MDOC_ROOT; + mdoc->next = MDOC_NEXT_CHILD; +} + + +/* + * Free up volatile resources (see mdoc_free1()) then re-initialises the + * data with mdoc_alloc1(). After invocation, parse data has been reset + * and the parser is ready for re-invocation on a new tree; however, + * cross-parse non-volatile data is kept intact. + */ +void +mdoc_reset(struct mdoc *mdoc) +{ + + mdoc_free1(mdoc); + mdoc_alloc1(mdoc); +} + + +/* + * Completely free up all volatile and non-volatile parse resources. + * After invocation, the pointer is no longer usable. + */ +void +mdoc_free(struct mdoc *mdoc) +{ + + mdoc_free1(mdoc); free(mdoc); } +/* + * Allocate volatile and non-volatile parse resources. + */ struct mdoc * -mdoc_alloc(void *data, int pflags, const struct mdoc_cb *cb) +mdoc_alloc(struct roff *roff, struct mparse *parse) { struct mdoc *p; - p = xcalloc(1, sizeof(struct mdoc)); + p = mandoc_calloc(1, sizeof(struct mdoc)); - p->data = data; - if (cb) - (void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb)); + p->parse = parse; + p->roff = roff; - p->last = xcalloc(1, sizeof(struct mdoc_node)); - p->last->type = MDOC_ROOT; - p->first = p->last; - p->pflags = pflags; - p->next = MDOC_NEXT_CHILD; - p->htab = mdoc_tokhash_alloc(); - + mdoc_hash_init(); + mdoc_alloc1(p); return(p); } +/* + * Climb back up the parse tree, validating open scopes. Mostly calls + * through to macro_end() in macro.c. + */ int -mdoc_endparse(struct mdoc *mdoc) +mdoc_endparse(struct mdoc *m) { - if (MDOC_HALT & mdoc->flags) - return(0); - if (NULL == mdoc->first) + assert( ! (MDOC_HALT & m->flags)); + if (mdoc_macroend(m)) return(1); - - assert(mdoc->last); - if ( ! macro_end(mdoc)) { - mdoc->flags |= MDOC_HALT; - return(0); - } - return(1); + m->flags |= MDOC_HALT; + return(0); } - -/* - * Main parse routine. Parses a single line -- really just hands off to - * the macro or text parser. - */ int -mdoc_parseln(struct mdoc *m, int ln, char *buf) +mdoc_addeqn(struct mdoc *m, const struct eqn *ep) { + struct mdoc_node *n; - /* If in error-mode, then we parse no more. */ + assert( ! (MDOC_HALT & m->flags)); - if (MDOC_HALT & m->flags) + /* No text before an initial macro. */ + + if (SEC_NONE == m->lastnamed) { + mdoc_pmsg(m, ep->ln, ep->pos, MANDOCERR_NOTEXT); + return(1); + } + + n = node_alloc(m, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN); + n->eqn = ep; + + if ( ! node_append(m, n)) return(0); - return('.' == *buf ? parsemacro(m, ln, buf) : - parsetext(m, ln, buf)); + m->next = MDOC_NEXT_SIBLING; + return(1); } - -void -mdoc_vmsg(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...) +int +mdoc_addspan(struct mdoc *m, const struct tbl_span *sp) { - char buf[256]; - va_list ap; + struct mdoc_node *n; - if (NULL == mdoc->cb.mdoc_msg) - return; + assert( ! (MDOC_HALT & m->flags)); - va_start(ap, fmt); - (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap); - va_end(ap); - (*mdoc->cb.mdoc_msg)(mdoc->data, ln, pos, buf); -} + /* No text before an initial macro. */ + if (SEC_NONE == m->lastnamed) { + mdoc_pmsg(m, sp->line, 0, MANDOCERR_NOTEXT); + return(1); + } -int -mdoc_verr(struct mdoc *mdoc, int ln, int pos, - const char *fmt, ...) -{ - char buf[256]; - va_list ap; + n = node_alloc(m, sp->line, 0, MDOC_MAX, MDOC_TBL); + n->span = sp; - if (NULL == mdoc->cb.mdoc_err) + if ( ! node_append(m, n)) return(0); - va_start(ap, fmt); - (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap); - va_end(ap); - return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf)); + m->next = MDOC_NEXT_SIBLING; + return(1); } +/* + * Main parse routine. Parses a single line -- really just hands off to + * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()). + */ int -mdoc_vwarn(struct mdoc *mdoc, int ln, int pos, - enum mdoc_warn type, const char *fmt, ...) +mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs) { - char buf[256]; - va_list ap; - if (NULL == mdoc->cb.mdoc_warn) - return(0); + assert( ! (MDOC_HALT & m->flags)); - va_start(ap, fmt); - (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap); - va_end(ap); - return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, type, buf)); -} + m->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_regisset(m->roff, REG_nS)) { + if (roff_regget(m->roff, REG_nS)) + m->flags |= MDOC_SYNOPSIS; + else + m->flags &= ~MDOC_SYNOPSIS; + } + return(mandoc_getcontrol(buf, &offs) ? + mdoc_pmacro(m, ln, buf, offs) : + mdoc_ptext(m, ln, buf, offs)); +} + int -mdoc_macro(struct mdoc *m, int tok, - int ln, int pp, int *pos, char *buf) +mdoc_macro(MACRO_PROT_ARGS) { + assert(tok < MDOC_MAX); - /* FIXME - these should happen during validation. */ + /* If we're in the body, deny prologue calls. */ if (MDOC_PROLOGUE & mdoc_macros[tok].flags && - SEC_PROLOGUE != m->lastnamed) - return(mdoc_perr(m, ln, pp, - "disallowed in document body")); + MDOC_PBODY & m->flags) { + mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY); + return(1); + } + /* If we're in the prologue, deny "body" macros. */ + if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && - SEC_PROLOGUE == m->lastnamed) - return(mdoc_perr(m, ln, pp, - "disallowed in prologue")); + ! (MDOC_PBODY & m->flags)) { + mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG); + if (NULL == m->meta.msec) + m->meta.msec = mandoc_strdup("1"); + if (NULL == m->meta.title) + m->meta.title = mandoc_strdup("UNKNOWN"); + if (NULL == m->meta.vol) + m->meta.vol = mandoc_strdup("LOCAL"); + if (NULL == m->meta.os) + m->meta.os = mandoc_strdup("LOCAL"); + if (NULL == m->meta.date) + m->meta.date = mandoc_normdate + (m->parse, NULL, line, ppos); + m->flags |= MDOC_PBODY; + } - if (1 != pp && ! (MDOC_CALLABLE & mdoc_macros[tok].flags)) - return(mdoc_perr(m, ln, pp, "not callable")); - - return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf)); + return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf)); } static int -mdoc_node_append(struct mdoc *mdoc, struct mdoc_node *p) +node_append(struct mdoc *mdoc, struct mdoc_node *p) { assert(mdoc->last); @@ -285,6 +364,25 @@ mdoc_node_append(struct mdoc *mdoc, struct mdoc_node * /* NOTREACHED */ } + p->parent->nchild++; + + /* + * Copy over the normalised-data pointer of our parent. Not + * everybody has one, but copying a null pointer is fine. + */ + + switch (p->type) { + case (MDOC_BODY): + /* FALLTHROUGH */ + case (MDOC_TAIL): + /* FALLTHROUGH */ + case (MDOC_HEAD): + p->norm = p->parent->norm; + break; + default: + break; + } + if ( ! mdoc_valid_pre(mdoc, p)) return(0); @@ -298,6 +396,8 @@ mdoc_node_append(struct mdoc *mdoc, struct mdoc_node * p->parent->tail = p; break; case (MDOC_BODY): + if (p->end) + break; assert(MDOC_BLOCK == p->parent->type); p->parent->body = p; break; @@ -306,155 +406,186 @@ mdoc_node_append(struct mdoc *mdoc, struct mdoc_node * } mdoc->last = p; + + switch (p->type) { + case (MDOC_TBL): + /* FALLTHROUGH */ + case (MDOC_TEXT): + if ( ! mdoc_valid_post(mdoc)) + return(0); + break; + default: + break; + } + return(1); } static struct mdoc_node * -mdoc_node_alloc(const struct mdoc *mdoc) +node_alloc(struct mdoc *m, int line, int pos, + enum mdoct tok, enum mdoc_type type) { struct mdoc_node *p; - p = xcalloc(1, sizeof(struct mdoc_node)); - p->sec = mdoc->lastsec; + p = mandoc_calloc(1, sizeof(struct mdoc_node)); + p->sec = m->lastsec; + p->line = line; + p->pos = pos; + p->tok = tok; + p->type = type; + /* Flag analysis. */ + + if (MDOC_SYNOPSIS & m->flags) + p->flags |= MDOC_SYNPRETTY; + else + p->flags &= ~MDOC_SYNPRETTY; + if (MDOC_NEWLINE & m->flags) + p->flags |= MDOC_LINE; + m->flags &= ~MDOC_NEWLINE; + return(p); } int -mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, int tok) +mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok) { struct mdoc_node *p; - assert(mdoc->first); - assert(mdoc->last); - - p = mdoc_node_alloc(mdoc); - - p->line = line; - p->pos = pos; - p->type = MDOC_TAIL; - p->tok = tok; - - return(mdoc_node_append(mdoc, p)); + p = node_alloc(m, line, pos, tok, MDOC_TAIL); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, int tok) +mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok) { struct mdoc_node *p; - assert(mdoc->first); - assert(mdoc->last); + assert(m->first); + assert(m->last); - p = mdoc_node_alloc(mdoc); - - p->line = line; - p->pos = pos; - p->type = MDOC_HEAD; - p->tok = tok; - - return(mdoc_node_append(mdoc, p)); + p = node_alloc(m, line, pos, tok, MDOC_HEAD); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, int tok) +mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok) { struct mdoc_node *p; - assert(mdoc->first); - assert(mdoc->last); - - p = mdoc_node_alloc(mdoc); - - p->line = line; - p->pos = pos; - p->type = MDOC_BODY; - p->tok = tok; - - return(mdoc_node_append(mdoc, p)); + p = node_alloc(m, line, pos, tok, MDOC_BODY); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_root_alloc(struct mdoc *mdoc) +mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok, + struct mdoc_node *body, enum mdoc_endbody end) { struct mdoc_node *p; - p = mdoc_node_alloc(mdoc); - - p->type = MDOC_ROOT; - - return(mdoc_node_append(mdoc, p)); + p = node_alloc(m, line, pos, tok, MDOC_BODY); + p->pending = body; + p->end = end; + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_SIBLING; + return(1); } int -mdoc_block_alloc(struct mdoc *mdoc, int line, int pos, - int tok, struct mdoc_arg *args) +mdoc_block_alloc(struct mdoc *m, int line, int pos, + enum mdoct tok, struct mdoc_arg *args) { struct mdoc_node *p; - p = mdoc_node_alloc(mdoc); - - p->pos = pos; - p->line = line; - p->type = MDOC_BLOCK; - p->tok = tok; + p = node_alloc(m, line, pos, tok, MDOC_BLOCK); p->args = args; - - if (args) + if (p->args) (args->refcnt)++; - return(mdoc_node_append(mdoc, p)); + switch (tok) { + case (MDOC_Bd): + /* FALLTHROUGH */ + case (MDOC_Bf): + /* FALLTHROUGH */ + case (MDOC_Bl): + /* FALLTHROUGH */ + case (MDOC_Rs): + p->norm = mandoc_calloc(1, sizeof(union mdoc_data)); + break; + default: + break; + } + + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos, - int tok, struct mdoc_arg *args) +mdoc_elem_alloc(struct mdoc *m, int line, int pos, + enum mdoct tok, struct mdoc_arg *args) { struct mdoc_node *p; - p = mdoc_node_alloc(mdoc); - - p->line = line; - p->pos = pos; - p->type = MDOC_ELEM; - p->tok = tok; + p = node_alloc(m, line, pos, tok, MDOC_ELEM); p->args = args; - - if (args) + if (p->args) (args->refcnt)++; - return(mdoc_node_append(mdoc, p)); + switch (tok) { + case (MDOC_An): + p->norm = mandoc_calloc(1, sizeof(union mdoc_data)); + break; + default: + break; + } + + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } - int -mdoc_word_alloc(struct mdoc *mdoc, - int line, int pos, const char *word) +mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p) { - struct mdoc_node *p; + struct mdoc_node *n; - p = mdoc_node_alloc(mdoc); + n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT); + n->string = mandoc_strdup(p); - p->line = line; - p->pos = pos; - p->type = MDOC_TEXT; - p->string = xstrdup(word); + if ( ! node_append(m, n)) + return(0); - return(mdoc_node_append(mdoc, p)); + m->next = MDOC_NEXT_SIBLING; + return(1); } -void +static void mdoc_node_free(struct mdoc_node *p) { + if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type) + free(p->norm); if (p->string) free(p->string); if (p->args) @@ -463,101 +594,341 @@ mdoc_node_free(struct mdoc_node *p) } +static void +mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n) +{ + + /* Adjust siblings. */ + + 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 (m && m->last == n) { + if (n->prev) { + m->last = n->prev; + m->next = MDOC_NEXT_SIBLING; + } else { + m->last = n->parent; + m->next = MDOC_NEXT_CHILD; + } + } + + if (m && m->first == n) + m->first = NULL; +} + + void -mdoc_node_freelist(struct mdoc_node *p) +mdoc_node_delete(struct mdoc *m, struct mdoc_node *p) { - if (p->child) - mdoc_node_freelist(p->child); - if (p->next) - mdoc_node_freelist(p->next); + while (p->child) { + assert(p->nchild); + mdoc_node_delete(m, p->child); + } + assert(0 == p->nchild); + mdoc_node_unlink(m, p); mdoc_node_free(p); } +#if 0 +/* + * Pre-treat a text line. + * Text lines can consist of equations, which must be handled apart from + * the regular text. + * Thus, use this function to step through a line checking if it has any + * equations embedded in it. + * This must handle multiple equations AND equations that do not end at + * the end-of-line, i.e., will re-enter in the next roff parse. + */ +static int +mdoc_preptext(struct mdoc *m, int line, char *buf, int offs) +{ + char *start, *end; + char delim; + while ('\0' != buf[offs]) { + /* Mark starting position if eqn is set. */ + start = NULL; + if ('\0' != (delim = roff_eqndelim(m->roff))) + if (NULL != (start = strchr(buf + offs, delim))) + *start++ = '\0'; + + /* Parse text as normal. */ + if ( ! mdoc_ptext(m, line, buf, offs)) + return(0); + + /* Continue only if an equation exists. */ + if (NULL == start) + break; + + /* Read past the end of the equation. */ + offs += start - (buf + offs); + assert(start == &buf[offs]); + if (NULL != (end = strchr(buf + offs, delim))) { + *end++ = '\0'; + while (' ' == *end) + end++; + } + + /* Parse the equation itself. */ + roff_openeqn(m->roff, NULL, line, offs, buf); + + /* Process a finished equation? */ + if (roff_closeeqn(m->roff)) + if ( ! mdoc_addeqn(m, roff_eqn(m->roff))) + return(0); + offs += (end - (buf + offs)); + } + + return(1); +} +#endif + /* * Parse free-form text, that is, a line that does not begin with the * control character. */ static int -parsetext(struct mdoc *mdoc, int line, char *buf) +mdoc_ptext(struct mdoc *m, int line, char *buf, int offs) { + char *c, *ws, *end; + struct mdoc_node *n; - if (SEC_PROLOGUE == mdoc->lastnamed) - return(mdoc_perr(mdoc, line, 0, - "text disallowed in prologue")); + /* No text before an initial macro. */ - if ( ! mdoc_word_alloc(mdoc, line, 0, buf)) + if (SEC_NONE == m->lastnamed) { + mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT); + return(1); + } + + assert(m->last); + n = m->last; + + /* + * Divert directly to list processing if we're encountering a + * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry + * (a MDOC_BODY means it's already open, in which case we should + * process within its context in the normal way). + */ + + if (MDOC_Bl == n->tok && MDOC_BODY == n->type && + LIST_column == n->norm->Bl.type) { + /* `Bl' is open without any children. */ + m->flags |= MDOC_FREECOL; + return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf)); + } + + if (MDOC_It == n->tok && MDOC_BLOCK == n->type && + NULL != n->parent && + MDOC_Bl == n->parent->tok && + LIST_column == n->parent->norm->Bl.type) { + /* `Bl' has block-level `It' children. */ + m->flags |= MDOC_FREECOL; + return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf)); + } + + /* + * 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 & m->flags) + continue; + break; + case '\\': + /* Skip the escaped character, too, if any. */ + if (c[1]) + c++; + /* FALLTHROUGH */ + default: + ws = NULL; + break; + } + end = c + 1; + } + *end = '\0'; + + if (ws) + mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE); + + if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) { + mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN); + + /* + * 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. + */ + if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL)) + return(0); + + m->next = MDOC_NEXT_SIBLING; + return(1); + } + + if ( ! mdoc_word_alloc(m, line, offs, buf+offs)) return(0); - mdoc->next = MDOC_NEXT_SIBLING; - return(1); -} + if (MDOC_LITERAL & m->flags) + return(1); + /* + * 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. + */ -static int -macrowarn(struct mdoc *m, int ln, const char *buf) -{ - if ( ! (MDOC_IGN_MACRO & m->pflags)) - return(mdoc_perr(m, ln, 1, "unknown macro: %s%s", - buf, buf[3] ? "..." : "")); - return(mdoc_pwarn(m, ln, 1, WARN_SYNTAX, - "unknown macro: %s%s", - buf, buf[3] ? "..." : "")); -} + assert(buf < end); + if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0)) + m->last->flags |= MDOC_EOS; + return(1); +} + /* * Parse a macro line, that is, a line beginning with the control * character. */ -int -parsemacro(struct mdoc *m, int ln, char *buf) +static int +mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs) { - int i, c; + enum mdoct tok; + int i, sv; char mac[5]; + struct mdoc_node *n; - /* Comments are quickly ignored. */ + /* Empty post-control lines are ignored. */ - if (buf[1] && '\\' == buf[1]) - if (buf[2] && '\"' == buf[2]) - return(1); + if ('"' == buf[offs]) { + mdoc_pmsg(m, ln, offs, MANDOCERR_BADCOMMENT); + return(1); + } else if ('\0' == buf[offs]) + return(1); - /* Copy the first word into a nil-terminated buffer. */ + sv = offs; - for (i = 1; i < 5; i++) { - if (0 == (mac[i - 1] = buf[i])) - break; - else if (isspace((unsigned char)buf[i])) - break; + /* + * Copy the first word into a nil-terminated buffer. + * Stop copying when a tab, space, or eoln is encountered. + */ + + i = 0; + while (i < 4 && '\0' != buf[offs] && + ' ' != buf[offs] && '\t' != buf[offs]) + mac[i++] = buf[offs++]; + + mac[i] = '\0'; + + tok = (i > 1 || i < 4) ? mdoc_hash_find(mac) : MDOC_MAX; + + if (MDOC_MAX == tok) { + mandoc_vmsg(MANDOCERR_MACRO, m->parse, + ln, sv, "%s", buf + sv - 1); + return(1); } - /* FIXME: be able to skip unknown macro lines! */ + /* Disregard the first trailing tab, if applicable. */ - mac[i - 1] = 0; + if ('\t' == buf[offs]) + offs++; - if (i == 5 || i <= 2) { - if ( ! macrowarn(m, ln, mac)) + /* Jump to the next non-whitespace word. */ + + while (buf[offs] && ' ' == buf[offs]) + offs++; + + /* + * Trailing whitespace. Note that tabs are allowed to be passed + * into the parser as "text", so we only warn about spaces here. + */ + + if ('\0' == buf[offs] && ' ' == buf[offs - 1]) + mdoc_pmsg(m, ln, offs - 1, MANDOCERR_EOLNSPACE); + + /* + * If an initial macro or a list invocation, divert directly + * into macro processing. + */ + + if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) { + if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf)) goto err; return(1); - } - - if (MDOC_MAX == (c = mdoc_tokhash_find(m->htab, mac))) { - if ( ! macrowarn(m, ln, mac)) + } + + n = m->last; + assert(m->last); + + /* + * If the first macro of a `Bl -column', open an `It' block + * context around the parsed macro. + */ + + if (MDOC_Bl == n->tok && MDOC_BODY == n->type && + LIST_column == n->norm->Bl.type) { + m->flags |= MDOC_FREECOL; + if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf)) goto err; return(1); } - /* The macro is sane. Jump to the next word. */ + /* + * 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. + */ - while (buf[i] && isspace((unsigned char)buf[i])) - i++; + if (MDOC_It == n->tok && MDOC_BLOCK == n->type && + NULL != n->parent && + MDOC_Bl == n->parent->tok && + LIST_column == n->parent->norm->Bl.type) { + m->flags |= MDOC_FREECOL; + if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf)) + goto err; + return(1); + } - /* Begin recursive parse sequence. */ + /* Normal processing of a macro. */ - if ( ! mdoc_macro(m, c, ln, 1, &i, buf)) + if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf)) goto err; return(1); @@ -566,4 +937,50 @@ err: /* Error out. */ m->flags |= MDOC_HALT; return(0); +} + +enum mdelim +mdoc_isdelim(const 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, "*(Ba")) + return(DELIM_MIDDLE); + + return(DELIM_NONE); }