-/* $Id: mdoc.c,v 1.167 2010/12/01 16:38:57 kristaps Exp $ */
+/* $Id: mdoc.c,v 1.197 2012/05/27 17:48:57 schwarze Exp $ */
/*
- * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
+ * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
* Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org>
*
* Permission to use, copy, modify, and distribute this software for any
#include <string.h>
#include <time.h>
+#include "mdoc.h"
#include "mandoc.h"
#include "libmdoc.h"
#include "libmandoc.h"
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 *m)
{
- return(MDOC_HALT & m->flags ? NULL : m->first);
+ assert( ! (MDOC_HALT & m->flags));
+ return(m->first);
}
mdoc_meta(const struct mdoc *m)
{
- return(MDOC_HALT & m->flags ? NULL : &m->meta);
+ assert( ! (MDOC_HALT & m->flags));
+ return(&m->meta);
}
free(mdoc->meta.vol);
if (mdoc->meta.msec)
free(mdoc->meta.msec);
+ if (mdoc->meta.date)
+ free(mdoc->meta.date);
}
mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
mdoc->first = mdoc->last;
mdoc->last->type = MDOC_ROOT;
+ mdoc->last->tok = MDOC_MAX;
mdoc->next = MDOC_NEXT_CHILD;
}
* Allocate volatile and non-volatile parse resources.
*/
struct mdoc *
-mdoc_alloc(struct regset *regs, void *data, mandocmsg msg)
+mdoc_alloc(struct roff *roff, struct mparse *parse, char *defos)
{
struct mdoc *p;
p = mandoc_calloc(1, sizeof(struct mdoc));
- p->msg = msg;
- p->data = data;
- p->regs = regs;
+ p->parse = parse;
+ p->defos = defos;
+ p->roff = roff;
mdoc_hash_init();
mdoc_alloc1(p);
mdoc_endparse(struct mdoc *m)
{
- if (MDOC_HALT & m->flags)
- return(0);
- else if (mdoc_macroend(m))
+ assert( ! (MDOC_HALT & m->flags));
+ if (mdoc_macroend(m))
return(1);
m->flags |= MDOC_HALT;
return(0);
}
+int
+mdoc_addeqn(struct mdoc *m, const struct eqn *ep)
+{
+ struct mdoc_node *n;
+
+ assert( ! (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);
+
+ m->next = MDOC_NEXT_SIBLING;
+ return(1);
+}
+
+int
+mdoc_addspan(struct mdoc *m, const struct tbl_span *sp)
+{
+ struct mdoc_node *n;
+
+ assert( ! (MDOC_HALT & m->flags));
+
+ /* No text before an initial macro. */
+
+ if (SEC_NONE == m->lastnamed) {
+ mdoc_pmsg(m, sp->line, 0, MANDOCERR_NOTEXT);
+ return(1);
+ }
+
+ n = node_alloc(m, sp->line, 0, MDOC_MAX, MDOC_TBL);
+ n->span = sp;
+
+ if ( ! node_append(m, n))
+ return(0);
+
+ m->next = MDOC_NEXT_SIBLING;
+ return(1);
+}
+
/*
* Main parse routine. Parses a single line -- really just hands off to
mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
{
- if (MDOC_HALT & m->flags)
- return(0);
+ assert( ! (MDOC_HALT & m->flags));
m->flags |= MDOC_NEWLINE;
* whether this mode is on or off.
* Note that this mode is also switched by the Sh macro.
*/
- if (m->regs->regs[(int)REG_nS].set) {
- if (m->regs->regs[(int)REG_nS].v.u)
+ if (roff_regisset(m->roff, REG_nS)) {
+ if (roff_regget(m->roff, REG_nS))
m->flags |= MDOC_SYNOPSIS;
else
m->flags &= ~MDOC_SYNOPSIS;
}
- return(('.' == buf[offs] || '\'' == buf[offs]) ?
+ return(mandoc_getcontrol(buf, &offs) ?
mdoc_pmacro(m, ln, buf, offs) :
mdoc_ptext(m, ln, buf, offs));
}
-
-int
-mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
- int ln, int pos, const char *fmt, ...)
-{
- char buf[256];
- va_list ap;
-
- va_start(ap, fmt);
- vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
- va_end(ap);
-
- return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
-}
-
-
int
mdoc_macro(MACRO_PROT_ARGS)
{
/* If we're in the body, deny prologue calls. */
if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
- MDOC_PBODY & m->flags)
- return(mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY));
+ 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) &&
! (MDOC_PBODY & m->flags)) {
- if ( ! mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG))
- return(0);
+ 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.vol = mandoc_strdup("LOCAL");
if (NULL == m->meta.os)
m->meta.os = mandoc_strdup("LOCAL");
- if (0 == m->meta.date)
- m->meta.date = time(NULL);
+ if (NULL == m->meta.date)
+ m->meta.date = mandoc_normdate
+ (m->parse, NULL, line, ppos);
m->flags |= MDOC_PBODY;
}
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);
mdoc->last = p;
switch (p->type) {
+ case (MDOC_TBL):
+ /* FALLTHROUGH */
case (MDOC_TEXT):
if ( ! mdoc_valid_post(mdoc))
return(0);
p->args = args;
if (p->args)
(args->refcnt)++;
+
+ 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;
p->args = args;
if (p->args)
(args->refcnt)++;
+
+ 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 *m, int line, int pos, const char *p)
{
struct mdoc_node *n;
- size_t sv, len;
-
- len = strlen(p);
n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
- n->string = mandoc_malloc(len + 1);
- sv = strlcpy(n->string, p, len + 1);
-
- /* Prohibit truncation. */
- assert(sv < len + 1);
+ n->string = roff_strdup(m->roff, p);
if ( ! node_append(m, n))
return(0);
mdoc_node_free(struct mdoc_node *p)
{
- /*
- * XXX: if these end up being problematic in terms of memory
- * management and dereferencing freed blocks, then make them
- * into reference-counted double-pointers.
- */
-
- if (MDOC_Bd == p->tok && MDOC_BLOCK == p->type)
- if (p->data.Bd)
- free(p->data.Bd);
- if (MDOC_Bl == p->tok && MDOC_BLOCK == p->type)
- if (p->data.Bl)
- free(p->data.Bl);
- if (MDOC_Bf == p->tok && MDOC_HEAD == p->type)
- if (p->data.Bf)
- free(p->data.Bf);
-
+ if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
+ free(p->norm);
if (p->string)
free(p->string);
if (p->args)
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. */
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
char *c, *ws, *end;
struct mdoc_node *n;
- /* Ignore bogus comments. */
-
- if ('\\' == buf[offs] &&
- '.' == buf[offs + 1] &&
- '"' == buf[offs + 2])
- return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
-
/* No text before an initial macro. */
- if (SEC_NONE == m->lastnamed)
- return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
+ if (SEC_NONE == m->lastnamed) {
+ mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT);
+ return(1);
+ }
assert(m->last);
n = m->last;
*/
if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
- LIST_column == n->data.Bl->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->data.Bl->type) {
+ 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));
ws = NULL;
for (c = end = buf + offs; *c; c++) {
switch (*c) {
- case '-':
- if (mandoc_hyph(buf + offs, c))
- *c = ASCII_HYPH;
- ws = NULL;
- break;
case ' ':
if (NULL == ws)
ws = c;
*end = '\0';
if (ws)
- if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
- return(0);
+ mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
- if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
- return(0);
+ mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
/*
* Insert a `sp' in the case of a blank line. Technically,
mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
{
enum mdoct tok;
- int i, j, sv;
+ int i, sv;
char mac[5];
struct mdoc_node *n;
- /* Empty lines are ignored. */
-
- offs++;
+ /* Empty post-control lines are ignored. */
- if ('\0' == buf[offs])
+ if ('"' == buf[offs]) {
+ mdoc_pmsg(m, ln, offs, MANDOCERR_BADCOMMENT);
+ return(1);
+ } else if ('\0' == buf[offs])
return(1);
- i = offs;
-
- /* Accept tabs/whitespace after the initial control char. */
-
- if (' ' == buf[i] || '\t' == buf[i]) {
- i++;
- while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
- i++;
- if ('\0' == buf[i])
- return(1);
- }
-
- sv = i;
+ sv = offs;
/*
* Copy the first word into a nil-terminated buffer.
* Stop copying when a tab, space, or eoln is encountered.
*/
- j = 0;
- while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i])
- mac[j++] = buf[i++];
- mac[j] = '\0';
+ 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;
- tok = (j > 1 || j < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
if (MDOC_MAX == tok) {
- mdoc_vmsg(m, MANDOCERR_MACRO, ln, sv,
- "unknown macro: %s%s",
- buf, strlen(buf) > 3 ? "..." : "");
+ mandoc_vmsg(MANDOCERR_MACRO, m->parse,
+ ln, sv, "%s", buf + sv - 1);
return(1);
}
/* Disregard the first trailing tab, if applicable. */
- if ('\t' == buf[i])
- i++;
+ if ('\t' == buf[offs])
+ offs++;
/* Jump to the next non-whitespace word. */
- while (buf[i] && ' ' == buf[i])
- i++;
+ 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[i] && ' ' == buf[i - 1])
- if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
- goto err;
+ 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
*/
if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
- if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
+ if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf))
goto err;
return(1);
}
*/
if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
- LIST_column == n->data.Bl->type) {
+ LIST_column == n->norm->Bl.type) {
m->flags |= MDOC_FREECOL;
if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
goto err;
if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
NULL != n->parent &&
MDOC_Bl == n->parent->tok &&
- LIST_column == n->parent->data.Bl->type) {
+ LIST_column == n->parent->norm->Bl.type) {
m->flags |= MDOC_FREECOL;
if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
goto err;
/* Normal processing of a macro. */
- if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
+ if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf))
goto err;
return(1);
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);
+}
git.ckatri.com / mandoc.git / blobdiff