-/* $Id: mdoc_validate.c,v 1.32 2009/07/13 07:23:07 kristaps Exp $ */
+/* $Id: mdoc_validate.c,v 1.58 2010/02/17 19:28:11 kristaps Exp $ */
/*
* Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
*
* 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 <sys/types.h>
#include <assert.h>
#include <ctype.h>
+#include <limits.h>
#include <stdarg.h>
#include <stdlib.h>
#include <string.h>
static int err_count(struct mdoc *, const char *,
int, const char *, int);
-#ifdef __linux__
-extern size_t strlcat(char *, const char *, size_t);
-#endif
-
static int berr_ge1(POST_ARGS);
static int bwarn_ge1(POST_ARGS);
static int ebool(POST_ARGS);
static int eerr_eq0(POST_ARGS);
static int eerr_eq1(POST_ARGS);
static int eerr_ge1(POST_ARGS);
-static int eerr_le2(POST_ARGS);
-static int ewarn_eq0(POST_ARGS);
+static int eerr_le1(POST_ARGS);
static int ewarn_ge1(POST_ARGS);
static int herr_eq0(POST_ARGS);
static int herr_ge1(POST_ARGS);
static int hwarn_le1(POST_ARGS);
static int post_an(POST_ARGS);
-static int post_args(POST_ARGS);
static int post_at(POST_ARGS);
static int post_bf(POST_ARGS);
static int post_bl(POST_ARGS);
static int post_lb(POST_ARGS);
static int post_nm(POST_ARGS);
static int post_root(POST_ARGS);
+static int post_rs(POST_ARGS);
static int post_sh(POST_ARGS);
static int post_sh_body(POST_ARGS);
static int post_sh_head(POST_ARGS);
static int post_st(POST_ARGS);
+static int post_vt(POST_ARGS);
static int pre_an(PRE_ARGS);
static int pre_bd(PRE_ARGS);
static int pre_bl(PRE_ARGS);
static v_post posts_bf[] = { hwarn_le1, post_bf, NULL };
static v_post posts_bl[] = { bwarn_ge1, post_bl, NULL };
static v_post posts_bool[] = { eerr_eq1, ebool, NULL };
-static v_post posts_ex[] = { eerr_eq0, post_args, NULL };
static v_post posts_fo[] = { hwarn_eq1, bwarn_ge1, NULL };
-static v_post posts_in[] = { eerr_eq1, NULL };
static v_post posts_it[] = { post_it, NULL };
static v_post posts_lb[] = { eerr_eq1, post_lb, NULL };
static v_post posts_nd[] = { berr_ge1, NULL };
static v_post posts_nm[] = { post_nm, NULL };
static v_post posts_notext[] = { eerr_eq0, NULL };
-static v_post posts_pf[] = { eerr_eq1, NULL };
-static v_post posts_pp[] = { ewarn_eq0, NULL };
-static v_post posts_rv[] = { eerr_eq0, post_args, NULL };
+static v_post posts_rs[] = { berr_ge1, herr_eq0, post_rs, NULL };
static v_post posts_sh[] = { herr_ge1, bwarn_ge1, post_sh, NULL };
+static v_post posts_sp[] = { eerr_le1, NULL };
static v_post posts_ss[] = { herr_ge1, NULL };
static v_post posts_st[] = { eerr_eq1, post_st, NULL };
static v_post posts_text[] = { eerr_ge1, NULL };
+static v_post posts_text1[] = { eerr_eq1, NULL };
+static v_post posts_vt[] = { post_vt, NULL };
static v_post posts_wline[] = { bwarn_ge1, herr_eq0, NULL };
static v_post posts_wtext[] = { ewarn_ge1, NULL };
-static v_post posts_xr[] = { eerr_ge1, eerr_le2, NULL };
+static v_post posts_xr[] = { eerr_ge1, NULL };
static v_pre pres_an[] = { pre_an, NULL };
static v_pre pres_bd[] = { pre_display, pre_bd, NULL };
static v_pre pres_bl[] = { pre_bl, NULL };
{ pres_os, NULL }, /* Os */
{ pres_sh, posts_sh }, /* Sh */
{ pres_ss, posts_ss }, /* Ss */
- { NULL, posts_pp }, /* Pp */
+ { NULL, posts_notext }, /* Pp */
{ pres_d1, posts_wline }, /* D1 */
{ pres_d1, posts_wline }, /* Dl */
{ pres_bd, posts_bd }, /* Bd */
{ NULL, NULL }, /* Dv */
{ pres_er, posts_text }, /* Er */
{ NULL, NULL }, /* Ev */
- { pres_ex, posts_ex }, /* Ex */
+ { pres_ex, NULL }, /* Ex */
{ NULL, NULL }, /* Fa */
{ pres_fd, posts_wtext }, /* Fd */
{ NULL, NULL }, /* Fl */
{ NULL, posts_text }, /* Fn */
{ NULL, posts_wtext }, /* Ft */
{ NULL, posts_text }, /* Ic */
- { NULL, posts_in }, /* In */
+ { NULL, posts_text1 }, /* In */
{ NULL, NULL }, /* Li */
{ NULL, posts_nd }, /* Nd */
{ NULL, posts_nm }, /* Nm */
{ NULL, posts_wline }, /* Op */
{ NULL, NULL }, /* Ot */
{ NULL, NULL }, /* Pa */
- { pres_rv, posts_rv }, /* Rv */
+ { pres_rv, NULL }, /* Rv */
{ NULL, posts_st }, /* St */
{ NULL, NULL }, /* Va */
- { NULL, posts_text }, /* Vt */
+ { NULL, posts_vt }, /* Vt */
{ NULL, posts_xr }, /* Xr */
{ NULL, posts_text }, /* %A */
- { NULL, posts_text }, /* %B */
- { NULL, posts_text }, /* %D */
+ { NULL, posts_text }, /* %B */ /* FIXME: can be used outside Rs/Re. */
+ { NULL, posts_text }, /* %D */ /* FIXME: check date with mandoc_a2time(). */
{ NULL, posts_text }, /* %I */
{ NULL, posts_text }, /* %J */
{ NULL, posts_text }, /* %N */
{ NULL, posts_text }, /* %O */
{ NULL, posts_text }, /* %P */
{ NULL, posts_text }, /* %R */
- { NULL, posts_text }, /* %T */
+ { NULL, posts_text }, /* %T */ /* FIXME: can be used outside Rs/Re. */
{ NULL, posts_text }, /* %V */
{ NULL, NULL }, /* Ac */
{ NULL, NULL }, /* Ao */
{ NULL, NULL }, /* Nx */
{ NULL, NULL }, /* Ox */
{ NULL, NULL }, /* Pc */
- { NULL, posts_pf }, /* Pf */
+ { NULL, posts_text1 }, /* Pf */
{ NULL, NULL }, /* Po */
{ NULL, posts_wline }, /* Pq */
{ NULL, NULL }, /* Qc */
{ NULL, NULL }, /* Qo */
{ NULL, posts_wline }, /* Qq */
{ NULL, NULL }, /* Re */
- { NULL, posts_wline }, /* Rs */
+ { NULL, posts_rs }, /* Rs */
{ NULL, NULL }, /* Sc */
{ NULL, NULL }, /* So */
{ NULL, posts_wline }, /* Sq */
{ NULL, NULL }, /* Fr */
{ NULL, posts_notext }, /* Ud */
{ pres_lb, posts_lb }, /* Lb */
- { NULL, posts_pp }, /* Lp */
- { NULL, NULL }, /* Lk */
+ { NULL, posts_notext }, /* Lp */
+ { NULL, posts_text }, /* Lk */
{ NULL, posts_text }, /* Mt */
{ NULL, posts_wline }, /* Brq */
{ NULL, NULL }, /* Bro */
{ NULL, NULL }, /* En */
{ NULL, NULL }, /* Dx */
{ NULL, posts_text }, /* %Q */
+ { NULL, posts_notext }, /* br */
+ { NULL, posts_sp }, /* sp */
+ { NULL, posts_text1 }, /* %U */
};
CHECK_CHILD_DEFN(warn, lt, <) /* warn_child_lt() */
CHECK_BODY_DEFN(ge1, warn, warn_child_gt, 0) /* bwarn_ge1() */
CHECK_BODY_DEFN(ge1, err, err_child_gt, 0) /* berr_ge1() */
-CHECK_ELEM_DEFN(eq0, warn, warn_child_eq, 0) /* ewarn_eq0() */
CHECK_ELEM_DEFN(ge1, warn, warn_child_gt, 0) /* ewarn_gt1() */
CHECK_ELEM_DEFN(eq1, err, err_child_eq, 1) /* eerr_eq1() */
-CHECK_ELEM_DEFN(le2, err, err_child_lt, 3) /* eerr_le2() */
+CHECK_ELEM_DEFN(le1, err, err_child_lt, 2) /* eerr_le1() */
CHECK_ELEM_DEFN(eq0, err, err_child_eq, 0) /* eerr_eq0() */
CHECK_ELEM_DEFN(ge1, err, err_child_gt, 0) /* eerr_ge1() */
CHECK_HEAD_DEFN(eq0, err, err_child_eq, 0) /* herr_eq0() */
case (MDOC_Inset):
/* FALLTHROUGH */
case (MDOC_Column):
- if (-1 != type)
+ if (type >= 0)
return(mdoc_nerr(mdoc, n, EMULTILIST));
type = n->args->argv[pos].arg;
break;
+ case (MDOC_Compact):
+ if (type < 0 && ! mdoc_nwarn(mdoc, n, ENOTYPE))
+ return(0);
+ break;
case (MDOC_Width):
- if (-1 != width)
+ if (width >= 0)
return(mdoc_nerr(mdoc, n, EARGREP));
+ if (type < 0 && ! mdoc_nwarn(mdoc, n, ENOTYPE))
+ return(0);
width = n->args->argv[pos].arg;
break;
case (MDOC_Offset):
- if (-1 != offset)
+ if (offset >= 0)
return(mdoc_nerr(mdoc, n, EARGREP));
+ if (type < 0 && ! mdoc_nwarn(mdoc, n, ENOTYPE))
+ return(0);
offset = n->args->argv[pos].arg;
break;
default:
break;
}
- if (-1 == type)
+ if (type < 0)
return(mdoc_nerr(mdoc, n, ELISTTYPE));
/*
switch (type) {
case (MDOC_Tag):
- if (-1 == width && ! mdoc_nwarn(mdoc, n, EMISSWIDTH))
+ if (width < 0 && ! mdoc_nwarn(mdoc, n, EMISSWIDTH))
return(0);
break;
case (MDOC_Column):
/* FALLTHROUGH */
case (MDOC_Diag):
/* FALLTHROUGH */
+ case (MDOC_Ohang):
+ /* FALLTHROUGH */
case (MDOC_Inset):
/* FALLTHROUGH */
case (MDOC_Item):
- if (-1 != width && ! mdoc_nwarn(mdoc, n, ENOWIDTH))
+ if (width >= 0 && ! mdoc_nwarn(mdoc, n, ENOWIDTH))
return(0);
break;
default:
for (i = 0, err = type = 0; ! err &&
i < (int)n->args->argc; i++)
switch (n->args->argv[i].arg) {
+ case (MDOC_Centred):
+ /* FALLTHROUGH */
case (MDOC_Ragged):
/* FALLTHROUGH */
case (MDOC_Unfilled):
case (MDOC_Filled):
/* FALLTHROUGH */
case (MDOC_Literal):
- /* FALLTHROUGH */
- case (MDOC_File):
if (0 == type++)
break;
return(mdoc_nerr(mdoc, n, EMULTIDISP));
pre_dt(PRE_ARGS)
{
+ /* FIXME: make sure is capitalised. */
+
if (0 == mdoc->meta.date || mdoc->meta.os)
if ( ! mdoc_nwarn(mdoc, n, EPROLOOO))
return(0);
return(1);
else if (0 == strcmp(p, "Li"))
return(1);
- else if (0 == strcmp(p, "Sm"))
+ else if (0 == strcmp(p, "Sy"))
return(1);
return(mdoc_nerr(mdoc, head, EFONT));
}
+static int
+post_vt(POST_ARGS)
+{
+ const struct mdoc_node *n;
+
+ /*
+ * The Vt macro comes in both ELEM and BLOCK form, both of which
+ * have different syntaxes (yet more context-sensitive
+ * behaviour). ELEM types must have a child; BLOCK types,
+ * specifically the BODY, should only have TEXT children.
+ */
+
+ if (MDOC_ELEM == mdoc->last->type)
+ return(eerr_ge1(mdoc));
+ if (MDOC_BODY != mdoc->last->type)
+ return(1);
+
+ for (n = mdoc->last->child; n; n = n->next)
+ if (MDOC_TEXT != n->type)
+ if ( ! mdoc_nwarn(mdoc, n, EBADCHILD))
+ return(0);
+
+ return(1);
+}
+
+
static int
post_nm(POST_ARGS)
{
if (mdoc->last->args) {
if (NULL == mdoc->last->child)
return(1);
- return(mdoc_nerr(mdoc, mdoc->last, ELINE));
+ return(mdoc_nerr(mdoc, mdoc->last, ENOLINE));
}
if (mdoc->last->child)
}
-static int
-post_args(POST_ARGS)
-{
-
- if (mdoc->last->args)
- return(1);
- return(mdoc_nerr(mdoc, mdoc->last, ELINE));
-}
-
-
static int
post_it(POST_ARGS)
{
c = mdoc->last->child;
for (i = 0; c && MDOC_HEAD == c->type; c = c->next)
i++;
- if (i == cols)
+
+ if (i < cols || i == (cols + 1)) {
+ if ( ! mdoc_vwarn(mdoc, mdoc->last->line,
+ mdoc->last->pos, "column "
+ "mismatch: have %d, want %d",
+ i, cols))
+ return(0);
break;
- return(mdoc_verr(mdoc, mdoc->last->line, mdoc->last->pos,
- "column mismatch (have %d, want %d)",
- i, cols));
+ } else if (i == cols)
+ break;
+
+ return(mdoc_verr(mdoc, mdoc->last->line,
+ mdoc->last->pos, "column mismatch: "
+ "have %d, want %d", i, cols));
default:
break;
}
if (NULL == mdoc->last->child)
return(1);
+ /*
+ * We only allow certain children of `Bl'. This is usually on
+ * `It', but apparently `Sm' occurs here and there, so we let
+ * that one through, too.
+ */
+
/* LINTED */
for (n = mdoc->last->child; n; n = n->next) {
- if (MDOC_BLOCK == n->type)
- if (MDOC_It == n->tok)
- continue;
- return(mdoc_verr(mdoc, n->line, n->pos,
- "bad child of parent %s",
- mdoc_macronames[mdoc->last->tok]));
+ if (MDOC_BLOCK == n->type && MDOC_It == n->tok)
+ continue;
+ if (MDOC_Sm == n->tok)
+ continue;
+ return(mdoc_nerr(mdoc, n, EBADCHILD));
}
return(1);
}
+static int
+post_rs(POST_ARGS)
+{
+ struct mdoc_node *nn;
+
+ if (MDOC_BODY != mdoc->last->type)
+ return(1);
+
+ for (nn = mdoc->last->child; nn; nn = nn->next)
+ switch (nn->tok) {
+ case(MDOC__U):
+ /* FALLTHROUGH */
+ case(MDOC__Q):
+ /* FALLTHROUGH */
+ case(MDOC__C):
+ /* FALLTHROUGH */
+ case(MDOC__A):
+ /* FALLTHROUGH */
+ case(MDOC__B):
+ /* FALLTHROUGH */
+ case(MDOC__D):
+ /* FALLTHROUGH */
+ case(MDOC__I):
+ /* FALLTHROUGH */
+ case(MDOC__J):
+ /* FALLTHROUGH */
+ case(MDOC__N):
+ /* FALLTHROUGH */
+ case(MDOC__O):
+ /* FALLTHROUGH */
+ case(MDOC__P):
+ /* FALLTHROUGH */
+ case(MDOC__R):
+ /* FALLTHROUGH */
+ case(MDOC__T):
+ /* FALLTHROUGH */
+ case(MDOC__V):
+ break;
+ default:
+ return(mdoc_nerr(mdoc, nn, EBADCHILD));
+ }
+
+ return(1);
+}
+
+
static int
post_sh(POST_ARGS)
{
return(0);
}
+ assert(n);
if (MDOC_BLOCK == n->type && MDOC_Nd == n->tok)
return(1);
return(mdoc_nwarn(mdoc, mdoc->last, ENAMESECINC));