Proper leading spaces for new- and old-groff in -man -Tascii mode.

[mandoc.git] / mdoc.c

diff --git a/mdoc.c b/mdoc.c
index 6bcea3a862123de824a6879fa148d41bceb4265b..aa6891fd7f00a314699eb26f7b584b70af3daa46 100644 (file)
--- a/mdoc.c
+++ b/mdoc.c
@@ -1,4 +1,4 @@
-/*     $Id: mdoc.c,v 1.98 2009/07/29 08:52:24 kristaps Exp $ */
+/*     $Id: mdoc.c,v 1.127 2010/05/08 10:25:27 kristaps Exp $ */

 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
  *
 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
  *


@@ -14,20 +14,27 @@
  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  */
  * 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 <stdarg.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <assert.h>
 #include <ctype.h>
 #include <stdarg.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>

+#include <time.h>

 
 #include "libmdoc.h"
 
 #include "libmdoc.h"

+#include "libmandoc.h"

 
 const  char *const __mdoc_merrnames[MERRMAX] = {                
        "trailing whitespace", /* ETAILWS */
        "unexpected quoted parameter", /* EQUOTPARM */
        "unterminated quoted parameter", /* EQUOTTERM */
 
 const  char *const __mdoc_merrnames[MERRMAX] = {                
        "trailing whitespace", /* ETAILWS */
        "unexpected quoted parameter", /* EQUOTPARM */
        "unterminated quoted parameter", /* EQUOTTERM */

-       "system: malloc error", /* EMALLOC */

        "argument parameter suggested", /* EARGVAL */
        "macro disallowed in prologue", /* EBODYPROL */
        "macro disallowed in body", /* EPROLBODY */
        "argument parameter suggested", /* EARGVAL */
        "macro disallowed in prologue", /* EBODYPROL */
        "macro disallowed in body", /* EPROLBODY */


@@ -74,8 +81,10 @@ const        char *const __mdoc_merrnames[MERRMAX] = {
        "unclosed explicit scope", /* EOPEN */
        "unterminated quoted phrase", /* EQUOTPHR */
        "closure macro without prior context", /* ENOCTX */
        "unclosed explicit scope", /* EOPEN */
        "unterminated quoted phrase", /* EQUOTPHR */
        "closure macro without prior context", /* ENOCTX */

-       "invalid whitespace after control character", /* ESPACE */
-       "no description found for library" /* ELIB */
+       "no description found for library", /* ELIB */
+       "bad child for parent context", /* EBADCHILD */
+       "list arguments preceding type", /* ENOTYPE */
+       "deprecated comment style", /* EBADCOMMENT */

 };
 
 const  char *const __mdoc_macronames[MDOC_MAX] = {              
 };
 
 const  char *const __mdoc_macronames[MDOC_MAX] = {              


@@ -90,11 +99,11 @@ const       char *const __mdoc_macronames[MDOC_MAX] = {
        "Nm",           "Op",           "Ot",           "Pa",
        "Rv",           "St",           "Va",           "Vt",
        /* LINTED */
        "Nm",           "Op",           "Ot",           "Pa",
        "Rv",           "St",           "Va",           "Vt",
        /* LINTED */

-       "Xr",           "\%A",          "\%B",          "\%D",
+       "Xr",           "%A",           "%B",           "%D",

        /* LINTED */
        /* LINTED */

-       "\%I",          "\%J",          "\%N",          "\%O",
+       "%I",           "%J",           "%N",           "%O",

        /* LINTED */
        /* LINTED */

-       "\%P",          "\%R",          "\%T",          "\%V",
+       "%P",           "%R",           "%T",           "%V",

        "Ac",           "Ao",           "Aq",           "At",
        "Bc",           "Bf",           "Bo",           "Bq",
        "Bsx",          "Bx",           "Db",           "Dc",
        "Ac",           "Ao",           "Aq",           "At",
        "Bc",           "Bf",           "Bo",           "Bq",
        "Bsx",          "Bx",           "Db",           "Dc",


@@ -111,9 +120,11 @@ const      char *const __mdoc_macronames[MDOC_MAX] = {
        "Fr",           "Ud",           "Lb",           "Lp",
        "Lk",           "Mt",           "Brq",          "Bro",
        /* LINTED */
        "Fr",           "Ud",           "Lb",           "Lp",
        "Lk",           "Mt",           "Brq",          "Bro",
        /* LINTED */

-       "Brc",          "\%C",          "Es",           "En",
+       "Brc",          "%C",           "Es",           "En",
+       /* LINTED */
+       "Dx",           "%Q",           "br",           "sp",

        /* LINTED */
        /* LINTED */

-       "Dx",           "\%Q",          "br",           "sp"
+       "%U"

        };
 
 const  char *const __mdoc_argnames[MDOC_ARG_MAX] = {            
        };
 
 const  char *const __mdoc_argnames[MDOC_ARG_MAX] = {            


@@ -125,27 +136,24 @@ const     char *const __mdoc_argnames[MDOC_ARG_MAX] = {
        "ohang",                "inset",                "column",        
        "width",                "compact",              "std",   
        "filled",               "words",                "emphasis",
        "ohang",                "inset",                "column",        
        "width",                "compact",              "std",   
        "filled",               "words",                "emphasis",

-       "symbolic",             "nested"
+       "symbolic",             "nested",               "centered"

        };
 
 const  char * const *mdoc_macronames = __mdoc_macronames;
 const  char * const *mdoc_argnames = __mdoc_argnames;
 
        };
 
 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_free1(struct mdoc *);

-static int               mdoc_alloc1(struct mdoc *);
+static void              mdoc_alloc1(struct mdoc *);

 static struct mdoc_node *node_alloc(struct mdoc *, int, int, 
 static struct mdoc_node *node_alloc(struct mdoc *, int, int, 

-                               int, enum mdoc_type);
+                               enum mdoct, enum mdoc_type);

 static int               node_append(struct mdoc *, 
                                struct mdoc_node *);
 static int               node_append(struct mdoc *, 
                                struct mdoc_node *);

-static int               parsetext(struct mdoc *, int, char *);
-static int               parsemacro(struct mdoc *, int, char *);
+static int               mdoc_ptext(struct mdoc *, int, char *);
+static int               mdoc_pmacro(struct mdoc *, int, char *);

 static int               macrowarn(struct mdoc *, int, const char *);
 static int               macrowarn(struct mdoc *, int, const char *);

-static int               pstring(struct mdoc *, int, int, 
-                               const char *, size_t);
-
-#ifdef __linux__
-extern size_t            strlcpy(char *, const char *, size_t);
-#endif

 
 
 const struct mdoc_node *
 
 
 const struct mdoc_node *


@@ -172,7 +180,7 @@ mdoc_free1(struct mdoc *mdoc)
 {
 
        if (mdoc->first)
 {
 
        if (mdoc->first)

-               mdoc_node_freelist(mdoc->first);
+               mdoc_node_delete(mdoc, mdoc->first);

        if (mdoc->meta.title)
                free(mdoc->meta.title);
        if (mdoc->meta.os)
        if (mdoc->meta.title)
                free(mdoc->meta.title);
        if (mdoc->meta.os)


@@ -189,21 +197,17 @@ mdoc_free1(struct mdoc *mdoc)
 /*
  * Allocate all volatile resources (parse tree, meta-data, fields).
  */
 /*
  * Allocate all volatile resources (parse tree, meta-data, fields).
  */

-static int
+static void

 mdoc_alloc1(struct mdoc *mdoc)
 {
 
 mdoc_alloc1(struct mdoc *mdoc)
 {
 

-       bzero(&mdoc->meta, sizeof(struct mdoc_meta));
+       memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));

        mdoc->flags = 0;
        mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
        mdoc->flags = 0;
        mdoc->lastnamed = mdoc->lastsec = SEC_NONE;

-       mdoc->last = calloc(1, sizeof(struct mdoc_node));
-       if (NULL == mdoc->last)
-               return(0);
-
+       mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));

        mdoc->first = mdoc->last;
        mdoc->last->type = MDOC_ROOT;
        mdoc->next = MDOC_NEXT_CHILD;
        mdoc->first = mdoc->last;
        mdoc->last->type = MDOC_ROOT;
        mdoc->next = MDOC_NEXT_CHILD;

-       return(1);

 }
 
 
 }
 
 


@@ -213,12 +217,12 @@ mdoc_alloc1(struct mdoc *mdoc)
  * and the parser is ready for re-invocation on a new tree; however,
  * cross-parse non-volatile data is kept intact.
  */
  * and the parser is ready for re-invocation on a new tree; however,
  * cross-parse non-volatile data is kept intact.
  */

-int
+void

 mdoc_reset(struct mdoc *mdoc)
 {
 
        mdoc_free1(mdoc);
 mdoc_reset(struct mdoc *mdoc)
 {
 
        mdoc_free1(mdoc);

-       return(mdoc_alloc1(mdoc));
+       mdoc_alloc1(mdoc);

 }
 
 
 }
 
 


@@ -231,8 +235,6 @@ mdoc_free(struct mdoc *mdoc)
 {
 
        mdoc_free1(mdoc);
 {
 
        mdoc_free1(mdoc);

-       if (mdoc->htab)
-               mdoc_hash_free(mdoc->htab);

        free(mdoc);
 }
 
        free(mdoc);
 }
 


@@ -245,22 +247,17 @@ mdoc_alloc(void *data, int pflags, const struct mdoc_cb *cb)
 {
        struct mdoc     *p;
 
 {
        struct mdoc     *p;
 

-       if (NULL == (p = calloc(1, sizeof(struct mdoc))))
-               return(NULL);
+       p = mandoc_calloc(1, sizeof(struct mdoc));
+

        if (cb)
        if (cb)

-               (void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb));
+               memcpy(&p->cb, cb, sizeof(struct mdoc_cb));

 
        p->data = data;
        p->pflags = pflags;
 
 
        p->data = data;
        p->pflags = pflags;
 

-       if (NULL == (p->htab = mdoc_hash_alloc())) {
-               free(p);
-               return(NULL);
-       } else if (mdoc_alloc1(p))
-               return(p);
-
-       free(p);
-       return(NULL);
+       mdoc_hash_init();
+       mdoc_alloc1(p);
+       return(p);

 }
 
 
 }
 
 


@@ -283,7 +280,7 @@ mdoc_endparse(struct mdoc *m)
 
 /*
  * Main parse routine.  Parses a single line -- really just hands off to
 
 /*
  * Main parse routine.  Parses a single line -- really just hands off to

- * the macro (parsemacro()) or text parser (parsetext()).
+ * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).

  */
 int
 mdoc_parseln(struct mdoc *m, int ln, char *buf)
  */
 int
 mdoc_parseln(struct mdoc *m, int ln, char *buf)


@@ -292,8 +289,8 @@ mdoc_parseln(struct mdoc *m, int ln, char *buf)
        if (MDOC_HALT & m->flags)
                return(0);
 
        if (MDOC_HALT & m->flags)
                return(0);
 

-       return('.' == *buf ? parsemacro(m, ln, buf) :
-                       parsetext(m, ln, buf));
+       return('.' == *buf ? mdoc_pmacro(m, ln, buf) :
+                       mdoc_ptext(m, ln, buf));

 }
 
 
 }
 
 


@@ -348,19 +345,33 @@ mdoc_err(struct mdoc *m, int line, int pos, int iserr, enum merr type)
 
 
 int
 
 
 int

-mdoc_macro(struct mdoc *m, int tok, 
+mdoc_macro(struct mdoc *m, enum mdoct tok, 

                int ln, int pp, int *pos, char *buf)
 {
                int ln, int pp, int *pos, char *buf)
 {

-       /*
-        * If we're in the prologue, deny "body" macros.  Similarly, if
-        * we're in the body, deny prologue calls.
-        */
+       assert(tok < MDOC_MAX);
+
+       /* If we're in the body, deny prologue calls. */
+

        if (MDOC_PROLOGUE & mdoc_macros[tok].flags && 
                        MDOC_PBODY & m->flags)
                return(mdoc_perr(m, ln, pp, EPROLBODY));
        if (MDOC_PROLOGUE & mdoc_macros[tok].flags && 
                        MDOC_PBODY & m->flags)
                return(mdoc_perr(m, ln, pp, EPROLBODY));

+
+       /* If we're in the prologue, deny "body" macros.  */
+

        if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && 
        if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && 

-                       ! (MDOC_PBODY & m->flags))
-               return(mdoc_perr(m, ln, pp, EBODYPROL));
+                       ! (MDOC_PBODY & m->flags)) {
+               if ( ! mdoc_pwarn(m, ln, pp, EBODYPROL))
+                       return(0);
+               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 (0 == m->meta.date)
+                       m->meta.date = time(NULL);
+               m->flags |= MDOC_PBODY;
+       }

 
        return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf));
 }
 
        return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf));
 }


@@ -431,41 +442,37 @@ node_append(struct mdoc *mdoc, struct mdoc_node *p)
 
 
 static struct mdoc_node *
 
 
 static struct mdoc_node *

-node_alloc(struct mdoc *m, int line, 
-               int pos, int tok, enum mdoc_type type)
+node_alloc(struct mdoc *m, int line, int pos, 
+               enum mdoct tok, enum mdoc_type type)

 {
        struct mdoc_node *p;
 
 {
        struct mdoc_node *p;
 

-       if (NULL == (p = calloc(1, sizeof(struct mdoc_node)))) {
-               (void)mdoc_nerr(m, m->last, EMALLOC);
-               return(NULL);
-       }
-
+       p = mandoc_calloc(1, sizeof(struct mdoc_node));

        p->sec = m->lastsec;
        p->line = line;
        p->pos = pos;
        p->tok = tok;
        p->sec = m->lastsec;
        p->line = line;
        p->pos = pos;
        p->tok = tok;

-       if (MDOC_TEXT != (p->type = type))
-               assert(p->tok >= 0);
+       p->type = type;

 
        return(p);
 }
 
 
 int
 
        return(p);
 }
 
 
 int

-mdoc_tail_alloc(struct mdoc *m, int line, int pos, int tok)
+mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)

 {
        struct mdoc_node *p;
 
        p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 {
        struct mdoc_node *p;
 
        p = node_alloc(m, line, pos, tok, MDOC_TAIL);

-       if (NULL == p)
+       if ( ! node_append(m, p))

                return(0);
                return(0);

-       return(node_append(m, p));
+       m->next = MDOC_NEXT_CHILD;
+       return(1);

 }
 
 
 int
 }
 
 
 int

-mdoc_head_alloc(struct mdoc *m, int line, int pos, int tok)
+mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)

 {
        struct mdoc_node *p;
 
 {
        struct mdoc_node *p;
 


@@ -473,86 +480,80 @@ mdoc_head_alloc(struct mdoc *m, int line, int pos, int tok)
        assert(m->last);
 
        p = node_alloc(m, line, pos, tok, MDOC_HEAD);
        assert(m->last);
 
        p = node_alloc(m, line, pos, tok, MDOC_HEAD);

-       if (NULL == p)
+       if ( ! node_append(m, p))

                return(0);
                return(0);

-       return(node_append(m, p));
+       m->next = MDOC_NEXT_CHILD;
+       return(1);

 }
 
 
 int
 }
 
 
 int

-mdoc_body_alloc(struct mdoc *m, int line, int pos, int tok)
+mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)

 {
        struct mdoc_node *p;
 
        p = node_alloc(m, line, pos, tok, MDOC_BODY);
 {
        struct mdoc_node *p;
 
        p = node_alloc(m, line, pos, tok, MDOC_BODY);

-       if (NULL == p)
+       if ( ! node_append(m, p))

                return(0);
                return(0);

-       return(node_append(m, p));
+       m->next = MDOC_NEXT_CHILD;
+       return(1);

 }
 
 
 int
 mdoc_block_alloc(struct mdoc *m, int line, int pos, 
 }
 
 
 int
 mdoc_block_alloc(struct mdoc *m, int line, int pos, 

-               int tok, struct mdoc_arg *args)
+               enum mdoct tok, struct mdoc_arg *args)

 {
        struct mdoc_node *p;
 
        p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
 {
        struct mdoc_node *p;
 
        p = node_alloc(m, line, pos, tok, MDOC_BLOCK);

-       if (NULL == p)
-               return(0);

        p->args = args;
        if (p->args)
                (args->refcnt)++;
        p->args = args;
        if (p->args)
                (args->refcnt)++;

-       return(node_append(m, p));
+       if ( ! node_append(m, p))
+               return(0);
+       m->next = MDOC_NEXT_CHILD;
+       return(1);

 }
 
 
 int
 mdoc_elem_alloc(struct mdoc *m, int line, int pos, 
 }
 
 
 int
 mdoc_elem_alloc(struct mdoc *m, int line, int pos, 

-               int tok, struct mdoc_arg *args)
+               enum mdoct tok, struct mdoc_arg *args)

 {
        struct mdoc_node *p;
 
        p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 {
        struct mdoc_node *p;
 
        p = node_alloc(m, line, pos, tok, MDOC_ELEM);

-       if (NULL == p)
-               return(0);

        p->args = args;
        if (p->args)
                (args->refcnt)++;
        p->args = args;
        if (p->args)
                (args->refcnt)++;

-       return(node_append(m, p));
+       if ( ! node_append(m, p))
+               return(0);
+       m->next = MDOC_NEXT_CHILD;
+       return(1);

 }
 
 
 }
 
 

-static int
-pstring(struct mdoc *m, int line, int pos, const char *p, size_t len)
+int
+mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)

 {
        struct mdoc_node *n;
 {
        struct mdoc_node *n;

-       size_t            sv;
-
-       n = node_alloc(m, line, pos, -1, MDOC_TEXT);
-       if (NULL == n)
-               return(mdoc_nerr(m, m->last, EMALLOC));
+       size_t            sv, len;

 
 

-       n->string = malloc(len + 1);
-       if (NULL == n->string) {
-               free(n);
-               return(mdoc_nerr(m, m->last, EMALLOC));
-       }
+       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);
 
        sv = strlcpy(n->string, p, len + 1);
 
        /* Prohibit truncation. */
        assert(sv < len + 1);
 

-       return(node_append(m, n));
-}
-
-
-int
-mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
-{
+       if ( ! node_append(m, n))
+               return(0);

 
 

-       return(pstring(m, line, pos, p, strlen(p)));
+       m->next = MDOC_NEXT_SIBLING;
+       return(1);

 }
 
 
 }
 
 


@@ -560,8 +561,6 @@ void
 mdoc_node_free(struct mdoc_node *p)
 {
 
 mdoc_node_free(struct mdoc_node *p)
 {
 

-       if (p->parent)
-               p->parent->nchild--;

        if (p->string)
                free(p->string);
        if (p->args)
        if (p->string)
                free(p->string);
        if (p->args)


@@ -570,16 +569,53 @@ mdoc_node_free(struct mdoc_node *p)
 }
 
 
 }
 
 

-void
-mdoc_node_freelist(struct mdoc_node *p)
+static void
+mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)

 {
 
 {
 

-       if (p->child)
-               mdoc_node_freelist(p->child);
-       if (p->next)
-               mdoc_node_freelist(p->next);
+       /* 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;
+       }
+
+       /* 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_delete(struct mdoc *m, struct mdoc_node *p)
+{
+
+       while (p->child) {
+               assert(p->nchild);
+               mdoc_node_delete(m, p->child);
+       }

        assert(0 == p->nchild);
        assert(0 == p->nchild);

+
+       mdoc_node_unlink(m, p);

        mdoc_node_free(p);
 }
 
        mdoc_node_free(p);
 }
 


@@ -589,74 +625,79 @@ mdoc_node_freelist(struct mdoc_node *p)
  * control character.
  */
 static int
  * control character.
  */
 static int

-parsetext(struct mdoc *m, int line, char *buf)
+mdoc_ptext(struct mdoc *m, int line, char *buf)

 {
 {

-       int              i, j;
+       int              i;
+
+       /* Ignore bogus comments. */
+
+       if ('\\' == buf[0] && '.' == buf[1] && '\"' == buf[2])
+               return(mdoc_pwarn(m, line, 0, EBADCOMMENT));
+
+       /* No text before an initial macro. */

 
        if (SEC_NONE == m->lastnamed)
                return(mdoc_perr(m, line, 0, ETEXTPROL));
 
        if (SEC_NONE == m->lastnamed)
                return(mdoc_perr(m, line, 0, ETEXTPROL));

+
+       /* Literal just gets pulled in as-is. */

        
        

-       /*
-        * If in literal mode, then pass the buffer directly to the
-        * back-end, as it should be preserved as a single term.
-        */
+       if (MDOC_LITERAL & m->flags)
+               return(mdoc_word_alloc(m, line, 0, buf));
+
+       /* Check for a blank line, which may also consist of spaces. */

 
 

-       if (MDOC_LITERAL & m->flags) {
-               if ( ! mdoc_word_alloc(m, line, 0, buf))
+       for (i = 0; ' ' == buf[i]; i++)
+               /* Skip to first non-space. */ ;
+
+       if ('\0' == buf[i]) {
+               if ( ! mdoc_pwarn(m, line, 0, ENOBLANK))
+                       return(0);
+
+               /*
+                * Insert a `Pp' 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, 0, MDOC_Pp, NULL))

                        return(0);
                        return(0);

+

                m->next = MDOC_NEXT_SIBLING;
                return(1);
        }
 
                m->next = MDOC_NEXT_SIBLING;
                return(1);
        }
 

-       /* Disallow blank/white-space lines in non-literal mode. */
-
-       for (i = 0; ' ' == buf[i]; i++)
-               /* Skip leading whitespace. */ ;
-       if (0 == buf[i])
-               return(mdoc_perr(m, line, 0, ENOBLANK));
-
-       /*
-        * Break apart a free-form line into tokens.  Spaces are
-        * stripped out of the input.
+       /* 
+        * Warn if the last un-escaped character is whitespace. Then
+        * strip away the remaining spaces (tabs stay!).   

         */
 
         */
 

-       for (j = i; buf[i]; i++) {
-               if (' ' != buf[i])
-                       continue;
+       i = (int)strlen(buf);
+       assert(i);

 
 

-               /* Escaped whitespace. */
-               if (i && ' ' == buf[i] && '\\' == buf[i - 1])
-                       continue;
+       if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
+               if (i > 1 && '\\' != buf[i - 2])
+                       if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS))
+                               return(0);

 
 

-               buf[i++] = 0;
-               if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
-                       return(0);
-               m->next = MDOC_NEXT_SIBLING;
+               for (--i; i && ' ' == buf[i]; i--)
+                       /* Spin back to non-space. */ ;

 
 

-               for ( ; ' ' == buf[i]; i++)
-                       /* Skip trailing whitespace. */ ;
+               /* Jump ahead of escaped whitespace. */
+               i += '\\' == buf[i] ? 2 : 1;

 
 

-               j = i;
-               if (0 == buf[i])
-                       break;
+               buf[i] = '\0';

        }
 
        }
 

-       if (j != i && ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
-               return(0);
+       /* Allocate the whole word. */

 
 

-       m->next = MDOC_NEXT_SIBLING;
-       return(1);
+       return(mdoc_word_alloc(m, line, 0, buf));

 }
 
 
 }
 
 

-
-

 static int
 macrowarn(struct mdoc *m, int ln, const char *buf)
 {
        if ( ! (MDOC_IGN_MACRO & m->pflags))
 static int
 macrowarn(struct mdoc *m, int ln, const char *buf)
 {
        if ( ! (MDOC_IGN_MACRO & m->pflags))

-               return(mdoc_verr(m, ln, 0, 
-                               "unknown macro: %s%s", 
+               return(mdoc_verr(m, ln, 0, "unknown macro: %s%s", 

                                buf, strlen(buf) > 3 ? "..." : ""));
        return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s",
                                buf, strlen(buf) > 3 ? "..." : ""));
                                buf, strlen(buf) > 3 ? "..." : ""));
        return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s",
                                buf, strlen(buf) > 3 ? "..." : ""));


@@ -668,43 +709,53 @@ macrowarn(struct mdoc *m, int ln, const char *buf)
  * character.
  */
 int
  * character.
  */
 int

-parsemacro(struct mdoc *m, int ln, char *buf)
+mdoc_pmacro(struct mdoc *m, int ln, char *buf)

 {
 {

-       int               i, c;
-       char              mac[5];
+       enum mdoct      tok;
+       int             i, j;
+       char            mac[5];

 
        /* Empty lines are ignored. */
 
 
        /* Empty lines are ignored. */
 

-       if (0 == buf[1])
+       if ('\0' == buf[1])

                return(1);
 
                return(1);
 

-       if (' ' == buf[1]) {
-               i = 2;
+       i = 1;
+
+       /* Accept whitespace after the initial control char. */
+
+       if (' ' == buf[i]) {
+               i++;

                while (buf[i] && ' ' == buf[i])
                        i++;
                while (buf[i] && ' ' == buf[i])
                        i++;

-               if (0 == buf[i])
+               if ('\0' == buf[i])

                        return(1);
                        return(1);

-               return(mdoc_perr(m, ln, 1, ESPACE));

        }
 
        /* Copy the first word into a nil-terminated buffer. */
 
        }
 
        /* Copy the first word into a nil-terminated buffer. */
 

-       for (i = 1; i < 5; i++) {
-               if (0 == (mac[i - 1] = buf[i]))
+       for (j = 0; j < 4; j++, i++) {
+               if ('\0' == (mac[j] = buf[i]))

                        break;
                else if (' ' == buf[i])
                        break;
                        break;
                else if (' ' == buf[i])
                        break;

+
+               /* Check for invalid characters. */
+
+               if (isgraph((u_char)buf[i]))
+                       continue;
+               return(mdoc_perr(m, ln, i, EPRINT));

        }
 
        }
 

-       mac[i - 1] = 0;
+       mac[j] = 0;

 
 

-       if (i == 5 || i <= 2) {
+       if (j == 4 || j < 2) {

                if ( ! macrowarn(m, ln, mac))
                        goto err;
                return(1);
        } 
        
                if ( ! macrowarn(m, ln, mac))
                        goto err;
                return(1);
        } 
        

-       if (MDOC_MAX == (c = mdoc_hash_find(m->htab, mac))) {
+       if (MDOC_MAX == (tok = mdoc_hash_find(mac))) {

                if ( ! macrowarn(m, ln, mac))
                        goto err;
                return(1);
                if ( ! macrowarn(m, ln, mac))
                        goto err;
                return(1);


@@ -715,11 +766,20 @@ parsemacro(struct mdoc *m, int ln, char *buf)
        while (buf[i] && ' ' == buf[i])
                i++;
 
        while (buf[i] && ' ' == buf[i])
                i++;
 

+       /* 
+        * 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_pwarn(m, ln, i - 1, ETAILWS))
+                       goto err;
+

        /* 
         * Begin recursive parse sequence.  Since we're at the start of
         * the line, we don't need to do callable/parseable checks.
         */
        /* 
         * Begin recursive parse sequence.  Since we're at the start of
         * the line, we don't need to do callable/parseable checks.
         */

-       if ( ! mdoc_macro(m, c, ln, 1, &i, buf)) 
+       if ( ! mdoc_macro(m, tok, ln, 1, &i, buf)) 

                goto err;
 
        return(1);
                goto err;
 
        return(1);


@@ -729,3 +789,5 @@ err:        /* Error out. */
        m->flags |= MDOC_HALT;
        return(0);
 }
        m->flags |= MDOC_HALT;
        return(0);
 }

+
+