Merging patch by Ingo Schwarze.

[mandoc.git] / mdoc.c

diff --git a/mdoc.c b/mdoc.c
index e382e6929678eb0a244cbabd33bef237f5bab89e..7e1d0f4b7a0ed1eafe4af7806c3ad1146c1a5797 100644 (file)
--- a/mdoc.c
+++ b/mdoc.c
@@ -1,4 +1,4 @@
-/*     $Id: mdoc.c,v 1.103 2009/08/20 11:44:47 kristaps Exp $ */
+/*     $Id: mdoc.c,v 1.119 2010/04/03 13:02:35 kristaps Exp $ */

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


@@ -14,6 +14,12 @@
  * 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 <assert.h>
 #include <ctype.h>
 #include <stdarg.h>


@@ -22,12 +28,12 @@
 #include <string.h>
 
 #include "libmdoc.h"
 #include <string.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 */


@@ -74,7 +80,9 @@ 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 */

-       "no description found for library" /* ELIB */
+       "no description found for library", /* ELIB */
+       "bad child for parent context", /* EBADCHILD */
+       "list arguments preceding type", /* ENOTYPE */

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


@@ -89,11 +97,11 @@ 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",


@@ -110,9 +118,11 @@ 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 */
        /* LINTED */

-       "Dx",           "\%Q",          "br",           "sp"
+       "Dx",           "%Q",           "br",           "sp",
+       /* LINTED */
+       "%U"

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


@@ -124,16 +134,16 @@ 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;
 
 static void              mdoc_free1(struct mdoc *);
        };
 
 const  char * const *mdoc_macronames = __mdoc_macronames;
 const  char * const *mdoc_argnames = __mdoc_argnames;
 
 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               parsetext(struct mdoc *, int, char *);
 static int               node_append(struct mdoc *, 
                                struct mdoc_node *);
 static int               parsetext(struct mdoc *, int, char *);


@@ -142,11 +152,6 @@ static     int               macrowarn(struct mdoc *, int, const char *);
 static int               pstring(struct mdoc *, int, int, 
                                const char *, size_t);
 
 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 *
 mdoc_node(const struct mdoc *m)
 {
 const struct mdoc_node *
 mdoc_node(const struct mdoc *m)
 {


@@ -188,21 +193,17 @@ 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);

 }
 
 
 }
 
 


@@ -212,12 +213,12 @@ 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);

 }
 
 
 }
 
 


@@ -230,8 +231,6 @@ mdoc_free(struct mdoc *mdoc)
 {
 
        mdoc_free1(mdoc);
 {
 
        mdoc_free1(mdoc);

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

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


@@ -244,22 +243,17 @@ 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);

 }
 
 
 }
 
 


@@ -347,9 +341,11 @@ 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)
 {

+
+       assert(tok < MDOC_MAX);

        /*
         * If we're in the prologue, deny "body" macros.  Similarly, if
         * we're in the body, deny prologue calls.
        /*
         * If we're in the prologue, deny "body" macros.  Similarly, if
         * we're in the body, deny prologue calls.


@@ -430,35 +426,28 @@ 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)
-               return(0);

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


@@ -467,7 +456,7 @@ mdoc_tail_alloc(struct mdoc *m, int line, int pos, int tok)
 
 
 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;
 


@@ -475,8 +464,6 @@ 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)
-               return(0);

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


@@ -485,13 +472,11 @@ mdoc_head_alloc(struct mdoc *m, int line, int pos, int tok)
 
 
 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)
-               return(0);

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


@@ -501,13 +486,11 @@ mdoc_body_alloc(struct mdoc *m, int line, int pos, int tok)
 
 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)++;


@@ -520,13 +503,11 @@ mdoc_block_alloc(struct mdoc *m, int line, int pos,
 
 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)++;


@@ -544,15 +525,7 @@ pstring(struct mdoc *m, int line, int pos, const char *p, size_t len)
        size_t            sv;
 
        n = node_alloc(m, line, pos, -1, MDOC_TEXT);
        size_t            sv;
 
        n = node_alloc(m, line, pos, -1, MDOC_TEXT);

-       if (NULL == n)
-               return(mdoc_nerr(m, m->last, EMALLOC));
-
-       n->string = malloc(len + 1);
-       if (NULL == n->string) {
-               free(n);
-               return(mdoc_nerr(m, m->last, EMALLOC));
-       }
-
+       n->string = mandoc_malloc(len + 1);

        sv = strlcpy(n->string, p, len + 1);
 
        /* Prohibit truncation. */
        sv = strlcpy(n->string, p, len + 1);
 
        /* Prohibit truncation. */


@@ -609,6 +582,7 @@ static int
 parsetext(struct mdoc *m, int line, char *buf)
 {
        int              i, j;
 parsetext(struct mdoc *m, int line, char *buf)
 {
        int              i, j;

+       char             sv;

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


@@ -625,8 +599,19 @@ parsetext(struct mdoc *m, int line, char *buf)
 
        for (i = 0; ' ' == buf[i]; i++)
                /* Skip leading whitespace. */ ;
 
        for (i = 0; ' ' == buf[i]; i++)
                /* Skip leading whitespace. */ ;

-       if (0 == buf[i])
-               return(mdoc_perr(m, line, 0, ENOBLANK));
+
+       if ('\0' == buf[i]) {
+               if ( ! mdoc_pwarn(m, line, 0, ENOBLANK))
+                       return(0);
+               /*
+                * Assume that a `Pp' should be inserted 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);
+       }

 
        /*
         * Break apart a free-form line into tokens.  Spaces are
 
        /*
         * Break apart a free-form line into tokens.  Spaces are


@@ -641,15 +626,30 @@ parsetext(struct mdoc *m, int line, char *buf)
                if (i && ' ' == buf[i] && '\\' == buf[i - 1])
                        continue;
 
                if (i && ' ' == buf[i] && '\\' == buf[i - 1])
                        continue;
 

-               buf[i++] = 0;
+               sv = buf[i];
+               buf[i++] = '\0';
+

                if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
                        return(0);
 
                if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
                        return(0);
 

+               /* Trailing whitespace?  Check at overwritten byte. */
+
+               if (' ' == sv && '\0' == buf[i])
+                       if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS))
+                               return(0);
+

                for ( ; ' ' == buf[i]; i++)
                        /* Skip trailing whitespace. */ ;
 
                j = i;
                for ( ; ' ' == buf[i]; i++)
                        /* Skip trailing whitespace. */ ;
 
                j = i;

-               if (0 == buf[i])
+
+               /* Trailing whitespace? */
+
+               if (' ' == buf[i - 1] && '\0' == buf[i])
+                       if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS))
+                               return(0);
+
+               if ('\0' == buf[i])

                        break;
        }
 
                        break;
        }
 


@@ -662,7 +662,6 @@ parsetext(struct mdoc *m, int line, char *buf)
 
 
 
 
 
 

-

 static int
 macrowarn(struct mdoc *m, int ln, const char *buf)
 {
 static int
 macrowarn(struct mdoc *m, int ln, const char *buf)
 {


@@ -687,7 +686,7 @@ parsemacro(struct mdoc *m, int ln, char *buf)
 
        /* Empty lines are ignored. */
 
 
        /* Empty lines are ignored. */
 

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

                return(1);
 
        i = 1;
                return(1);
 
        i = 1;


@@ -698,17 +697,23 @@ parsemacro(struct mdoc *m, int ln, char *buf)
                i++;
                while (buf[i] && ' ' == buf[i])
                        i++;
                i++;
                while (buf[i] && ' ' == buf[i])
                        i++;

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

                        return(1);
        }
 
        /* Copy the first word into a nil-terminated buffer. */
 
        for (j = 0; j < 4; j++, i++) {
                        return(1);
        }
 
        /* Copy the first word into a nil-terminated buffer. */
 
        for (j = 0; j < 4; j++, i++) {

-               if (0 == (mac[j] = buf[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[j] = 0;
        }
 
        mac[j] = 0;


@@ -719,7 +724,7 @@ parsemacro(struct mdoc *m, int ln, char *buf)
                return(1);
        } 
        
                return(1);
        } 
        

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

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


@@ -730,6 +735,12 @@ parsemacro(struct mdoc *m, int ln, char *buf)
        while (buf[i] && ' ' == buf[i])
                i++;
 
        while (buf[i] && ' ' == buf[i])
                i++;
 

+       /* Trailing whitespace? */
+
+       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.