Remove predefined strings from the chars.in file, as they're now local

[mandoc.git] / chars.c

diff --git a/chars.c b/chars.c
index 2653983256671ccc22cf116dcbf67b88086d7a06..b32c5bbaa261861de2a1550a33c5eddd1c89eac4 100644 (file)
--- a/chars.c
+++ b/chars.c
@@ -1,4 +1,4 @@
-/*     $Id: chars.c,v 1.36 2011/04/29 22:18:12 kristaps Exp $ */
+/*     $Id: chars.c,v 1.45 2011/05/24 21:24:16 kristaps Exp $ */

 /*
  * Copyright (c) 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2011 Ingo Schwarze <schwarze@openbsd.org>
 /*
  * Copyright (c) 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2011 Ingo Schwarze <schwarze@openbsd.org>


@@ -20,11 +20,13 @@
 #endif
 
 #include <assert.h>
 #endif
 
 #include <assert.h>

+#include <ctype.h>

 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 
 #include "mandoc.h"
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 
 #include "mandoc.h"

+#include "libmandoc.h"

 
 #define        PRINT_HI         126
 #define        PRINT_LO         32
 
 #define        PRINT_HI         126
 #define        PRINT_LO         32


@@ -34,20 +36,12 @@ struct      ln {
        const char       *code;
        const char       *ascii;
        int               unicode;
        const char       *code;
        const char       *ascii;
        int               unicode;

-       int               type;
-#define        CHARS_CHAR       (1 << 0)
-#define        CHARS_STRING     (1 << 1)
-#define CHARS_BOTH      (CHARS_CHAR | CHARS_STRING)

 };
 
 };
 

-#define        LINES_MAX         353
+#define        LINES_MAX         325

 
 #define CHAR(in, ch, code) \
 
 #define CHAR(in, ch, code) \

-       { NULL, (in), (ch), (code), CHARS_CHAR },
-#define STRING(in, ch, code) \
-       { NULL, (in), (ch), (code), CHARS_STRING },
-#define BOTH(in, ch, code) \
-       { NULL, (in), (ch), (code), CHARS_BOTH },
+       { NULL, (in), (ch), (code) },

 
 #define        CHAR_TBL_START    static struct ln lines[LINES_MAX] = {
 #define        CHAR_TBL_END      };
 
 #define        CHAR_TBL_START    static struct ln lines[LINES_MAX] = {
 #define        CHAR_TBL_END      };


@@ -55,13 +49,11 @@ struct      ln {
 #include "chars.in"
 
 struct mchars {
 #include "chars.in"
 
 struct mchars {

-       enum mcharst      type;

        struct ln       **htab;
 };
 
        struct ln       **htab;
 };
 

-static inline int        match(const struct ln *,
-                               const char *, size_t, int);
-static const struct ln  *find(struct mchars *, const char *, size_t, int);
+static inline int        match(const struct ln *, const char *, size_t);
+static const struct ln  *find(struct mchars *, const char *, size_t);

 
 void
 mchars_free(struct mchars *arg)
 
 void
 mchars_free(struct mchars *arg)


@@ -72,7 +64,7 @@ mchars_free(struct mchars *arg)
 }
 
 struct mchars *
 }
 
 struct mchars *

-mchars_init(enum mcharst type)
+mchars_alloc(void)

 {
        struct mchars    *tab;
        struct ln       **htab;
 {
        struct mchars    *tab;
        struct ln       **htab;


@@ -103,7 +95,6 @@ mchars_init(enum mcharst type)
        }
 
        tab->htab = htab;
        }
 
        tab->htab = htab;

-       tab->type = type;

        return(tab);
 }
 
        return(tab);
 }
 


@@ -116,7 +107,7 @@ mchars_spec2cp(struct mchars *arg, const char *p, size_t sz)
 {
        const struct ln *ln;
 
 {
        const struct ln *ln;
 

-       ln = find(arg, p, sz, CHARS_CHAR);
+       ln = find(arg, p, sz);

        if (NULL == ln)
                return(-1);
        return(ln->unicode);
        if (NULL == ln)
                return(-1);
        return(ln->unicode);


@@ -131,33 +122,43 @@ mchars_res2cp(struct mchars *arg, const char *p, size_t sz)
 {
        const struct ln *ln;
 
 {
        const struct ln *ln;
 

-       ln = find(arg, p, sz, CHARS_STRING);
+       ln = find(arg, p, sz);

        if (NULL == ln)
                return(-1);
        return(ln->unicode);
 }
 
        if (NULL == ln)
                return(-1);
        return(ln->unicode);
 }
 

-

 /*
 /*

- * Numbered character to literal character,
- * represented as a null-terminated string for additional safety.
+ * Numbered character string to ASCII codepoint.
+ * This can only be a printable character (i.e., alnum, punct, space) so
+ * prevent the character from ruining our state (backspace, newline, and
+ * so on).
+ * If the character is illegal, returns '\0'.

  */
  */

-const char *
+char

 mchars_num2char(const char *p, size_t sz)
 {
        int               i;
 mchars_num2char(const char *p, size_t sz)
 {
        int               i;

-       static char       c[2];

 
 

-       if (sz > 3)
-               return(NULL);
-       i = atoi(p);
-       if (i < 0 || i > 255)
-               return(NULL);
-       c[0] = (char)i;
-       c[1] = '\0';
-       return(c);
+       if ((i = mandoc_strntou(p, sz, 10)) < 0)
+               return('\0');
+       return(isprint(i) ? i : '\0');

 }
 
 }
 

+/*
+ * Hex character string to Unicode codepoint.
+ * If the character is illegal, returns '\0'.
+ */
+int
+mchars_num2uc(const char *p, size_t sz)
+{
+       int               i;
+
+       if ((i = mandoc_strntou(p, sz, 16)) < 0)
+               return('\0');
+       /* FIXME: make sure we're not in a bogus range. */
+       return(i > 0x80 && i <= 0x10FFFF ? i : '\0');
+}

 
 /* 
  * Special character to string array.
 
 /* 
  * Special character to string array.


@@ -167,7 +168,7 @@ mchars_spec2str(struct mchars *arg, const char *p, size_t sz, size_t *rsz)
 {
        const struct ln *ln;
 
 {
        const struct ln *ln;
 

-       ln = find(arg, p, sz, CHARS_CHAR);
+       ln = find(arg, p, sz);

        if (NULL == ln)
                return(NULL);
 
        if (NULL == ln)
                return(NULL);
 


@@ -175,7 +176,6 @@ mchars_spec2str(struct mchars *arg, const char *p, size_t sz, size_t *rsz)
        return(ln->ascii);
 }
 
        return(ln->ascii);
 }
 

-

 /* 
  * Reserved word to string array.
  */
 /* 
  * Reserved word to string array.
  */


@@ -184,7 +184,7 @@ mchars_res2str(struct mchars *arg, const char *p, size_t sz, size_t *rsz)
 {
        const struct ln *ln;
 
 {
        const struct ln *ln;
 

-       ln = find(arg, p, sz, CHARS_STRING);
+       ln = find(arg, p, sz);

        if (NULL == ln)
                return(NULL);
 
        if (NULL == ln)
                return(NULL);
 


@@ -193,7 +193,7 @@ mchars_res2str(struct mchars *arg, const char *p, size_t sz, size_t *rsz)
 }
 
 static const struct ln *
 }
 
 static const struct ln *

-find(struct mchars *tab, const char *p, size_t sz, int type)
+find(struct mchars *tab, const char *p, size_t sz)

 {
        struct ln        *pp, *prev;
        struct ln       **htab;
 {
        struct ln        *pp, *prev;
        struct ln       **htab;


@@ -219,7 +219,7 @@ find(struct mchars *tab, const char *p, size_t sz, int type)
                return(NULL);
 
        for (prev = NULL; pp; pp = pp->next) {
                return(NULL);
 
        for (prev = NULL; pp; pp = pp->next) {

-               if ( ! match(pp, p, sz, type)) {
+               if ( ! match(pp, p, sz)) {

                        prev = pp;
                        continue;
                }
                        prev = pp;
                        continue;
                }


@@ -237,11 +237,9 @@ find(struct mchars *tab, const char *p, size_t sz, int type)
 }
 
 static inline int
 }
 
 static inline int

-match(const struct ln *ln, const char *p, size_t sz, int type)
+match(const struct ln *ln, const char *p, size_t sz)

 {
 
 {
 

-       if ( ! (ln->type & type))
-               return(0);

        if (strncmp(ln->code, p, sz))
                return(0);
        return('\0' == ln->code[(int)sz]);
        if (strncmp(ln->code, p, sz))
                return(0);
        return('\0' == ln->code[(int)sz]);