Simplify chars.c---there's really no need for extra code to reorder the

hash chain or an extra function for checking matches.

1 files changed, 8 insertions, 61 deletions

diff --git a/chars.c b/chars.c
index 5158612a..d0fd8724 100644
--- a/chars.c
+++ b/chars.c
@@ -1,4 +1,4 @@
-/*	$Id: chars.c,v 1.46 2011/05/24 21:31:23 kristaps Exp $ */
+/*	$Id: chars.c,v 1.47 2011/07/07 06:41:50 kristaps Exp $ */
 /*
  * Copyright (c) 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2011 Ingo Schwarze <schwarze@openbsd.org>
@@ -21,7 +21,6 @@
 
 #include <assert.h>
 #include <ctype.h>
-#include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 
@@ -52,7 +51,6 @@ struct	mchars {
 	struct ln	**htab;
 };
 
-static	inline int	  match(const struct ln *, const char *, size_t);
 static	const struct ln	 *find(struct mchars *, const char *, size_t);
 
 void
@@ -74,8 +72,7 @@ mchars_alloc(void)
 	/*
 	 * Constructs a very basic chaining hashtable.  The hash routine
 	 * is simply the integral value of the first character.
-	 * Subsequent entries are chained in the order they're processed
-	 * (they're in-line re-ordered during lookup).
+	 * Subsequent entries are chained in the order they're processed.
 	 */
 
 	tab = mandoc_malloc(sizeof(struct mchars));
@@ -98,10 +95,6 @@ mchars_alloc(void)
 	return(tab);
 }
 
-
-/* 
- * Special character to Unicode codepoint.
- */
 int
 mchars_spec2cp(struct mchars *arg, const char *p, size_t sz)
 {
@@ -113,13 +106,6 @@ mchars_spec2cp(struct mchars *arg, const char *p, size_t sz)
 	return(ln->unicode);
 }
 
-/*
- * 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'.
- */
 char
 mchars_num2char(const char *p, size_t sz)
 {
@@ -130,10 +116,6 @@ mchars_num2char(const char *p, size_t sz)
 	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)
 {
@@ -145,9 +127,6 @@ mchars_num2uc(const char *p, size_t sz)
 	return(i > 0x80 && i <= 0x10FFFF ? i : '\0');
 }
 
-/* 
- * Special character to string array.
- */
 const char *
 mchars_spec2str(struct mchars *arg, const char *p, size_t sz, size_t *rsz)
 {
@@ -164,52 +143,20 @@ mchars_spec2str(struct mchars *arg, const char *p, size_t sz, size_t *rsz)
 static const struct ln *
 find(struct mchars *tab, const char *p, size_t sz)
 {
-	struct ln	 *pp, *prev;
-	struct ln	**htab;
+	struct ln	 *pp;
 	int		  hash;
 
 	assert(p);
-	if (0 == sz)
-		return(NULL);
 
-	if (p[0] < PRINT_LO || p[0] > PRINT_HI)
+	if (0 == sz || p[0] < PRINT_LO || p[0] > PRINT_HI)
 		return(NULL);
 
-	/*
-	 * Lookup the symbol in the symbol hash.  See ascii2htab for the
-	 * hashtable specs.  This dynamically re-orders the hash chain
-	 * to optimise for repeat hits.
-	 */
-
 	hash = (int)p[0] - PRINT_LO;
-	htab = tab->htab;
 
-	if (NULL == (pp = htab[hash]))
-		return(NULL);
-
-	for (prev = NULL; pp; pp = pp->next) {
-		if ( ! match(pp, p, sz)) {
-			prev = pp;
-			continue;
-		}
-
-		if (prev) {
-			prev->next = pp->next;
-			pp->next = htab[hash];
-			htab[hash] = pp;
-		}
-
-		return(pp);
-	}
+	for (pp = tab->htab[hash]; pp; pp = pp->next)
+		if (0 == strncmp(pp->code, p, sz) && 
+				'\0' == pp->code[(int)sz])
+			return(pp);
 
 	return(NULL);
 }
-
-static inline int
-match(const struct ln *ln, const char *p, size_t sz)
-{
-
-	if (strncmp(ln->code, p, sz))
-		return(0);
-	return('\0' == ln->code[(int)sz]);
-}