ruby-changes:2240
From: ko1@a...
Date: 17 Oct 2007 06:18:23 +0900
Subject: [ruby-changes:2240] nobu - Ruby:r13731 (trunk): * encoding.c (rb_enc_alias, rb_enc_find_index): changed
nobu 2007-10-17 06:18:09 +0900 (Wed, 17 Oct 2007)
New Revision: 13731
Modified files:
trunk/ChangeLog
trunk/encoding.c
trunk/st.c
Log:
* encoding.c (rb_enc_alias, rb_enc_find_index): changed
enc_table_alias to a name-to-index hash.
* encoding.c (rb_enc_init): use upper case names for aliases to use as
constant names.
* encoding.c (enc_find): allow symbols.
* encoding.c (Init_Encoding): define encoding constants.
* st.c (strcasehash): fix wrong code range condition.
http://svn.ruby-lang.org/cgi-bin/viewvc.cgi/trunk/ChangeLog?r1=13731&r2=13730
http://svn.ruby-lang.org/cgi-bin/viewvc.cgi/trunk/encoding.c?r1=13731&r2=13730
http://svn.ruby-lang.org/cgi-bin/viewvc.cgi/trunk/st.c?r1=13731&r2=13730
Index: encoding.c
===================================================================
--- encoding.c (revision 13730)
+++ encoding.c (revision 13731)
@@ -13,6 +13,7 @@
#include "ruby/ruby.h"
#include "ruby/encoding.h"
#include "regenc.h"
+#include <ctype.h>
static ID id_encoding;
static VALUE rb_cEncoding;
@@ -152,12 +153,12 @@
if (!enc_table_alias) {
enc_table_alias = st_init_strcasetable();
}
- while ((idx = rb_enc_find_index(orig)) < 0) {
+ if ((idx = rb_enc_find_index(orig)) < 0) {
if (!st_lookup(enc_table_alias, (st_data_t)orig, &data))
return -1;
- orig = (const char *)data;
+ idx = (int)data;
}
- st_insert(enc_table_alias, (st_data_t)alias, (st_data_t)orig);
+ st_insert(enc_table_alias, (st_data_t)alias, (st_data_t)idx);
return idx;
}
@@ -170,10 +171,10 @@
ENC_REGISTER(ONIG_ENCODING_SJIS);
ENC_REGISTER(ONIG_ENCODING_UTF8);
#undef ENC_REGISTER
- rb_enc_alias("ascii", rb_enc_name(ONIG_ENCODING_ASCII));
- rb_enc_alias("binary", rb_enc_name(ONIG_ENCODING_ASCII));
- rb_enc_alias("us-ascii", rb_enc_name(ONIG_ENCODING_ASCII)); /* will be defined separately in future. */
- rb_enc_alias("sjis", rb_enc_name(ONIG_ENCODING_SJIS));
+ rb_enc_alias("ASCII", rb_enc_name(ONIG_ENCODING_ASCII));
+ rb_enc_alias("BINARY", rb_enc_name(ONIG_ENCODING_ASCII));
+ rb_enc_alias("US-ASCII", rb_enc_name(ONIG_ENCODING_ASCII)); /* will be defined separately in future. */
+ rb_enc_alias("SJIS", rb_enc_name(ONIG_ENCODING_SJIS));
}
rb_encoding *
@@ -198,7 +199,6 @@
if (!enc_table) {
rb_enc_init();
}
- find:
for (i=0; i<enc_table_size; i++) {
if (strcasecmp(name, enc_table[i].name) == 0) {
return i;
@@ -206,8 +206,7 @@
}
if (!alias && enc_table_alias) {
if (st_lookup(enc_table_alias, (st_data_t)name, &alias)) {
- name = (const char *)alias;
- goto find;
+ return (int)alias;
}
}
return -1;
@@ -490,7 +489,9 @@
static VALUE
enc_find(VALUE klass, VALUE enc)
{
- int idx = rb_enc_find_index(StringValueCStr(enc));
+ int idx;
+ if (SYMBOL_P(enc)) enc = rb_id2str(SYM2ID(enc));
+ idx = rb_enc_find_index(StringValueCStr(enc));
if (idx < 0) {
rb_raise(rb_eArgError, "unknown encoding name - %s", RSTRING_PTR(enc));
}
@@ -540,9 +541,59 @@
return rb_primary_encoding;
}
+static void
+set_encoding_const(const char *name, rb_encoding *enc)
+{
+ VALUE encoding = rb_enc_from_encoding(enc);
+ char *s = (char *)name;
+ int haslower = 0, valid = 0;
+
+ if (ISUPPER(*s)) {
+ while (*++s && (ISALNUM(*s) || *s == '_')) {
+ if (ISLOWER(*s)) haslower = 1;
+ }
+ }
+ if (!*s) {
+ valid = 1;
+ rb_define_const(rb_cEncoding, name, encoding);
+ }
+ if (!valid || haslower) {
+ int len = strlen(name) + 1;
+ if (!haslower) {
+ while (!ISLOWER(*s) && *++s);
+ if (*s) haslower = 1;
+ }
+ MEMCPY(s = ALLOCA_N(char, len), name, char, len);
+ name = s;
+ if (!valid) {
+ if (ISLOWER(*s)) *s = ONIGENC_ASCII_CODE_TO_UPPER_CASE((int)*s);
+ for (; *s; ++s) {
+ if (!ISALNUM(*s)) *s = '_';
+ }
+ rb_define_const(rb_cEncoding, name, encoding);
+ }
+ if (haslower) {
+ for (s = (char *)name; *s; ++s) {
+ if (ISLOWER(*s)) *s = ONIGENC_ASCII_CODE_TO_UPPER_CASE((int)*s);
+ }
+ rb_define_const(rb_cEncoding, name, encoding);
+ }
+ }
+}
+
+static int
+set_encoding_alias(st_data_t name, st_data_t orig, st_data_t arg)
+{
+ rb_encoding *enc = rb_enc_from_index((int)orig);
+ set_encoding_const((const char *)name, enc);
+ return ST_CONTINUE;
+}
+
void
Init_Encoding(void)
{
+ int i;
+
rb_cEncoding = rb_define_class("Encoding", rb_cObject);
rb_undef_alloc_func(rb_cEncoding);
rb_define_method(rb_cEncoding, "to_s", enc_to_s, 0);
@@ -557,4 +608,10 @@
rb_primary_encoding = rb_enc_from_encoding(rb_enc_from_index(0));
rb_define_singleton_method(rb_cEncoding, "primary_encoding", get_primary_encoding, 0);
rb_define_singleton_method(rb_cEncoding, "primary_encoding=", set_primary_encoding, 1);
+
+ for (i = 0; i < enc_table_size; ++i) {
+ rb_encoding *enc = enc_table[i].enc;
+ if (enc) set_encoding_const(rb_enc_name(enc), enc);
+ }
+ st_foreach(enc_table_alias, set_encoding_alias, 0);
}
Index: ChangeLog
===================================================================
--- ChangeLog (revision 13730)
+++ ChangeLog (revision 13731)
@@ -1,3 +1,17 @@
+Wed Oct 17 06:18:06 2007 Nobuyoshi Nakada <nobu@r...>
+
+ * encoding.c (rb_enc_alias, rb_enc_find_index): changed
+ enc_table_alias to a name-to-index hash.
+
+ * encoding.c (rb_enc_init): use upper case names for aliases to use as
+ constant names.
+
+ * encoding.c (enc_find): allow symbols.
+
+ * encoding.c (Init_Encoding): define encoding constants.
+
+ * st.c (strcasehash): fix wrong code range condition.
+
Wed Oct 17 05:07:18 2007 Nobuyoshi Nakada <nobu@r...>
* encoding.c (rb_primary_encoding): added Encoding.primary_encoding.
Index: st.c
===================================================================
--- st.c (revision 13730)
+++ st.c (revision 13731)
@@ -842,7 +842,7 @@
*/
while (*string) {
unsigned int c = (unsigned char)*string++;
- if ((unsigned int)(c - 'A') > ('Z' - 'A')) c += 'a' - 'A';
+ if ((unsigned int)(c - 'A') <= ('Z' - 'A')) c += 'a' - 'A';
hval ^= c;
/* multiply by the 32 bit FNV magic prime mod 2^32 */
--
ML: ruby-changes@q...
Info: http://www.atdot.net/~ko1/quickml