summaryrefslogtreecommitdiff
path: root/enc/trans
diff options
context:
space:
mode:
authornaruse <naruse@b2dd03c8-39d4-4d8f-98ff-823fe69b080e>2010-11-23 20:49:56 +0000
committernaruse <naruse@b2dd03c8-39d4-4d8f-98ff-823fe69b080e>2010-11-23 20:49:56 +0000
commit7f38397b6c01b65219deb622c027e177bdc23ede (patch)
treee798702443d0680dbf47e809820b34a4dec328ed /enc/trans
parent3ab82a65d701db47cea2ccbe9e7821468e71dd90 (diff)
downloadruby-7f38397b6c01b65219deb622c027e177bdc23ede.tar.gz
* enc/trans/utf_16_32.trans: add a convert from UTF-8 to UTF-16.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@29892 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Diffstat (limited to 'enc/trans')
-rw-r--r--enc/trans/utf_16_32.trans30
1 files changed, 28 insertions, 2 deletions
diff --git a/enc/trans/utf_16_32.trans b/enc/trans/utf_16_32.trans
index 2d7005993e..01caffe02c 100644
--- a/enc/trans/utf_16_32.trans
+++ b/enc/trans/utf_16_32.trans
@@ -289,7 +289,7 @@ fun_si_from_utf_16(void *statep, const unsigned char *s, size_t l)
}
break;
case BE:
- if (s[0] < 0xD8 && 0xDF < s[0]) {
+ if (s[0] < 0xD8 || 0xDF < s[0]) {
return (VALUE)FUNso;
}
else if (s[0] <= 0xDB) {
@@ -297,7 +297,7 @@ fun_si_from_utf_16(void *statep, const unsigned char *s, size_t l)
}
break;
case LE:
- if (s[1] < 0xD8 && 0xDF < s[1]) {
+ if (s[1] < 0xD8 || 0xDF < s[1]) {
return (VALUE)FUNso;
}
else if (s[1] <= 0xDB) {
@@ -321,6 +321,19 @@ fun_so_from_utf_16(void *statep, const unsigned char *s, size_t l, unsigned char
return 0;
}
+static ssize_t
+fun_so_to_utf_16(void *statep, const unsigned char *s, size_t l, unsigned char *o, size_t osize)
+{
+ unsigned char *sp = statep;
+ if (*sp == 0) {
+ *o++ = 0xFE;
+ *o++ = 0xFF;
+ *sp = 1;
+ return 2 + fun_so_to_utf_16be(statep, s, l, o, osize);
+ }
+ return fun_so_to_utf_16be(statep, s, l, o, osize);
+}
+
static const rb_transcoder
rb_from_UTF_16BE = {
"UTF-16BE", "UTF-8", from_UTF_16BE,
@@ -429,6 +442,18 @@ rb_from_UTF_16 = {
NULL, fun_si_from_utf_16, NULL, fun_so_from_utf_16
};
+static const rb_transcoder
+rb_to_UTF_16 = {
+ "UTF-8", "UTF-16", from_UTF_8,
+ TRANSCODE_TABLE_INFO,
+ 1, /* input_unit_length */
+ 4, /* max_input */
+ 4, /* max_output */
+ asciicompat_encoder, /* asciicompat_type */
+ 1, state_init, NULL, /* state_size, state_init, state_fini */
+ NULL, NULL, NULL, fun_so_to_utf_16
+};
+
void
Init_utf_16_32(void)
{
@@ -441,4 +466,5 @@ Init_utf_16_32(void)
rb_register_transcoder(&rb_from_UTF_32LE);
rb_register_transcoder(&rb_to_UTF_32LE);
rb_register_transcoder(&rb_from_UTF_16);
+ rb_register_transcoder(&rb_to_UTF_16);
}