diff --git a/lib/modules/Sql.pmod/mysql.pike b/lib/modules/Sql.pmod/mysql.pike
index 08e2fabb8423170778a69655d8ebb5453d7a9b28..3e3e440f2052dee4fbb6fcb3e3ff8753ea4a809f 100644
--- a/lib/modules/Sql.pmod/mysql.pike
+++ b/lib/modules/Sql.pmod/mysql.pike
@@ -1,5 +1,5 @@
/*
- * $Id: mysql.pike,v 1.29 2006/08/15 14:50:39 grubba Exp $
+ * $Id: mysql.pike,v 1.30 2006/08/22 11:27:08 grubba Exp $
*
* Glue for the Mysql-module
*/
@@ -16,6 +16,13 @@ inherit Mysql.mysql;
#define UNICODE_DECODE_MODE 1 // Unicode decode mode
#define LATIN1_UNICODE_ENCODE_MODE 2 // Unicode encode mode with latin1 charset
#define UTF8_UNICODE_ENCODE_MODE 4 // Unicode encode mode with utf8 charset
+#define BINARY_LATIN1_MODE 8 // Don't special-case latin1 control chars
+
+#ifdef MYSQL_CHARSET_DEBUG
+#define CH_DEBUG(X...) werror("Sql.mysql: " + X)
+#else
+#define CH_DEBUG(X...)
+#endif
// Set to the above if the connection is in utf8-mode. Enable latin1
// unicode encode mode by default; it should be compatible with
@@ -23,9 +30,9 @@ inherit Mysql.mysql;
static int utf8_mode;
// The charset, either "latin1" or "utf8", currently assigned to
-// character_set_client when unicode encode mode is enabled. Zero when
-// the connection charset has been set to something else than "latin1"
-// or "unicode".
+// character_set_client and character_set_connection when unicode
+// encode mode is enabled. Zero when the connection charset has been
+// set to something else than "latin1" or "unicode".
static string send_charset;
static void update_unicode_encode_mode_from_charset (string charset)
@@ -35,16 +42,19 @@ static void update_unicode_encode_mode_from_charset (string charset)
utf8_mode |= LATIN1_UNICODE_ENCODE_MODE;
utf8_mode &= ~UTF8_UNICODE_ENCODE_MODE;
send_charset = "latin1";
+ CH_DEBUG("Entering latin1 mode.\n");
break;
case "unicode":
utf8_mode |= UTF8_UNICODE_ENCODE_MODE;
utf8_mode &= ~LATIN1_UNICODE_ENCODE_MODE;
send_charset = "utf8";
+ CH_DEBUG("Entering utf8 mode.\n");
break;
default:
// Wrong charset - the mode can't be used.
utf8_mode |= LATIN1_UNICODE_ENCODE_MODE|UTF8_UNICODE_ENCODE_MODE;
send_charset = 0;
+ CH_DEBUG("Entering other mode.\n");
break;
}
}
@@ -58,10 +68,10 @@ int(0..1) set_unicode_encode_mode (int enable)
//!
//! Unicode encode mode works as follows: Eight bit strings are sent
//! as @expr{latin1@} and wide strings are sent using @expr{utf8@}.
-//! @[big_query] sends @expr{SET character_set_client@} statements as
-//! necessary to update the charset on the server side. If the server
-//! doesn't support that then it fails, but the wide string query
-//! would fail anyway.
+//! @[big_query] sends @expr{SET character_set_client@} and @expr{SET
+//! character_set_connection@} statements as necessary to update the
+//! charset on the server side. If the server doesn't support that
+//! then it fails, but the wide string query would fail anyway.
//!
//! To make this transparent, string literals with introducers (e.g.
//! @expr{_binary 'foo'@}) are excluded from the UTF-8 encoding. This
@@ -95,6 +105,12 @@ int(0..1) set_unicode_encode_mode (int enable)
//! will get UTF-8 encoded by the server.
//!
//! @note
+//! When unicode encode mode is enabled, the connection charset
+//! will mirror the client charset. This is necessary for unicode
+//! characters to survive for wide queries, and for binary data
+//! to survive for narrow queries in a transparent manner.
+//!
+//! @note
//! When unicode encode mode is enabled and the connection charset
//! is @expr{latin1@}, the charset accepted by @[big_query] is not
//! quite Unicode since @expr{latin1@} is based on @expr{cp1252@}.
@@ -107,11 +123,13 @@ int(0..1) set_unicode_encode_mode (int enable)
//! @seealso
//! @[set_unicode_decode_mode], @[set_charset]
{
- if (enable)
+ if (enable) {
+ CH_DEBUG("Enabling unicode encode mode.\n");
update_unicode_encode_mode_from_charset (lower_case (get_charset()));
- else {
+ } else {
utf8_mode &= ~(LATIN1_UNICODE_ENCODE_MODE|UTF8_UNICODE_ENCODE_MODE);
send_charset = 0;
+ CH_DEBUG("Disabling unicode encode mode.\n");
}
return !!send_charset;
}
@@ -154,10 +172,12 @@ void set_unicode_decode_mode (int enable)
//! @[set_unicode_encode_mode]
{
if (enable) {
+ CH_DEBUG("Enabling unicode decode mode.\n");
::big_query ("SET character_set_results = utf8");
utf8_mode |= UNICODE_DECODE_MODE;
}
else {
+ CH_DEBUG("Disabling unicode decode mode.\n");
::big_query ("SET character_set_results = " + get_charset());
utf8_mode &= ~UNICODE_DECODE_MODE;
}
@@ -240,6 +260,8 @@ void set_charset (string charset)
{
charset = lower_case (charset);
+ CH_DEBUG("Setting charset to %O.\n", charset);
+
::set_charset (charset == "unicode" ? "utf8" : charset);
if (charset == "unicode" ||
@@ -312,6 +334,7 @@ string quote(string s)
string latin1_to_utf8 (string s)
//! Converts a string in MySQL @expr{latin1@} format to UTF-8.
{
+ CH_DEBUG("Converting latin1 query to utf8.\n");
return string_to_utf8 (replace (s, ([
"\x80": "\u20AC", /*"\x81": "\u0081",*/ "\x82": "\u201A", "\x83": "\u0192",
"\x84": "\u201E", "\x85": "\u2026", "\x86": "\u2020", "\x87": "\u2021",
@@ -546,9 +569,11 @@ int decode_datetime (string timestr)
string restore_charset; \
if (charset) { \
restore_charset = send_charset || get_charset(); \
- if (charset != restore_charset) \
- ::big_query ("SET character_set_client=" + charset); \
- else \
+ CH_DEBUG("Restore charset is %O.\n", restore_charset); \
+ if (charset != restore_charset) { \
+ ::big_query("SET character_set_client=" + charset); \
+ ::big_query("SET character_set_connection=" + charset); \
+ } else \
restore_charset = 0; \
} \
\
@@ -572,14 +597,19 @@ int decode_datetime (string timestr)
new_send_charset = "utf8"; \
} \
} \
+ CH_DEBUG("New send charset is %O.\n", new_send_charset); \
\
if (new_send_charset != send_charset) { \
- if (mixed err = \
- ::big_query ("SET character_set_client=" + new_send_charset)) { \
+ CH_DEBUG("Send charset was %O.\n", send_charset); \
+ if (mixed err = catch { \
+ ::big_query("SET character_set_client=" + new_send_charset); \
+ ::big_query("SET character_set_connection=" + \
+ new_send_charset); \
+ }) { \
if (new_send_charset == "utf8") \
predef::error ("The query is a wide string " \
"and the MySQL server doesn't support UTF-8: %s\n", \
- describe_error (err)); \
+ describe_error(err)); \
else \
throw(err); \
} \
@@ -587,18 +617,24 @@ int decode_datetime (string timestr)
} \
} \
\
+ CH_DEBUG("Sending query %O.\n", query); \
+ \
int|object res = ::do_query(query); \
\
if (restore_charset) { \
if (send_charset && (<"latin1", "utf8">)[charset]) \
send_charset = charset; \
- else \
- ::big_query ("SET character_set_client=" + restore_charset); \
+ else { \
+ CH_DEBUG("Restoring charset to %O.\n", restore_charset); \
+ ::big_query("SET character_set_client=" + restore_charset); \
+ ::big_query("SET character_set_connection=" + restore_charset); \
+ } \
} \
\
if (!objectp(res)) return res; \
\
if (utf8_mode & UNICODE_DECODE_MODE) { \
+ CH_DEBUG("Adding UnicodeWrapper.\n"); \
return .sql_util.UnicodeWrapper(res); \
} \
return res;