Fail pgwin32_message_to_UTF16() for SQL_ASCII messages.

nmisch · nmisch · commit 409f5303ced6 · 2019-05-12T10:33:08.000-07:00
The function had been interpreting SQL_ASCII messages as UTF8, throwing an error when they were invalid UTF8. The new behavior is consistent with pg_do_encoding_conversion(). This affects LOG_DESTINATION_STDERR and LOG_DESTINATION_EVENTLOG, which will send untranslated bytes to write() and ReportEventA(). On buildfarm member bowerbird, enabling log_connections caused an error whenever the role name was not valid UTF8. Back-patch to 9.4 (all supported versions). Discussion: https://postgr.es/m/20190512015615.GD1124997@rfd.leadboat.com
diff --git a/src/backend/utils/mb/mbutils.c b/src/backend/utils/mb/mbutils.c
@@ -1057,11 +1057,16 @@ GetMessageEncoding(void)
 WCHAR *
 pgwin32_message_to_UTF16(const char *str, int len, int *utf16len)
 {
+	int			msgenc = GetMessageEncoding();
 	WCHAR	   *utf16;
 	int			dstlen;
 	UINT		codepage;
 
-	codepage = pg_enc2name_tbl[GetMessageEncoding()].codepage;
+	if (msgenc == PG_SQL_ASCII)
+		/* No conversion is possible, and SQL_ASCII is never utf16. */
+		return NULL;
+
+	codepage = pg_enc2name_tbl[msgenc].codepage;
 
 	/*
 	 * Use MultiByteToWideChar directly if there is a corresponding codepage,
@@ -1086,7 +1091,7 @@ pgwin32_message_to_UTF16(const char *str, int len, int *utf16len)
 		{
 			utf8 = (char *) pg_do_encoding_conversion((unsigned char *) str,
 													  len,
-													  GetMessageEncoding(),
+													  msgenc,
 													  PG_UTF8);
 			if (utf8 != str)
 				len = strlen(utf8);
diff --git a/src/bin/pg_dump/t/010_dump_connstr.pl b/src/bin/pg_dump/t/010_dump_connstr.pl
@@ -14,10 +14,8 @@
 	plan tests => 14;
 }
 
-# In a SQL_ASCII database, pgwin32_message_to_UTF16() needs to
-# interpret everything as UTF8.  We're going to use byte sequences
-# that aren't valid UTF-8 strings, so that would fail.  Use LATIN1,
-# which accepts any byte and has a conversion from each byte to UTF-8.
+# We're going to use byte sequences that aren't valid UTF-8 strings.  Use
+# LATIN1, which accepts any byte and has a conversion from each byte to UTF-8.
 $ENV{LC_ALL}           = 'C';
 $ENV{PGCLIENTENCODING} = 'LATIN1';
 
diff --git a/src/bin/scripts/t/200_connstr.pl b/src/bin/scripts/t/200_connstr.pl
@@ -7,10 +7,8 @@
 
 # Tests to check connection string handling in utilities
 
-# In a SQL_ASCII database, pgwin32_message_to_UTF16() needs to
-# interpret everything as UTF8.  We're going to use byte sequences
-# that aren't valid UTF-8 strings, so that would fail.  Use LATIN1,
-# which accepts any byte and has a conversion from each byte to UTF-8.
+# We're going to use byte sequences that aren't valid UTF-8 strings.  Use
+# LATIN1, which accepts any byte and has a conversion from each byte to UTF-8.
 $ENV{LC_ALL}           = 'C';
 $ENV{PGCLIENTENCODING} = 'LATIN1';
 
