1 files changed, 45 insertions, 16 deletions

diff --git a/src/pl/plperl/plperl_helpers.h b/src/pl/plperl/plperl_helpers.h
index 1b6648be1d1..ed99194ed1e 100644
--- a/src/pl/plperl/plperl_helpers.h
+++ b/src/pl/plperl/plperl_helpers.h
@@ -3,21 +3,29 @@
 
 /*
  * convert from utf8 to database encoding
+ *
+ * Returns a palloc'ed copy of the original string
  */
 static inline char *
-utf_u2e(const char *utf8_str, size_t len)
+utf_u2e(char *utf8_str, size_t len)
 {
 	int			enc = GetDatabaseEncoding();
-
-	char	   *ret = (char *) pg_do_encoding_conversion((unsigned char *) utf8_str, len, PG_UTF8, enc);
+	char	   *ret;
 
 	/*
-	 * when we are a PG_UTF8 or SQL_ASCII database pg_do_encoding_conversion()
-	 * will not do any conversion or verification. we need to do it manually
-	 * instead.
+	 * When we are in a PG_UTF8 or SQL_ASCII database
+	 * pg_do_encoding_conversion() will not do any conversion (which is good)
+	 * or verification (not so much), so we need to run the verification step
+	 * separately.
 	 */
 	if (enc == PG_UTF8 || enc == PG_SQL_ASCII)
-		pg_verify_mbstr_len(PG_UTF8, utf8_str, len, false);
+	{
+		pg_verify_mbstr_len(enc, utf8_str, len, false);
+		ret = utf8_str;
+	}
+	else
+		ret = (char *) pg_do_encoding_conversion((unsigned char *) utf8_str,
+												 len, PG_UTF8, enc);
 
 	if (ret == utf8_str)
 		ret = pstrdup(ret);
@@ -27,11 +35,15 @@ utf_u2e(const char *utf8_str, size_t len)
 
 /*
  * convert from database encoding to utf8
+ *
+ * Returns a palloc'ed copy of the original string
  */
 static inline char *
 utf_e2u(const char *str)
 {
-	char	   *ret = (char *) pg_do_encoding_conversion((unsigned char *) str, strlen(str), GetDatabaseEncoding(), PG_UTF8);
+	char	   *ret =
+		(char *) pg_do_encoding_conversion((unsigned char *) str, strlen(str),
+										   GetDatabaseEncoding(), PG_UTF8);
 
 	if (ret == str)
 		ret = pstrdup(ret);
@@ -41,6 +53,8 @@ utf_e2u(const char *str)
 
 /*
  * Convert an SV to a char * in the current database encoding
+ *
+ * Returns a palloc'ed copy of the original string
  */
 static inline char *
 sv2cstr(SV *sv)
@@ -51,7 +65,9 @@ sv2cstr(SV *sv)
 
 	/*
 	 * get a utf8 encoded char * out of perl. *note* it may not be valid utf8!
-	 *
+	 */
+
+	/*
 	 * SvPVutf8() croaks nastily on certain things, like typeglobs and
 	 * readonly objects such as $^V. That's a perl bug - it's not supposed to
 	 * happen. To avoid crashing the backend, we make a copy of the sv before
@@ -63,18 +79,27 @@ sv2cstr(SV *sv)
 		(SvTYPE(sv) > SVt_PVLV && SvTYPE(sv) != SVt_PVFM))
 		sv = newSVsv(sv);
 	else
-
+	{
 		/*
 		 * increase the reference count so we can just SvREFCNT_dec() it when
 		 * we are done
 		 */
 		SvREFCNT_inc_simple_void(sv);
+	}
 
-	val = SvPVutf8(sv, len);
+	/*
+	 * Request the string from Perl, in UTF-8 encoding; but if we're in a
+	 * SQL_ASCII database, just request the byte soup without trying to make it
+	 * UTF8, because that might fail.
+	 */
+	if (GetDatabaseEncoding() == PG_SQL_ASCII)
+		val = SvPV(sv, len);
+	else
+		val = SvPVutf8(sv, len);
 
 	/*
-	 * we use perl's length in the event we had an embedded null byte to
-	 * ensure we error out properly
+	 * Now convert to database encoding.  We use perl's length in the event we
+	 * had an embedded null byte to ensure we error out properly.
 	 */
 	res = utf_u2e(val, len);
 
@@ -88,16 +113,20 @@ sv2cstr(SV *sv)
  * Create a new SV from a string assumed to be in the current database's
  * encoding.
  */
-
 static inline SV *
 cstr2sv(const char *str)
 {
 	SV		   *sv;
-	char	   *utf8_str = utf_e2u(str);
+	char	   *utf8_str;
+
+	/* no conversion when SQL_ASCII */
+	if (GetDatabaseEncoding() == PG_SQL_ASCII)
+		return newSVpv(str, 0);
+
+	utf8_str = utf_e2u(str);
 
 	sv = newSVpv(utf8_str, 0);
 	SvUTF8_on(sv);
-
 	pfree(utf8_str);
 
 	return sv;