Workaround for flaw in musl 1.1.8's CP437 char mapping

FIXME: Remove this patch after musl upgrade.

See: http://www.openwall.com/lists/musl/2017/11/22/2
--- a/tests/at-charset.c
+++ b/tests/at-charset.c
@@ -93,15 +93,6 @@
 	test_result(mywstrncmp(ubuffer, latin1text, sizeof(latin1text) / 2) == TRUE);
 
 #ifdef ICONV_FOUND
-	Priv->Charset = AT_CHARSET_PCCP437;
-	error = ATGEN_EncodeText(s, latin1text, sizeof(latin1text) / 2, buffer, sizeof(buffer), &result);
-	gammu_test_result(error, "Encode - 3");
-	strconv_test_result(latin1cp437, buffer, result);
-
-	error = ATGEN_DecodeText(s, buffer, result, ubuffer, sizeof(ubuffer), FALSE, FALSE);
-	gammu_test_result(error, "Decode - 3");
-	test_result(mywstrncmp(ubuffer, latin1text, sizeof(latin1text) / 2) == TRUE);
-
 	Priv->Charset = AT_CHARSET_ISO88591;
 	error = ATGEN_EncodeText(s, latin1text, sizeof(latin1text) / 2, buffer, sizeof(buffer), &result);
 	gammu_test_result(error, "Encode - 4");