When the default locale is set to the Turkish language, using the XmlStreamReader constructor on an XML stream with a prolog including the ISO-8859-1 charset name in lowercase as its encoding throws a UnsupportedEncodingException (java.io.UnsupportedEncodingException: İSO-8859-1).
Example XML prolog : <?xml version="1.0" encoding="iso-8859-1"?>
This is apparently because the XmlStreamReader class uses String.toUpperCase() in its getXmlProlog() function. It should rather use toUpperCase(Locale.ROOT) or toUpperCase(Locale.US) as already done in the getContentTypeEncoding() function. Otherwise the behaviour can be different depending on the default locale, as the dotted lower case i becomes a dotted upper case i in the Turkish language, which not the case with other languages.