Missing UTF-32 charset should fail only in case UTF-32 is detected

[danielkec]

Right now UTF-32 chaset is looked up for unicode version detection during class initialization, no matter whether there will be ever any use of it:

parsson/impl/src/main/java/org/eclipse/parsson/UnicodeDetectingInputStream.java

Lines 32 to 35 in 981cd2f

	class UnicodeDetectingInputStream extends FilterInputStream {
	private static final Charset UTF_32LE = Charset.forName("UTF-32LE");
	private static final Charset UTF_32BE = Charset.forName("UTF-32BE");

Actual lookup should be done lazily when UTF-32 is actually detected to support runtimes without UTF-32, like GraalVM native image oracle/graal/issues/1294

[Captain1653]

@lukasj This issue was fixed in January, but wasn't released as I can see. Is it possible to create the new version with this fix?

[lukasj]

@Captain1653 if you have a question, use new issue/mail/discussion post for it