[core] RulesetFactoryCompatibility stores the whole ruleset file in memory as a string

[oowekyala]

See here:

pmd/pmd-core/src/main/java/net/sourceforge/pmd/RuleSetFactoryCompatibility.java

Lines 108 to 116 in 0c1afc4

	public Reader filterRuleSetFile(InputStream stream) throws IOException {
	byte[] bytes = IOUtils.toByteArray(stream);
	String encoding = determineEncoding(bytes);
	String ruleset = new String(bytes, encoding);
	ruleset = applyAllFilters(ruleset);
	return new StringReader(ruleset);
	}

applyAllFilters performs a bunch of regex replacements on the entire String, which creates many copies of it. We could do better by only filtering the ref attributes when parsing a rule reference node, this just seems very inefficient.

@jsotuyod This may offset your performance comparison of DOM and SAX for #724

[adangel]

Fixed with PMD 7 via #2969 .