[MIME4J-174] Refactor org.mime4j.util.CharsetUtil to lazily determine supported encodings/decodings

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.7
• Assignee: Stefano Bagnara
• Reporter: Richard von Keyserling
• Created: 04/Mar/10 19:39
• Updated: 14/Mar/10 23:05
• Resolved: 14/Mar/10 23:05

Description

On initialization CharsetUtil determines all supported endcodings and decodings by attempting to encode and decode a dummy string with every entry in JAVA_CHARSETS. This loads a lot of classes into the JVM which in turn uses up a lot of permGen.

Moving the decoding and encoding tests into isDecodingSupported() and isEncodingSupported() and adding positive results to the decodingSupported and encodingSupported treeSets from those methods would allow the class to only load encoders and decoders the application needs.

Issue Links

• is related to: MIME4J-159 Update CharsetUtil to java5. Some of CharsetUtil is no more "useful" as it was a java 1.3 backport of java2 1.4-5 stuff.