For some reason, cldc compact is not being picked up by the stream reader.


Actually, there appears to be a logical error in the ZIP stream reader, which I fixed. It was failing on the max supported extract version (2.0) due to the greater than or equal sign.


The namespace browser would probably be best if it were to be given output streams for namespaces. This would be compared to having it in another class as before, while the functionality would be similar.


it would also be best if the JIT just wrote to OutputStreams instead of just juggling the potential for it to be written to memory or having specific cache forms. This simplifies, plus the resulting namespace cache could be directly executable anyway.


I could also probably have a better content management system for the basic output compared to the mess that the generic output was.


I should also have different areas for classes and resources. Before I used a type flag. But with two different tables I can have the constraints for classes be different than that of resources.


Extended data output stream can use align, since previously I have used it much in the generic writer.


When the launcher/JVM is created it can specify the terminal and LUI interface to use for sub-processes. This way they use the same class which just forwards calls to an implementation that can handle switching between applications.


One thing I can do is have an interface similar to TerminalScreen for the LUI code except have it for characters and such. It would be best if it became a managed set of character buffers.


Basically it can hide the characters internally and instead provide reading from it as a CharSequence. Then toString() can be cached so that if any text does not change at all, it does not waste memory.


I was interrupted and now I am back.