I suppose today will be a break.


Ok, so previously I noted that the namespace writer would not work because the output is unknown. However it would work because there is config which contains the cache creator. For other JIT parts that are directly executable, the namespace can write using a similar means so to speak.


There is no PrintWriter, however PrintStream works the same except that special work is needed to prevent characters from being directly written.


So since namespaces must be visible in the source code, the resource data in a given resource looks like this: jvm_1ejar___META_1dINF_1fMANIFEST_1eMF. Any classes would also share that prefix also. So what I need to do now is start writing the class file data, as in actually performing the JIT compile.


I will need some base structure information so that namespace lookup and the basic initialization of the C machine works.


The JITOutput would likely need a kind of final link, so that it can be given a bunch of files which refer to namespace caches and determine what to do with them.