DISCLAIMER: These notes are from the defunct k8 project which precedes SquirrelJME. The notes for SquirrelJME start on 2016/02/26! The k8 project was effectively a Java SE 8 operating system and as such all of the notes are in the context of that scope. That project is no longer my goal as SquirrelJME is the spiritual successor to it.


Awake, sort of. Now to rewrite the processor code so generation of source files is much cleaner.


I could probably use a file which automatically creates files as I desire with my create script when it comes to Java and properties files.


Was doing real life stuff, but this better create script should save me some time from copying and pasting since I do create many files. Should have thought of it months ago.


For SQ my plan was to make things isometric, but I am going to stick to tiles and then have everything like of like a cliff side view of things. So essentially the cliffs of edges are on the bottom of tiles and all tiles will be square and can be walked on top of or not. It would be much simpler to draw and require less graphics too.


Using a special interface I can make my char arrays be CharSequence and have them actually be Appendable too.


Looks like hairball with annotation processing does not run the processor if the parent package has already been built.


In fact, it seems as if the processor is no longer called despite being called before all the time.


Changed the code to load from the package list, but still the processor is not called at all. Maybe it is an optimization. No, because I use annotations in my code. The only one I really use is @Override however. As for the order, before annotation processing has always run before compiling everything else.


From what I understand of the processing code it only does processing if there are annotations present, so even with a fake annotation it does not appear to run still when there are definitely annotation present.


Checking revisions:


The only thing I did between those revisions is fix the constant pool and other such errors. However I have removed all instances and usages of EntryType in the ConstantPool class. Maybe that is why.




Instead of an annotation processor I am just going to modify hairball so that instead of a processor there is a source generation mechanism which I can use to generate source code with and such. Then I would know if code is called. That would slightly complicate the build process however, but it could work. That would have to wait until tomorrow though since I am tired, fizzled, and somewhat angry about this. However my code does not really use any annotation bits at all so moving the code over will be very simple. I just need to write a generator interface that can be used for file output and adding extra files to the Java source input.


Suppose I will make a META-HAIRBALL directory in the packages (which is not included in the JAR), if classes exist there they will be compiled and loaded. Then a hairball specific generation pass will be used which can modify inputs, add source files, remove them, etc. Then with this, the generation stuff can be stored in the same package and ran without doing much magical dependency stuff that may exist.