I believe I like this new class decoding code straight in the JIT much better. It just gets right to the point.


One thing I need to consider are virtual method calls. Due to the default by nature execution of methods. It would either be a read of a field or a double jump to the correct method area. So basically any call to a virtual method will read some kind of offset and then read the pointer at that offset which then points to a method that should be the one that is executed. That is a way to easily implement virtual calls. Static and special calls however can just directly execute methods.


Interfaces would require interface tables. But interface lookup should be fast. I want to completely avoid having linear checks for classes and interfaces. Just want it to be a simple jump.


Another thing that can exist in the cards is the check and comparison nature of values. So if a null pointer was checked then it does not have to be checked. But if there is a jump from an unchecked position to a checked one then the check will have to be performed or delayed at that point. That can lead to inconsistent exceptions around branches around code which should execute, so that is a no-go. Unless I have a potential alternative path. But actually I can generate all of those checks and then optimize them away for example. There can be a dependency detection of things and such to determine which checks could be potentially removed. This would then mean that I am definitely going for an IL. So then faces which implement IL operations should be abstract classes which are extended by the architecture specific code. So there is an addition face for example. Then the MIPS one just knows how to generate a machine code instruction from that. I could however take a hybrid approach. Have a default face set that utilizes registers and such. Then have faces that are architecture specific such as clearing interrupts or whatever. But it would probably be the simplest to just have standard cards which perform standard operations on registers. Then on output to machine code, just translate those cards to native operations as needed. That would make the JIT implementations simple as they would just translate the faces to machine code accordingly. The faces would be modeled after MIPS because that is the simplest. Other architectures would map well to it. The faces used for one target will not be compatible with another target however. This is because of the registers and data types. This would allow me to support many architectures more easily because I would only have to write the final stage. But for final stage reusing, the standard cards will be replaced by architecture specific cards. That way there can be a final architecture specific optimization pass. Then it would also enable me to use the card system like I plan to when generating internal use methods.


As for invokeinterface. I could do a large number of static references. For example say method a from class A invokes the method z in interface Z. This can be determined at linking time. Basically a would do something similar to the virtual invoke. But would link to a special dispatcher. But pretty much a dispatcher must be used. However every class would have interface tables which point to associated interfaces. Also since tons of objects will use the same virtual method links, they will be in the class export table for those classes. So class B replacing a method in A. For A's class table that points to the code for that method, while for class B in that same index (in A's part of the code) will point to the method it is replaced by. So every class would have unique export tables.


The deck should be completely type safe with known program control flow in the cards. This would mean that no cards can dangle. Future cards in the chain would morph based on the previous cards. This means for state though, there would be recursive state determination based on registers and such however. I would basically have basic blocks in a way. I would need backlinks, aliases, and some other things too. If state is stored then I would have to check if any dependency has changed before returning it, which would be insanely slow. And that would require lots of code and compution time to calculate. I will need multiple parts of methods. prolog, epilog, main body, exception jumpers. Then whenever a card is placed in the output it becomes fixed and cannot be changed. So once a card is placed it is immutable and has a purely fixed state that will never change. Forward jumps to undeclared cards define state while backward jumps to declared cards must forward state. So the exception jumpers would purely be state cycling. The state cyclers would have a given input set of registers and a given set of target registers.


If an existing transition from one state does not match another then a new one will be created. The prolog would require that the state of the input at the end of all of its execution matches the main body entry point. Jumps to the prolog and epilog would be illegal. The epilog may only be jumped to in the main body or the exception checks. I would need an exception check for instances. Then if no exceptions are handled it would just go straight into the epilog and return the exception for handling by the calling method. So once card is placed, it is immutable. The prolog state can be checked at the end to make sure it is valid. I then also continue the route of treating exceptions as just another variable to be checked along with the secondary return value.


Immutibility would be easier for states, less compicated.


Ok so for removing things such as bounds checks and null pointer checks, I can have potential byes.


Basically, when a variable is modified there can be a bye chain for each variable representation which has a source value. This means getting rid of aliases and having it completely being value based. So basically on entry of a method, variables are completely washed away and everything is just a value and where it is located. So you could for example copy a local variable to the stack then set the local variable to some other value, or duplicate it. That value is never actually lost. So arguments are called by value. Then each value at a specific position has flags which determine if the value as changed or similar. However, code generated has to be done in a way where byes do not exist. The removing of operations is only a final step optimization.


So if there is a jump back where the value is not compatible then it will perform those checks. This means that loops would have many many fewer checks if the inputs are correct. I would also need to make sure that variables get keyed with lengths also. So there would be branch conditions where a variable is always less than a given amount. Then if some values are associated, then the checks can be removed.


So this is much more complex now. There would still be weak bindings to variable slots because that is how Java operates, via the stack, rather than discreet values. Because jump backs can modify variables as they are placed. So unless the value is copied back, it would have to split and make it unknown.


So actually I do not have to worry about the slots in a way. Basically the slots are windows to values.


So basically the accessing of values is done via pointers and such in a way.


I believe for interfaces that calls should be treated as if they were virtual, but there would be virtualized methods for each method in the interface which dispatches them accordingly.