2016/10/07

09:44

Going to work on Squirrel Digger for a bit and take a break from the JIT compiler since I just cannot seem to continue it at this time. I suppose that it is burn out and Squirrel Digger is completely different.

09:57

One thing to consider is the limit to the world size. For chunk IDs I can use 32-bit or 64-bit values. However since 64-bit is quite large I can just limit it to 32-bit. This would be the most sane because this is only for chunk references. If I divide by 3 I can have 10-bits for the chunk position. This means 1024x1024x1024 chunks are possible.

10:04

The chunk size can then be derived from the 64-bit value. That divided by 3 gives me 21-bits. Since 10 are used for the chunk, that leaves 11 for the chunk size and the sub-block precision. A chunk size of 8 gives me 3-bits which means that the sub-lock is divided into 256 parts. If I increase the chunk size to 16 then it gets divided by 128 parts. Since I want to target some very small systems that may be rather constrained, keeping the chunk size of 8 could be quite easy when it comes to memory.

12:17

Must find a JAR with JSR271.

12:18

JSR271 is defined, however so far I can only find SDKs with MIDP 2.1 at the most. MIDP 3.0 is far better than 2.1.

12:26

The SDK has JSR 135, which is the sound API which MIDP 3 partially uses.

12:51

I suppose what I would have to do is take the base MIDP 2.1 and then just add anything that is defined by MIDP 3 manually.

12:52

So yes, this will have to be done. Take MIDP 2.1 and just add anything that is in MIDP 3 manually by hand from the documentation. It would not be completely from scratch at least.

12:55

I will start with the media protocol since MIDP 3 uses that and there is a JAR I can extract symbols from.

14:06

Soon I will see exactly what the difference between MIDP 2.1 and 3 is when it comes to LCDUI.

14:10

Probably the best thing to do would be to perform a graphical diff of the HTML JavaDoc rather than going by it one by one.

15:26

Cannot find any JARs and diffing HTML is a bit ugly. So I will have to go through it by hand.

20:41

Looks like that the file chooser uses javax.microedition.io.StreamConnection however it is not in the documentation or the JAR.

21:21

Actually, the I/O classes were moved into the GCF so I must handle that accordingly.