I can not download the source code

[badly]

I use EGIT for eclipse to download the source, but its maybe the owner was disable the git.

This is the git link: http://replicaisland.googlecode.com/svn/trunk/ replicaisland-read-only

Anyone help ?

[Chris Pruett]

I didn't intentionally disable git, but if I remember correctly I had to select between svn and git, and I went with svn.

Cheers,

Chris

--
You received this message because you are subscribed to the Google Groups "ReplicaIsland Coding Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to replica-island-coding...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Joe Da Silva]

Hi Chris,
Would it be okay if I posted a copy of replica island on github with source code all aligned-up?

Earlier - I printed out the code to look at (so that I could see more code at once and also try to see interactions between different functions) - was very difficult to see and the alignment definitely didn't help.
I've aligned the same code as going 2-in and changed a couple of remarks to fit within 80chars so more code fits on a page before it wraps around.

If you're okay with me posting it, I'll put it up on gitgub unless you want it elsewhere.
If not, I'm okay with that too since this is code to learn from - just means more work for the end user if they have to end up doing the same themselves.

[Chris Pruett]

Hi Joe,

It's fine with me!  Just make sure you note that it's a branch and link back to the source.google.com page, please.

--

You received this message because you are subscribed to the Google Groups "ReplicaIsland Coding Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to replica-island-coding...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Joe Da Silva]

Thanks Chris,

It took a bit of effort to align it since there is a Lot of code, then a quick verify to ensure I didn't add errors, but it should be really close to original code, plus some patches I've found to bring it up bit closer to date :-)

I was going to begin tinkering with changing some floats to ints, but decided best not to introduce significant changes at this patch-point.

Code is located at https://github.com/JoesCat/ReplicaIsland for anyone interested in an aligned-copy of Chris' code.

[Chris Pruett]

Cheers!

I tested float vs int way back when I wrote this thing, and even on the very low-end FPU-less devices, there was no obvious win. 

Chris

[Joe Da Silva]

I happen to think from a lower-level perspective, like assembler on tiny chips like the microchip pic, motorola 68xx, even the zilog Z80 if anyone recalls those, and one thing that sticks is that there is a significant amount more CPU work to handle floats versus ints, since ints in assembler are simply jumps, add/subtract, shifts, tests.

I can see your point, after all, Java is quite a high level language, therefore working with floats is likely not going to introduce much extra delay relative to other code.

Glad to hear you checked floats vs ints, but for me old habits die hard, so I'll likely stick with ints - just a preference thing at this point I suppose. ;-)

[Chris Pruett]

Hi Joe,

Yeah, I mean, I got my industry start writing code for the ARM9 for the GameBoy Advance, and was 100% fixed-point integer math too.

But in Java it makes no sense.  If I remember correctly, fixed math was actually slightly slower on the G1 in Java because the VM wasn't able to optimize it as effectively.

Anyway, at the time (2009) when I ran some benchmarks, there was no performance improvement and there was significant room for error (e.g. when doing fixed divides, etc), so I left it as float.  I think today it matters even less.

Cheers!

Chris

[Joe Da Silva]

That is a really important bit of info.
Hehehe - Seems old dog will learn new trick after all ;-)