aaaah ok, well it was more for Kindermoumoute, who wanted to know how to copy a sprite from a pointer to another to no avail.

That's the copy command Conj().  You can use it to copy a sprite,  into for instance the L1 address, by doing conj(Pic1,L1,8)

Really? for the grayscale in the games I've made, I always use 3/8 on for light gray (37.5%) and 5/8 for dark gray (62.5%).  Do other 4 level grayscale games use 1/4 and 3/4 for gray levels?  How would these be interwoven?

I would think it would have to be something like this for best results:
11101110
10111011
11011101
01110111

It seems hard to create this kind of cycle though.

I think you're thinking of "VPutS" instead of "PutS".  You also change "PutC" to "VPutMap" and instead of using CurCol and CurRow you use PenCol and PenRow.  I'm surprised this worked in small font before as a non-app, that's strange...

The sprites not working now but working before sounds like you might still have some self modifying code.  Either that, or you're using some other bcall that can't be executed in the $4000 address.

@ztrumpet The [on] interrupt will come eventually.  But keep in mind that only protects you from infinite loops and NOT from writing data in the wrong places.  Also, it does not allow you to to use your own interrupts in this mode, however, you can still add quitting code in the custom interrupt yourself.


Well, I'm rewriting the way number parsing works.  I have to or else I can't implement some new optimizations.  I'm grouping every single routine that can possibly return a constant into a single read-number routine.  That includes Integers, Hex, Binary, Characters, Tokens(new), Pointers, and everything else.  The goal is to make all expressions that evaluate to constants be single constants, not just pointers.

Now, I ran into some really cool syntax problems.  I was going to include multiplication and division in these optimized routines which can really help sometimes.  For instance, when using *256 coordinates you can type 44*256 to make 44 the coordinate and Axe automatically multiplies this precompile time.  But then I realized there was a slight technical problem with doing this.  What if you typed A/2/2 to do faster division?  Axe would automatically optimize this to A/4 which although it is actually optimized in terms of size, it definitely isn't in terms of speed.

So, I think I will only optimize multiplication and division if the first number is a constant rather than if the first operation is a constant operation.  However, I will still optimize addition and subtraction both ways since you never have these problems with those operations becasue there's no speed loss from grouping individual operations.

This might take a while to re-code a bunch of routines to handle the new method, but I've already done the bulk of it and its turning out pretty good so far.

Hmm... I like the idea of having a command to return the VAT pointer, but it seems like it would be too low level to work with.  You would need to know all the equates to change the type for instance and know how that data structure looks in general to change anything.  Although it would be easy to add, I think it would just require too much assembly knowledge which is what I'm trying to avoid in this project.

If I ever do add it, it would be at the very end of the project where I just add a bunch of "undocumented" instructions (which actually are documented separately) but for assembly programmers who want to do some really low level stuff.

Oh yeah, I completely forgot about that!  I actually did add it, but don't use it becasue I forgot to make the arguments in order.  Its currently sortA(size,ptr) but I will change it to sortA(ptr,size) probably.  Or I don't know, which one do you think is more intuitive?

He's talking about the new error scrolling feature, not something wrong with the calculator.

It sounds like what's happening is that the program pointer is changing to a different location so its scrolling through garbage to the location of the error instead of the actual program.  I'll look around for something that might cause that, but I guess this feature is just temporary anyway until I have the real thing working.

I haven't started working on it yet, but probably all the stuff I was going to do last week but didn't have time for.