E2 User Defined Functions

[jacoby6000]

I think his is more like action script :3

also
HOW FRAKING LONG HAVE YOU BEEN WORKING ON THIS WITHOUT TELLING US!?!?!?!?!?!?!?

[Colonel Thirty Two]

About a month, maybe. It actually wasn't too hard.

[mattwd0526]

So Syranide was lying to us the whole time?

[Divran]

Nice work on the functions.
Just wanted to tell you that the thing able to run Lua code snippets will definitely not be added to the official SVN. If you do that, you might as well add a crashServer() command while you're at it.

[Colonel Thirty Two]

It doesn't run Lua snippets, it runs E2 code.
However, I think there might be a way to hack it to run arbitrary lua code involving glon, I'll have to fix...

[Divran]

It doesn't run Lua snippets, it runs E2 code.

ohhhhhhhh my bad
EDIT: Heh you're right. Even the first post says "runs strings of E2 code"... I failed
EDIT: In that case, nice work on all of it

[Riddler]

Er...

C++:
returntype name(parameters) {...}

Lua:
function name(parameters) {...}

Neither of those start with the name of the function, and neither of them involve defining a variable (except C++ defining the type)

Actualy, if you want to be nit-picky about it they do just not in a way that we would tend perceve them as variables. As far as what you see in lua/c++ that tends to be true. However as far as the 'inner workings' of it go and what is actualy given out when it's compiled that's wrong.

Function names are stored in a table and when you call a 'function' it looks it up in the table, the function name is a veriable and if your compiler permits it, you can overwrite it, set it to a normal number, whatever and treat it exactly like a variable. This is because when it's compiled there is no variables or function names anymore, it's just set/call memory cell X, so "functionname()" in code is actualy replaced with "goto memory cell X". So while memory cell X could be a 'veriable' you could rewrite it to be a 'function' the computer won't care, as long as it's used in the right way.

Thats why you get errors from C++ compilers about the function table, it's just a table where the function names are a veriable containing a pointer to it's code. WireCPU is a good example, if you've ever used it. labels are pointers to adresses in memory you can put a 'function' after them and call it or put a veriable after it, and change it to the other later, as long as it's used correctly the cpu dosn't care. The variable system is there and seperate from the function system because it's convienient and nicer to work with, nothing more. I use the word function and variable there because I can't think of a better word, to the processor the two are not seperate but I needed words to diferentiate for the purposes of the example.
That's oversimplified quite a bit, but you get the idea.

In the start, theres both, in the middle there is only one, in the end there is nither.

[Dav1d]

Pinch me, I think I'm dreaming.

[mattwd0526]

@Riddler, yeah, but I'm talking about what we would see

[itsbth]

Er...

C++:
returntype name(parameters) {...}

Lua:
function name(parameters) {...}

Neither of those start with the name of the function, and neither of them involve defining a variable (except C++ defining the type)

Actually, function foo() end is just syntactic sugar for foo = function() end in Lua in its current form.

On topic: Nice work, though I doubt it'll make it into the Wire SVN in its current form.