[E2] gvar overhaul

[turck3]

Wow. Ridiculously better.

[feha]

(From another thread)

The question is though, is your gvars breaking the current e2's? XD

No, old E2s work fine. I've already tested it.
I did, however, remove a few functions as well (but not the ones you removed).
I removed the gSet*(*), gGet*(), and gDelete*(*) functions (the ones with no index arguments) because they were utterly useless. They saved the gvar to a single variable rather than in a table at the specified index. This means 3 extra functions for only 1 variable (per type).

LIEZ, I tested them, and apparently they do break. My code:

First E2:

gShare(1)
gSetStr("test","hello world")

Second E2:

gShare(1)
print(gGetStr("test"))

[Divran]

(From another thread)





LIEZ, I tested them, and apparently they do break. My code:

First E2:

gShare(1)
gSetStr("test","hello world")

Second E2:

gShare(1)
print(gGetStr("test"))

Works fine for me. Tried that exact same code.

[Buggzie]

sv: Expression 2 (GunShop Console Screen Controller (Networked)): entities/gmod_wire_expression2/core/globalvars.lua:58: attempt to index global 'gravs' (a nil value)

Breaks when i do this:

        Group = gTable("GunShop")
        CostArray[1,number] = 150
        CostArray[2,number] = 250
        CostArray[3,number] = 300
        CostArray[4,number] = 400
        CostArray[5,number] = 450
        Group["Cost",array] = CostArray

post merge:

scratch that.

If you have gTable("WHOOPDYDOOHERESMYTABLE") your e2 will crash because you have no following number... gTable("Table",0)

quick little line edit?

[Divran]

sv: Expression 2 (GunShop Console Screen Controller (Networked)): entities/gmod_wire_expression2/core/globalvars.lua:58: attempt to index global 'gravs' (a nil value)

Breaks when i do this:

        Group = gTable("GunShop")
        CostArray[1,number] = 150
        CostArray[2,number] = 250
        CostArray[3,number] = 300
        CostArray[4,number] = 400
        CostArray[5,number] = 450
        Group["Cost",array] = CostArray

post merge:

scratch that.

If you have gTable("WHOOPDYDOOHERESMYTABLE") your e2 will crash because you have no following number... gTable("Table",0)

quick little line edit?

Make sure you have the latest SVN version of Wiremod.

[Buggzie]

Make sure you have the latest SVN version of Wiremod.

i update daily

[Selig]

Am I the only one not very fond of the Blah["Index", type] syntax? It just "Feels" wrong to me. In most cases the parser should be able to guess the type of variable you are setting/getting, shouldn't it? But then the problem is with how to return that type, since there is no type conversion (From what I recall). I might look into how this can be solved sometime. ~Adds to her massive TODO List~

EDIT:
Am I right in the guess that this is the type of thing initrd.gz is trying to solve with the "generic" type he is making? (Though it's not quite there, you still have to pull the type info out yourself)

[Divran]

i update daily

G = gTable("test")

Tested, works fine.

[Syranide]

Am I the only one not very fond of the Blah["Index", type] syntax? It just "Feels" wrong to me. In most cases the parser should be able to guess the type of variable you are setting/getting, shouldn't it? But then the problem is with how to return that type, since there is no type conversion (From what I recall). I might look into how this can be solved sometime. ~Adds to her massive TODO List~

It's impossible to solve the way the language currently works, and it's done that way to be fast and simple. Setting the an index doesn't NEED the type, but it's there so that people don't mess up and it keeps the syntax consistent. Issue is not even that there is no type conversion, there can be no implicit type conversion when assigning to variables unless one changes the entire scope-model in E2 and a bunch of other stuff. E2 is statically typed and deduces variable types from assignment, hence, assignment to variables must always have a predefined type... which can only be known if you specify it or return a generic type that is implicitly cast when used (but that opens up a lot of ambiguity and unnecessarily complicates the language).

Long story short, E2 is statically typed, there is nothing one can do about it without adversely affecting the language.

[Buggzie]

G = gTable("test")

Tested, works fine.

not egpTable

gTable.