Faulty restoration of super global variables after for loop Topic is solved

[Helgef]

Example,

Code: Select all

k(){
	for k in [1]
		(1)
}
global k := 1
k
msgbox k ; blank

compare to

Code: Select all

global k := 1
k(){
	for k in [1]
		(1)
}
k
msgbox k ; 1

and

Code: Select all

k(){
	global
	for k in [1]
		(1)
}
k := 1
k
msgbox k ; 1

First observed here (see the spoiler).

Cheers.

[iseahound]

What does (1) do?

[swagfag]

its just a garbage forced expression, evaluating to true, then discarding the result as there's nothing to capture it, intended to make the forloop run without the need for calling actual function or placing braces

id rewrite it tbh:

Code: Select all

fun(){
	for var in [1] {
	}
}
global var := 1
fun()
msgbox(var) ; blank

Code: Select all

global var := 1
fun(){
	for var in [1] {
	}
}
fun()
msgbox(var) ; 1

Code: Select all

fun(){
	global
	for var in [1] {
	}
}
var := 1
fun()
msgbox(var) ; 1

[iseahound]

That is very cool. Appreciate the explanation. It might be better to use (0) instead to denote the null process.

EDIT: or just continue

[lexikos]

Fixed in v2.0-a097.

The bug applied to all aliases; that is: ByRef, free variables (variables captured by a closure, both in a closure and in the outer function), and "local references" to super-global variables. The last one is an implementation hack; the variable references are resolved when they are encountered, so a later stage converts each undeclared local variable to an alias if it corresponds to a super-global.

[Helgef]

Thanks for the updates.

Sorry for the confusing examples.

id rewrite it tbh:

yes, that is better.

Cheers.