Confusion with "type vs typeof" usage

[XandrosDarkstorm]

Hello.
I am trying to understand the difference between calling type and typeof in Squirrel 3.0.7 and i am having a hard time.
Here is what i gathered so far:

type

• Is a function (script/user can redefine it, "=type" shows "function")
• According to the official documentation is supposed to return "raw" type of the object (without calling _typeof() metamethod of the object) [citing: "return the 'raw' type of an object without invoking the metatmethod '_typeof'."]. Which is exactly what it does.

typeof

• Is an operator (script/user cannot redefine it, "=typeof" results in an error "error expression expected".
• According to the official documentation is supposed to return the type of the object [citing: "returns the type name of a value as string."]. Considering that "type" function is supposed to return "raw" type of the object, it is assumed that "typeof" must return the "full" type (that is, call _typeof() and return its result). However, this is not what is happening -- it returns the same "raw" type, like "type" does.

I used this code:

s <- {
	_typeof = function()
	{
	return "mycooltable"
	}
}

local t = {}
local a = ["first","second","third"]
//creates a weakref to the array and assigns it to a table slot
t.thearray <- a.weakref();

print("_typeof test\n")
print(type(s) + "\n")
print((typeof s) + "\n")

print("=====\nweakref test\n")
print(type(t.thearray) + "\n")
print((typeof t.thearray) + "\n")

Please, help me understand: am i doing something wrong here or is it a bug?

[zeromus]

You didn't explain what you think the bugs are, but I can guess.
Bug 1 (line 3 of output): you are not setting the metamethods correctly. setdelegate() is required. This is squirrel 101
Bug 2 (lines 5 and 6): t.thearray is an array because the table get dereferences the weak reference so you can use it without knowing that it's a weak reference. if it were otherwise, t.thearray[0] would be nonsense because myweakref[0] is nonsense. What you want is t.thearray.weakref() which seems a bit crazy (is it a weakref to a weakref?) but no, it's the weakref itself (check SQRefCount::GetWeakRef() which returns the thing itself it it's actually a weakref). I don't blame you for getting this one wrong.

[XandrosDarkstorm]

@zeromus Thank you for the explanation. I think i now understand what was wrong with my code.
Weakref stuff was directly copy-pasted from the official documentation. Maybe this weakref.weakref() stuff needs to be mentioned there/corrected.
Anyways, thanks again for your help.

[zeromus]

hmmmm can you quote the exact part of docs at which URL?

[XandrosDarkstorm]

http://www.squirrel-lang.org/doc/squirrel3.html#weakrefs

The code examples from gray zones never show "t.thearray.weakref()". It feels as if t.thearray is supposed to be its own object with "weakref" type.

[zeromus]

Yes, that's what one would expect. But the docs say otherwise:

The table slot 'thearray' contains a weak reference to an array. The following line prints "first", because tables(and all other containers) always return the object pointed by a weak ref

t.theArray is "the object pointed to by a weak ref", not, "the weak ref".

This is SNEAKY and the docs could probably explain it better. But they do explain it. They would explain it better with a weakref.weakref() example which would alert pretty much any reader to the complication of it all.

Sorry to all, I don't have suggested alternative docs at this moment.