Trying to understand JavaScript from Ruby's view

[Sam Kong]

Hi Rubyists,

I have a question which is not directly about ruby but about other
language (but from ruby's view).
I may have to post to other newsgroups.
So if you think this posting is inappropriate, forgive me.

I understand how ruby implements value types (Fixnum, TrueClass,
FalseClass, NilClass) and reference types (other classes).
To separate them, the last 2 bits (or 3?) of variables are used as
flags (integer is 31-bit instead of 32-bit).
And I think that's a really genuis design.

I wanted to compare that design to JavaScript.
In JavaScript, numbers(8 bytes), booleans, null, and undefined are
immediate values and all others are references.
I want to know how they are handled in memory (like variable size, bit
flags, etc.) especially how differently they are implemented from
ruby's implementation.
I couldn't find a good document about it.
Yesterday, I posted the question to comp.lang.javascript but nobody
seems to care so far.
So I'm begging some help to rubyists who also understand JavaScript
well.

If possible, I also want to hear about python.
I read some postings regarding this issue in comp.lang.python, it seems
like everything in python is reference type unlike ruby.
Is it true?


I'm nervous that somebody might throw me a stone for this off-topic
posting.
I hesitated a lot before posting.
But I thought this was the best to get the answer.

Thanks.

Sam

 

[Martin Coxall]

I wanted to compare that design to JavaScript.

In JavaScript, numbers(8 bytes), booleans, null, and undefined are
immediate values and all others are references.
I want to know how they are handled in memory (like variable size, bit
flags, etc.) especially how differently they are implemented from
ruby's implementation.
I couldn't find a good document about it.

It would be implementation specific. In Rhino, they are Classes (ref
types), and in KJS they are structs. So, the rather unhelpful answer
is, it depends. It's not specified in the ECMA standard.

Martin