GPVM

/proggles/, would you agree if I were to say that high level languages based on MREs are both fragmentated in their implementation of MRE and pityfully unstandardized, in terms of VM?
What if someone were to write a ``General Purpose Virtual Machine'' that both standardized, and provided facilities to execute code generated by compilers for said high-level languages? The benifits would be great, no? If something like this were to catch on to the point of a community, with growing support for languages in the form of compilers, it would be easy for, say, browsers, to support something like this; for it could run perl, dart, python, PHP, javascript or any other language's code if there were compilers for those languages, and it would be good.

What does /plague/ think? I personally think this should have been done a long time ago, for a few reasons; MREs take a long time and a lot of testing to implement, and their adoption rate is massive. With the same adoption rate, and a single VM in use, language developers could focus on the quality of their languages, so we don't have more shitty language like Python.

http://parrot.org/
Check out Parrot.

>>1
There are already Python clones. See Falcon. Pig disgusting.

``Parrot is a virtual machine designed to efficiently compile and execute bytecode for dynamic languages.''

>dynamic languages
>implying VMS should care about typing

The only reason a VM would need to know the type of an object is type specific operations, such as dividing a float as opposed to a long. Even that can be approximated by a compiler for a typeless language. I don't see why these dynamicfags see themselves as a paradigm.

A VM should be able to allocate any amount of bytes for any variable; types should be completely irrelevant to the VM, only addressing modes and how the memory is used in operations.

>>4

uh, you are judging based on the introduction sentence. Parrot is "designed for dynamic languages" in the sense that ... fuck just read the wikipedia article you lazy twit

I'm not sure what you are even complaining about.

>>2
That looks really nice. Too bad I haven't seen anyone seriously adopting it.

>>6
Perl6 is using it now. (it is called Rakudo)

Also, this: http://pl.parrot.org/ (Parrot embedded in PostgreSQL)

(Mental note)
TODO:

* LISP implemented on ParrotVM

>>8
It won't be too long before someone implements use Lisp; in Perl 6. Actually it will probably be a new setting. Hacking the Perl 6 grammar is too easy.

>>1
There are already Python clones. See Falcon. Pig disgusting.

DubsVM

>>10
No, PIG DISGUSTING!

nice dubs. >>22, >>33, >>44, >>55,>>66, >>77, >>88 and >>99

>>13
Thanks.

>>14
Your'e welcom'e

The only reason a VM would need to know the type of an object is type specific operations, such as dividing a float as opposed to a long.

Unrelated. Does anyone knows why LLVM has separate instructions for ints and float as add and fadd?

because llvm is a fadd!