[ / / / / / / / / / / / / / ] [ dir / 8teen / abdl / fur / hentai / htg / madchan / newbrit / strek ][Options][ watchlist ]

/tech/ - Technology

You can now write text to your AI-generated image at https://aiproto.com It is currently free to use for Proto members.
Name
Email
Subject
Comment *
File
Select/drop/paste files here
* = required field[▶ Show post options & limits]
Confused? See the FAQ.
Expand all images

[–]

 No.807956>>807989 >>808013 >>808016 >>808255 >>808265 >>808316 >>808344 >>808773 [Watch Thread][Show All Posts]

HolyJit: A New Hope

tl;dr: We believe there is a safer and easier way of writing a Jit.

Current State

Today, all browsers’ Jits share a similar design. This design makes extending the language or improving its performance time-consuming and complex, especially while avoiding security issues.

For instance, at the time of this writing, our Jit relies upon ~15000 lines of carefully crafted, hand-written assembly code (~36000 in Chromium’s v8). The Jit directory represents 5% of all the C++ code of Firefox, and contains 3 of the top 20 largest files of Firefox, all written by hand.

Interestingly, these files all contain code that is derived by hand from the Interpreter and a limited set of built-in functions of the JavaScript engine. But why do it by hand, when we could automatize the process, saving time and risk? HolyJit is exploring this possibility.

Introducing HolyJit (prototype)

This week, during the JS Team meetup, we have demonstrated the first prototype of a Rust meta-Jit compiler, named HolyJit. If our experiment proves successful, we believe that employing a strategy based on HolyJit will let us avoid many potential bugs and let us concentrate on strategic issues. This means more time to implement JavaScript features quickly and further improve the speed of our Jit.

https://blog.mozilla.org/javascript/2017/10/20/holyjit-a-new-hope/

https://github.com/nbp/holyjit

discuss

 No.807963>>807965

>rust

ew


 No.807965>>808479

>>807963

i put a trigger warning right in the title just for faggots like you. why do you feel the need to reply to the thread if you have nothing to say?

at least sage this shit


 No.807968

Sweet. Rust's developers may be shot through with SJW weasels that would be the first up against the wall in case of civil war, but it does seem to be capable of honestly improving developer productivity and code safety, which is huge for a native compiled non-GC language.


 No.807982

>mozarella

kill yourself sjwnigger


 No.807989

>>807956 (OP)

>discuss

rust rust rust poop


 No.807996

>executing javascript


 No.808012

Is this it?

Is this the year of Rust?

Is this the timeline of Rust?


 No.808013>>808074

>>807956 (OP)

>but is unlikely to be part of the final solution.

>final solution

OY VEYYY

someone create an issue and complain about this choice of words


 No.808016

File (hide): 51dae817ebaa889⋯.jpg (108.23 KB, 500x616, 125:154, 51dae817ebaa889d9b9aae93ef….jpg) (h) (u)

>>807956 (OP)

>This week, during the JS Team meetup, we have demonstrated the first prototype of a Rust meta-Jit compiler, named HolyJit. If our experiment proves successful, we believe that employing a strategy based on HolyJit will let us avoid many potential bugs and let us concentrate on strategic issues.

You are correct. Why haven't you already have done this?

>This means more time to implement JavaScript features quickly and further improve the speed of our Jit.

You are correct again. But why in the actual fuck are you using this to improve the unredeemable trash that is pajeetscript? With javascript you NEED undefined functions and functions that have memory overflowed by default. Otherwise you break web backwards compatibility. So unless you are literally disabling memory safety in rust defeating the whole purpose of why you wanted to use it in the first place then you can't have a web compatible javascript interpreter. If you are going to even bother doing this, then port the seamonkey interpreter and have a more solid foundation

>our Jit relies upon ~15000 lines of carefully crafted, hand-written assembly code (~36000 in (((Chromium)))’s v8).

>(((our)))

Soros funded giga kikes get out.


 No.808074

>>808013

>someone create an issue and complain about this choice of words

I know you're just shitposting, but if anyone wants to raise an issue make a suggestion that can also be triggering, then someone else raises another issue with another triggering suggestion, and so on.


 No.808078

AOT here, JIT is shit.


 No.808255

File (hide): c7689af486e1d28⋯.png (614.03 KB, 620x253, 620:253, no.png) (h) (u)

>>807956 (OP)

>rust


 No.808265>>808296

>>807956 (OP)

I'm sick of rust threads it's bean more than a year and I wish more than ever that it would never had bean created.

Also the winning about JS can easily be solved by not using any JS.

Remove JS and you solve 99% of the problems of the web.


 No.808279>>808296

>Rust

Where are my double linked lists?


 No.808296

>>808265

i will never stop making rust threads :^^^^^^^^^^^^^)

>>808279

right here: std::collections::LinkedList


 No.808302

I fucking love rust


 No.808316

>>807956 (OP)

Kill yourself.


 No.808344

>>807956 (OP)

>our browser is abysmally slow

>let's put a compiler notorious for being slow into it and run it on every single page load


 No.808394>>808468 >>808573

File (hide): 4d3c0e617168b56⋯.gif (1.23 MB, 269x202, 269:202, 8cd397fa2731.gif) (h) (u)

>>JIT library for Rust


 No.808468

>>808394

Actually, it's for JS. It's written in Rust.


 No.808479

>>807965

Right wing trigger warnings? Sneaky schlomo.


 No.808573

>>808394

That's probably the best gif you could have chosen for reasons you won't understand here


 No.808576>>808577

Rust can still be stopped.

We just need to get some big name to declare 2018 "The Year of Rust on the Desktop"


 No.808577>>808595

>>808576

It's ok anon, we've got Mozilla on the case; everything they touch turns to absolute shit. As for the Rustroach infestation here, spraying them with Ada helps.


 No.808595>>808601 >>808614

>>808577

quick Ada question: how do you use buffered output?


 No.808601>>808645

>>808595

No one has actually ever used Ada.


 No.808614>>808623 >>808625 >>808685

File (hide): 200ab8320ccb54e⋯.png (140.96 KB, 785x541, 785:541, developer_responsibility_i….png) (h) (u)

>>808595

If you're asking whether the standard Ada library has buffered output, then no, you have to define it yourself. Fortunately buffers are trivial shit.


 No.808623>>808627

>>808614

That's a weird graph to have with that answer, but OK: when I looked around I saw hints of "oh gnat has a switch" or such, so I kept looking.


 No.808625>>808627

>>808614

How does a huge pile of corporates and government generate a massive language spec yet somehow not even handle simple "Hello, World!" buffering?


 No.808627>>808641

>>808623

Meh, I switch out Ada-related pics from time to time, and it's kinda hard to find a pic that's both about Ada and about I/O buffering.

Have some documentation:

https://docs.adacore.com/gnat_rm-docs/html/gnat_rm/gnat_rm/the_implementation_of_standard_i_o.html

>There is no internal buffering of any kind at the Ada library level.

>>808625

>Ada

>massive language

The spec is excessively detailed and precise, as it should be, but massive? Nah. There are a couple of things that probably should be in the standard library that aren't, like bignums.


 No.808641>>808657

>>808627

>massive? Nah

The spec is a thousand pages long. C++17 is around 1300 without the index and is considered to be impossibly bloated.


 No.808645>>808648

>>808601

I use it at work


 No.808648>>808682 >>808747

>>808645

Are you one of the many embedded devs that frequent /tech/?

We sure have a lot of those.


 No.808657

>>808641

The Ada standard also covers a lot more than C++ in a more detailed, consistent, and saner way. For further comparison, consider that the C11 standard is about 700 pages and the Common Lisp standard is about 1100 pages.


 No.808682

>>808648

Sorry, Javag, but the FP circlejerking board is on halfchan.


 No.808685

>>808614

Problem with this graph is the tools have to be done by a developer, so really you've achieved nothing.


 No.808747

>>808648

Not really, it is for simulation in defense


 No.808773

>>807956 (OP)

>a new hope

>naming your shit after a stupid b-movie that eventually was artificially memed into popularity




[Return][Go to top][Catalog][Screencap][Nerve Center][Cancer][Update] ( Scroll to new posts) ( Auto) 5
38 replies | 5 images | Page ?
[Post a Reply]
[ / / / / / / / / / / / / / ] [ dir / 8teen / abdl / fur / hentai / htg / madchan / newbrit / strek ][ watchlist ]