JS-Git | Hacker News Reader

JS-Git

github.com

117 pointsdbaupp22 comments

Check out my pure javascript git implementation. https://github.com/ryanackley/git-html5.js

It has a complete workflow at the moment. You can see it in action here: http://tin.cr/tailor.html.

It looks like you're both using the MIT licence (although your LICENSE.md file doesn't explicity state it), so it should be possible to reuse code between the projects. Although whether the architectures you have chosen are amenable to that might make it difficult.

"JavaScript has farther reach than any other programming language out there."

The ghost of Dennis Ritchie would like to have a word with you...

#include "necromancy.h"

    #include <Python.h>
    int main() {
        Py_Initialize();
        PyImport_ImportModule("necromancy");

require 'necromancy'

fixed it for you

        return 0;  // <stdlib.h> not #imported
    }

Seeing as the js-git project is maturing quickly, I was thinking it might be time to integrate into npm.

https://github.com/isaacs/npm/issues/3765

The inexorable march of Atwood Law.

I can't wait to see the Javascript implementation of Linux.

see http://bellard.org/jslinux. He implemented a PC emulator in javascript and then just compiled a version of Linux to run on the emulator.

http://bellard.org/jslinux/

Previous discussion (2011):

https://news.ycombinator.com/item?id=2555349

I can't wait to hear Linus' reaction. He about blew a gasket when someone wanted to use C++ in the main git project. Then again, this is a wholly separate project so he might just shrug and move on, but I still think he might have some colorful comments.

I don't personally buy his claims that performance is so critical it needs to be written in C. I'd be interested in any speed comparisons the js-git authors might have made. It would surprise me if js-git was anywhere near as fast, but I can't imagine it's impossible to make it at least usable.

Speed is definitely a huge benefit of git. When you use it heavily every day, you become thankful that cloning, merging, branching etc is very fast. If it were a couple of seconds slower for each of these tasks, I think it would make a big difference.

Of course, but whether a JS implementation is necessarily a couple seconds slower for typical (or even extreme) scenarios remains to be seen.

Mozilla seems to really want this.

I can completely see why Adobe (Edge Code/Brackets) and Mozilla want this. It's a missing piece for coding completely in the browser. But I think it goes way beyond this, and once people start playing around with having git sandboxed inside the browser we're gonna see some really, really interesting ideas come out of it. I was sad that his second funding wasn't seemingly headed for success, and it's pretty awesome Mozilla laid down $25k to revert that.

What are some interesting ideas you can think of that could come out of this?

from the site:

ChromeOS IDE for developing on Chromebooks.

Node.JS blog engine with git as the database.

Custom Git hosting using custom storage back-ends.

GIT CLI for restricted environments.

Standalone GIT GUI desktop app.

Git based deployment tools.

JavaScript package management for server and client.

Whatever else you come up with.

As a guy that works for one of the major banks in the US as a developer...the restrictions placed on even our development boxes is rediculous...HOWEVER they don't seem to give two craps about javascript...so this would be a GREAT capability for me to have and extend my workstations capabilities in a myriad of ways...

I can't wait for this! I might have to try and contribute if I can actually...

We're also very interested in having this for http://noflojs.org/. Our current git components just run the CLI version, but obviously pure JS library would give us more control.

Reply on news.ycombinator.com