News for hackers. Stuff that matters.
Re:Inside the Go Playground by akanet on Thursday December 12, 2013 @10:24PM (#6898900)
If anyone's interested, https://coderpad.io has this exact feature as well - you can give an interview candidate a problem in Go, and watch them solve it in realtime. As you go, you can execute any of the code written so far.

My sandboxing uses Docker instead of NaCl.

Re:Inside the Go Playground by thomasahle on Friday December 13, 2013 @08:26AM (#6900348)
How do you handle things like filesystem, threading, networks and time?
Re:Inside the Go Playground by ithkuil on Friday December 13, 2013 @08:39AM (#6900386)
The time thing is ingenious. It reduces backend memory usage and allows caching.

However, it's possible to achieve the same advantage of cacheability, by executing a unmodified binary and recording the output in a time stamped manner, save it in a cache and replay it later.

Re:Inside the Go Playground by jitl on Thursday December 12, 2013 @09:40PM (#6898747)
I'm very interested in the NaCl approach to server-side jails:

  > To isolate user programs from Google's infrastructure, the back 
  > end runs them under Native Client (or "NaCl"), a technology 
  > developed by Google to permit the safe execution of x86 programs 
  > inside web browsers. The back end uses a special 
  > version of the gc tool chain that generates NaCl executables
I'm a bit sad that the article didn't discuss using NaCl for constrained execution environments a bit more. Docker/linux containers are still a bit heavy-weight for low-budget ARM servers, and it'd be interesting to check out NaCl as a more lightweight, per-executable sandbox.
Re:Inside the Go Playground by comex on Friday December 13, 2013 @12:59AM (#6899307)
I'm annoyed that NaCl's SFI is being repurposed so blatantly uselessly. The overhead isn't that much, but when seccomp provides equal or better security without any expense, it really seems like reinventing the wheel.
Re:Inside the Go Playground by krasin on Friday December 13, 2013 @01:37AM (#6899391)
It's always better to have multiple layers of defense. I won't be surprised, if Russ Cox has chosen sel_ldr_seccomp as the loader: https://src.chromium.org/viewvc/native_client/trunk/src/nati...
Re:Inside the Go Playground by enneff on Thursday December 12, 2013 @10:23PM (#6898899)
I don't think the NaCl port will do ARM code generation.

I'll write more about Go's NaCl support once it's actually in the tree. :-)

Re:Inside the Go Playground by jbarham on Thursday December 12, 2013 @07:45PM (#6898380)
And by the way, Go 1.3 will include support for Native Client (aka "NaCL").

Also: Russ Cox is the Chuck Norris of Go programming.

Re:Inside the Go Playground by pokstad on Thursday December 12, 2013 @07:49PM (#6898391)
So this is how Go will make its appearance on Chrome Browser & Chrome OS? Pretty cool. I was a little scared off from NaCl from the C/C++ requirement, but Go will make that a little easier to handle. When is 1.3 suppose to be released?
Re:Inside the Go Playground by jbarham on Thursday December 12, 2013 @07:54PM (#6898412)
> When is 1.3 suppose to be released?

Around 7 months from now according to http://blog.golang.org/go12.

Re:Inside the Go Playground by kkowalczyk on Thursday December 12, 2013 @09:50PM (#6898791)
Not really, at least not in 1.3.

As http://golang.org/s/go13nacl describes, supporting Chrome is not a goal for 1.3 release, only being able to run nacl-sandboxed executables from command line.

Re:Inside the Go Playground by AYBABTME on Friday December 13, 2013 @09:53AM (#6900737)
Can you expand on the Russ Norris part? Just out of curiosity as I also have much respect for Chuck Cox.
Re:Inside the Go Playground by mseepgood on Thursday December 12, 2013 @06:16PM (#6897965)
Runnable code in documentation, talk slides, blog posts and as a pastebin/playground is extremely convenient and user friendly.
Re:Inside the Go Playground by ykumar6 on Thursday December 12, 2013 @07:49PM (#6898392)
check out runnable.com :)