First stab at implementing shell command.

[Oshuma]

Issue #196

The only thing left is to steam stdout/stderr to Jupyter. I couldn't quite figure out how to do that, so if anyone could give me some pointers/examples, I can implement it.

[cosmos72]

Looks good, thanks :)

To stream stdout/stderr to Jupyter, you need four things:

1. get the stdout/stderr of the command as Go streams: you can use the methods Cmd.StdoutPipe() and Cmd.StderrPipe() of the struct os/exec.Cmd you are already using
2. modify Kernel.handleExecuteRequest to create the two JupyterStreamWriter immediately - it currently creates them inside goroutines - and pass them to doEval() and further down to your evalShellCommand.
3. a goroutine reading from Cmd.StdoutPipe() and writing into the first JupyterStreamWriter. You can copy and adapt the code in kernel.go, lines 406-411
4. another goroutine, reading from Cmd.StderrPipe() and writing into the second JupyterStreamWriter. You can copy and adapt the code in kernel.go, lines 413-417

Note: be careful to wait for the command to exit and for the goroutines to finish before returning from evalShellCommand

[Oshuma]

@cosmos72 Sounds good. I'll implement that as soon as I get a chance and update this PR.

[Oshuma]

@cosmos72 Can you verify/merge when you get a chance?

[cosmos72]

merged, with some little improvements.
Thanks!