carthage/reactivetask

Flexible, stream-based abstraction for launching processes

ReactiveTask

ReactiveTask is a Swift framework for launching shell tasks (processes), built using ReactiveSwift.

let strings = [ "foo\n", "bar\n", "buzz\n", "fuzz\n" ]
let input = SignalProducer<Data, NoError>(values: strings.map { $0.data(using: .utf8)! })
let task = Task("/usr/bin/sort")

// Run the task, ignoring the output, and do something with the final result.
let result: Result<String, TaskError>? = task.launch(standardInput: input)
    .ignoreTaskData()
    .map { String(data: $0, encoding: .utf8) }
    .ignoreNil()
    .single()
print("Output of `\(task)`: \(result?.value ?? "")")

// Start the task and print all the events, which includes all the output
// that was received.
task.launch(standardInput: input)
    .flatMapTaskEvents(.concat) { data in
        return SignalProducer(value: String(data: data, encoding: .utf8))
    }
    .startWithNext { (event: TaskEvent) in
        switch event {
        case let .launch(task):
            print("launched task: \(task)")

        case let .standardError(data):
            print("stderr: \(data)")

        case let .standardOutput(data):
            print("stdout: \(data)")

        case let .success(string):
            print("value: \(string ?? "")")
        }
    }

For examples of how to use ReactiveTask, see the Xcode and Git integration code from the CarthageKit framework.

License

ReactiveTask is released under the MIT license.

reactivetask's Issues

Better Readme

The README.md is a little bit confusing.
I created this iOS Open source Readme Template so you can take a look on how to better organize.
If you want, I can help you to organize the README.md.

What are your thoughts?

Add Swift Package Manager support?

as title

xctool is required for bootstrap script

Possibly outdated scripts.

Providing interactive prompt input

Hi, I'm trying to figure out if there is a way to simulate interactive input as a reaction to the standard output.

For example, imagine calling $ git add -p with a bunch of unstaged changes, normally it'll produce something like

diff --git a/Hello world! b/Hello world!
index d45509a..b8f14f0 100644
--- a/Hello world!  
+++ b/Hello world!  
@@ -1,7 +1,3 @@
-
-
-
-
        Hello world!


Stage this hunk [y,n,q,a,d,/,e,?]? 

Let's say I want to response with y\n after analyzing output. Am I correct that currently there is no way to provide such a response? From what I understood, normally, I'd write into the standard input of the task. Do you think it's a good idea to add a case to TaskEvent providing you with a chance to write into the task after output was provided?
I didn't find anything similar in the carthage, but I might've missed it

Update it seems like it might be impossible according to several sources: http://www.cocoabuilder.com/archive/cocoa/64719-nstask-nspipe-and-interactive-unix-command.html

Subprocesses are not terminated when the parent process exits

Originally reported and discussed in Carthage/Carthage#89. The most notable comment is this:

From a lot of Googling and consulting with @keithduncan, it appears that NSTask is a special 💩❄️ in this scenario, and spawns processes outside of the parent's process group. The same problem supposedly wouldn't arise with fork() or posix_spawn() directly.

New Release

I really want to get the recent API changes and the RAC updates into Carthage. Would anyone object to me releasing master as 0.9?