Swiftpack.co - danramteke/swish as Swift Package

Swiftpack.co is a collection of thousands of indexed Swift packages. Search packages.
danramteke/swish 0.9.16
Who wants to use a Ruby script to upload your iOS project? Not me. Let's use Swift.
⭐️ 4
🕓 2 weeks ago
.package(url: "https://github.com/danramteke/swish.git", from: "0.9.16")


Who wants to use a Ruby scripts to maintain your iOS project? Not me. Let's use Swift.


Bash scripts have gotten us pretty far, but it's difficult reasoning about control flow. And there's no type safety.

Another motivation is a script fails, and you make some fixes, and you want to run only the command that failed. However, the command that failed had lots of $VARIABLE substitutions that are difficult to recreate.

Many command line tools already have decent interfaces, it's just the control flow of bash (and typing!) that could use some improvements.


Swish is a Swift task runner that lets you reason about your script in Swift, easily calling shell commands and using their output in your Swift program. It also

  • keeps track of all commands run, and all output.

  • logs the out of the shell commands to the terminal, for easy copy-pasting for re-running failed commands.

  • can parse output into Swift types that conform to either ShellOutputInitable or StdOutputInitible

  • clears previous logs (by default) to keep a clean logs directory

  • Swish does not intend to offer a Domain Specific Language to different shell commands.

  • Swish does not intend to be used a #! point-of-entry for scripts


  • Enable calling command line tools easily from Swift, since Swift's type system is nice.
  • A swifty replacement for make
  • allow easy variable substitution in shell calls, and what was run in the shell is logged to the log, for easy copy-paste
  • allow easy calls to all the commands declared. Sometimes a shell script fails halfway, or you simply want to run part of the script.
  • scripts don't really want to be compiled since then their contents are opaque. Scripts want to be compiled just-in-time on every run

This package specifically does not try to provide a domain specific language for various tools. For those looking for a DSL, there is Fastlane or Puma.


There are two example projects in the Examples folder

  • The DockerDev example sets up and tears down some docker containers.

  • The AppIconRender example renders iOS App Icons using ImageMagick. The AppIconRender example also showcases Swish's make-like targets to avoid regenerating large assets when they are not needed.


  • Create a swift package with your scripts models and logic, importing SwishKit.
  • use sh to create a command and execute it immediately. (e.g. let gitRoot: Path = try sh("git rev-parse --show-toplevel"))
  • use cmd to create a command, and save it for later. This can be useful when setting up longer scripts. For example:
let name = "Example"
let checkIfExampleIsRunning = cmd("docker ps -aq -f name=\(name) --format \"{{ .Names }}\"", .equalsTrimming(name))`

You can then compose these building blocks into your own script. For example, we can compose a function to check if a docker container named name is currently running:

func containerIsRunning(name: String) -> ConcreteBooleanShellQuery {
  cmd("docker ps -aq -f name=\(name) --format \"{{ .Names }}\"", .equalsTrimming(name))

and it would be used like this: let isExampleRunning: Bool = containerIsRunning(name: "example").execute()

Or we can subclass ConcreteBooleanShellQuery to achieve the same thing:

class ContainerIsRunning: ConcreteBooleanShellQuery {
  let name: String
  init(name: String) {
    self.name = name
    let text = "docker ps -aq -f name=\(name) --format \"{{ .Names }}\""
    super.init(text, .equalsTrimming(name))

which would be used like let isExampleRunning: Bool = ContainerIsRunning(name: "Example").execute()

  • Finally, create a Swish file, in either JSON, YAML or Swift format (swift format not yet implmented). This should feel familiar to users of Yarn.
  startContainers: swift run --package-path swift-scripts startContainers
  • And run that script like this:
$ swish startContainers

Callable as function

All Commands and Querys implement Swift's callAsFunction(), which simply delegates to execute. This can aid slightly in readability. Consider

let startExampleContainer = cmd("docker run -d -p 8080:80 --name Example example/example:1.0")
let checkIfExampleIsRunning = containerIsRunning(name: "Example")
// ...
if !checkIfExampleIsRunning() {


There are two things you typically do with a shell. Run Commands and ask Querys. An example of a Command might be xcodebuild ..., where we aren't necessarily concerned about the output, just if it passes or fails. An example of a Query might be git status --porcelain, where we are going to parse the output and do something with it. A second kind of query might be a BooleanShellQuery where we translate the output of git status --porcelain into a boolean value: whether or not the git working copy is clean, in this example.

The sh and cmd helpers reference the SwishContext.default global object, which provides context to the shell invocations, such as where to write the logs.

Not everything you want to needs to be done as a shell command. For example, you may want to use Alamofire instead of shelling out to curl. Just make your type adhere to Command or Query if needed.


A Docker snippet to confirm that it works on Linux.

docker run --rm -it  -v `pwd`:`pwd` -w `pwd` --platform linux/amd64 swift:5.4.1-focal
apt-get update && apt-get install -y imagemagick


Stars: 4
Last commit: 2 weeks ago

Ad: Job Offers

iOS Software Engineer @ Perry Street Software
Perry Street Software is Jack’d and SCRUFF. We are two of the world’s largest gay, bi, trans and queer social dating apps on iOS and Android. Our brands reach more than 20 million members worldwide so members can connect, meet and express themselves on a platform that prioritizes privacy and security. We invest heavily into SwiftUI and using Swift Packages to modularize the codebase.

Submit a free job ad (while I'm testing this). The analytics numbers for this website are here.

Swiftpack is being maintained by Petr Pavlik | @ptrpavlik | @swiftpackco | API | Analytics