Cacher Run Server

The Cacher Run Server is a standalone Socket.io-based server that allows clients to run shell commands via Websocket messages. It is used primarily to run code snippets on the user's local machine via Cacher.

Demo of a Cacher snippet file executing on a Run Server:

Starting the Server
- Built-in Mode
- Standalone Mode
Editing the Configuration
Contributing
- Development
- Pull Requests
Libraries Used
Author / License

Starting the Server

Built-in Mode

The Cacher Desktop Client comes with the Run Server built-in. Open the Run Server dialog and click Start Server:

Standalone Mode

Start the Run Server in standalone mode if you require more control over its configuration.

Using the Cacher CLI

Start server locally

npm i -g @cacherapp/cli
cacher run-server:start

Starting the server on a remote machine

Running a remote server ensures all your developers are able to run snippet file commands against the same environment.

Note: Since the commands will be run using the shell account from which you launch the CLI, we recommend you use only machines which are for testing or are ephemeral (i.e. Docker instances).

Example of launching with a secure tunnel (ngrok).

npm i -g @cacherapp/cli
cacher run-server:start -p 39135 -t 4D5dzRGliafhGg~btNlR9 -o file:// -v
ngrok http 39135

Start via custom script

Install run-server package for your Node project.

npm i --save @cacherapp/run-server

Examples of starting the Run Server:

// examples.js
const RunServer = require('@cacherapp/run-server').RunServer;

// Let server pick random port and token
(new RunServer(
  { origin: 'https://app.cacher.io' }
)).start();

// Specify port and token, log verbose output to file
(new RunServer(
  {
    origin: 'file://',
    port: 43193,
    token: 'secret_token',
    verbose: true,
    logToFile: true
  }
)).start();

See server.model.ts for details on options.

Connecting to the standalone server

You'll need either the Cacher Web App or the Desktop Client in order to connect to a standalone Run Server.

Once you've logged in, open the Run Server dialog. Input your server's port and token and press Connect.

Editing the Configuration

The Run Server uses two Javascript-based config files to define the set of rules which govern how different file extensions are mapped to shell commands:

config.default.js - The default set of file types -> shell command mappings. While this config is copied to the Cacher directory as ~/.cacher/run-server.config.js, you should never edit this file. If you would like to add a file handler to this config (and let everyone use it), please submit a pull request instead. See the Contributing section below.
~/.cacher/run-server.user-config.js - Edit this file to add or overwrite file extension handlers on your local machine. Here are some examples of functions you might add to handle new file types:

Add command for .awesome file extension.

// ~/.cacher/run-server.user-config.js

return {
  rules: [
    {
      pattern: '\.awesome$',
      run: (command, filepath) => `awesome-compiler "${filepath}"`
    },
    ...
  ]
}

Compile .md (Markdown) file into HTML, then display in default browser.

// ~/.cacher/run-server.user-config.js

// Requires `npm -g markdown` to be run first.
return {
  rules: [
    {
      pattern: '\.md$',
      run: (command, filepath, args) => {
        const outputHtmlFile = `${args.runDir}/${args.baseFilename}.html`;
        childProcess.execSync(`md2html ${filepath} > ${outputHtmlFile}`);
        opn(outputHtmlFile);

        // Must return a shell command
        return `echo "Generated '${outputHtmlFile}'" and opened in default browser`;
      }
    },
    ...
  ]
}

Match Javascript files by .js extension and file content. Calls nvm use [version] before execution.

// ~/.cacher/run-server.user-config.js

return {
  rules: [
    {
      pattern: (command) => {
        return /\.js$/.test(command.file.filename)
          && command.file.content.indexOf('//nvm:') >= 0;
      },
      run: (command, filepath) => {
        let nvmVersion = command.file.content.match(/(\/\/nvm: v(.+))/)[0]
          .replace('//nvm: ', '');

        return `export NVM_DIR=~/.nvm ` +
          `&& source ~/.nvm/nvm.sh` +
          `&& nvm use ${nvmVersion} ` +
          `&& node "${filepath}"`;
      }
    },
    ...
  ]
}

Rule Properties

pattern: (String|Function)

Indicates whether to execute the run: callback for the given command.

Pass it either:

Regex string that matches against the filename.
Function that takes a command object and returns a boolean to indicate whether the rule should match.

Examples:

// Match files which end in .rb
pattern: '\.rb$'

// Match command which has "foobar" in the file's content
pattern: (command) => command.file.content.indexOf('foobar') >= 0

run: (Function)

Takes as input command, filepath and args and returns a string to be executed on the user's shell. Rules matched against interpreted (non-compiled) languages generally return [interpreter] "${filepath}". Rules matched with compiled languages will compile the source file first before executing the binary.

Examples:

// Run .dart file with interpreter
run: (command, filepath) => `dart "${filepath}"`

// Compile C++ file and run binary
run: (command, filepath, args) => {
  const outputFile = `${args.runDir}/${args.baseFilename}.out`;
  return `g++ "${filepath}" -o "${outputFile}" && cd "${args.runDir}" && ./"${args.baseFilename}.out"`;
}

timestamp: (Boolean)

Set to true to append an epoch timestamp to the end of the filename. This helps to prevent files from being overwritten in the ~/.cacher/run folder. Set to false for languages which require a strict match between the filename and the compiled binary. (i.e. Java, Haxe)

Arguments

command (Object)

The object that is sent to the pattern: and run: callback functions.

Properties:

channel - WebSocket channel for the command
file.filename - Filename with extension (i.e. example_code.cs)
file.filetype - Type of the file (i.e. markdown, lisp)
file.content - The file's content which will be executed

filepath (String)

Passed to the run: callback. The full path for the file to be run (in the ~/.cacher/run folder).

args (Object)

Additional arguments sent to the run: callback function.

Properties:

runDir - The fully-resolved path to ~/.cacher/run.
baseFilename - The filename without file extension. This is sometimes necessary to execute the compiled binary. (i.e. java "${args.baseFilename}")

Testing your changes

With built-in server on Cacher's Desktop Client

Stop the built-in server (if running).

Make your changes to ~/.cacher/run-server.config.js.
Start the built-in server again.

Debugging

If you run into issues with the server not starting or if your file handler does not perform as expected, debug the problem via the log file: ~/.cacher/logs/run-server.log.

With standalone server

Install the Cacher CLI globally:

npm i -g @cacherapp/cli

Make your changes to ~/.cacher/run-server.config.js.
Start the server:

cacher run-server:start --verbose

Make note of the server port and token, then connect with the Cacher client.

Contributing

Development

Grab the source code for this repo:

git clone git@github.com:CacherApp/cacher-run-server.git

Install Typescript globally:

npm i -g typescript

Install packages and start the development server:

npm i
npm start https://app.cacher.io

Pull Requests

We are happy to review any code changes, especially for new file rules added to config.default.js. Please adhere to contribution guidelines while drafting your pull request.

Libraries Used

socket.io - Websocket server implementation
nanoid - Unique string ID generator
winston - Logger for everything
shelljs - Portable Unix shell commands
opn - Better node-open
chalk - Terminal string styling
get-port - Get available TCP port

Author / License

Released under the MIT License by Rui Jiang of Cacher.

CacherApp/cacher-run-server