Skip to content

Latest commit

 

History

History
192 lines (143 loc) · 6.9 KB

new-run-all.md

File metadata and controls

192 lines (143 loc) · 6.9 KB
index new-run-all run-s run-p Node API

new-run-all command

Usage:
    $ new-run-all [--help | -h | --version | -v]
    $ new-run-all [tasks] [OPTIONS]

    Run given npm-scripts in parallel or sequential.

    <tasks> : A list of npm-scripts' names and Glob-like patterns.

Options:
    --aggregate-output   - - - Avoid interleaving output by delaying printing of
                               each command's output until it has finished.
    -c, --continue-on-error  - Set the flag to continue executing
                               other/subsequent tasks even if a task threw an
                               error. 'new-run-all' itself will exit with
                               non-zero code if one or more tasks threw error(s)
    --max-parallel <number>  - Set the maximum number of parallelism. Default is
                               unlimited.
    --npm-path <string>  - - - Set the path to npm. Default is the value of
                               environment variable npm_execpath.
                               If the variable is not defined, then it's "npm."
                               In this case, the "npm" command must be found in
                               environment variable PATH.
    -l, --print-label  - - - - Set the flag to print the task name as a prefix
                               on each line of output. Tools in tasks may stop
                               coloring their output if this option was given.
    -n, --print-name   - - - - Set the flag to print the task name before
                               running each task.
    -p, --parallel <tasks>   - Run a group of tasks in parallel.
                               e.g. 'new-run-all -p foo bar' is similar to
                                    'npm run foo & npm run bar'.
    -r, --race   - - - - - - - Set the flag to kill all tasks when a task
                               finished with zero. This option is valid only
                               with 'parallel' option.
    -s, --sequential <tasks> - Run a group of tasks sequentially.
        --serial <tasks>       e.g. 'new-run-all -s foo bar' is similar to
                                    'npm run foo && npm run bar'.
                               '--serial' is a synonym of '--sequential'.
    --silent   - - - - - - - - Set 'silent' to the log level of npm.

Examples:
    $ new-run-all --serial clean lint build:**
    $ new-run-all --parallel watch:**
    $ new-run-all clean lint --parallel "build:** -- --watch"
    $ new-run-all -l -p start-server start-browser start-electron

npm-scripts

It's "scripts" field of package.json. For example:

{
    "scripts": {
        "clean": "rimraf dist",
        "lint":  "eslint src",
        "build": "babel src -o lib"
    }
}

We can run a script with npm run command. On the other hand, this new-run-all command runs multiple scripts in parallel or sequential.

Run scripts sequentially

$ new-run-all clean lint build

This is same as npm run clean && npm run lint && npm run build.

Note: If a script exited with non zero code, the following scripts are not run. If --continue-on-error option is given, this behavior will be disabled.

Run scripts in parallel

$ new-run-all --parallel lint build

This is similar to npm run lint & npm run build.

Note1: If a script exited with a non-zero code, the other scripts and those descendant processes are killed with SIGTERM (On Windows, with taskkill.exe /F /T). If --continue-on-error option is given, this behavior will be disabled.

Note2: & operator does not work on Windows' cmd.exe. But new-run-all --parallel works fine there.

Run a mix of sequential and parallel scripts

$ new-run-all clean lint --parallel watch:html watch:js
  1. First, this runs clean and lint sequentially / serially.
  2. Next, runs watch:html and watch:js in parallel.
$ new-run-all a b --parallel c d --sequential e f --parallel g h i

or

$ new-run-all a b --parallel c d --serial e f --parallel g h i
  1. First, runs a and b sequentially / serially.
  2. Second, runs c and d in parallel.
  3. Third, runs e and f sequentially / serially.
  4. Lastly, runs g, h, and i in parallel.

Glob-like pattern matching for script names

We can use glob-like patterns to specify npm-scripts. The difference is one -- the separator is : instead of /.

$ new-run-all --parallel watch:*

In this case, runs sub scripts of watch. For example: watch:html, watch:js. But, doesn't run sub-sub scripts. For example: watch:js:index.

$ new-run-all --parallel watch:**

If we use a globstar **, runs both sub scripts and sub-sub scripts.

new-run-all reads the actual npm-script list from package.json in the current directory, then filters the scripts by glob-like patterns, then runs those.

Run with arguments

We can enclose a script name or a pattern in quotes to use arguments. The following 2 commands are similar.

$ new-run-all --parallel "build:* -- --watch"
$ npm run build:aaa -- --watch & npm run build:bbb -- --watch

When we use a pattern, arguments are forwarded to every matched script.

Argument placeholders

We can use placeholders to give the arguments preceded by -- to scripts.

$ new-run-all build "start-server -- --port {1}" -- 8080

This is useful to pass through arguments from npm run command.

{
    "scripts": {
        "start": "new-run-all build \"start-server -- --port {1}\" --"
    }
}
$ npm run start 8080

> example@0.0.0 start /path/to/package.json
> new-run-all build "start-server -- --port {1}" -- "8080"

There are the following placeholders:

  • {1}, {2}, ... -- An argument. {1} is the 1st argument. {2} is the 2nd.
  • {@} -- All arguments.
  • {*} -- All arguments as combined.

Those are similar to Shell Parameters. But please note arguments are enclosed by double quotes automatically (similar to npm).

Known Limitations

  • If --print-label option is given, some tools in scripts might stop coloring their output. Because some coloring library (e.g. chalk) will stop coloring if process.stdout is not a TTY. new-run-all changes the process.stdout of child processes to a pipe in order to add labels to the head of each line if --print-label option is given.
    For example, eslint stops coloring under new-run-all --print-label. But eslint has --color option to force coloring, we can use it. For anything chalk based you can set the environment variable FORCE_COLOR=1 to produce colored output anyway.