dogma-io / nodely Goto Github PK

View Code? Open in Web Editor NEW

0.0 2.0 2.0 1.14 MB

Write Node code using whatever syntax you want.

License: MIT License

JavaScript 100.00%

babel node nodejs es6 es2015 transpile transpiler multiprocessing workers master

nodely's Introduction

nodely

Write Node code using whatever syntax you want.

Installation
Documentation
Code of Conduct
Contributing
License

Installation

npm

npm install nodely

yarn

yarn add nodely

Documentation

The most basic usage is achieved with the following command, replacing the source path with the path to your source code to transform and the output path with the path to where you want the transformed code to be written.

nodely -s ./path/to/source -o ./path/to/output

NOTE: If you have a .babelrc.js or .babelrc.json configuration file for Babel, nodely will use that instead of it's own built-in configuration. It's built-in configuration supports ES6, Flow types, and React.

Options

Include

By default Nodely will copy/transform all files from the source directory into the output directory. However if you want to limit which files are copied/transformed you can provide a regular expression to match included files against.

nodely -s ./path/to/source -o ./path/to/output -i "\.js$"

Note: In the above example only Javascript files will be transformed into the output directory.

Target

By default Nodely outputs code in ES5 which should work for both Node and browsers (as long as you aren't using target specific API's such as fs). If you'd like your build to target a specific Node version you can set the target like so:

nodely -s ./path/to/source -o ./path/to/output -t 8

Note: This option will have no effect if you have your own Babel configuration defined, this only applies when you are relying on nodely's built-in configuration.

Verbose

By default nodely will inform you when files fail to transform but won't provide much detail. If you want entire stack-traces when the failures occur you can use the -v flag like so:

nodely -s ./path/to/source -o ./path/to/output -v

Watcher

If you want to leave the nodely server running and have it transform modified files on the fly then you simply need to add the -w flag like so:

nodely -s ./path/to/source -o ./path/to/output -w

Workers

By default the server will spawn a worker process for all but one CPU, reserving the last CPU for the master process. If you want to spawn less workers you can use the -n flag like so:

nodely -s ./path/to/source -o ./path/to/output -n 3 # This will spawn 3 workers

Code of Conduct

Please see the code of conduct.

Contributing

Please see the contributing guide.

License

MIT

nodely's People

Contributors

Watchers

Forkers

sandersky job13er

nodely's Issues

Switch from babel-core to @babel/core in source code

Include Flow types in package

Look into flow gen-flow-files.

Add option to not use clustering

Add ability to set workers to zero which will just execute everything in a single process.

Add CLA

See React CLA as an example.

Generated flow types breaking flow checking in downstream repo

Now that this project is generating Flow type files at lib/*.flow, they are breaking Flow type checking in downstream repos with the following error:

Error: node_modules/nodely/lib/master.js.flow:21
 21: declare export default function(argv: $Exact<{output: string, source: string, target: string, verbose: boolean, watch: boolean, workerCount: number}>): {erred: boolean, isWatching: boolean, queue: Array<$Exact<{filePath: string, type: any | any}>>, workers: Array<$Exact<{idle: boolean, worker: Class0}>>};
                                    ^ Unexpected token (

Error: node_modules/nodely/lib/worker.js.flow:9
  9: declare export default function(argv: $Exact<{output: string, source: string, target: string, verbose: boolean, watch: boolean, workerCount: number}>): void;
                                    ^ Unexpected token (

See: facebook/flow#2830

Add more tests

Add more tests so we get to 100% code coverage.

Add Node 6 to Travis Config

Add code coverage

Enable Jest's code coverage features so we can ensure we are fully covering our source code with tests.

Handle memory better

In some CircleCI builds the build VM has a lot of CPU's but running a worker on every CPU can cause the VM to run out of memory, due to Babel doing a lot of processing (one Babel instance per worker process). See if we can evaluate heap usage and not process files on every worker when heap usage is high. This way we can spawn all of the workers but can auto-adjust our processing in real-time to avoid running out of memory.

Changelog not being updated on PR merge

Better handling of file permissions

When using nodely on a Node project, it'd be nice to maintain the same file permissions on the output files that are found on the source files. For example when writing a Node script where src/index.js is actually going to be run via something like node ./lib/index.js, we'd typically do something like chmod +x ./lib/index.js to make it executable. However seeing as the lib directory isn't being checked into repos using this project, every dev that wants to do a build has to manually run the chmod command. Instead the chmod command should be done on the source file, which is checked into the repo. Nodely should then respect these file permissions when writing the output files, making it so devs never have to run chmod after the initial source file has been chmod'd and checked in.

Use fb-watchman

Consider using fb-watchman as a replacement for glob and node-watch. Benchmark change to make sure it is as performant if not better. Even if the performance is the same the amount of code required in the master process should become less.

Wire up to CI

Get continuous integration setup to automatically bump version and append to changelog on pull requests merges. Also need to run tests before allowing pull requests to be merged and verify coverage doesn't drop.