Our team has built a small CLI used for maintenance. The package.json specifies a path for with the bin property, and everything works great; "bin": { "eddy": "./dist/src/cli/entry.js"}
Autocompletion is achived by using yargs#17.0.1. However we recently converted the project to use es6 modules, because of a migration to Sveltekit, i.e. the package.json now contains type: module. Because of this, the CLI now only works if we run with:
what works
node --experimental-specifier-resolution=node ./dist/src/cli/entry.js help
However, if we run this without the flag, we get an error "module not found":
Error [ERR_MODULE_NOT_FOUND]: Cannot find module...
So the question is
Can we somehow "always" add the experimental-specifier-resolution=node to the CLI - so we can continue to use the shorthand eddy, and utilize auto completion?
There are two probable solutions here.
Solution 1
Your entry.js file should start with a shebang like #!/usr/bin/env node. You cannot specify the flag directly here, however, if you could provide the absolute path to node directly in the shebang, you can specify the flag.
Assuming you have node installed in /usr/bin/node, you can write the shebang in entry.js like:
#!/usr/bin/node --experimental-specifier-resolution=node
(Use which node to find the absolute path)
However, this is not a very portable solution. You cannot always assume everyone has node installed in the same path. Also some may use nvm to manage versions and can have multiple version in different path. This is the reason why we use /usr/bin/env to find the required node installation in the first place. This leads to the second solution.
Solution 2
You can create a shell script that would intern call the cli entry point with the required flags. This shell script can be specified in the package.json bin section.
The shell script (entry.sh) should look like:
#!/usr/bin/env bash
/usr/bin/env node --experimental-specifier-resolution=node ./entry.js "$#"
Then, in your package.json, replace bin with:
"bin": { "eddy": "./dist/src/cli/entry.sh"}
So when you run eddy, it will run the entry.js using node with the required flag. The "$#" in the command will be replaced by any arguments that you pass to eddy.
So eddy help will translate to /usr/bin/env node --experimental-specifier-resolution=node ./entry.js help
Just add a script to your package.json:Assuming index.js is your entry point and package.json is in the same directory
{
"scripts": {
"start": "node --experimental-specifier-resolution=node index.js"
}
}
Then you can just run on your console:
npm start
By default ESLint will print it's results to standard output. If you add the output option, it will redirect the output to a file. So far so good, but is there a way for it to do both?
We need the file output for GitLab to parse the results and display them in the UI, but some of our developers can't be bothered to change the way they do things and want to look at the output instead.
Is there an out-of-the-box way to get both or is my only chance to write my own script for running ESLint using the CLIEngine Node stuff mentioned in their documentation?
Thanks in advance.
So, after some research I think I found the answer myself.
There's basically 2 ways to have output on both console and file at the same time:
The easy way is by using this JavaScript package called ESLint Output.
The more complicated way is basically building said package yourself. You have to create a JavaScript file which you will run instead of ESLint and import/require ESLint in that file. By importing ESLint as a package instead of running it directly you can then run ESLint using Node's CLIEngine on your files to be linted and store the output in a variable. That output can then be saved to a file and printed again.
Hopefully, to no one's surprise, the methodology described in Option 2 is exactly what the package in Option 1 does: wrapping option 2 in an easy to understand config file and gathering all the configuration from the default eslintrc file for you.
Just write ESLint result into file and then read it by using one of ESLint-formatters:
npx eslint ./cartridges/ --format json --output-file result.json
node -p "require('./node_modules/eslint/lib/cli-engine/formatters/unix.js')(require('./result.json'))"
You can achieve it with eslint-html-reporter. When you'd like to have results from Terminal and export this into an .html file then the output from Terminal becomes a bit less pretty (which is not the case if it's done only in one of those), but it's doable by running:
./node_modules/.bin/eslint --ext .jsx . & ./node_modules/.bin/eslint --ext .jsx . -f node_modules/eslint-html-reporter/reporter.js -o report.html (in this example it looks recursively for .jsx file in your project's folder).
I had the same issue, and worked around by running eslint twice.
Created a shell script with
#!/usr/bin/env bash
node_modules/.bin/eslint --quiet \
--max-warnings 100 \
-o eslint-output.$1 \
-f $1 '*/**/*.{js,ts,tsx}'
Then call the script twice thru Jenkins:
steps {
echo 'Rodando ESLint...'
sh("""
yarn install
./eslint.sh junit || true
./eslint.sh codeframe || true
""")
junit(
allowEmptyResults: true,
healthScaleFactor: 0.0,
testResults: 'eslint-output.junit'
)
}
To generate the file for report:
eslint src/ --format json --output-file eslintreport.json
I just installed ESLint and I can successfully run it by doing this at the terminal window:
./node_modules/.bin/eslint app
Note: app is the root folder I want lint to inspect.
I then put that exact command in my package.json:
"scripts": {
"lint": "./node_modules/.bin/eslint app"
}
I expected to be able to run it in the terminal window like this:
npm run lint
Note: I know how to fix the no-undef error. My question is about the many errors lines after that.
It actually works, but it also produces a bunch of errors after showing me the correct output:
Why is that happening?
This is the default way of how the npm script runner handles script errors (i.e. non-zero exit codes). This will always happen, even if you only run a script like exit 1.
I'm not sure why this feature exists, it seems annoying and useless in most cases.
If you don't want to see this, you can add || true at the end of your script.
Example:
lint: "eslint app || true"
As you might've noticed, I've omitted the part to the eslint binary. The npm script runner already includes local binaries as part of the path when trying to run the script so there is no need to use the full path.
document is a global, so eslint thinks you are missing an import somewhere. For those cases, you can adapt your config so that the error is not reported, Something like this:
module.exports = {
"globals": {
"document": true
}
}
this should be saved as .eslintrc.js and be at the same level where your package.json is
I have a package of custom rules which I want to pull in (eslint-config-common), then use an .eslintrc file to override some of them.
extends: common
rules:
no-invalid-this: 0 # override a rule in common
If I run it directly, it all works as expected:
./node_modules/.bin/eslint src/**/*.js
Though if I run it as an .sh file or through an NPM script like lint: eslint src/**/*.js, it doesn't pull in the extends rules. It only runs using the rules found directly in .eslintrc. In my case, that's really bad since my .eslintrc is generally just turning off or down rules I don't want to use.
I've run it with DEBUG:eslint:* and it finds and loads the proper extends file, it just doesn't seem to apply the rules.
I found a bug similar to this, which they seem to say is fixed: https://github.com/eslint/eslint/issues/2754
This bug seems similar, so I'm not sure if I'm doing something wrong, or if there is still a bug.
I'm using the latest version of eslint 3.17.1
It turned out I was using a glob without wrapping it in quotation marks. It seems like NPM chewed on it a bit and it seems like it only went down as deep as one directory.
Instead of this:
"script": {
"lint": "eslint path/to/code/**/*.js"
}
Use this:
"script": {
"lint": "eslint 'path/to/code/**/*.js'"
}
I'm having a surprisingly hard time finding an answer to this. With plain Node.JS, you can run any js file with node path/to/file.js, with CoffeeScript it's coffee hello.coffee and ES6 has babel-node hello.js. How do I do the same with Typescript?
My project has a tsconfig.json which is used by Webpack/ts-loader to build a nice little bundle for the browser. I have a need for a build step run from the console before that, though, that would use some of the .ts files used in the project to generate a schema, but I can't seem to be able to run a single Typescript file without compiling the whole project.
How do I do the same with Typescript
You can leave tsc running in watch mode using tsc -w -p . and it will generate .js files for you in a live fashion, so you can run node foo.js like normal
TS Node
There is ts-node : https://github.com/TypeStrong/ts-node that will compile the code on the fly and run it through node 🌹
npx ts-node src/foo.ts
Run the below commands and install the required packages globally:
npm install -g ts-node typescript '#types/node'
Now run the following command to execute a typescript file:
ts-node typescript-file.ts
We have following steps:
First you need to install typescript
npm install -g typescript
Create one file helloworld.ts
function hello(person){
return "Hello, " + person;
}
let user = "Aamod Tiwari";
const result = hello(user);
console.log("Result", result)
Open command prompt and type the following command
tsc helloworld.ts
Again run the command
node helloworld.js
Result will display on console
To add to #Aamod answer above, If you want to use one command line to compile and run your code, you can use the following:
Windows:
tsc main.ts | node main.js
Linux / macOS:
tsc main.ts && node main.js
Edit: May 2022
ts-node now has an --esm flag use it.
Old Answer:
None of the other answers discuss how to run a TypeScript script that uses modules, and especially modern ES Modules.
First off, ts-node doesn't work in that scenario, as of March 2020. So we'll settle for tsc followed by node.
Second, TypeScript still can't output .mjs files. So we'll settle for .js files and "type": "module" in package.json.
Third, you want clean import lines, without specifying the .js extension (which would be confusing in .ts files):
import { Lib } from './Lib';
Well, that's non-trivial. Node requires specifying extensions on imports, unless you use the experimental-specifier-resolution=node flag. In this case, it would enable Node to look for Lib.js or Lib/index.js when you only specify ./Lib on the import line.
Fourth, there's still a snag: if you have a different main filename than index.js in your package, Node won't find it.
Transpiling makes things a lot messier than running vanilla Node.
Here's a sample repo with a modern TypeScript project structure, generating ES Module code.
I created #digitak/esrun, a thin wrapper around esbuild and that executes a TypeScript file almost instantly. esrun was made because I was disappointed with ts-node: too slow, and just didn't work most of the time.
Advantages of esrun over ts-node include:
very fast (uses esbuild),
can import ESM as well as CJS (just use the libraries of your choice and esrun will work out of the box),
there is an included watch mode, run your script with the --watch option and any change to your entry file or any of its dependencies will re-trigger the result
you can use esrun in inspect mode to use the DevTools console instead of your terminal console.
After installing, just run:
npx #digitak/esrun file.ts
Just helpful information - here is newest TypeScript / JavaScript runtime Deno.
It was created by the creator of node Ryan Dahl, based on what he would do differently if he could start fresh.
You can also try tsx.
tsx is a CLI command (alternative to node) for seamlessly running TypeScript, its build upon esbuild so its very fast.
https://github.com/esbuild-kit/tsx
Example:
npx tsx ./script.ts
As of May 2022 ts-node does support es modules
npx ts-node --esm file.ts
you will likely need to add "type": "module", to your package.json. And some of the imports might be wonky unless you turn on experimental-specifier-resolution=node
npmjs.com/package/ts-node#commonjs-vs-native-ecmascript-modules
For linux / mac you can add the ts-node-script shebang.
Install typescript / ts-node globally (see 1 below for non global install):
npm install ts-node typescript --save-dev --global
Add this as the first line in your .ts file:
#!/usr/bin/env ts-node-script
Then make the file executable:
$ chmod +x ./your-file.ts
You can then run the file directly from the command line:
$ ./your-file.ts
Notes:
1 For non global install you can install local to your project
npm install ts-node typescript --save-dev
and add the relative path to the shebang script eg:
#!/usr/bin/env ./node_modules/.bin/ts-node-script
2 Support for shebangs was officially added in ts-node v8.9.0.
Like Zeeshan Ahmad's answer, I also think ts-node is the way to go. I would also add a shebang and make it executable, so you can just run it directly.
Install typescript and ts-node globally:
npm install -g ts-node typescript
or
yarn global add ts-node typescript
Create a file hello with this content:
#!/usr/bin/env ts-node-script
import * as os from 'os'
function hello(name: string) {
return 'Hello, ' + name
}
const user = os.userInfo().username
console.log(`Result: ${hello(user)}`)
As you can see, line one has the shebang for ts-node
Run directly by just executing the file
$ ./hello
Result: Hello, root
Some notes:
This does not seem to work with ES modules, as Dan Dascalescu has pointed out.
See this issue discussing the best way to make a command line script with package linking, provided by Kaspar Etter. I have improved the shebang accordingly
Update 2020-04-06: Some changes after great input in the comments: Update shebang to use ts-node-script instead of ts-node, link to issues in ts-node.
Write yourself a simple bash wrapper may helps.
#!/bin/bash
npx tsc $1 && node ${1%%.ts}
For environments such as Webstorm where the node command cannot be changed to ts-node or npx:
npm install ts-node typescript (Install dependencies)
node --require ts-node/register src/foo.ts (Add --require ts-node/register to "Node parameters")
This answer may be premature, but deno supports running both TS and JS out of the box.
Based on your development environment, moving to Deno (and learning about it) might be too much, but hopefully this answer helps someone in the future.
Just in case anyone is insane like me and wants to just run typescript script as though it was a .js script, you can try this. I've written a hacky script that appears to execute the .ts script using node.
#!/usr/bin/env bash
NODEPATH="$HOME/.nvm/versions/node/v8.11.3/bin" # set path to your node/tsc
export TSC="$NODEPATH/tsc"
export NODE="$NODEPATH/node"
TSCFILE=$1 # only parameter is the name of the ts file you created.
function show_usage() {
echo "ts2node [ts file]"
exit 0
}
if [ "$TSCFILE" == "" ]
then
show_usage;
fi
JSFILE="$(echo $TSCFILE|cut -d"." -f 1).js"
$TSC $TSCFILE && $NODE $JSFILE
You can do this or write your own but essentially, it creates the .js file and then uses node to run it like so:
# tsrun myscript.ts
Simple. Just make sure your script only has one "." else you'll need to change your JSFILE in a different way than what I've shown.
Install ts-node node module globally.
Create node runtime configuration (for IDE) or use node in command line to run below file js file (The path is for windows, but you can do it for linux as well)
~\AppData\Roaming\npm\node_modules\ts-node\dist\bin.js
Give your ts file path as a command line argument.
Run Or Debug as you like.
Create your TypeScript file (ex. app.ts)
npm i -D typescript ts-node -> to install the dev dependencies local
npx nodemon app.ts
Using nodemon, automatically recompile app.ts every time you change the file
Here is the command
tsc index.ts --outDir .temp && node .temp/index.js && rm -rf .temp
<<<<<<<<< Compile >>>>>>>>> <<<<<<< Run >>>>>>> << Clean >>
There is also an option to run code directly from the CLI, not the *.ts file itself.
It's perfectly described in the ts-node manual.
As a first step, install ts-node globally via npm, yarn, or whatever you like.
...and now just use ts-node -e 'console.log("Hello, world!")' (you may also add the -p flag for printing code)
This little command is perfect for checking, does everything installed fine. And for finding some other error, relevant with tsconfig.json options.
We can run it using nodemon as well
nodemon ./filepath/filename.ts
This question was posted in 2015. In 2018, node recognizes both .js and .ts. So, running node file.ts will also run.