Numbro 2.0.0 by BenjaminVanRyseghem · Pull Request #305 · BenjaminVanRyseghem/numbro

BenjaminVanRyseghem · 2017-08-12T20:11:11Z

a full rewrite of numbro

Todo:

New features:

the provided format can now be an object
defaults are now language specific and not global anymore
new options to language files:
- spaceSeparated: should a space be inserted between the number and an abbreviation/currency symbol
- currencyDefaults: defaults value for formatting currency
- byteDefaults: defaults value for formatting byte
- ordinalDefaults: defaults value for formatting ordinal
- percentDefaults: defaults value for formatting percent
- thousandsSize: should that be renamed characteristicGroupSize?
cleanup of the API (unformat is now a static method instead of an instance method)

BenjaminVanRyseghem · 2017-08-12T20:12:48Z

@NicolasPetton @herodrigues @gwynjudd @Ben305 it's not ready yet, but if you want to have a look 😄

ArmorDarks · 2017-08-12T21:00:03Z

Wow, quite cool!

Didn't look into code yet, though...

Just for the information, it might be worth to look at Standard as a replacement for pure ESLint, since it comes with reasonable styling and more powerful auto --fix , have much more built-in features than ESLint-only solution and remove the burden of maintaining its own ESLint config.

Another alternative is Prettier coupled with Prettier-based ESLint config.

Also, it is a matter of taste, but most recent Jest and it's watch mode, snapshot testing, built-in coverage are really cool. We've used it on our projects and never looked back.

BenjaminVanRyseghem · 2017-08-12T21:03:35Z

@ArmorDarks thanks for the input 😄

I know Standard and Prettier but I tend to find them too strict. I like that eslint can be really customized and extended.

I've never tried Jest, I will have a look 😄

ArmorDarks · 2017-08-12T21:05:17Z

-                verbose: true, // If you need output with rule names http://jscs.info/overview.html#verbose
-                validateIndentation: 4
-            }
+            all: ["tests-nodeunits/**/*.js"]


Grunt is cool, but you really do not need it for this task.

Instead, it is much easier to run nodeunit directly from scripts in package.json

"scripts": { "test": "nodeunit tests-nodeunits/**/*.js" }

and then run in console

npm test

grunt is to be removed, I kept it so far to run old tests 😄

it's already replaced with gulp

ArmorDarks · 2017-08-12T21:12:45Z

+        .pipe(plugins.eslint())
+        .pipe(plugins.eslint.format("unix"))
+        .pipe(plugins.eslint.failAfterError());
+});


Here it comes again. Gulp is overkill for that task.

"scripts": { "lint": "eslint \"src/**/*.js,tests/**/*.js,languages/**/*.js\"", "test": "lint && nodeunit tests-nodeunits/**/*.js" }

and then run in console

npm test

With Standard this might be even simpler:

"scripts": { "lint": "standard", "test": "lint && nodeunit tests-nodeunits/**/*.js" }

thanks 😄

I usually try to use gulp as much as I can, but I don't really know why 😄

There is really no point in doing that, since you basically keep adding additional abstraction layers for really simple things.

Do not mistake me. I have very large experience with Grunt (Kotsu) and it took us some time too to realize, that by doring everything exclusively through Grunt\Gulp we only make things harder, dealing with abstractions, and complicating things when it comes to CI integration.

Grunt and Gulp are cool, but they are for more difficult tasks, really.

ArmorDarks · 2017-08-12T21:15:28Z

+    return gulp.src(["./src/**/*.js", "./languages/**/*.js"])
+        .pipe(plugins.istanbul())
+        .pipe(plugins.istanbul.hookRequire());
+});


This one better to live in scripts too, Gulp here doesn't help at all

However, with Jest it would be even easier, since Istanbul comes as part of Jest:

"scripts": { "lint": "eslint \"src/**/*.js,tests/**/*.js,languages/**/*.js\"", "test": "lint && jest --coverage && nodeunit tests-nodeunits/**/*.js" }

ArmorDarks · 2017-08-12T21:17:39Z

+        }))
+        .pipe(plugins.istanbul.writeReports());
+    // .pipe(plugins.istanbul.enforceThresholds({thresholds: {global: 90}}));
+});


This one could be moved to scripts too, effectively reducing dependency on 3rd-party plugins and eleminating unnecessary abstraction.

ArmorDarks · 2017-08-12T21:18:28Z

+// Coverage
+
+gulp.task("basic-coverage", () => {
+    return gulp.src("").pipe(plugins.shell("istanbul cover ./node_modules/.bin/jasmine --captureExceptions")


Oh my. This one just cries to move directly into scripts, since you're basically hacking Gulp to make it work.

ahah 😄

I think all of that is mainly copy/pasted without too much thinking (also it was made in a train during a 7h trip 😄 )

ArmorDarks · 2017-08-12T21:20:02Z

+
+// Release
+
+gulp.task("release", () => {


You already have npm publish. Really, there is no need to replicate it functionality, since it does same thing.

I just took the grunt tasks and mimic them 😄

It looks like all of that really need some love 😉

I try to provide some ❤️ :)

herodrigues · 2017-08-12T21:24:59Z

@BenjaminVanRyseghem nice! I'll take a closer look tomorrow. I can help to finish those TODO tasks.

ArmorDarks · 2017-08-12T21:32:45Z

I know Standard and Prettier but I tend to find them too strict. I like that eslint can be really customized and extended.

Well, that's the point of such tools. It's just widely accepted by community standards, you just plug them and stop worrying about all those little things. But yeap, it takes some time to embrace them, and at first it was hard for us to submit to some of their choices too.

But after all, it makes all JS codebase much more uniform. Anyway, it is up to you. I'm just mentioning.

I've never tried Jest, I will have a look

Jest is basically Jasmin on steroids. At first it was completely Jasmin-based, but now they switched to their own engine due to performance limitations of Jasmine, but API mostly same with some cool additions, like snapshots testing. And you really should check their watch. It is really cool.

But in same time since they interchangeable, this isn't really critical choice. Jasmin is cool. Jest too, but tries to be cooler. And it is trendy. Choose whatever your heart desires.

ArmorDarks · 2017-08-12T21:35:00Z

Also, if CI build time worries you, I'd recommend to switch to CircleCI. I can help with configuration.

BenjaminVanRyseghem · 2017-08-12T21:37:38Z

But in same time since they interchangeable, this isn't really critical choice. Jasmin is cool. Jest too, but tries to be cooler. And it is trendy. Choice whatever your heart desires.

I had a quick look, it indeed looks close, but the mocking library is a bit different. And honestly I don't feel like rewriting all the tests again 😄

ArmorDarks · 2017-08-12T21:40:56Z

I had a quick look, it indeed looks close, but the mocking library is a bit different. And honestly I don't feel like rewriting all the tests again 😄

I don't think that most part of tests will require any rewriting, but I can fix those which will require it.

But yeap, mocking library is indeed different. And I don't know how much is it different, since I'm trying not to use mocking at all, thus never worked with Jest-one.

Well, if you'll decide that switching to Jest worth it, I can look into mocking part and decide, is it easily portable or not. Meanwhile, will learn Jest's mocking capabilities better...

ArmorDarks · 2017-08-13T09:31:59Z

+    forceSign: "boolean"
+};
+
+const validLanguage = {


Too late for me, but I will mention it anyway.

It would be more logical to use Flow for those checks, it will greatly reduce amount of code.

If you're not ok with Flow, there is also a great library tcomb. To be honest, I've been completely charmed with its elegancy and how refinements works.

Here is example of tcomb validation.

One of the side benefits is that tcomb can be used during runtime (on contrary to Flow), and in most desperate situations there is a babel plugin which transpires Flow to tcomb.

Anyway, it is probably too late, but I'd stand for already implemented validation solution, since it will be much easier to contribute to it for newcomers (because, well, of well-known technologies).

Flow and tcomb seems to be type checkers, where here we need to validate more that just the type, so I am not sure it will fit

But I will give a shot to Validate.js I think

Yeap, Flow is a type checker, but you can validate much more with it.

Tcomb is slightly more. It is type checker, but also a validator. With it refinements you can validate almost anything based on provided types, and API for refinements is very elegant. In some ways, it is much more flexible than Flow, but they are just for slightly different purposes, since Flow is static type checker, while tcomb is for runtime type checking, validation and introspection. I'd say, usually they live quite good together.

Here is example with some tcomb refinaments, which validates much more complex things. Just for getting impression how flexible it is.

I don't want to push on you any tools, for sure, sorry if it looks like this. Just sharing my experience — my life before starting using tcomb been definitely harder, since I wasted a lot of time for writing boilerplate just to validate what I needed.

~~Regarding Validate.js — it is quite cool, but for me it was too heavy, with too much things I'd never need.~~ Oops, I've actually mixed up Validate.js with Joi. Well, we discarded Validate.js just because writing precise refinements which suited our needs for tcomb was easier and felt more powerful, more flexible.

It might be also worth check this issue https://github.com/LotusTM/Kotsu/issues/165 where I conducted some research on possible validation solutions. Our internal discussions, unfortunately, didn't make it to that issue, but at least you can see list of most popular solutions.

I don't want to push on you any tools, for sure, sorry if it looks like this. Just sharing my experience — my life before starting using tcomb been definitely harder, since I wasted a lot of time for writing boilerplate just to validate what I needed.

Sorry if I gave you this impression... I am glad you're sharing your experience here 😄 I will definitely dig into this. Maybe we could create a ticket to discuss that further more?

Sure, just mention me if you'll come to that, I'll be glad to help.

herodrigues · 2017-08-19T15:15:31Z

@BenjaminVanRyseghem what can be done about this task rewrite all languages files?

BenjaminVanRyseghem · 2017-08-19T16:23:56Z

@herodrigues If you have a look at src/en-US.js you will see the format changed a bit.
We need to adjust all the other language files.

- replace grunt by gulp - replace jshint by eslint - use jasmine for tests - add integration tests support

It keeps backward compatibility as much as possible Breaking feature: - previously, not specifying a mantissa size meant to have none, now it means to keep it untouched: numbro(100.12).format("0")` // -> `"100.12"` New features: - the provided format can/should now be an object - defaults are now language specific and not global anymore - new options to language files: - spaceSeparated: should a space be inserted between the number and an abbreviation/currency symbol - currencyDefaults: defaults value for formatting currency - byteDefaults: defaults value for formatting byte - ordinalDefaults: defaults value for formatting ordinal - percentDefaults: defaults value for formatting percent - thousandsSize: how many digits per group in the characteristic - cleanup of the API (unformat is now a static method instead of an instance method)

Carniatto · 2020-02-10T13:00:50Z

@BenjaminVanRyseghem or others, is there any migration guide from v1 to v2?

herodrigues · 2020-02-10T13:12:32Z

@Carniatto this library is a bit outdated and it hasn't been maintained for a while.

If you just want to perform simple format/unformat operations, then you can achieve that by using plain JavaScript.

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Number/toLocaleString

BenjaminVanRyseghem · 2020-02-12T09:35:09Z

@Carniatto this library is a bit outdated and it hasn't been maintained for a while.

ouch

@BenjaminVanRyseghem or others, is there any migration guide from v1 to v2?

v2 should be fully backward compatible to v1, so no stress migrating.

If that's not the case, I'll be glad to help you.

herodrigues · 2020-02-12T10:51:59Z

@BenjaminVanRyseghem I didn't mean to be rude.

The last feature PR 014d56c is indeed a bit old in terms of JS programming.

Also, there are open issues from 2018.

ArmorDarks reviewed Aug 12, 2017

View reviewed changes

BenjaminVanRyseghem force-pushed the numbro2.0 branch 2 times, most recently from bb92d81 to 0df4e4e Compare August 12, 2017 23:41

ArmorDarks reviewed Aug 13, 2017

View reviewed changes

BenjaminVanRyseghem force-pushed the numbro2.0 branch 3 times, most recently from 872fd7c to 4cc780b Compare August 17, 2017 00:28

jrowen mentioned this pull request Sep 4, 2017

Number formatting jrowen/rhandsontable#88

Closed

BenjaminVanRyseghem added this to the 2.0 milestone Sep 18, 2017

BenjaminVanRyseghem mentioned this pull request Sep 18, 2017

Why I can't set percent format with percent sign at the beginning? #307

Closed

BenjaminVanRyseghem force-pushed the numbro2.0 branch 5 times, most recently from 8b4d17e to c2d6e32 Compare September 23, 2017 23:27

BenjaminVanRyseghem force-pushed the numbro2.0 branch from 801cb39 to 02cc991 Compare November 22, 2017 00:34

BenjaminVanRyseghem added 3 commits November 22, 2017 09:19

Rewrite numbro boilerplate

c319c20

- replace grunt by gulp - replace jshint by eslint - use jasmine for tests - add integration tests support

Add unit and integration tests

902acf3

BenjaminVanRyseghem force-pushed the numbro2.0 branch from 0850b51 to 902acf3 Compare November 22, 2017 08:27

BenjaminVanRyseghem merged commit b5c4b63 into develop Jan 15, 2018

BenjaminVanRyseghem deleted the numbro2.0 branch January 15, 2018 09:58


		// Release

		gulp.task("release", () => {

Uh oh!

Conversation

BenjaminVanRyseghem commented Aug 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

BenjaminVanRyseghem commented Aug 12, 2017

Uh oh!

ArmorDarks commented Aug 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

BenjaminVanRyseghem commented Aug 12, 2017

Uh oh!

ArmorDarks Aug 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ArmorDarks Aug 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ArmorDarks Aug 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

herodrigues commented Aug 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ArmorDarks commented Aug 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ArmorDarks commented Aug 12, 2017

Uh oh!

BenjaminVanRyseghem commented Aug 12, 2017

Uh oh!

ArmorDarks commented Aug 12, 2017

Uh oh!

ArmorDarks Aug 13, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ArmorDarks Aug 14, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ArmorDarks Aug 14, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

herodrigues commented Aug 19, 2017

Uh oh!

BenjaminVanRyseghem commented Aug 19, 2017

Uh oh!

Carniatto commented Feb 10, 2020

Uh oh!

BenjaminVanRyseghem commented Aug 12, 2017 •

edited

Loading

ArmorDarks commented Aug 12, 2017 •

edited

Loading

ArmorDarks Aug 12, 2017 •

edited

Loading

ArmorDarks Aug 12, 2017 •

edited

Loading

ArmorDarks Aug 12, 2017 •

edited

Loading

herodrigues commented Aug 12, 2017 •

edited

Loading

ArmorDarks commented Aug 12, 2017 •

edited

Loading

ArmorDarks Aug 13, 2017 •

edited

Loading

ArmorDarks Aug 14, 2017 •

edited

Loading

ArmorDarks Aug 14, 2017 •

edited

Loading