Implement missing transformers

[ForbesLindesay]

This project aims to replace transformers with a large collection of separate libraries, each of which able to be separately maintained, tested and updated. Each library will implement some subset of the API in this module, and may optionally return a string body rather than the {body: String, dependencies: Array.<String>} format (lots of file formats don't actually have dependencies so it wouldn't really make sense). This module can then be used to fully normalise that interface, including implementing the methods that are otherwise not implemented.

Here is a list of all the modules that need implementing

• [x] coffeecup
• [x] coffeekup
• [x] haml-coffee
• [x] jazz
• [ ] jqtpl
• [x] just
• [x] mote
• [x] qejs
• [x] templayed
• [x] walrus
• [x] whiskers
• [x] atpl
• [x] dot
• [x] dust
• [x] haml
• [x] handlebars
• [x] liquor -- done but unusable and not published because of chjj/liquor#5
• [x] markdown-js
• [x] toffee
• [x] cdata
• [x] cdata-css
• [x] cdata-js
• [x] coffee-script
• [x] cson
• [x] eco
• [x] ect
• [x] ejs
• [x] escape-html
• [x] highlight.js
• [x] hogan
• [x] jade
• [x] less
• [x] livescript
• [x] markdown
• [x] markdown-it
• [x] marked
• [x] mustache
• [x] plates
• [x] remarkable
• [x] sass
• [x] scss
• [x] stylus
• [x] supermarked
• [x] swig
• [x] uglify-css -> rename this to "clean-css" and use the clean-css module
• [x] uglify-js
• [x] uglify-json -> rename this to just "json"
• [x] underscore
• [x] verbatim

  Potential Extras

• [x] JsRender
• [x] Marko
• [ ] CSS Longhand
• [x] autoprefixer built with postcss
• [ ] coffee-stylesheets
• [ ] cssnext built with postcss
• [ ] doiuse built with postcss
• [ ] fest xml to html transformation
• [ ] gaikan
• [ ] hyperscript
• [ ] j140
• [x] stylecow
• [ ] String-Template
• [x] Duo.js
• [ ] Coffee React
• [x] octet
• [ ] renderkid
• [ ] rtlcss built with postcss
• [x] slm-lang jade-like template engine for node and browser
• [x] svgo Nodejs-based tool for optimizing SVG vector graphics files
• [ ] ccss bugs, need PRs there
• [ ] coffee-templates tricky shit, atm
• [x] absurdjs
• [x] babel
• [x] browserify
• [x] es6-template-strings
• [x] gray-matter
• [x] ini
• [x] jsx
• [x] lodash
• [x] mini-handlebars
• [x] myth
• [x] postcss
• [x] resin built with rework
• [x] rework
• [x] toml
• [x] webpack
• [x] xml2js
• [x] styl
• [x] jscss bugs, need PRs there

Once all of these have been implemented, we can switch jade to look for these rather than transformers. I'll need as many collaborators as possible for this, so if you're willing to contribute a transformer, let me know and I'll add you to the organisation. P.S. note that we add jstransformer to the keywords in package.json for all transformers, so they can be found at https://www.npmjs.com/browse/keyword/jstransformer

[tunnckoCore]

:sparkles: postcss :sparkles: transform CSS to JSON AST using postcss.parse()

[tunnckoCore]

:sparkles: myth :sparkles: done

• render
• renderFile

[TimothyGu]

@tunnckoCore, I am not exactly satisfied with the quality of the new modules.

1. jstransformer-css has a very confusing name. It is impossible to tell that it is a CSS parser just from its name. Even though css is the name of the module I think it would make more sense if a more descriptive name is used.
2. Stuff like this
3. Stuff like this
4. I also don't feel comfortable with https://github.com/jstransformers/jstransformer-postcss#author. That kind of destroys the unity of jstransfomers as an organization.

I would also suggest looking at the existing modules to get a feel on how everything works, to preserve some sense of unity among all the modules: https://github.com/jstransformers/jstransformer-jade/blob/master/package.json#L16-L20

I also do not think we should do things like:

matrix:
  allow_failures:
    - node_js: "0.10"

or

node_js:
  - 0.11

If we don't support 0.10 why test it at all? Plus nobody uses Node.js 0.11 (or at least nobody encourages using 0.11 now that 0.12 is out). Those are just plain useless.

Also, these files are also up for discussion:

• .editorconfig
• .gitattributes (which also contains stuff we would not EVER need like say rules for .mp3 files)
• .jshint*

[tunnckoCore]

1. its better to be than jstransformer-reworkcss, actually it is just transform from css to json, so dont see the problem
2. i cant understand whats the problem is. just describes that pkg is published as v0.1.1
3. always add it because I dont want in every change to go to the shitty coveralls.io website to regen it and get it and etc, and etc.

If we don't support 0.10 why test it at all? Plus nobody uses Node.js 0.11 (or at least nobody encourages using 0.11 now that 0.12 is out). Those are just plain useless.

templates, templates, templates. I must change them soon. Also they are the reason for the author and etc stuff. I just forgot that.

[TimothyGu]

1. its better to be than jstransformer-reworkcss

Have you considered jstransformer-css-parse?

3. always add it because I dont want in every change to go to the shitty coveralls.io website to regen it and get it and etc, and etc.

See how it is already done in other modules. You do NOT need to add the token to the Git repo, as coveralls now support Travis CI out-of-the-box. No tokens needed.

Posting the token unencrypted is a security hole, and can be exploited. See for example davidbau/seedrandom#17 and proof of concept.

templates, templates, templates. I must change them soon. Also they are the reason for the author and etc stuff. I just forgot that.

Yes I understand. I am planning on create a boilerplates repo under the org to make it easier to create new modules.

[tunnckoCore]

No tokens needed.

hm, thanks, good to know.

[ForbesLindesay]

As part of the www I was thinking of creating a not so we could have a button that anyone (within reason) could create a new repo with appropriate boilerplate set up and be given admin rights.

As for naming I'd much prefer css-parse but I can see why you went with css. The convention is to use the module name, but in this case that isn't very descriptive.

What I had in mind for rework was using rework and taking a list of plugins to do css to css transformation.

[ForbesLindesay]

For versioning, don't duplicate the version. You already put that in the heading of the changelig anyway. Look at github.com/jadejs/jade for an example of how the change log should look.

We may want to create tooling around change logs or an additional www page in the future so we should be careful to keep that consistent.

[tunnckoCore]

What I had in mind for rework was using rework and taking a list of plugins to do css to css transformation.

cool. and seems it is better that I didn't get the jstransformer-reworkcss name :)

[tunnckoCore]

habits, bite me, lol

It was better you start with these things before any else.

[tunnckoCore]

Can I use different promise library?

[ForbesLindesay]

It should just be jstransformer-rework as that's the module name. We don't generally append the output format to the name.

Also postcss should output the transformed css rather than an ast. We could have a separate postcss-parser transform.

[tunnckoCore]

You're added to the npm, guys. :)

btw, about

I'll write a bot at some point to add all of the owners on GitHub as owners on npm :)

I called it package-ownership https://github.com/tunnckoCore/blankr/issues/23

[TimothyGu]

@tunnckoCore

I called it package-ownership tunnckoCore/blankr#23

BTW, the bot itself literally takes only 6 lines of Bash code. Problem is we don't know where to store the owner information.

[tunnckoCore]

I try it, but not works for me on centos.

Problem is we don't know where to store the owner information.

meta repository as in regexps, datetime and there we store all discussions, questions and etc.

[hemanth]

A big :+1: for a meta repository, it's proven to be very useful multiple number of times.

[TimothyGu]

^^ +1 too. Created: https://github.com/jstransformers/meta

[TimothyGu]

@tunnckoCore said:

I try it, but not works for me on centos.

Yes, because that depends on the up-to-dateness of the ownership of the jstransformer package, which right now only has one owner.

[tunnckoCore]

toml and ini added to the list.

[tunnckoCore]

@ForbesLindesay feel free to change purpose of postcss and css transformers, your ideas seems better

[tunnckoCore]

:sparkles: toml :sparkles: done

[tunnckoCore]

https://github.com/Leonidas-from-XIV/node-xml2js - added to the list

[tunnckoCore]

:sparkles: toffee :sparkles: done :sparkles: xml2js :sparkles: done :sparkles: mini-handlebars :sparkles: done

:sparkles: j140 :sparkles: added, btw mega powerful for just 140bytes lib, but using with :sparkles: octet :sparkles: added

:sparkles: handlebars :sparkles: PR added more tests and render methods

[tunnckoCore]

:sparkles: gray-matter :sparkles: done

[tunnckoCore]

:sparkles: lodash :sparkles: added

[ForbesLindesay]

Don't forget to push your code before commenting here

[tunnckoCore]

@ForbesLindesay yea. With "added" I mean "added to the list" :)

btw, about template cruft, no problems I'll update all my libs in org as soon as I create basic project generator for jstransformers - workin on that now :) thought that isn't a problem? :)

[TimothyGu]

@tunnckoCore please add the new libraries to the "potential extras" instead of the main list. Also, you can try porting the ones from transformers first THEN adding extras.

[tunnckoCore]

please add the new libraries to the "potential extras"

@TimothyGu sorry, done and okey.

[stoeffel]

I created a transformer for jsx. If you like I can pass it over to you guys. https://github.com/stoeffel/jstransformer-jsx

[hemanth]

Nice @stoeffel.

Update: Sorted the list and looks like jqtpl is depreciated, so stroked it.

[tunnckoCore]

@jstransformers/owners btw, I reimplement the rework and postcss transforms to match @ForbesLindesay expectations and ideas - css to css transformation. I thought the following would be the best:

• jstransformer-rework - css to css transformation
• jstransformer-postcss - css to css transformation
• jstransformer-css - css to json ast, using reworkcss/css parser
• jstransformer-postcss-parser - css to json ast, using postcss

[tunnckoCore]

will commit them tomorrow :)

[ForbesLindesay]

@tunnckoCore That sounds ideal

[stoeffel]

I have moved jstransformer-jsx from stoeffel/jstransformer-jsx to jstransformers/jstransformer-jsx.

[tunnckoCore]

I'm back again :) All is done, going to normalize/update my other libs in org.

[RobLoach]

Thoughts on jstransformer-es6-templates? It's a bit hacky, so I'm not sure if it's something for /jstransformers.

[tunnckoCore]

:+1: I think for it before few days, but using https://github.com/medikoo/es6-template-strings instead.

[tunnckoCore]

@RobLoach seems to be better to use medikoo's package, its more up-to-date.

[RobLoach]

I think for it before few days, but using https://github.com/medikoo/es6-template-strings instead.

Thanks! Good find.... Updated and moved to here: https://github.com/jstransformers/jstransformer-es6-template-strings

[tunnckoCore]

@RobLoach see the commit comment :)

[RobLoach]

@stoeffel Used generator-jstransformer for jstransformer-es6-template-strings, jstransformer-whiskers and jstransformer-styl. It worked perfectly :+1: .

[tunnckoCore]

@RobLoach yep, but yeoman is too big and too shitty, same as gulp. lol

btw, should run add-owners for ur packages :)

[tunnckoCore]

:100: jstransformer-absurd :100: added to the list and almost finished, it is little tricky because not so node-friendly api.

[stoeffel]

@RobLoach nice :sunglasses:

[tunnckoCore]

hm.. both coffeekup and coffeecup looks very outdated and are similar, so i think we should implement only one of them?

[tunnckoCore]

coffeekup - done

[tunnckoCore]

list is sorted again added to the list

• autoprefixer
• cssnext
• doiuse
• fest
• gaikan
• rtlcss
• slm-lang
• resin

[ForbesLindesay]

I'm not sure about the explosion of postcss based transformers. I feel like maybe we should be taking the approach of the less transformer and accepting a plugins option instead of having a jstransformer per plugin. One very real downside of the approach we are taking at the moment is that a pipeline of postcss transformations could have very poor performance.

[tunnckoCore]

a plugins option

Yea, we will, I'll update postcss, rework and styl later today. But they are almost separate transformation tools, commonly used and it will be just sugar of doing install jstransformer-postcss, install cssnext, install autoprefixer separately.

This dont have bad side.

is that a pipeline of postcss transformations could have very poor performance.

And? This isn't jstransformer(s) problem.