Todo/Refactor

[alexander-akait]

• Migrate on prettier to avoid difference code style (Format with prettier #208).
• Enable recommitted eslint rules avoid stylistics rules (Format with prettier #208).
• Upgrade https://github.com/evilebottnawi/favicons/tree/gh-pages and publish their.
• Run test on windows using appveyor. (Setup Appveyor CI #207)
• Implement promise for generate (Return a promise if callback is undefined #210).
• Migrate from gulp to ava for testing. (Migrate tests to Ava #206)
• Migrate from gulp to rollup/webpack for browser support.
• ES6 refactor (Migrate codebase to modern JS #215).

Wait feedback 👍

//cc @brunocodutra

[brunocodutra]

I'd recommend ava instead of jest.

I tried both in brunocodutra/webapp-webpack-plugin and for some reason ava is several times faster than jest.

[brunocodutra]

I volunteer to improve the testing framework using ava if you agree this is the way to go

[alexander-akait]

@brunocodutra ava fix snapshot testing? it will be great for snapshot image testing to avoid generate broken images?

[brunocodutra]

@evilebottnawi yes it does snapshot testing much like jest

see https://github.com/avajs/ava#snapshot-testing

I'll start working something out and create a PR for us to better evaluate it

[alexander-akait]

@brunocodutra great 👍

[brunocodutra]

@evilebottnawi regarding the ES6 refactor I was thinking, what about dropping support for node 4, which is reaching EOL this April, and sticking to node > 6 so that we may reasonably get rid of gulp and babel altogether?

[alexander-akait]

@brunocodutra bit early, right now not April 😄

[brunocodutra]

Transpiling is not that big of a hassle, it's just that @jantimon convinced me it makes it harder for users that face issues to contribute fixes or even open better issue reports.

Either way we can decide about dropping node < 6 independently from the ES6 refactor.

[alexander-akait]

@brunocodutra also i think we should build bundle use webpack or rollup for browser usage 👍 Then we can do demo on gh-pages

[brunocodutra]

@evilebottnawi sounds reasonable to me

[alexander-akait]

@brunocodutra i added to list

[alexander-akait]

@brunocodutra what do you think about buy domain favicons.xyz for future gh-pages, favicons.io better but very expensive 😞

[brunocodutra]

@evilebottnawi to be honest I think gh-pages are fine

[alexander-akait]

@brunocodutra we will use gh-pages, github allow to add own domain on gh-pages

[brunocodutra]

@evilebottnawi I know, I mean the github domain is fine, is there such a demand for the website to justify owning a domain?

BTW, how about adding a link to it to the project description?

[alexander-akait]

@brunocodutra 👍

[brunocodutra]

@evilebottnawi while writing new tests to increase code coverage I realized we have some undocumented features.

One feature in particular updates/creates HTML files directly on disc only through the stream interface and only if the html handler is not provided and if params.html && !params.pipeHTML.

Do we really need this feature?

[alexander-akait]

@brunocodutra intresting, need investigate.

In near future i want move favicons to https://github.com/itgalaxy/ org, leaving you access to the repository.
No important PR is expected in the near future?

[brunocodutra]

@evilebottnawi go ahead, I don't have anything right now

[alexander-akait]

/cc @brunocodutra why we use @babel/polyfill as dep?

[brunocodutra]

@evilebottnawi because we need something that isn't available in node 4, although I can't remember what it is.

If we drop support for node < 6, which is quite reasonable since node 4 reached EOL, I believe we might be able to get rid of transpiring with Babel entirely.

[alexander-akait]

@brunocodutra i think we should drop node@4 👍