Remove babel and mjs

[segayuu]

No description provided.

[segayuu]

The travis CI error could not be confirmed locally.
It is possible that the test itself is broken in CI.

[JLHwung]

I can reproduce test failure on my machine. The test fixture is highly related to the output of hexo helpers and highlight.js.

Since hello-world does not contain any TeX definition, I suggest we rewrite the test and treat it as a hexo smoke test. For example we can test if it generates non-empty fixture.

As long as tag-escape and other TeX related test pass, it should be good to go.

Could you enable the skipped block and inline test?

[segayuu]

I don't know how to adapt the review.
If you get an error on this side, you can investigate, but it works in my environment.

• OS: win10
• Node: 12.8.0
• NPM: 6.10.3
• Git: 2.22.0.windows.1

[yoshinorin]

@segayuu
This test failure seems does not occur on native Windows.
But I can reproduce it using by WSL.

• Windows 10
• WSL (Ubuntu 18.04.1)
• Node.js v12.3.1
• npm 6.10.2
• git 2.17

Please be careful, npm script might be a failure on native Windows if execute npm install once on WSL.

Please delete node_modules and re-execute npm install on native Windows if you face a npm script failure on native Windows after execute npm install on WSL.

[curbengh]

This test failure seems does not occur on native Windows.

Line ending difference?

[yoshinorin]

Maybe I got it. The test failure caused by hexo-renderer-marked version.

The test success on Linux (WSL) if npm-script preset:clone change to following.

"pretest:clone": "git clone --recursive https://github.com/hexojs/hexo-starter.git ./test/site && cd ./test/site && git checkout a4d53738d734b1151576337ea427b6186115943a"

The commit hash a4d53738d734b1151576337ea427b6186115943a is parent commit of update hexo-renderer-marked to v1.0.1

Line ending difference?

Maybe no. But, I don't know why the test passes on native Windows. That's weird 🤔

[segayuu]

Thank @yoshinorin !

[curbengh]

Maybe I got it. The test failure caused by hexo-renderer-marked version.

similar issue with hexojs/hexo#3571?