Descriptions containing unicode characters gets encoded

[rigwild]

After some testing, I got unlucky with superb with a description containing an apostrophe.

Yeoman seems to encode Unicode characters to be HTML-safe.

My cat's meow module

package.json:

{
  "description": "My cat's meow module"
}

readme.md:

> My cat's meow module

It does not break the package.json file, nor the readme rendering though.

Related: #59 #60

[sindresorhus]

This is an edge-case, so I'm most likely not going to look into this. But a PR is always welcome.

[rigwild]

did some testing, checked which chars would break package.json

Tested:

& é ~ " # ' { ( [ - | è ` _ \ ç ^ à @ ) ] ° = + } ¨ $ £ ¤ ù % * µ < > , ? ; . : / ! § a ã ä â à

Ones that gets HTML-encoded:

& " ' < >

One that does not get encoded but still break

\

Emojis seems to work as expected.
There will always be some random unicode chars that will always break it, but that's too edgy, no one really care.

Escaping is not enough (they still get converted by yeoman when using fs.copyTpl

I can do a PR that simply replace these with a space.