stringify: incorrect entity encoding

[ysiw]

remark-stringify outputs incorrect escaped entites in table if options.entities is "escape"

$ cat a.md
| aaa        | aaa |
| ---------- | --- |
| http://abc |     |
$ remark a.md
| aaa             | aaa |
| --------------- | --- |
| http://abc |     |
a.md: no issues found
$ remark --setting '"entities":"escape"' a.md
| aaa                 | aaa |
| ------------------- | --- |
| http://abc |     |
a.md: no issues found
$

[wooorm]

What is the reason for using entities: 'escape'?

[wooorm]

This is indeed a bug. Double escaping is going on if entities: 'escape' is defined.
I think it makes the most sense to remove the option altogether.

[ysiw]

I prefer some entites are escaped, such as ✓ (✓), ✗ (✗). It is the only way to do it. entities: true is not an option for me, because it escapes all CJK characters.

[wooorm]

I prefer some entites are escaped, such as ✓ (✓), ✗ (✗)

That behavior is closer to entities: true than entities: 'escape'.
escape used to work but is now broken, and should do the inverse: https://github.com/wooorm/stringify-entities/tree/355100188e1f3661359e7404115cd7e04c75ed41#optionsescapeonly.

I think removing 'escape' makes the most sense.

[ysiw]

I suggest modifing behavior of entites: 'escape' to escape ONLY all character entites listed in https://dev.w3.org/html5/html-author/charref

There is a package character-entities which has the complete list.

Current implementation gives me no option to keep my document unchanged.

[wooorm]

I suggest modifing behavior of entites: 'escape' to escape ONLY all character entites listed in https://dev.w3.org/html5/html-author/charref

That does include many ASCII / other characters, such as a space, a tab, a line feed, +, -, :, and so many more!

---

As you are already invested in Unicode, with Chinese characters, I don’t really get why check and cross can’t be in unicode either?

I instead strongly think we should drop options on encoding character references. /CC @ChristianMurphy, what do you think?

[ChristianMurphy]

drop options on encoding character references

As in remove options.entities from the project entirely?

[wooorm]

Yes

[wooorm]

background: it was added because I recently made stringify-entities, which had those options. Now, years later, we definitely live in a unicode world, there is basically no reason anymore to use character references.

[ChristianMurphy]

Removing the option works for me.
The only character entity that I've used in projects recently is  , which I believe has a Unicode equivalent.