.html() is decoding entities when decodeEntities: false

[anagai]

I am using 1.0.0-rc.2 release version. I have html entity such as > and would not like that decoded when I use .html(). I had loaded the html object this way:

let obj = cheerio.load('>', {decodeEntities: false})

console.log(obj.html())

The output is decoded to >. This is not correct.

When I used the 0.22.0 release this issue does not occur. Can you please set the latest stable release as 0.22.0 please.

[AlynxZhou]

This is a bug because of different behavior of parse5 and htmlparse2, if you set decodeEntities to false, dom-serializer won't encode them, and htmlparse2 won't decode too. But when cheerio switches to parse2, it always decode entities, which leads into this problem. If you let dom-serializer encode entities, other chars like CJK will be encode too, which leads into other problems. Also, it's really confusing because sometimes cheerio chooses parse5 and sometimes cheerio chooses htmlparser2

This is a long-time problem, and I am feeling bad that no cheerio maintainer makes a decision to solve this, it's really annoying, we should keep a correct behavior instead of keeping compatible to old incorrect behavior, I have submitted a solution here but no one replied to my PR. In a word, to encode all entities is not a good choice, cheerio is used to parse like DOM, not encoder.

[fb55]

Thanks @AlynxZhou!