Unexpected escaping of characters since 2.1.1

[tripodsan]

Initial checklist

• I read the support docs
• I read the contributing guide
• I agree to follow the code of conduct
• I searched issues and couldn’t find anything (or linked relevant results below)

Affected packages and versions

mdast-util-to-markdown@2.1.1

Link to runnable example

No response

Steps to reproduce

the changes in mdast-util-to-markdown@2.1.1 cause unexpected escaping of normal characters, for example:

├─8  paragraph[3]
│    ├─0 strong[1]
│    │   └─0 link[3]
│    │       │ url: "https://www.example.com"
│    │       │ title: ""
│    │       ├─0 text "Curves on"
│    │       ├─1 emphasis[1]
│    │       │   └─0 text "the"
│    │       └─2 text " ipad. "

generates:

      **[Curves on_the_ ipad. ](https://www.example.com)**

it used to generate:

      **[Curves on_the_ ipad. ](https://www.example.com)**

another example:

├─2 paragraph[1]
│   └─0 emphasis[2]
│       ├─0 text "first"
│       └─1 strong[2]
│           ├─0 text "second"
│           └─1 emphasis[1]
│               └─0 link[1]
│                   │ url: "about:blank"
│                   │ title: "title"
│                   └─0 text "important"

generates:

_first**second_[important](about:blank "title")_**_

before:

_first**second_[important](about:blank "title")_**_

Expected behavior

it should not escape characters that are not needed to be escaped.

Actual behavior

it escapes unexpected characters.

Affected runtime and version

node v20.18.0

Affected package manager and version

No response

Affected OS and version

No response

Build and bundle tools

No response

[wooorm]

Hi! You note that a change happened in a version. Did you see the release notes / commits that were made? The changes are intentional. There are commits, there is talk, there are tests. Do you have arguments?

[tripodsan]

I looked at the changelog but afaiu, the fix should only affect "invalid" mdast trees that would produce whitespace inside formats (eg _ the _). but in my examples above, the escaping also happens for the valid cases, like:

**[Curves on_the_ ipad. ](https://www.example.com)**

Curves on_the_ ipad.

or

_first**second_[important](about:blank "title")_**_

first**secondimportant**

(well, the case above has problems with the **`, but this is another issue)

[wooorm]

This also indeed touches on “invalidness”. Although it isn’t that: all of this is valid.

Asterisks are different than underscores in markdown. There rules are simpler. They can form emphasis/strong more often, because they are less often used in natural language. Underscores, in words, as you perhaps know, cannot form emphasis. That is why this escape happens here. That _ is inside on and the. take for example: on_the_ipad: on_the_ipad. See? The underscores show, no emphasis. Not with on*the*ipad: ontheipad.

I can recommend two things:
a) use asterisks
b) there’s a space missing in the example, you likely want on instead of on

Either is fine?

[wooorm]

I will add that it is very hard to get emphasis/strong right for mdast-util-to-markdown. The rules are 17 points, and that’s vague ideas instead of actual concrete code: https://spec.commonmark.org/0.31.2/#can-open-emphasis

[tripodsan]

ok. thanks for the explanation. probably a good thing to switch to * everywhere :-)