[lexical-markdown][lexical-playground] Feature: Option to include blanklines in markdown render

[potatowagon]

Description

Describe the changes in this pull request

option to make the number of blank lines consistent between markdown and the rendered markdown

Closes: #

To enable:

$convertFromMarkdownString(
  ...,
  true // shouldIncludeBlankLines -> set to true,
)

$convertToMarkdownString(
   ...,
   false, // isNewlineDelimited -> set to false
);

These flags are optional. Defaults to old behavior of removing newlines on import and delimiting with newlines on export.

Test plan

Before

https://github.com/facebook/lexical/assets/19604232/108fdd80-0cde-4678-b4dd-b301132082fa

blank lines are ignored and removed

After

https://github.com/facebook/lexical/assets/19604232/b8ce7e72-e1c2-4b60-9b96-c92a95791faf

number of blank lines consistent between markdown and rendered markdown

[vercel[bot]]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
lexical	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 17, 2024 0:56am
lexical-playground	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 17, 2024 0:56am

[github-actions[bot]]

size-limit report 📦

Path	Size	Loading time (3g)	Running time (snapdragon)	Total time
packages/lexical/dist/Lexical.js	23.93 KB (0%)	479 ms (0%)	175 ms (+1113.49% 🔺)	654 ms
packages/lexical-rich-text/dist/LexicalRichText.js	34.59 KB (0%)	692 ms (0%)	207 ms (+521% 🔺)	899 ms
packages/lexical-plain-text/dist/LexicalPlainText.js	34.51 KB (0%)	691 ms (0%)	195 ms (+490.96% 🔺)	885 ms

[fantactuka]

LGTM, I believe there were some PRs in past with similar idea. More broader discussion (covered it a bit offline with Gerard & folks): it seems like best route for MD plugin is to eliminate out of the box import and export parsing and instead start relying on external library (e.g., remark), while plugin will only responsible to connect Lexical and 3rd party lib. That would allow delegating this kind of configuration to those libs.

Shortcuts is a bit more lightweight functionality (and handles simple formats only) so it could still remain a part of the plugin (that way we're not bloating editor with full MD parsing library for simple shortcuts).

[GermanJablo]

it seems like best route for MD plugin is to eliminate out of the box import and export parsing and instead start relying on external library (e.g., remark) @fantactuka

I'm not very familiar with Remark, but I have a hard time seeing how a generic library could correctly parse Lexical's editorState.

Separate topic, I do like this PR. In my app I use markdown for the tests, and to make the conversion idempotent I do something quite dirty: I add a specific string to the empty paragraphs, and after the conversion is done I delete them.

[potatowagon]

note: there is a change to blockquote behavior.

with shouldPreserveNewLines = false (old behavior)

hello

text

becomes

>hello

text

because of newline delimeter

---

with shouldPreserveNewLines = true (new optional behavior)

hello

text

^imagine no newlines seperating them

becomes

>hello
>text

tho this is consistent with other markdown surfaces such as github & dillinger where texts right below blockquotes get included in the blockquote