[bug] blocks will be reset on preprocess every time and will break .md in .md

[JounQin]

The will not work when writing markdown codes in code blocks.

Lifecycle:

preprocess('.md') -> preprocess('.md/0_0.md') -> postprocess('.md/0_0.md') -> postprocess('.md')

This was not a problem before because .md files are not linted by eslint before. By adding esling-plugin-mdx which support .md files, it will be a problem.

Originally posted by @JounQin in #178 (comment)

---

@btmills

It's clear to be an bug

https://github.com/eslint/eslint-plugin-markdown/blob/main/lib/processor.js#L219

[btmills]

Thanks for opening this so we can dig in. I think I follow what you're saying, but to be sure, can you provide an example Markdown document that triggers the bug?

[JounQin]

@btmills

This issue will only occur when integrating with other plugins like eslint-plugin-mdx which will lint original .md files by preprocess too.

Here is the fake version for testing.

https://github.com/eslint/eslint-plugin-markdown/pull/183/files#diff-83e0a2056017ee834945f51308072e4a883a4ed96193581fe4aabc41cb98710a

[JounQin]

So maybe I shouldn't flag it as bug but a better compatibility with other markdown related ESLint plugins based on eslint-plugin-markdown.

[btmills]

This seems related to https://github.com/eslint/eslint/pull/14227/files#r602802758. I can reproduce this today in a test if I configure this plugin to run on .md and .markdown files and have a README.md file containing a markdown code block. If #14227 changes ESLint to run processors directly recursively, this would be even easier to reproduce.

[JounQin]

Yes, I should simplify the test case in current PR.