Respec: Explore about replacing `normalizePadding()` with `reindent()`

[saschanaz]

Turns out that a simple replacement doesn't work, so explore about why.

[marcoscaceres]

Markdown is annoying... but normalizePadding() is not my proudest moment. Would love to see it be replaced with something simpler.

[saschanaz]

Looks like indentation processing is weird on normalizePadding(). For example:

<section id='sotd'><p>foo</p></section><section id='toc'></section>

    # Title
    Some text

Per the markdown spec, the text should be interpreted as <pre> because of its indentation. Instead, the function tends to interpret it as normal unindented markdown by trimming the text and all its indentations.

[marcoscaceres]

Correct... so, this is for legacy reasons. People were writing markdown inside the body indented like that:(

[saschanaz]

Is there a live example? I saw this on our tests that appends indented template string. I don't think the tests did it intentionally...

[marcoscaceres]

I’ll dig some up.

[marcoscaceres]

https://github.com/coremob/coremob-2012/blob/gh-pages/index.html

[marcoscaceres]

Hello from 2012!

[saschanaz]

🤦‍♀️

[marcoscaceres]

Heh, yep. Kinda stuck. But we could make a case to break things for sanity’s sake.

[marcoscaceres]

@saschanaz do you think we can do anything here? Or “abandon all hope”?

[saschanaz]

@marcoscaceres I think we should learn exactly what behavior differences exist, add some tests for them, and then we can revisit this. For now it's rather a black box...