White space handling looks inconsistent between HTML block element vs HTML inline element

[seiyab]

Hi. I'm curious about why white space preservation / removal differs between <div> and <a> in following case:

https://spec.commonmark.org/dingus/?text=%3Cdiv%20class%3D%22abc%0A%20%20%20%20d%20%20e%20f%22%3Etext%3C%2Fdev%3E%0A%0A%3Ca%20class%3D%22abc%0A%20%20%20%20d%20%20e%20f%22%3Etext%3C%2Fa%3E

Input:

<div class="abc
    d  e f">text</dev>

<a class="abc
    d  e f">text</a>

Output:

<div class="abc
    d  e f">text</dev>
<p><a class="abc
d  e f">text</a></p>

I've read https://spec.commonmark.org/0.31.2/ but it's not obvious whether it's a bug or intentional.
I came from https://github.com/orgs/micromark/discussions/211.

[jgm]

I suppose this isn't right, because removing that whitespace changes the meaning of the attribute.

I think it's a side effect of code in the parser that skips leading indentation before sending paragraphs to the inline parser.

[jgm]

Note that the same issue affects the C parser commonmark/cmark and the Haskell parser jgm/commonmark