strikethrough incorrect for 3+ tildes

[dwelle]

~~~text~~~

renders to (taken from markdown-it's repl):

<pre class="hljs"><code></code></pre>

GFM ignores the extra ~ and renders:

<p><del>text</del></p>

[puzrin]

See example. ~~~ at start of line is a fenced block, according to CM spec.

I'd recommend to discuss at commonmark forum https://talk.commonmark.org/t/a-problem-with-backtick-code-fences/1053

[puzrin]

Do you have example from real world when ~~~ needed at start of paragraph?

[dwelle]

E.g. when users mistype (when they want to cross-out paragraph or sth) and wonder why their text suddenly disappeared.

Edit: I thought it's a bug since I wasn't aware of that discussion and the fact it's actually in CM spec. Either way, I think overloading ~ operator for code blocks might not be a good idea for the above reason. If anything, I'd use ~~~ for block-strike-through, which is a requested feature, too.

WTBS, shouldn't the closing ~~~ be on new line for it to be considered fenced block, acc to CM spec?

[puzrin]

spec.commonmark.org/0.27/#fenced-code-blocks it's better to discuss spec-related things in other place.

closing as not a bug.

[dwelle]

Still, as I said above, acc to CM spec, the closing ~~~ must not be on the same line as the opening sequence, else it's not considered a code block.

Any text after opening fenced block sequence is considered info string, and if the line contains the closing sequence, it's not considered code block at all.

The spec is kinda not explicit about this, tho, but it can be verified at their dingus.

input:

~~~text~~~

CM output:

<p>~~~text~~~</p>

md-it output:

<pre class="hljs"><code></code></pre>

So currently, md-it is not spec compliant.

Markdown-it's fenced block using the ``` sequence does behave according to CM spec, so the ~~~ behavior is not even consistent in this regard.