@TTO: Pinging you as per your request. :)

Heh, thanks Andre! I have simply been too busy to work on this. I've set a reminder for Saturday.

Change 328887 had a related patch set uploaded (by TTO; owner: TTO):
[mediawiki/extensions/Poem@master] Switch <poem> to output HTML <pre> instead of <div>

https://gerrit.wikimedia.org/r/328887

@cscott @ssastry Here is what I have come up with:

My latest reimplementation of <poem> is indent-pre with the following differences:

• <poem> is displayed as normal text with white-space: pre-wrap, not in monospace with white-space: pre and a box around it. And because it's a tag, it can have inline styles, classes, etc.
• Multiple consecutive intra-line space characters are folded into a single space. (We generally agree that we want to get rid of this difference in the long term.)
• Block-level #*:; formatting and horizontal rules ---- are permitted.

It turns out to be quite difficult to achieve this behaviour with the PHP parser, because it loves to add <p> tags in all kinds of stupid places, even when you're trying your best to make everything an indent-pre. I'm not at all familiar with the Parsoid code base, but I hope it is much cleaner to implement this there. If the Parsoid implementation is nice, then I think it is OK to tolerate the slightly hacky approach in the legacy parser (can I call it "legacy" yet?).

My hope is that this is a refreshing, simple approach that resonates with you (even if the proposed implementation leaves a little bit to be desired). Let me know your thoughts.

In T194651#4294645, @TTO wrote:

My latest reimplementation of <poem> is indent-pre with the following differences:

• <poem> is displayed as normal text with white-space: pre-wrap, not in monospace with white-space: pre and a box around it. And because it's a tag, it can have inline styles, classes, etc.
• Multiple consecutive intra-line space characters are folded into a single space. (We generally agree that we want to get rid of this difference in the long term.)
• Block-level #*:; formatting and horizontal rules ---- are permitted.

This seems pretty reasonable from a semantics point-of-view, especially if we can remove the middle "consecutive intra-line space character" behavior (or move it into CSS, I think that's basically white-space: pre-line?). Then it's basically just a CSS style change and "allow markup which requires a start of line context".

Hopefully we can make good test cases and confirm that a Parsoid implementation of these semantics is "natural".