Add Multi-line Dictionary Definition (<dd>) rows.

[dbolack-ab]

Expands the existing syntax to allow/expect the option of multiple definitions by adding any number of :: Definition to an a DT/DD set.

Solves #2340.

Final Syntax and Behavior

This has been updated to a multi-description ( DD ) mode that can take multi-line input. This mode expects the term on one line, the description(s) on lines after, then a blank line before a new term. Two blank lines terminate the list.

This cannot be mixed with the existing inline style definition lists.

Example:

Term 1
:: Definition 1
:: Definition 2
is on multiple lines

Term 2:
:: Definition 1

Test case Brew at: https://homebrewery.naturalcrit.com/share/h-sHSQsRyUlz

[calculuschild]

Before going further with this, I think we need to have a little more discussion about where this syntax would actually be used. As I mentioned in this comment #2340 (comment), there was a specific use case we identified for the single-line syntax we currently use. I don't think we have any use case identified for multiple <dd> entries yet. How should we render these? What CSS should go along with them? Does it do something that we already do just as well with plain paragraphs? I don't think any of that was settled in #2340.

As another note, this PR unfortunately breaks existing <dl> rendering by adding a \n

[dbolack-ab]

Before going further with this, I think we need to have a little more discussion about where this syntax would actually be used. As I mentioned in this comment #2340 (comment), there was a specific use case we identified for the single-line syntax we currently use. I don't think we have any use case identified for multiple <dd> entries yet. How should we render these? What CSS should go along with them? Does it do something that we already do just as well with plain paragraphs? I don't think any of that was settled in #2340.

As another note, this PR unfortunately breaks existing <dl> rendering by adding a \n

I'm not spotting this?

                return `<dl>${token.definitions.reduce((html, def)=>{
-                       return `${html}<dt>${this.parser.parseInline(def.dt)}</dt>`
-                                   + `<dd>${this.parser.parseInline(def.dd)}</dd>\n`;
+                       const dds = def.dd.map((s)=>`<dd>${this.parser.parseInline(s)}</dd>`).join('\n');
+                       return `${html}<dt>${this.parser.parseInline(def.dt)}</dt>
+                               ${dds}`;
                }, '')}</dl>`;

Do you mean the linebreak between the close of the <dt> and the open of the first <dd> ?

[calculuschild]

Do you mean the linebreak between the close of the

and the open of the first

?

Yep. Try generating a monster stat block on this PR for example. We format the <dt> and <dd> to be on one line, but that \n breaks all the formatting.

[dbolack-ab]

Do you mean the linebreak between the close of the and the open of the first ?

Yep. Try generating a monster stat block on this PR for example. We format the <dt> and <dd> to be on one line, but that \n breaks all the formatting.

Weird. That is not what I remember to be the correct behavior for a DD/DT pair but it's been a spell. :)

[dbolack-ab]

Do you mean the linebreak between the close of the and the open of the first ?

Yep. Try generating a monster stat block on this PR for example. We format the <dt> and <dd> to be on one line, but that \n breaks all the formatting.

Weird. That is not what I remember to be the correct behavior for a DD/DT pair but it's been a spell. :)

I think this is good now.

[calculuschild]

This looks good! I'm just going to go through and see if the logic can be shortened up at all, then merge.

[dbolack-ab]

@calculuschild Is the multiline regression from your work or did I miss it?

[calculuschild]

@calculuschild Is the multiline regression from your work or did I miss it?

It's just a new test I added today for an edge case we missed.

Pretty sure I know where the fix is though. I'm planning to merge it in just a bit.

[dbolack-ab]

@calculuschild Is the multiline regression from your work or did I miss it?

It's just a new test I added today for an edge case we missed.

Pretty sure I know where the fix is though. I'm planning to merge it in just a bit.

Okay. The fail I'm seeing is indeed a fail that worked at last I remember.

      <dl><dt>Term 1</dt>
    - <dd>Definition 1 and more and more</dd>
    + <dd>Definition 1</dd>
    + <dt>and more and more</dt>
      <dd>Definition 2</dd>
      <dd>Definition 3</dd></dl>

[calculuschild]

No, it was broken before. It comes from this test which was written with bad syntax:

Term 1
::Definition 1
Term 2          <- Pandoc-style lists need a blank line between terms.
::Definition 1

This should output:

<dl>
<dt>Term 1</dt>
<dd>Definition 1 Term 2</dd> ("Term 2" is part of dd1 multiline)
<dd>Definition 1</dd>
</dl>

I added an extra newline between the Terms to correct that test, then added this new test which just changes Term 2 to more and more and added one more Definition 3 for good measure.

[dbolack-ab]

No, it was broken before. It comes from this test which was written with bad syntax:

Term 1
::Definition 1
Term 2          <- Pandoc-style lists need a blank line between terms.
::Definition 1

This should output:

<dl>
<dt>Term 1</dt>
<dd>Definition 1 Term 2</dd> ("Term 2" is part of dd1 multiline)
<dd>Definition 1</dd>
</dl>

I added an extra newline between the Terms to correct that test, then added this new test which just changes Term 2 to more and more and added one more Definition 3 for good measure.

Yeah, that seems right.

[calculuschild]

Thanks @dbolack-ab ! 💯 ⭐ 🎉

I found more edge cases, but it turns out just splitting each syntax into its own extension solved everything. And then I realized that let me simplify a lot too as a bonus.