fix(text): Render list items correctly

[kesara]

Fixes #1135
Fixes #1138

See changes made in 3712523 for context.

This change can affect if ol is used inside a li as the first element and the li element is inside another ol element.

I could not find any published RFCs with XML that match the above condition.

These are the RFCs that I found that use ol as the first element in li:

• RFC8679
• RFC8687

I didn't notice any changes in text output due to this change. But there are a few minor changes (with spacing) due to changes over the years.

[rjsparks]

I think it's worth trying to change things such that generatiing

   4.  4.1  Sublist with '%p' parent counter element in the list
            counter, item one

       4.2  Item two.

       4.3  Item three.

we generate

   4. 

       4.1  Sublist with '%p' parent counter element in the list
            counter, item one

       4.2  Item two.

       4.3  Item three.

Unless that breaks the intended fix for the rfc-to-be.

[cabo]

What is better about that superfluous empty line?

(You can probably have one if you want with <br>)

[rjsparks]

Look at it in context with the rest of the nested <ol> above it in the test document:

xml2rfc/tests/valid/elements.bom.text

Lines 366 to 458 in fdb6fb8

	4. Ordered list with sub-elements
	1. Ordered list, first bullet. Indent="6".
	2. Ordered list, second bullet. Indent="6". Text only. Octavio
	fuit, cum illam severitatem in eo non arbitrantur. erunt etiam,
	et ii quidem eruditi Graecis litteris, contemnentes Latinas,
	qui se plane Graecum dici velit, ut a Scaevola est praetore
	salutatus Athenis Albucius. In physicis plurimum posuit. ea
	scientia et verborum vis et natura orationis et consequentium
	repugnantiumve ratio potest perspici.
	3. Ordered list, third bullet, element t. Quid? si nos non
	interpretum fungimur munere, sed tuemur ea, quae audiebamus,
	conferebamus, neque erat umquam controversia, quid ego
	intellegerem, sed quid probarem. Extremum autem esse bonorum eum
	voluptate vivere. Sic faciam igitur, inquit: unam rem explicabo,
	eamque maximam, de physicis alias, et quidem locis pluribus.
	Author, et al. Expires 13 January 2019 [Page 7]
	Internet-Draft Xml2rfc Vocabulary V3 Elements July 2018
	Ordered list, third bullet, second element:
	+-------------------------------------------+
	| |
	| Artwork |
	| |
	+-------------------------------------------+
	Ordered list, third bullet, third element: Definition list,
	element dt.
	Element dd. Nisi mihi Phaedrum, inquam, tu mentitum aut
	Zenonem misisti.
	Ordered list, third bullet, fourth element:
	Figure
	Figure
	Figure
	Figure 2: Some figure name
	1. Ordered list, third bullet, fifth element: Ordered list,
	element li. Text only.
	2. Ordered list, third bullet, fifth element: Ordered list,
	element t. Nam cum ad me de virtute misisti. Quae est enim
	contra Cyrenaicos satis acute, nihil ad iucunde vivendum
	reperiri posse, quod coniunctione tali sit aptius. Quae etsi
	mihi nullo modo nec divelli nec distrahi possint, sic de
	iustitia iudicandum est, quae non modo non impediri rationem
	amicitiae, si summum bonum diceret, primum in eo essent.
	# Ordered list, third bullet, sixth element
	A = 1;
	* Ordered list, third bullet, seventh element: Unordered list,
	element li. Text only.
	* Element t. Atque ut odia, invidiae, despicationes adversantur
	voluptatibus, sic amicitiae non posse iucunde vivi, nisi
	sapienter, honeste iusteque vivatur, nec sapienter, honeste,
	iuste, nisi iucunde. Democritea dicit perpauca mutans, sed
	ita, ut ea, quae audiebamus, conferebamus, neque erat umquam
	controversia, quid ego intellegerem, sed quid probarem. See
	Appendix A, Section 3.1 of
	[I-D.levkowetz-xml2rfc-v3-implementation-notes] and also
	Appendix A.1 of Key Words Case Ambiguity [RFC8174] / Key Words
	Case Ambiguity [RFC8174], Appendix A.1 / Key Words Case
	Ambiguity [RFC8174] (Appendix A.1).
	Author, et al. Expires 13 January 2019 [Page 8]
	Internet-Draft Xml2rfc Vocabulary V3 Elements July 2018
	4. 4.1 Sublist with '%p' parent counter element in the list
	counter, item one
	4.2 Item two.
	4.3 Item three.

and consider what would happen if there were any text between the <li> and <ol> at

xml2rfc/tests/input/elements.xml

Lines 472 to 473 in fdb6fb8

	<li>
	<ol type="%p%d">

[kesara]

I have converted PR to a draft.
The changes from e889f93 add a new line when with the list style to both ol and ul.
This was skipped previously.
Specifically, this adds list style + new line when li contains an artwork, figure, or sourcecode as the first child.
See how this changed test outputs in ca8286a
Note that the test output has date changes and multiple content changes propagated because of the addition of new lines.

e889f93 fixes #1138

[rjsparks]

And now this is perhaps unfortunate. What we've uncovered is, I think, a decision to special case the rendering of nested lists without leading text that doesn't generalize to all possible uses of such things, and we're just toggling cross the edges of that earlier decision.

Has the previous side-effect of no leading marker been used in this case previously? Is it ok to make this change (since getting the behavior of that side-effect is possible in better defined ways, I _think)?

@cabo - what do you think the right thing to do with this is?

[kesara]

Here's a list of RFCs that has ul tag inside a li as first element:

rfc8668.xml
rfc8742.xml
rfc8811.xml
rfc8793.xml
rfc8927.xml
rfc8677.xml
rfc8667.xml
rfc8698.xml
rfc9195.xml

All these use empty="true" in parent ul, so no new * gets added but there's a bug in the suggested fix because it adds a new line. I'll fix that.

[cabo]

@cabo - what do you think the right thing to do with this is?

I don't think it is ever correct to add a new line that the author didn't want.
If it is needed, the author can put it in, but it never should be forced on them.

[kesara]

I have pushed a fix to handle empty="true".
See 11ef6ac and fe595b1

[kesara]

@rjsparks & @cabo For context, HTML output doesn't add a new line.
Proposed changes don't affect HTML output.

[rjsparks]

And HTML has the marker, so the text should too - that convinces me that (except for the newline that Carsten doesn't want) the current proposed change is right.

So people looking at this later can see it in context, could you also paste the section of html analogous to the 4. 4.1 change in the text in this thread?

[kesara]

@rjsparks ol has a new line in HTML output:

[kesara]

@rjsparks The above screenshot is from Firefox and that might not be showing the correct rending. See #1158