incorrect paragraph number section contains list, table, or figure

[rvanrheenen]

Description

<xref> to a <t> do not produce the correct paragraph number if the section contains a list, figure, or table prior to the desired paragraph.

Some examples from the files listed below:

<xref target="no-indef"/> produces “Section 5.1, Paragraph 9” in the output files, but I would count the paragraph with <t anchor="no-indef"> as Paragraph 5. Seems the two figures, the unordered list, and the table are counted as paragraphs.

<xref target="norm-nfc"/> produces "Section 3, Paragraph 7” in the output files, but I would count the paragraph with <t anchor="norm-nfc"> as Paragraph 6. Seems the unordered list is being counted as a paragraph.

<xref target="zone-id-issue"/> produces "Section 6.1, Paragraph 3" in the output files, but this is confusing as the paragraph with <t anchor="zone-id-issue"> seems to be Paragraph 9. Seems the entire definition list (that includes multiple paragraphs and an <aside> paragraph) is being counted as a single paragraph.

Files:
https://www.rfc-editor.org/v3test/rfc9988-TEST2.xml
https://www.rfc-editor.org/v3test/rfc9988-TEST2.txt
https://www.rfc-editor.org/v3test/rfc9988-TEST2.html
https://www.rfc-editor.org/v3test/rfc9988-TEST2.pdf

Related issue: #900

Code of Conduct

• I agree to follow the IETF's Code of Conduct

[cabo]

One problem is that if you are pointing to an unordered list, you need a paragraph number...

The fun part to read is:

            elif t.tag in [ 't', 'ul', 'ol', 'dl', ]:
                type, num, para = split_pn(t, pn)
                content = "%s %s, Paragraph %s" % (type, num, para)
            elif t.tag in [ 'li', ]:
                type, num, para = split_pn(t, pn)
                para, item = para.split('.', 1)
                content = "%s %s, Paragraph %s, Item %s" % (type, num, para, item)

I don't claim to fully understand this, but at least lists seem to be referred to as paragraphs.

[kesara]

This is because of the logic @cabo pasted above.
NOTE: I am using draft-ietf-core-href-30.xml from datatracker.

In the example of zone-id-issue, high-level structure is :

<section anchor="cri-to-uri">
  <name>..</name>
  <t>...</t>
  <dl>..</dl>
  <t anchor="host-ip-to-uri">..</t>
  <t anchor="zone-id-issue">..</td>
  ...
</section>

xml2rfc only counts t, dl elements in above as paragraphs. So <xref target="zone-id-issue"/> generates:

see [Section 6.1, Paragraph 4]