How can I have an uncaptioned table nested within a custom crossref environment?

[DylanSpicker]

Description

I have a custom cross-reference (float) environment defined, that works without concern. However, if I end up using an (uncaptioned) Markdown table within the div for the custom cross-reference, the cross-reference becomes unnumbered/unlabelled in the rendered HTML.

I am not sure whether this is a misunderstanding of the expected behaviour or whether there may be a workaround for this.

From my perspective, the expected behaviour would be to have cross-reference still labelled, with no caption required, as it is for the instances without tables within them. If I have uncaptioned figures, they work as expected (as do divs with no nesting of cross-reference environments).

Reproducible Example

---
title: "Minimal Example for Cross-Reference Problem"
format: html
crossref:
  custom:
    - kind: float
      reference-prefix: My Cross Ref
      key: mcr
mcr-cap-location: top
---

:::{#mcr-1}
I work as intended!
:::

::::{#mcr-2}
I do not work. I think because of this table:

|        Table         |
| :------------------: |
| Basic Table Here ... |

::::

:::{#mcr-3}
If I use two captions here, then the numbering returns. 

|        Table         |
| :------------------: |
| Basic Table Here ... |

: This is the table caption.
<!-- Anything written below the table caption stands in as a caption for the MCR -->
This is the MCR Caption 
:::

Details

Running Windows 11.

Quarto 1.6.32
[>] Checking environment information...
      Quarto cache location: C:\Users\dylan\AppData\Local\quarto
[>] Checking versions of quarto binary dependencies...
      Pandoc version 3.4.0: OK
      Dart Sass version 1.70.0: OK
      Deno version 1.46.3: OK
      Typst version 0.11.0: OK
[>] Checking versions of quarto dependencies......OK
[>] Checking Quarto installation......OK
      Version: 1.6.32
      Path: C:\Users\dylan\AppData\Local\Programs\Quarto\bin
      CodePage: 1252

[>] Checking tools....................OK
      TinyTeX: (not installed)
      Chromium: (not installed)

[>] Checking LaTeX....................OK
      Tex:  (not detected)

[>] Checking basic markdown render....OK

[>] Checking Python 3 installation....OK
      Version: 3.12.4
      Path: C:/Users/dylan/AppData/Local/Programs/Python/Python312/python.exe
      Jupyter: (None)

      Jupyter is not available in this Python installation.
      Install with py -m pip install jupyter

[>] Checking R installation...........OK
      Version: 4.4.1
      Path: C:/PROGRA~1/R/R-44~1.1
      LibPaths:
        - C:/Users/dylan/AppData/Local/R/win-library/4.4
        - C:/Program Files/R/R-4.4.1/library
      knitr: 1.48
      rmarkdown: 2.29

[>] Checking Knitr engine render......OK

[mcanouil]

It's not about the table, it's because your code is not a valid cross-reference div at least to get what you want.

A cross-reference div is:

::: {#prefix-label}
content

caption
:::

The line before the table is a content and so your table is the caption.
Make the two one single content to get the desired result.

[DylanSpicker]

Oh, excellent. Thank you for this!

This gives at least one way to work around this, but I have some additional questions to try to solidify my understanding of this.

1. If I remove the line before the table in the markup, I get the same results (which is to say, I do not get "My Cross Ref 1" output anywhere). In this case, is the table still being treated as the caption, or perhaps, it's simply not being parsed correctly because of the lack of a caption?

::::{#mcr-1}
|        Table         |
| :------------------: |
| Basic Table Here ... |

::::

2. I would suppose that inner divs are treated in a similar way. So, for instance, in the following example, the whole "inner-div" becomes the content, and the line afterwards becomes the caption. This is what is rendered, and that makes sense from your explanation.

:::{#mcr-1}
::: {.inner-div}
This is content. 
:::
This is the caption. 
:::

3. If I drop the "This is the caption." from the above example, the div is still treated as content (which is to say, it outputs "My Cross Ref 1", with "This is content." shown beneath that). This is in contrast to the example above with the table, or indeed, with the following example that puts the table inside of inner-div. (That is, the table is rendered, but there is not "My Cross Ref 1" anywhere). Is there something else going on with this example that I am missing that causes these differences?

:::{#mcr-1}
::: {.inner-div}
|        Table         |
| :------------------: |
| Basic Table Here ... | 
:::
:::

4. Is there a "correct" way to have a cross-reference div, without a caption? Most of the time, in my actual use case, they are of the form of the inner-div, with no caption following it. It was only the presence of the tables that started to give the undesired behaviour (missing the "My Cross Ref #"), which is where I figured the tables were to blame. But, perhaps there is a more correct way to achieve this desired result all together?

[mcanouil]

I don't have any issues.

Input

Output

--- title: "Minimal Example for Cross-Reference Problem" format: html crossref: custom: - kind: float reference-prefix: My Cross Ref key: mcr mcr-cap-location: top --- ::: {#mcr-1} I work as intended! ::: :::: {#mcr-2} ::: {#my-content-1} I do not work. I think because of this table: | Table | | :------------------: | | Basic Table Here ... | ::: :::: :::: {#mcr-3} ::: {#my-content-2} | Table | | :------------------: | | Basic Table Here ... | ::: :::: :::: {#mcr-4} | Table | | :------------------: | | Basic Table Here ... | ::::

[mcanouil]

Note that the current stable version is 1.6.43, the Release Candidate is 1.7.23 and that 1.6.32 is an old pre-release.
If you intend to use pre-releases, please ensure to keep it up to date.

[DylanSpicker]

Oh, great! Thank you.