Some issues with citations in footnotes

[denismaier]

This here gives me wrong results with the latest pandoc version:

---
references:
- type: book
  id: doeA
  author:
  - family: Doe
    given: John
  title: 'Title A'
- type: book
  id: doeB
  author:
  - family: Doe
    given: John
  title: 'Title B'
---

[^1] [^2] [^3] [^4]


[^1]: @doeA Some text in between. Also @doeB.

[^2]: @doeB

[^3]: @doeB

[^4]: @doeA.

This gives me these footnotes:

[^1]: 1 Some text in between. Also Doe, *Title B*, n.d.

[^2]: Doe,

[^3]: Doe,

[^4]: 1.  

Couple of issues:

1. There's a incorrect ibid in the second footnote. I think here we would want the short form.
2. The ibid in footnote 3 is correct, but note the comma.
3. The period in footnote 4 seems to start an enumeration.
4. Footnote 1 shows that the enumeration is actually started even without a period. (Is the @ starting a list here?)

[jgm]

What style are you using?
With current master, I get this (with pandoc --citeproc -t plain --csl chicago-fullnote-bibliography.csl):

[1] 1 Some text in between. Also John Doe, Title B, n.d.

[2] Doe, ibid

[3] Doe, ibid

[4] 1.  

[jgm]

Yes, @doeA. does conform to the spec for an example list item.
I guess you'd need to write @doeA\. to avoid this.

[jgm]

And the @doeA in your first note is turning into 1 because it's being treated as a reference to the example list item below...

[jgm]

With the backslash, I get:

[1] John Doe, Title A, n.d Some text in between. Also John Doe, Title B,
n.d.

[2] Doe, ibid

[3] Doe, ibid

[4] Doe, Title A, n.d.

which looks right except for the missing periods after "ibid". That does seem to be a bug, but in pandoc, not the citeproc library.

[jgm]

Oh, and yes, I see the point about the ibid in n. 2.
I guess that is a citeproc library issue.

[denismaier]

What style are you using?

Chicago-note-bibliography.csl
Sorry, forgot to include that info

[jgm]

OK, with the latest change I'm getting:

[1] Doe, Title A. Some text in between. Also Doe, Title B.

[2] Doe, ibid.

[3] Doe, ibid.

[4] Doe, Title A.

Everything okay now I think, except the ibid in n. 2.

[jgm]

Actually I think this is also a pandoc issue. We need to recognize when these citations occur inside a note and set citationNoteNum accordingly. Sorry, we are doing that correctly.

[jgm]

Actually I think we conform to the CSL spec: "the current cite is the first cite in the citation, and the previous citation consists of a single cite referencing the same item."

Note that you have two citations in the first note. So the "previous citation" is to @doeB.

I know, it's a bit unexpected, because if you'd done a regular note citation
[@doeA; @doeB] then this whole thing wolud be considered one citation and the ibid would not be triggered.

But this is all a result of in-text citations not being really integrated into the spec.

[denismaier]

Ok. Looks like we need to amend the spec here. The ibid in note 2 is not wanted. I think citeproc-js suppresses that already. See citation-style-language/documentation#121

[jgm]

I think I can get that behavior by modifying citeproc.

[jgm]

OK, I've fixed it in citeproc.