Bug with referencing inside a definition list

[adityam]

Consider the following markdown file

Assumption #big

:   Big assumption

Lemma #

:   A random lemma

Lemma #

:   A random lemma

Now we derive a result which uses [Assumption #](#assumption:big).

Lemma #

:   Under [Assumption #](#assumption:big), we have ... 

Processing it using pandoc --filter pandoc-numbering --to=markdown gives

[**Assumption 1**]{#assumption:big .pandoc-numbering-text .assumption}

:   Big assumption

[**Lemma 1**]{#lemma:1 .pandoc-numbering-text .lemma}

:   A random lemma

[**Lemma 2**]{#lemma:2 .pandoc-numbering-text .lemma}

:   A random lemma

Now we derive a result which uses [Assumption 1](#assumption:big).

[**Lemma 3**]{#lemma:3 .pandoc-numbering-text .lemma}

:   Under [Assumption 3](#assumption:big), we have ...

Note that in Lemma 3, the "Assumption #" got translated to "Assumption 3" rather than "Assumption 1".

[aubertc]

I can reproduce your bug.
You can side-step it using the short-cut notation.
I'm assuming it really is a bug, and not a feature, since the "usual" way of referencing your assumption (cf. this page) still creates a bug.

Assumption #big

:   Big assumption

Lemma #

:   A random lemma

Lemma #

:   A random lemma

Now we derive a result which uses [%D %n](#assumption:big).

Lemma #

:   Under 

    - [Assumption #](#assumption:big), your example, does not work
    - [%D %n](#assumption:big), this solution should work, but does not either.
    - @assumption:big, this works.

produces:

Assumption 1

Big assumption

Lemma 1

A random lemma

Lemma 2

A random lemma

Now we derive a result which uses Assumption 1.

Lemma 3

Under

• Assumption 3: your example, does not work
• Lemma 3, this solution should work, but does not either.
• Assumption 1, this works.

Note that if you are using pandoc-citeproc at the same time, the order of the filters is important.

[adityam]

Thanks! This workaround is sufficient for me.

[chdemko]

This is a bug. I have to look at it

[adityam]

Thanks a lot!