Hi:
* version: pandoc 2.1.3, duplicated on pandoc HEAD
* files: https://gist.github.com/rbdixon/490e2bc39be0b93efc2bc79ade5d497b
I put tables in YAML metadata blocks as multiline strings to aid integration of tools that generate Markdown that is then processed by Pandoc. I've been chasing an odd behavior that I've narrowed down to pandoc's paring of tables in YAML metadata blocks.
This example (test_passes.md) works:
---
table: |-
Table: Title
| A | B |
|:--|:--|
| A | B |
| A | B |
| A | B |
x
...
Text
When I inspect the Pandoc JSON I see that it has parsed a table with a header and three rows. This example (test_fails.md) does not work:
---
table: |-
Table: Title
| A | B |
|:--|:--|
| A | B |
| A | B |
| A | B |
...
Text
The difference is the removal of "\nx\n" at the end of the multiline string. When I inspect the Pandoc JSON the last row of the table is split off into its own `Plain` stanza rather than being included, as expected, in the `Table` stanza.
Any advice on what I might be doing wrong here? Perhaps I'm abusing multiline strings and their markdown interpretation a bit.
Thanks,
Brad