Pandoc 2 changes the behavior of knitr chunk option indent for html_document

[jdblischak]

By filing an issue to this repo, I promise that

• I have fully read the issue guide at https://yihui.name/issue/.
• I have provided the necessary information about my issue.
  • If I'm asking a question, I have already asked it on Stack Overflow or RStudio Community, waited for at least 24 hours, and included a link to my question there.
  • If I'm filing a bug report, I have included a minimal, self-contained, and reproducible example, and have also included xfun::session_info('rmarkdown'). I have upgraded all my packages to their latest versions (e.g., R, RStudio, and R packages), and also tried the development version: remotes::install_github('rstudio/rmarkdown').
  • If I have posted the same issue elsewhere, I have also mentioned it in this issue.
• I have learned the Github Markdown syntax, and formatted my issue correctly.

The hidden knitr chunk option indent behaves differently for html_document() between pandoc 1 and 2. I realize this is a minor issue since indent is only supposed to be used for markdown output. Also, it could be argued that the pandoc 2 behavior is preferred. But since I observed it, I wanted to share the result in case anyone depended on the pandoc 1 behavior.

Reproducible Rmd file

The examples below use the following Rmd file, which I saved as test-indent.Rmd:

---
output:
  html_document:
    keep_md: true
---

```{r example, indent="  "}
1 + 1
```

pandoc 1 preserves the indentation in the HTML file

Rendering this document with pandoc 1.19.2.1 preserves the indentation in the HTML file:

$ Rscript -e 'rmarkdown::render("test-indent.Rmd", quiet = TRUE)'
$ grep '1 + 1' test-indent.*
test-indent.html:<p><code>r   1 + 1</code></p>
test-indent.md:  1 + 1
test-indent.Rmd:1 + 1

which causes the code chunks to be malformed (screenshot below):

> xfun::session_info(c("knitr", "rmarkdown"))
R version 3.5.1 (2018-07-02)
Platform: x86_64-conda_cos6-linux-gnu (64-bit)
Running under: Ubuntu 18.04.2 LTS

Locale:
  LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
  LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
  LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
  LC_PAPER=en_US.UTF-8       LC_NAME=C                 
  LC_ADDRESS=C               LC_TELEPHONE=C            
  LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

Package version:
  base64enc_0.1.3 digest_0.6.18   evaluate_0.13   glue_1.3.1     
  graphics_3.5.1  grDevices_3.5.1 highr_0.8       htmltools_0.3.6
  jsonlite_1.6    knitr_1.22      magrittr_1.5    markdown_0.9   
  methods_3.5.1   mime_0.6        Rcpp_1.0.1      rmarkdown_1.12 
  stats_3.5.1     stringi_1.4.3   stringr_1.4.0   tinytex_0.11   
  tools_3.5.1     utils_3.5.1     xfun_0.5        yaml_2.2.0     

Pandoc version: 1.19.2.1

For conda users, I created this environment with the following command:

conda create -y -n test-pandoc1 -c conda-forge pandoc=1.19.2.1 r-rmarkdown

pandoc 2 removes the indentation in the HTML file

Rendering this document with pandoc 2.7.1 removes the indentation in the HTML file:

$ Rscript -e 'rmarkdown::render("test-indent.Rmd", quiet = TRUE)'
$ grep '1 + 1' test-indent.*
test-indent.html:<pre class="r"><code>1 + 1</code></pre>
test-indent.md:  1 + 1
test-indent.Rmd:1 + 1

which causes the code chunks to properly formatted (screenshot below):

> xfun::session_info(c("knitr", "rmarkdown"))
R version 3.5.1 (2018-07-02)
Platform: x86_64-conda_cos6-linux-gnu (64-bit)
Running under: Ubuntu 18.04.2 LTS

Locale:
  LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
  LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
  LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
  LC_PAPER=en_US.UTF-8       LC_NAME=C                 
  LC_ADDRESS=C               LC_TELEPHONE=C            
  LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

Package version:
  base64enc_0.1.3 digest_0.6.18   evaluate_0.13   glue_1.3.1     
  graphics_3.5.1  grDevices_3.5.1 highr_0.8       htmltools_0.3.6
  jsonlite_1.6    knitr_1.22      magrittr_1.5    markdown_0.9   
  methods_3.5.1   mime_0.6        Rcpp_1.0.1      rmarkdown_1.12 
  stats_3.5.1     stringi_1.4.3   stringr_1.4.0   tinytex_0.11   
  tools_3.5.1     utils_3.5.1     xfun_0.5        yaml_2.2.0     

Pandoc version: 2.7.1

For conda users, I created this environment with the following command:

conda create -y -n test-pandoc2 -c conda-forge pandoc=2 r-rmarkdown

[atusy]

I guess the behavior in Pandoc 2 is the way supposed to be.
knitr manual says

Currently this option is only used to indent markdown output, because leading white spaces have special meanings in markdown. https://yihui.name/knitr/options/#code-decoration

Utilizing "special meaning" by indent = " " (4 spaces) will output markdown like

    ```r
    1 + 1
    ```

and output html preserving fences like below.

```r
1 + 1
```

I guess Pandoc 1.x is outputing something like

```r
    1 + 1
```

If so, there is no way to use "special meaning" of indents.

Would you check indents of fences?

[jdblischak]

Would you check indents of fences?

@atusy I'm not exactly sure what you are asking. Do you want to know how the backticks are indented in the Markdown file? This result has not changed between Pandoc 1 and 2. No matter which you use, the output Markdown file will have everything indented: the code chunk, the results it creates, and the backticks around each. See below.

Is that what you wanted to know, or did you have another question?

Pandoc 1

$ pandoc -v | head -n 1
pandoc 1.19.2.1
$ cat test-indent.md 
---
output:
  html_document:
    keep_md: true
---

  
  ```r
  1 + 1
  ```
  
  ```
  ## [1] 2
  ```

Pandoc 2

$ pandoc -v | head -n 1
pandoc 2.7.2
$ cat test-indent.md 
---
output:
  html_document:
    keep_md: true
---

  
  ```r
  1 + 1
  ```
  
  ```
  ## [1] 2
  ```

[atusy]

Yes, that's what I meant.

It seems like a change in Pandoc 2.0.2

Markdown reader:

Allow fenced code blocks to be indented 1-3 spaces (#4011). This brings our handling of them into alignment with CommonMark’s.

https://pandoc.org/releases.html#pandoc-2.0.2-12-november-2017

[jdblischak]

It seems like a change in Pandoc 2.0.2

@atusy Thanks for investigating! It does appear that the change in the behavior of indent arose from the changes introduced in response to Issue jgm/pandoc#4011.

Since this change in behavior for processing indented code fences is documented in this thread and the main pandoc repository, hopefully this is sufficient explanation to anyone searching for this.

[github-actions]

This old thread has been automatically locked. If you think you have found something related to this, please open a new issue by following the issue guide (https://yihui.org/issue/), and link to this old issue if necessary.