Add RMarkdown support to ESS layer.

[cpaulik]

No description provided.

[izahn]

Why just RMarkdown? polymode supports Sweave, org-mode, and several other
literate programming systems. And it is not specific to R either, so
polymode should probably have its own layer.

I'm not against merging this as a temporary measure, but I do think it
needs some thought and a coherent way to integrate polymode into spacemacs
generally.
On Feb 18, 2016 10:52 AM, "Christoph Paulik" notifications@github.com
wrote:

---

You can view, comment on, or merge this pull request online at:

#5126
Commit Summary

• Add RMarkdown support to ESS layer.

File Changes

• M layers/+lang/ess/README.org
  https://github.com/syl20bnr/spacemacs/pull/5126/files#diff-0 (8)
• M layers/+lang/ess/packages.el
  https://github.com/syl20bnr/spacemacs/pull/5126/files#diff-1 (11)

Patch Links:

• https://github.com/syl20bnr/spacemacs/pull/5126.patch
• https://github.com/syl20bnr/spacemacs/pull/5126.diff

—
Reply to this email directly or view it on GitHub
#5126.

[cpaulik]

I was just playing around with RMarkdown and saw that it was not supported. Since it seems more clumsy than using org mode directly I have no real need for it. So no strong opinion on this from my side.

[roryk]

polymode doesn't work very well with evil-mode: polymode/polymode#41

[izahn]

@roryk I can't reproduce the problems described in polymode/polymode#41. Can you? It is true that insert/normal state is not synchronized across indirect buffers, but I don't really see that as a problem, and certainly not something that should block integration in spacemacs.

[roryk]

Yeah, I totally can, I enabled polymode again to edit a Rmarkdown file to check to see if it was fixed and disabled it a couple of days ago because it wasn't. It seems like it works fine at first but has problems. I'm not sure what is exactly how to reproduce the behavior, but you can get into a state where undo gets destroyed. It might be you have to get to a more complicated undo history like undoing in the two different blocks before it happens.

[roryk]

I can't seem to reproduce it though, I'll keep it turned on for a while and see if I can figure out when and why it is happening.

The insert/normal state not being synced is pretty brutal. I can't even make named blocks:

```{r

traps me in a look up the function mode in spacemacs.

[izahn]

On Thu, Feb 25, 2016 at 1:40 PM, Rory Kirchner notifications@github.com
wrote:

I can't seem to reproduce it though, I'll keep it turned on for a while
and see if I can figure out when and why it is happening.

The insert/normal state not being synced is pretty brutal. I can't even
make named blocks:

traps me in a look up the function mode in spacemacs.

​For me it just switches to ​normal state as soon as polymode recognizes
the delimiter, requiring an extra "i" to get back to insert state. Is that
what you are describing, or another problem? For me this is annoying but
not a deal breaker.

—
Reply to this email directly or view it on GitHub
https://github.com/syl20bnr/spacemacs/pull/5126#issuecomment-188922942.

[roryk]

Yup, you are right.

[vspinu]

The issue has been fixed on the polymode side, at least to the extent I can test it as a non evil user.

[biocyberman]

Org-mode seems to be more stable than knitr; it supports many other languages; and managing document structure with it is very pleasing. So I tried to use org-mode to write org+R documents as a replacement for rmarkdown. However, here are the reasons I switched back to knitr (rmarkdown) for now:

• I could not find out how to highlight R code with colors.
• Knitr supporst better when I want to include plots from R code chunks, adding figure captions.

So yes, I still want to have support for Rmarkdown/knitr in ess. Implementing this pull request (enabling polymode) with some slight modifications, I find that polymode does not play well with evil-mode when:

• opening a code chunk. After ````{r`, the whole line start jumping to the right on any keystroke.
• yanking and pasting. If the text content contains a whole code chunk, i.e. including the start and the end ```, the paste is truncated to the first line of the yanked region.

Switching to emacs key (by looping with Ctrl+z) eliminates these problems. But you know that, vim key bindings rock. I can't live without evil.

[cpaulik]

For syntax highlighting you must have the ess layer installed and probably also do (add-to-list 'org-babel-load-languages '(R . t)) somewhere, actually this should be done by the ESS layer but isn't at the moment. I don't know about the figure captions but this may help. What exactly was the problem with the figure captions?

[biocyberman]

@cpaulik Clarification: Syntax highlight while editing is working fine. I was taking about syntax highlighting for code chunks in rendered PDF or HTML documents.

[cpaulik]

@biocyberman Strange. HTML syntax highlighting should work fine out of the box AFAIK. For latex you have to use the minted package.

Something like

      (setq org-latex-listings 'minted)
      ;; setup minted to have frame, small text and line numbers
      (setq org-latex-minted-options
                 '(("frame" "lines")
                   ("fontsize" "\\scriptsize")
                   ("linenos" "")))

      ;; setup of latex processing
      (setq org-latex-pdf-process '("latexmk -synctex=1 -shell-escape -pdflatex=xelatex -pdf %f"))

in your config should do if you use latexmk. Otherwise change the org-latex-pdf-process variable accordingly.

[izahn]

I can reproduce the evil-yank problem, I guess it should be reported upstream. I don't see anything like

After ```{r, the whole line start jumping to the right on any keystroke.

@biocyberman can you give more detailed reproduction steps for this one?

[biocyberman]

@cpaulik and @izahn Don't want to hit and run, just because I don't work with R lately so I do not have input to reproduce this problem.

[Compro-Prasad]

@cpaulik @biocyberman Any updates on this PR?

[Compro-Prasad]

@duianto This seems a simple one. Can it be merged?
@cpaulik Please add an entry to CHANGELOG.develop and rebase.

[duianto]

I tried opening a .rmd example file:
https://raw.githubusercontent.com/tgirke/GEN242/gh-pages/_vignettes/07_Rbasics/sample.Rmd

But it opens in fundamental mode and the messages buffer shows:

Loading c:/Users/username/.emacs.d/elpa/26.2/develop/julia-mode-20190407.2119/julia-latexsubs...done
File mode specification error: (error Autoloading file c:/Users/username/.emacs.d/elpa/26.2/develop/ess-20190616.1211/ess-site.elc failed to define function poly-markdown+r-mode)

I found this page:
https://www.reddit.com/r/spacemacs/comments/9ciefe/polymode_for_rmd/

Where the comment says:

It looks like Polymode received a major update. You need to specify now the poly-modes you want to install (see here)

Thus, for .Rmd file support it should be

polymode
poly-R
poly-noweb
poly-markdown

When I add the last three packages (polymode is added to the ess layer by this PR).

Now the same .rmd file opens in markdown major mode with poly-markdown+r-mode minor mode.

All three packages might not be needed, because the same thing happens when only poly-R is added.

Syntax highlighting issue

The author of the reddit thread seems to have had some syntax highlighting issues:

If I open a .Rmd, some chunks have proper syntax highlighting while others don't, and adding more text/code results in somewhat sporadic behavior. Code highlights/unhighlights unexplicably.

And there's a small syntax highlighting issue for me as well. The highlighting doesn't appear until the cursor is moved. I even tried waiting 30 seconds but nothing happened until the cursor was moved, so it doesn't seem to be delayed.

When the highlighting does appear then it seems to be fine. Except for one place where it seems like there's an attempt at adding an indented code block within a code block:

    ```
	```\{r code_chunk_name, eval=FALSE\}
	x <- 1:10
	```
    ```

(everything in the code block is indented 4 spaces to make the backticks visible)
Then the highlighting marks everything until the start of the next code block as code. But this is unrelated to this PR, it seems to be a limitation (unsupported feature) in markdown-mode.

System Info 💻

• OS: windows-nt
• Emacs: 26.2
• Spacemacs: 0.300.0
• Spacemacs branch: develop (rev. 4b20bb791)
• Graphic display: t
• Distribution: spacemacs
• Editing style: vim
• Completion: helm
• Layers:

(autohotkey colors emacs-lisp ess git helm markdown multiple-cursors org spell-checking syntax-checking treemacs version-control)

• System configuration features: XPM JPEG TIFF GIF PNG RSVG SOUND NOTIFY ACL GNUTLS LIBXML2 ZLIB TOOLKIT_SCROLL_BARS THREADS LCMS2

[vspinu]

Issues with polymode are better reported on polymode tracker.

I found this page:

Polymode has been developed actively for the past months, so most of the issues people had in the past are either resolved or are no longer relevant. Setups and configurations in older posts/blogs might no longer be correct either.

All three packages might not be needed, because the same thing happens when only poly-R is added.

Only poly-R is needed. All others are dependencies of poly-R. Assignment into auto-mode-alist is not needed either, it's done automatically.

The highlighting doesn't appear until the cursor is moved.

This could come from the emacs re-display engine. In emacs 26 I have started seeing delays in re-display on various occasions in non-polymode buffer. I have never seen them in polymode buffers though. User input trigers the re-display so it's not a such a big deal. I haven't seen such issues with emacs 27 though. So maybe it has been solved already.

Except for one place where it seems like there's an attempt at adding an indented code block within a code block:

Fenced code blocks cannot be nested. It doesn't make much sense in general. Indented blocks do work.

[duianto]

Issues with polymode are better reported on polymode tracker.

This seems to just be a setup issue in Spacemacs.

When I add poly-R to the ess layers ess-packages variable, and an entry for:

(defun ess/init-poly-R ())

Then it works without having to manually add the package to the dotspacemacs-additional-packages variable in .spacemacs.

---

Only poly-R is needed. All others are dependencies of poly-R.

Confirmed, it still works when these are commented out:

• polymode from the ess-packages variable
• and the ess/init-polymode function

(defun ess/init-polymode ()
  (use-package poly-R
    :defer t)
  (use-package poly-markdown
    :defer t)
  )

---

This could come from the emacs re-display engine. In emacs 26 I have started seeing delays in re-display on various occasions in non-polymode buffer. I have never seen them in polymode buffers though. User input trigers the re-display so it's not a such a big deal. I haven't seen such issues with emacs 27 though. So maybe it has been solved already.

Confirmed, it seems to be an Emacs 26 issue, syntax highlighting are present instantly when the sample.Rmd file opens with the latest snapshot:
GNU Emacs 27.0.50 (build 1, x86_64-w64-mingw32) of 2019-06-09

---

Fenced code blocks cannot be nested. It doesn't make much sense in general. Indented blocks do work.

Then it's just an anomaly by the author of the sample file:
https://raw.githubusercontent.com/tgirke/GEN242/gh-pages/_vignettes/07_Rbasics/sample.Rmd

[github-actions]

This PR has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Please let us know if this PR is still need merging!