fenced code blocks go missing

[ramnathv]

fenced code blocks go missing

[ramnathv]

Can you provide me with a reproducible example? A link to your Rmd file would do.

[piccolbo]

Just checked in

https://github.com/RevolutionAnalytics/RHadoop/blob/rmr-1.3/rmr/pkg/docs/introduction-to-vectorized-API.Rmd

Replace all occurrences of ### with --- and that's what I was trying to
slidify

Thanks

Antonio

On Wed, Jun 20, 2012 at 7:22 PM, Ramnath Vaidyanathan <
[email protected]

wrote:

Can you provide me with a reproducible example? A link to your Rmd file
would do.

---

Reply to this email directly or view it on GitHub:
#43 (comment)

[ramnathv]

Code blocks are showing up for me. I wonder what the issue then is. Just to be clear, did you insert --- before every slide? Can you try installing the latest version of slidify from github and tell me what happens? Send me links to the .md and .html files generated so that I can see what is going wrong

[piccolbo]

Hi,
I am totally open to the idea that I didn't put the --- in the right spots
or something, but the end results is that I can generate slides and they
break exactly as they should, but no code. I tried again and the code
blocks are missing from the md file. But if I generate it from Rstudio,
then the code blocks are there. I think it could be a problem with
directories. The current directory in R is RHadoop but the Rmd is in
RHadoop/rmr/pkg/docs/. It includes code chunks that are in
RHadoop/rmr/pkg/tests. When I knitr it from RStudio, the md file is
generate in the docs directory, but when I run slidify it is generated at
the R cwd, that is RHadoop, but that would make it hard for knitr to find
the code chunks. I am attaching the original Rmd, the file with the code
and the md generated by slidify at top level and the one generated by knitr
in the same dir as the md, but without the directory structure it may be
hard to repro unless my theory is completely wrong.

Antonio

On Thu, Jun 21, 2012 at 2:48 AM, Ramnath Vaidyanathan <
[email protected]

wrote:

Code blocks are showing up for me. I wonder what the issue then is. Just
to be clear, did you insert --- before every slide? Can you try
installing the latest version of slidify from github and tell me what
happens? Send me links to the .md and .html files generated so that I
can see what is going wrong

---

Reply to this email directly or view it on GitHub:
#43 (comment)

[piccolbo]

I see the attachment have not made it, let me figure this out.

[ramnathv]

I think your theory is correct. Currently, all paths are specified with respect to the root containing the source file. Since you are using external chunks, it is leading to an issue. I will implement more flexible directory specifications in the future, so that it is easy to accommodate cases like yours.

And yes, you don't need to specify --- explicitly, since I wrote a short internal script that automatically inserts --- before <h3> tags (not customizable currently to use other header levels) before running the slidify script.

Let me know if you were able to resolve your issue.

[ramnathv]

Can switch to the directory in which the Rmd file is residing and then run slidify. If that works, then the issue is clearly directories.

[piccolbo]

Positive, setwd to where the Rmd is fixes the issue. By the way, I was thinking about making sections the implicit slide delimiters too. Makes a lot of sense. h3 is arbritrary, any h level should be the default. If there is a title -> new slide.

[piccolbo]

Well I closed it because there is a viable workaround. You may want to reopen because it shouldn't be necessary.

[ramnathv]

The reason I chose <h3> is because it is the default slide heading in html5slides which is the default for slidify. It is arbitrary and I plan to make it more generic. The easiest way is for the user to specify the header level that should be used to split slides. One complication is the use of comments inside code blocks, which can lead to incorrect splitting. I need to think a little before I implement it.

Any h level leading to a slide limits the ability to use sub-headers within a slide. I am trying to keep things compatible with pandoc which follows a similar approach.

Let me know if you have any other suggestions or feedback.

[piccolbo]

My slide style is very simple: only graphics, math and code, no text. The text comes from the speaker's voice and channel interference is a bane to attention. Having sub-headers within a slide is therefore not possible. So this auto-segementation feature is useless to me and I think it supports a repurposing of documents that should be discouraged. I suspect the majority of your potential users to differ at least in part with me on this issue, but I am probably a more experienced public speaker than your average user (http://piccolboni.info/speaking). So it depends whether you agree with me and whether you want to write a more "opinionated" package like ggplot2, which enforces a number of personal ideas as to what good statistical graphics is, or a more neutral package that gives more rope, ahem, choice to its users. If it were for me, I would move on to other features (presenter display!).

[ramnathv]

I see your point. Let me think about it. To be frank, my style is to use very little text as well. However, I use headers of levels 4 and above to keep the styling easy.

Presenter display is available in some of the frameworks. I am not well versed in javascript, so will have to depend on existing frameworks that already do it. I will write up some documentation as soon as I get the chance.