[R-pkg-devel] R CMD check not finding my vignettes

[David Hugh-Jones]

Hi Duncan,

Thank you very much for taking the time to look at this.

I tried rebuilding the tar file so as to include only the .Rmd files, not
the HTML files, in 'vignettes':

drwxr-xr-x  0 david  staff       0 20 Apr 09:21 huxtable/vignettes/

-rw-r--r--  0 david  staff    1633  6 Apr 16:26
huxtable/vignettes/comparison.csv

-rw-r--r--  0 david  staff    6697  6 Apr 14:44
huxtable/vignettes/design-principles.Rmd

-rw-r--r--  0 david  staff    5521  6 Apr 14:30
huxtable/vignettes/huxreg.Rmd

-rw-r--r--  0 david  staff   20552  6 Apr 16:19
huxtable/vignettes/huxtable.Rmd

-rw-r--r--  0 david  staff      22 17 Mar 00:19
huxtable/vignettes/placeins-header.tex


But when I run R CMD check, I still get the same warning:

* checking package vignettes in ‘inst/doc’ ... WARNING

Package vignettes without corresponding PDF/HTML:

   ‘design-principles.Rmd’

   ‘huxreg.Rmd’

   ‘huxtable.Rmd’

So, the warning does not seem to be related to the presence of HTML files
in vignettes.

I also tried manually removing .Rmd files from inst/doc (leaving them only
in vignettes) but this still gave the same error.

My goal here is that for package-specific reasons, I would like both pdf
and HTML versions of my vignettes to be available (I am writing code that
prints HTML and LaTeX tables and want my users to have examples of how both
output formats work). This is why I build those files manually, and place
them in inst/doc.

Side comment: at the moment, I feel as if I am running through the
combinatorics of including and excluding files from vignettes and inst/doc,
without much insight into what I am doing. Would it be fair to say that the
current system is not very easy to comprehend?

Cheers,

David

On 19 April 2017 at 21:24, Duncan Murdoch <murdoch.duncan at gmail.com> wrote:

> On 19/04/2017 1:00 PM, David Hugh-Jones wrote:
>
>> Hi Uwe,
>>
>> I'm not sure if you ever got my off-list message with my tarball or
>> subsequent messages. I can't send a tarball on-list - it gets rejected as
>> too large - but here is a dropbox link. If you could confirm receipt, that
>> would be extremely helpful!
>>
>> https://www.dropbox.com/s/179jrm19kx9o7dz/huxtable_0.2.0.tar.gz?dl=0
>>
>>
> Not sure if Uwe has had a chance to look, but I just ran R CMD check on
> your tarball using the latest version of R-devel.  It reported the
> following problems:
>
> * checking CRAN incoming feasibility ... NOTE
> Maintainer: ‘David Hugh-Jones <davidhughjones at gmail.com>’
>
> Found the following (possibly) invalid URLs:
>   URL: http://cran.rstudio.com/web/packages/huxtable/index.html
>     From: README.md
>     CRAN URL not in canonical form
>   Canonical CRAN.R-project.org URLs use https.
>
> The URL for your package should be in the form
>
> https://CRAN.R-project.org/package=huxtable
>
> * checking top-level files ... NOTE
> Non-standard file/directory found at top level:
>   ‘multirow.rds’
>
> I'm not sure where that file belongs, but not there.
>
> * checking files in ‘vignettes’ ... WARNING
> Files in the 'vignettes' directory newer than all files in 'inst/doc':
>   ‘huxreg.html’, ‘huxtable.html’
>
> Those files shouldn't be in the vignettes directory, they are products of
> building the vignettes.  Only the source should normally be in the
> vignettes directory.
>
> * checking package vignettes in ‘inst/doc’ ... WARNING
> Package vignettes without corresponding PDF/HTML:
>    ‘design-principles.Rmd’
>    ‘huxreg.Rmd’
>    ‘huxtable.Rmd’
>
> I think this is the warning you were asking about.  If I look at just the
> first of those files, I see that design-principles.Rmd exists in vignettes
> and is set to produce PDF output (only the first output entry in the YAML
> counts).  You also have a .html file in vignettes; it shouldn't be there.
> In inst/doc, you have source, .html and .pdf versions of that file.
>
> The warning isn't very helpful, but I think it is triggered by the fact
> that you've got the .html file in your vignettes directory.  If I remove
> everything but the source from that directory, and everything from the
> inst/doc directory, then all the vignette warnings go away.
>
> (I haven't traced through the code, but I think the warning may be
> literally correct.  Since you had .html files that looked like vignette
> outputs but weren't, you have vignettes without *corresponding* HTML. It
> would have been nicer if it suggested how to fix this, but that vignette
> code is quite tricky, because any file could be source, and any html, tex
> or pdf file could be output.  At some point I may try to clean it up a bit
> and then maybe the error messages will be less obscure.)
>
> Duncan Murdoch
>

	[[alternative HTML version deleted]]