[R-SIG-Mac] Problem with CHECK --as-cran on Catalina

[Jeff Newmiller]

AFAIK tgz is one possible file name extension equivalent to .tar.gz, which implies gzip compression, so I think Bob was just being a bit sloppy when he wrote tar.tgz.

Key point is not that a particular compression makes a difference, but that the check needs to be run on a source archive rather than the in-place source you have been editing. He was confirming that you are making that file and specifying it when checking.

On June 6, 2020 11:59:24 AM PDT, Carl Witthoft <carl using witthoft.com> wrote:
>I plead ignorance here: I don't see an option to create .tgz . The
>R CMD BUILD --help just lists "gzip" (default), "none", "bzip2", "xz" 
>.Does one of those create the .tgz -type compression?
>
>Carl
>
>On 6/6/20 2:37 PM, Bob Rudis wrote:
>> Does it still do it if you build a tar.tgz archive and run it on
>that?
>> 
>> I could not dup your problem, but when I tried to run it on a couple
>> of my packages that work fine with tar.gz (I use Authors using R) in dir
>(`R
>> CMD check --as-cran .` or parent dir and specify the pkg dir) I get:
>> 
>>      Required field missing or empty:
>>        ‘Author’
>> 
>> (running last R 4.0.1 beta before the release of R 4.0.1, latest
>> Catalina beta AND R 4.0.0 on ubuntu)
>> 
>> So, def something strange going on.
>> 
>> On Sat, Jun 6, 2020 at 12:47 PM Carl Witthoft <carl using witthoft.com>
>wrote:
>>>
>>> Hi,
>>> I recently upgraded to a new iMac w/ Catalina.  Something really
>strange
>>> is happening when I try to run "R CMD CHECK --as-cran  " .
>>> If I run it on various older packages that have previously passed
>the
>>> check, all is well.
>>>
>>> I am writing a new package, and here's the response I get:
>>>
>>>
>>> * using log directory ‘/Users/cgw/Rgames/fitConic.Rcheck’
>>> * using R version 4.0.0 (2020-04-24)
>>> * using platform: x86_64-apple-darwin17.0 (64-bit)
>>> * using session charset: UTF-8
>>> * using option ‘--as-cran’
>>> * checking for file ‘fitConic/DESCRIPTION’ ... OK
>>> * this is package ‘fitConic’ version ‘1.0’
>>> * checking CRAN incoming feasibility ...Error:  file './DESCRIPTION'
>>> does not exist
>>> Execution halted
>>>
>>>
>>> This happens whether I run R CMD CHECK  from the parent dirctory
>>> /Users/cgw/Rgames  or from the source directory
>/Users/cgw/Rgames/fitConic .
>>>
>>> I verified I can edit the DESCRIPTION file in my previous packages &
>>> still pass the check.  I even took one of those old DESCRIPTION
>files
>>> and put it into the fitConic directory, and changed only the
>"Package: "
>>> line  .
>>>
>>> I'm guessing that last error message is not telling me what actually
>>> went wrong, but I'm at a loss as to why this new package fails. 
>I've
>>> created a new directory and moved all my files (DESCRIPTION,
>NAMESPACE,
>>> and the R & man folders) over. No luck.
>>>
>>> Is this a hangup somewhere in R 4.0 for OSX, or is there something
>else
>>> going on?   (do I need, for example to do something to give XQuartz
>more
>>> permissions, or some other called app? )
>>>
>>> thanks
>>> --
>>> Carl Witthoft
>>> carl using witthoft.com
>>> resume: https://app.box.com/file/498153801347
>>>
>>> _______________________________________________
>>> R-SIG-Mac mailing list
>>> R-SIG-Mac using r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac

-- 
Sent from my phone. Please excuse my brevity.