[R-SIG-Mac] CRAN LaTeX infrastructure for packages

[Simon Urbanek]

Ben,

it is not an oversight in that I'm not aware of any agreed set of optional packages. I'm also not aware of the discussion you quote from 2017, but if there is anything that you need that is not in the standard set then all you need to do is tell me that you require it.

The change is that the most recent builds are now done on VMs that are now setup fully automatically (=no manual intervention) and use TinyTeX instead of MacTeX to make maintenance easier - FWIW  see
https://github.com/R-macos/vm-scripts-mini-r/blob/master/shared/setup.sh

If you need any packages that are not a part of the standard TinyTeX bundle then you just need to tell me (or you can ask TinyTeX to include it so it is available to others as well).

Cheers,
Simon


> On 3/12/2025, at 1:10 PM, Ben Bolker <bbolker using gmail.com> wrote:
> 
>  I'm not a regular reader of r-sig-mac, but have a question about CRAN LaTeX infrastructure for MacOS.
> 
>  This may be a question specifically for Simon Urbanek, but thought I'd check here first in case anyone has insights.
> 
>  I recently sent a new version of the lme4 package (1.1-38) to CRAN; it passed all the incoming tests, but I'm now seeing warnings (at https://cran.r-project.org/web/checks/check_results_lme4.html) from the r-release-macos-arm64  and r-oldrel-macos-arm64  (but *not* from the x86_64 platforms):
> 
> >   ! LaTeX Error: File `blkarray.sty' not found.
> 
> This hasn't happened before.  Furthermore, when I go back in the git log for the package I see:
> 
> commit 3140cc0f1b394c6888d7704ee3d5de5f45df2886
> Author: Ben Bolker <bbolker using gmail.com>
> Date:   Thu Aug 10 10:28:55 2017 -0400
> 
>    .Rbuildignore blkarray.sty as per CRAN request
> 
> In particular, Kurt Hornik sent a mass e-mail in 2017 to many package maintainers saying:
> 
> >  These [packages] contain copies of standard LaTeX .sty files which should be excluded from the package (as shipping these copies should not be necessary, and may even cause trouble in case the local copies become outdated [as just happened for a recent CRAN submission]).
> 
>  So my best guess is that this is an oversight/platform glitch on the arm64 platforms ...
> 
>  Thoughts?
> 
>  sincerely
>   Ben Bolker
> 
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>