[R-pkg-devel] NOTE: Installation took CPU time 3.5 times elapsed time

[Dirk Eddelbuettel]

On 27 July 2025 at 07:57, Andrew Johnson wrote:
| When submitting a package to CRAN, I�m seeing a note for the installation CPU time being greater than the elapsed time. Normally I�d assume this would be due to some configuration running the compilation in parallel, but I haven�t set this in my package at all. Is this something that anyone has seen before and would know what to change?
| 
| Check result is here: https://win-builder.r-project.org/incoming_pretest/boostmath_1.0.1_20250727_090857/specialChecks/gcc-san/package/00check.log

[ Aside: Interestingly that Windows check does not warn about excessive test time,
  whereas it does on my machine. I presume you opted to skip tests as CRAN to
  keep within their limit? ]

For tests and examples we know we are asked to over options(Ncpus) and alike;
I am not aware we can override a MAKEFLAGS="-j SOMELARGENUMBER" if it is set
so I am a little stunned.

The only thing that is different in your package is the use of UseLTO you now
added in this release.  If that this new in 1.0.1, maybe it is the difference
maker?  [ And I just validated it: removing that last line from DESCRIPTION
removes the compilation time NOTE. ] So that seems to be something you may
need to sort out with CRAN and their setup.
 
| And package source is here: https://github.com/andrjohns/boostmath

I sent you a courtesy PR as I noticed that we may be able to improve the
link(s) to Boost documentation.

Dirk

-- 
dirk.eddelbuettel.com | @eddelbuettel | edd using debian.org