[R-pkg-team] Bug#1041556: What is the architecture name in R what we call armel/armhf

[Dirk Eddelbuettel]

On 17 August 2023 at 07:24, Graham Inggs wrote:
| On Thu, 17 Aug 2023 at 05:52, Andreas Tille <andreas at an3as.eu> wrote:
| > I tried to skip two tests in r-cran-checkmate with this patch[1]
| > which is based on
| >
| >   arch <- R.version$arch
| >   identical(arch, "i386") || identical(arch, "i686") || identical(arch, "armel") || identical(arch, "armhf")
| 
| Stack Overflow suggested to check .Machine$sizeof.pointer
| which will return the pointer size in bytes,
| so 8 on 64-bit and 4 on 32-bit architectures.

Seconded to detect 'word size'. Closest to an R-provided 'is it 32bit' ?

The failed logs for armhf and armel also clearly show that the value from
R.version$arch is 'arm', see

   https://ci.debian.net/data/autopkgtest/testing/armel/r/r-cran-checkmate/36601942/log.gz

and its displayed text:

  68s R version 4.3.1 (2023-06-16) -- "Beagle Scouts"
  68s Copyright (C) 2023 The R Foundation for Statistical Computing
  68s Platform: arm-unknown-linux-gnueabi (32-bit)
                ^^^

That is the value returned from R.version$arch.  'arm', not 'armel' or 'armhf'. 

For completeness, the armhf run

   https://ci.debian.net/data/autopkgtest/testing/armhf/r/r-cran-checkmate/36588004/log.gz

has

  56s R version 4.3.1 (2023-06-16) -- "Beagle Scouts"
  56s Copyright (C) 2023 The R Foundation for Statistical Computing
  56s Platform: arm-unknown-linux-gnueabihf (32-bit)

R.version$platform has the full triplet, ie 'x86_64-pc-linux-gnu' for me here.

Dirk

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