[R-pkg-devel] Debugging M1mac problems

Ivan Krylov |kry|ov @end|ng |rom |@@er@chem@m@u@ru
Sat Mar 27 15:26:16 CET 2021


Dear R-package-devel,

My package "albatross" doesn't contain any compiled code and has been
checked on both a POSIX-compatible system and on Windows, so I wasn't
expecting any problems after seeing it pass CRAN "pretest" checks.
Imagine my surprise when it failed a test on M1mac while reading some
files that the test creates in a temporary directory:

https://www.stats.ox.ac.uk/pub/bdr/M1mac/albatross.out

Admittedly, I missed a chance to provide a useful test failure message
here, so I only know that the names differ from my expectations, not
what the differences are. I changed the function to output the actual
differences, but so far wasn't able to reproduce the failure;
apparently it's not a Mac-specific problem, since the package passed
the test on R-hub's "CRAN-like" Mac setup:

https://artifacts.r-hub.io/albatross_0.3-0.tar.gz-6df95a00d916dc9c5ff28bdfe2a1e217/albatross.Rcheck/

I had an idea to rent some CPU time on an M1 machine (buying my own not
being an option). The Web is full of ads for Scaleway offering exactly
this service, but when I checked, they turned out to be out of stock.

I'm probably doing something stupid with the file names that usually
works but isn't guaranteed to work in all cases, but I cannot figure it
out. What should I do to debug this problem? I really don't want to be
bothering CRAN with small updates that only help diagnose the problem
instead of fixing it.

Please keep me Cc:-d in your replies if possible.

-- 
Best,
Ivan



More information about the R-package-devel mailing list