We use the gptk package to produce the initialization fit of GP to the data. The latest R devel version has made a change to how the class() function works. Since gptk uses class(), and we use gptk, this led to the following issue as reported by the CRAN maintainers:
Specifically, see the problems shown for the r-devel Debian checks.
These can be reproduced by checking with --as-cran using current
r-devel, which for now sets
For now only active when environment variable _R_CLASS_MATRIXARRAY
is set to something true, but planned to be the new unconditional
behavior when R 4.0.0 is released:
matrix objects now also inherit from class "array", namely, e.g.,
class(diag(1)) is c("matrix", "array") which invalidates code
assuming that length(class(obj)) == 1, an incorrect assumption that
is less frequently fulfilled now.
S3 methods for "array", i.e., .array(), are now also
dispatched for matrix objects.
Apparently your package no longer works correctly when
class(matrix(...)) gives a vector of length two: please fix as
necessary.
For now this only produces a warning, so I have worked around it by disabling the unit tests that checked that we were getting no warnings, but we need a longer-term solution. Is there another GP fitting package that we could use instead?
We use the
gptk
package to produce the initialization fit of GP to the data. The latest R devel version has made a change to how theclass()
function works. Sincegptk
usesclass()
, and we usegptk
, this led to the following issue as reported by the CRAN maintainers:For now this only produces a warning, so I have worked around it by disabling the unit tests that checked that we were getting no warnings, but we need a longer-term solution. Is there another GP fitting package that we could use instead?