[R-pkg-team] Bug#1070238: r-cran-irlba: Please rebuild under updated Matrix package

[Dirk Eddelbuettel]

Package: r-cran-irlba
Version: 2.3.5.1-3
Severity: important

CRAN package Matrix had a new release 1.7.0 bringing in a new
SuiteSparse API which requires a rebuild if (and only if) the
Matrix headers are used. Your package is one of those that do,
and therefore needs a rebuild.

This was reasonably well circulated earlier (by upstream in [1],
and I followed up on debian-devel) but it was then decided to tie 
this 1.7-0 release to the R 4.4.0 release last week.

To recap we can look at the total dependencies of Matrix at CRAN (1300+)
and the ones using the headers via LinkingTo (15) in R via 

  > db <- tools::CRAN_package_db()
  > matrixrevdep <- tools::package_dependencies("Matrix", reverse=TRUE, db=db)[[1]]
  > length(matrixrevdep)    # the vector 'matrixrevdep' list all
  [1] 1349
  > tools::package_dependencies("Matrix", which = "LinkingTo", reverse = TRUE)[[1L]]
   [1] "ahMLE"               "bayesWatch"          "cplm"                "GeneralizedWendland"
   [5] "geostatsp"           "hibayes"             "irlba"               "lme4"               
   [9] "mcmcsae"             "OpenMx"              "PRIMME"              "PUlasso"            
  [13] "robustlmm"           "spGARCH"             "TMB"                
  > 

But among these 15 affected only five are in Debian:
 
   irlba    r-cran-irlba
   lme4     r-cran-lme4
   OpenMx   r-cran-openmx
   TMP      r-cran-tmp
   bcSeq    r-bioc-bcseq

and lme4 is my package, and I already rebuilt it.

You should see a message about 'Matrix API 1, needs 2' in case of
a mismatch, if you rebuild then the `library(...)` call in R will 
be quiet as usual.

All it takes is a rebuild: for r-cran-lme4 I just adjusted (Build-)
Depends for r-cran-matrix to r-cran-matrix (>= 1.7-0) (and I also 
adjusted r-base-dev to depend on 4.4.0 or greater, but that is both
optional, and implied via Matrix).

It would be terrific if you could update the package in the next
few days. If you are unable I could do a binary-only NMU -- just
let me know.

Many thanks,  Dirk

[1] https://stat.ethz.ch/pipermail/r-package-devel/2024q1/010463.html

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