rhdf5 dependency issue // bioRad scheduled for archiving

[adokter]

@peterdesmet @bart1 @iskandari FYI bioRad is scheduled for archiving on Feb 2 2025, because of issues in bioconductor dependency package rhdf5, see grimbough/rhdf5#154. If you identify a fix please post in that issue.

[peterdesmet]

Do the issues exist in previous versions of rhdf5? If not, we could specify a version in DESCRIPTION:

rhdf5 (<= x.x.x)

Or is rhdf5 going to be archived on bioconductor? If so, you could install from another source. R-universe is designed for this: https://bioc.r-universe.dev/rhdf5. A user can do this as such:

install.packages("rhdf5", repos = "https://bioc.r-universe.dev")

But I forgot how to specify that in DESCRIPTION though.

[bart1]

It seems there is work on the main issue in rhdf5 so I guess that should be resolved this time.

[adokter]

Fixes to rhdf5 should propagate to bioconductor over the next few days, hopefully that will solve this issue

[grimbough]

I don't see the "UBSAN" report for the CRAN check on bioRad, so I think the changes I made to rhdf have propagated and been successful.

The offending code has been in rhdf5 for many years, so I guess Prof Ripley has stepped up his tracking and eforcement of these issues for CRAN packages. Something further comes up, please don't hesitate to open another issue with rhdf5.

[adokter]

Hi @grimbough, thanks so much for your help and quick fix for this, really appreciate it!

You're right, CRAN has been adding more and more checks over the last few years, regularly triggering requests for fixes for code that has been around for a long time. It's sometimes a challenge when the time given to upload a fix is short ...

[adokter]

fixed by grimbough/rhdf5#154