Bump irlba version in renv

[sjspielman]

May or may not close #907

In this PR, I'm updating irlba to the most recent version. This time, I did the update on the RStudio Server, and there were no other changes besides irlba versioning in the snapshot....except for that one record which got changed from RSPM -> CRAN, but it is my understanding from comments in #905 that we accept our fate for that one.

To confirm, the normalization notebook which gave us the error in #906 is still running fine in this updated environment on the server for me.

Once this updated Docker image gets pushed, we can try again to run the GHA in #906 and see if this made an impact.

[sjspielman]

I'm not sure why this is now failing at the apt step...seems like it could be intermittent? Will come back and re-trigger in a bit to see if it sorts itself out. If not, I'll revert 603abb2 and try again (although it looks like boost libraries are the problem and this new dependency was grabbed just fine).

[jashapiro]

Since the failure was in the R package compilation, this actually needs to be added to the final build phase, (after FROM bioconductor/...) at line ~60.

[sjspielman]

arm continues to not be happy here. (...do we need to keep that build around?)

[jashapiro]

arm continues to not be happy here. (...do we need to keep that build around?)

We should keep it around, yes. It makes for a much smoother experience for Mac users who might want to use the image locally.

I updated to the latest RSQLite (2.4.6) and the image built find locally. So I would do that. As long as we are updating a lot of packages for this image and the new R version, we might as well have everything as up to date as possible. A round of renv::update() here should be fine.

[sjspielman]

no longer an RSQLite dependency

[jashapiro]

👍🏼