Thoughts on migrating from `read_sparse_block()` to SVTs about delayedarray

[hpages]

None of this is really expected to work properly at the moment. There's a long road before we can get there (see https://github.com/Bioconductor/DelayedArray/blob/devel/TODO). I'll coordinate with you when it's time to make the switch from SparseArraySeed to SVT_SparseArray. In the meantime, please ignore extract_sparse_array() and read_block_as_sparse().

[LTLA]

Oh. I started using it in BioC-devel a few weeks ago. Seemed okay.

More specifically, beachmat now uses extract_sparse_array() to extract SVT_SparseMatrix objects for abstract R matrices that the new tatami interface doesn't otherwise know how to handle. SingleR - the only package currently updated to use the new interface - seems pretty happy with it, so I assumed we were ready to go.

Guess I'll just leave it in there for now. Looks mature enough that it'll be in good shape for the next BioC release.