new test-ADlaplace error in use of `chm_factor_ldetL2` via `lmer`

[paciorek]

In addition to a reoccurrence of the "VECTOR_ELT" error we've seen before we now also have this:

── 2. Error ('test-ADlaplace.R:1154:3'): simple LME case works ─────────────────
505
Error in `initializePtr()`: function 'chm_factor_ldetL2' not provided by package 'Matrix'
506
Backtrace:
507
     ▆
508
  1. └─lme4::lmer(y ~ x + (1 + x || g), REML = FALSE) at test-ADlaplace.R:1154:3
509
  2.   ├─base::do.call(...)
510
  3.   └─lme4 (local) `<fn>`(...)
511
  4.     ├─base::do.call(...)
512
  5.     └─methods (local) `<rfMthdDf>`(...)
513
  6.       └─methods::new(def, ...)
514
  7.         ├─methods::initialize(value, ...)
515
  8.         └─methods::initialize(value, ...)
516
  9.           └─.Object$initialize(...)
517
 10.             └─lme4 (local) initializePtr()
518

[paciorek]

Not clear if this has been addressed by the vector_elt-related work done by Perry. I don't think we've seen this recently but possibly masked by the other test-ADlaplace Heisenbug.

[paciorek]

Closing this as not seen recently.