Deal with reduced dependency fallout

[hadley]

• riverdist — Add tibble dependency mbtyers/riverdist#22
• TreeRingShape — Add tibble dependency ishidamgm/TreeRingShape#3
• usethis: Use lifecycle instead of tibble usethis#2080
• Only use bit4 path if installed (httr2, filearray)

expect_vector():

• PLindleyROC — CRAN failure ErtanSU/PLindleyROC#1
• bigD — CRAN failure rstudio/bigD#14
• charcuterie — CRAN failure jonocarroll/charcuterie#1
• drape — CRAN failure harveyklyne/drape#1
• erfe — CRAN failure amadoudiogobarry/erfe#1
• gKRLS — CRAN failure mgoplerud/gKRLS#3
• kgrams — CRAN failure vgherard/kgrams#32
• mlr3fselect — CRAN failure mlr-org/mlr3fselect#125
• mobsim — CRAN failure MoBiodiv/mobsim#34
• r2r — CRAN failure vgherard/r2r#9
• ravetools — CRAN failure dipterix/ravetools#10
• serp — CRAN failure ejikeugba/serp#12
• translated — CRAN failure ttscience/translated#22
• vglmer — Failure on CRAN mgoplerud/vglmer#16

Not on GitHub:

• PGSD: [email protected]
• good: [email protected]
• sleev: [email protected]
• stepSplitReg: [email protected]

@olivroy unfortunately #196 ended up breaking a bunch of packages on CRAN because they used vctrs or tibble and got them implicitly via testthat + waldo. AND unfortunatley our revdepcheck server didn't set the right env var to pick this up 😞

[olivroy]

I was afraid this could happen, but was surprised that the revdep check didn't catch it...

[hadley]

Message I'm copying and pasting:

Unfortunately the recent waldo release broke your package on CRAN because it no longer imports tibble, and your package was implicitly depending on it. The easiest fix is to just make the dependency explicit.

Apologies for not discovering this before CRAN release, but we had a buglet in the code that runs our revdep checks 🙁.