Closed wleoncio closed 6 months ago
Hard deadline on 2024-03-10, as the package is failing on CRAN:
Dear maintainer,
Please see the problems shown on https://cran.r-project.org/web/checks/check_results_TruncExpFam.html.
The ERRORs for r-devel are from
r85979 | luke | 2024-02-24 01:18:27 +0100 (Sat, 24 Feb 2024) | 4 lines Move towards UseMethod no longer forwarding local variables from the generic. For now they are forwarded with promises that signal an error i forced.
Please correct before 2024-03-10 to safely retain your package on CRAN.
Best, -k
Example of useful error message:
══ Failed tests ════════════════════════════════════════════════════════════════ ── Error ('test-direct-sampling.R:7:3'): Original attributes are retrieved ───── Error in `mget(ls())`: getting UseMethod variable 'parms' from generic 'rtrunc_direct'; this is no longer supported Backtrace: ▆ 1. └─TruncExpFam::rtrunc(1e+06, mean = 1, sd = 2, faster = TRUE) at test-direct-sampling.R:7:3 2. ├─TruncExpFam::rtrunc_direct(n, family, ...) 3. └─TruncExpFam:::rtrunc_direct.normal(n, family, ...) 4. ├─TruncExpFam:::truncated_q(...) 5. │ └─base::paste0("trunc_", class(parms[["n"]])) 6. └─base::mget(ls())
See more here. Problem seems to be contained to gcc, as R-devel on clang doesn't fail. Regardless, must address within the fortnight.
To reproduce:
rhub::check_with_rdevel(env_vars = c("R_USEMETHOD_FORWARD_LOCALS" = "error"))
CI tests failing on develop
but passing on release
, so something probably went wrong in merging the hotfix branch.
Keep an eye on the CRAN check results for 1.1.1 before really closing this. Fix on develop
is independent of all this.
All good, closing:
This was received on 2024-02-23 as a message forwarded by Luke Tierney from the R-devel mailing list:
I'll study and try to reproduce the issue―so far our R-devel checks are fine, but as Luke said they are bound to fail soon―and release a fix ASAP.