[Bug]: Change in emmeans package causes an error in tern.gee

[rvlenth]

What happened?

The emmeans package supports both glmgee and geeglm objects, but it was inconsistent in that the latter used infinite d.f. while the former used finite d.f. I have a new update to emmeans that makes them consistent with one another and both now use the finite d.f. (See emmeans issue 496.)This changes some column names in the summaries -- specifically, the confidence limits are now named lower.CL and upper.CL instead of asymp.LCL and asymp.UCL, as is expected in the third line in the code for lsmeans.tern_gee_logistic().

Because the d.f. change was a bug fix in emmeans, I don't want to revert the change, so I am requesting that you change gee.tern so it is compatible.

Additional comment

You have an lsmeans() function that has the same name -- but is incompatible with -- a function of the same name in emmeans. This will confuse your users because they will get different results depending on which package is earlier in the search path:

> tern.gee::lsmeans(fev_fit, data = fev_data)
Error in `[.data.frame`(data.frame(stats::confint(prop_emm)), , c(object$vars$arm,  : 
  undefined columns selected
> # This is the current error message due to the issue reported here

> emmeans::lsmeans(fev_fit, data = fev_data)
Error in emmfcn(...) : argument "specs" is missing, with no default
> # Due to incompatibility between the two functions

Which version(s) of R were you using?

No response

Relevant log output

(No additional material needed)

Code of Conduct

• I agree to follow this project's Code of Conduct.

Contribution Guidelines

• I agree to follow this project's Contribution Guidelines.

[rvlenth]

OK, 1.10.4 is now on CRAN

[shajoezhu]

Thanks so much @rvlenth , I will prep for a CRAN release