Fix as_gt() examples

[jdblischak]

The as_gt() examples are skipped during R CMD check since they generate a lot of extraneous output (#334, #338). However, this means they are not regularly tested. While working on #497, I discovered that the examples with gs_power_ahr() and gs_power_wlr() were broken.

Error in gs_power_wlr() : 
  gs_power_wlr: please input the total_spend to the futility spending function.

I did a git bisect to determine that the example broke in a6a0166 when we updated gs_power_ahr() and gs_power_wlr() to require that users specify total_spend (#396, #398).

I fixed the examples as was done for as_rtf() in #398 and also updated the examples to be run during R CMD check to prevent this from happening again. The downsides are that gsDesign2.Rcheck/gsDesign2-Ex.Rout gets bloated with all the as_gt() output, and the examples are also decently long-running. If CRAN complains about either of these on the next submission, then we can re-evaluate the strategy. An alternative would be to copy the examples into a test file, and then return to skipping the examples during R CMD check.

[nanxstats]

Looking great, thanks for identifying this problematic behavior and the neat correction.

One small issue: CRAN will almost surely flag this check note as a problem, considering their build machine has even lower single-core performance than mine:

* checking examples ... [27s/15s] NOTE
Examples with CPU (user + system) or elapsed time > 5s
       user system elapsed
as_gt 8.552  0.254   4.681

You probably know that the standard approach to avoid this note is by wrapping the examples in \donttest{} - although they will still be ran.

[jdblischak]

@nanxstats I have wrapped the as_gt.gs_design() examples in \donttest{}, and now as_gt() does not appear in any of the GitHub Actions R CMD check results.

There are some other example sections that occasionally run over 5s in some of the GitHub Actions jobs, but those weren't affected by my PR. I propose we wait and clean up any remaining longer-running examples in one-go prior to the next release (or the next time they are updated).

[nanxstats]

Thanks! Looking good to me.