search

ropensci-review-tools / pkgstats

Historical statistics of every R package ever
https://docs.ropensci.org/pkgstats/
17 stars 1 forks source link

[BUG] in check_vignette #28

Closed mpadge closed 3 years ago

mpadge commented 3 years ago

@steffilazerte Your weathercan package is a very helpful example indeed! The vignettes checks reports 12, yet you only have 10, so something is awry there too. Ping @maelle, author of original vignette check.

mpadge commented 3 years ago

Aha, actually comes from the number extracted by pkgstats, which needs to be updated to reflect @maelle's superior code.

mpadge commented 3 years ago

@maelle Vignette counting has been updated here to reflect the same code in the vignette checks of pkgcheck (which also had to be updated, because of @steffitlazerte's fine example with sub-directories in the vignettes dir). The reported number is now the overall number of .Rmd files BUT this count does not do the is_html check. The main pkgcheck itself thus remains a confirmation that there is at least one HTML vignette, while the number here, and therefore in the extended pkgcheck output, is simply that regardless of rmarkdown::default_output_format(). Just so that we all understand here :smile:

maelle commented 3 years ago

Thanks for the clarification!

If I remember correctly my code came from pkgdown.

mpadge commented 3 years ago

@maelle Just another FYI here that this actually has to be reverted to previous version, because pkgstats must remain truly general and applicable to the entire history of R packages long before .Rmd was even a thing. I'll then modify downstream to ensure the numbers reported by pkgcheck are indeed correct.

mpadge commented 3 years ago

Actually even easier to just modify here - count .Rmd files if there are any, otherwise count .R files. That code doesn't apply to CRAN tarballs from the mirror anyway, so should be perfectly okay in that form.