WIP for reprex_scrape()

[crew102]

Putting comments for this PR in the #67 thread

[crew102]

Was this the use of ghurl that you had in mind?

[krlmlr]

This is what reprex:::reprex_scrape("https://github.com/tidyverse/dplyr/issues/3826") gets me:

#' From https://github.com/tidyverse/dplyr/issues/3545#issuecomment-385150055, should we skip evaluating cases where there is no match?
library(tidyverse)
noisy_case_when <- function(..., .default = NA_character_) {
  case_when(..., TRUE ~ {
    warning("case not covered")
    .default
  })
}
noisy_case_when(
  TRUE ~ "case covered"
)
#' <sup>Created on 2018-09-15 by the [reprex package](https://reprex.tidyverse.org) (v0.2.1)</sup>
#' 

Amazing stuff!

Minor remarks:

• reprex_scrape() doesn't seem to be exported yet
• Do we want to keep the "Created on ..." ad? Do we want to add our own comment, like <sup>Scraped on ... from ...</sup>, instead or in addition to?
• This branch now conflicts with master

[crew102]

Hey @krlmlr , I was waiting until this PR was reviewed before documenting reprex_scrape() (though perhaps I should of at least exported it with @export...I'll export when doing rest of roxygen tags).

Re: how to handle the advertisement tag, I don't have a strong opinion either way.

I'll rebase on top of master when I come back to this PR after #204 is addressed.

[jennybc]

Re: keeping the ad. This is a good question. I played with "undo" some, while building up to this weekend's patch release. Currently, roundtrips lead to a steady accumulation of ad instances, which is gross. I think they should probably be stripped by default (?). If you are undoing and immediately calling reprex, then presence of the add suggests re-calling with advertise = TRUE.