Two clarifications on the progress bar implementation

[DavisVaughan]

• I'd like to check and make sure that it was purposeful that cli_progress_set() uses i rather than i + 1 here in the loop

  purrr/src/map.c

  Line 40 in 8b2c3cf

  cli_progress_set(bar, i);

• I'd also like to see if we should consider putting cli_progress_set() after the call to .f (brought about by thinking about Vignette: Increment progressor after iteration futureverse/furrr#248)

In particular, this example has two quirks:

library(purrr)

x <- 1:5

map(x, .progress = TRUE, function(elt) {
  Sys.sleep(4)
})

• The progress bar doesn't show up at all until after the first iteration (i.e. about 4 seconds in). I don't know if this is a consequence of show_after or the fact that we are setting i = 0 on the first iteration.
• The progress bar never actually hits 100%. On the last iteration we set i = 4, which corresponds to 80% which does show up, but then we do the last iteration and exit the loop. I assume that calling cli_progress_done() will mark it as 100%, and that does seem to be the case if you set clear = FALSE, but I still wonder if we should be using i + 1 in the loop.

CC @gaborcsardi

[hadley]

That reasoning makes sense to me.

[gaborcsardi]

I'd like to check and make sure that it was purposeful that cli_progress_set() uses i rather than i + 1 here in the loop

I'd also like to see if we should consider putting cli_progress_set() after the call to .f (brought about by thinking about futureverse/furrr#248)

Yes, it is on purpose. We set it to i because we haven't done the ith step yet. I.e. we set it to i+1 after the ith step, just like you want.

We set it before the call to .f, so if the developer or the user sets show_after = 0, then it is shown before a potentially lengthy first iteration.

The progress bar doesn't show up at all until after the first iteration (i.e. about 4 seconds in). I don't know if this is a consequence of show_after or the fact that we are setting i = 0 on the first iteration.

Yes, it is because of show_after.

The progress bar never actually hits 100%. On the last iteration we set i = 4, which corresponds to 80% which does show up, but then we do the last iteration and exit the loop. I assume that calling cli_progress_done() will mark it as 100%, and that does seem to be the case if you set clear = FALSE, but I still wonder if we should be using i + 1 in the loop.

Yes, cli_progress_done() will mark it as 100%, and another cli_progress_set() call would be redundant.

[DavisVaughan]

Okay, that all makes sense to me. I agree that it is fine as is.

[gthlor]

does the progress bar works if there is a mutate in the map function?