Force eval_facets() to return a data.frame, not a tibble

[yutannihilation]

Closes #3489 (I think it should be fixed on tibble's side, though)

As we saw #3048, using tibble internally causes unexpected errors because the existing codes expects a bare data.frame. I think #3489 is just a bug on tibble's side, but, I believe it's better for ggplot2 internals to stick with a bare data.frame for consistency.

Note that this pull request leaves these two tibble() uses as is because they don't seem the cases when the result is used internally.

ggplot2/R/summarise-plot.R

Lines 66 to 70 in b842024

	layout <- tibble(
	panel = l$layout$PANEL,
	row = l$layout$ROW,
	col = l$layout$COL
	)

ggplot2/R/summarise-plot.R

Lines 134 to 136 in b842024

	tibble(
	mapping = layer_mappings
	)

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/