Can `data_color` provide a deprecation message through `lifecycle` or `rlang`?

[mrcaseb]

gt/R/data_color.R

Lines 567 to 584 in e263689

	if (is.character(rlang::eval_tidy(fn))) {
	palette <- rlang::eval_tidy(fn)
	fn <- NULL
	cli::cli_warn(c(
	"Since gt v0.9.0, the `colors` argument has been deprecated.",
	"*" = "Please use the `palette` argument to define a color palette."
	))
	} else {
	cli::cli_warn(c(
	"Since gt v0.9.0, the `colors` argument has been deprecated.",
	"*" = "Please use the `fn` argument instead."
	))
	}

gt 0.9.0 deprecates the colors argument in data_color and prints the warning on every call.

Can this be changed to a "once every 8 hours" deprecation warning?

It's possible to do this with either lifcycle (which is a reverse dependency through ggplot2) or by passing .frequency and .frequency_id arguments in the cli_warning like

cli::cli_warn(
  c(
    "Since gt v0.9.0, the `colors` argument has been deprecated.",
    "*" = "Please use the `palette` argument to define a color palette."
  ), 
  .frequency = "regularly",
  .frequency_id = "data_color_warning"
)
#> Warning: Since gt v0.9.0, the `colors` argument has been deprecated.
#> • Please use the `palette` argument to define a color palette.
#> This warning is displayed once every 8 hours.

[rich-iannone]

Thank you for this issue! This is an excellent suggestion so I'll apply this frequency level to all deprecation messages in the package.

[rich-iannone]

I tried this out and discovered quickly that many tests that check for warnings fail (says they do not actually elicit warnings). Until that’s corrected with some sort of workaround, this cannot be done.

[mrcaseb]

I tried this out and discovered quickly that many tests that check for warnings fail (says they do not actually elicit warnings). Until that’s corrected with some sort of workaround, this cannot be done.

I don't know if this would break anything else but it is possible to set the option
options("rlib_warning_verbosity" = "verbose")
in the test environment to make rlang warn every time
See https://rlang.r-lib.org/reference/abort.html#muffling-and-silencing-conditions

[rich-iannone]

I tried this out and discovered quickly that many tests that check for warnings fail (says they do not actually elicit warnings). Until that’s corrected with some sort of workaround, this cannot be done.

I don't know if this would break anything else but it is possible to set the option options("rlib_warning_verbosity" = "verbose") in the test environment to make rlang warn every time See https://rlang.r-lib.org/reference/abort.html#muffling-and-silencing-conditions

This helped, thank you! I'm now using local_options("rlib_warning_verbosity" = "verbose") in selected test_that({...}) blocks to ensure that expect_warning() and expect_no_warning() work correctly.