[WIP] Introduce line-breaks in multiline arguments

[IndrajeetPatil]

cf. tidyverse/style#209

Creating an early PR because I'm having a hard time figuring out the scope of this change. The motivating issue and the style guide don't go into too much depth as to when one should introduce line breaks in the context of multi-line arguments. But this should act as a good starting point.

---

This is an adapted version of a similar feature in grkstyle.

Examples

styler::style_text("
list(
  c(
    a
  ), c(
    b
  )
)
")
#> list(
#>   c(
#>     a
#>   ),
#>   c(
#>     b
#>   )
#> )

styler::style_text("c(1, 2, 3)")
#> c(1, 2, 3)

styler::style_text("c(1, 
                   2, 3)")
#> c(
#>   1,
#>   2,
#>   3
#> )


'tribble(
  ~x,       ~d,
  "b",      4:6
)' -> code
styler::style_text(code)
#> tribble(
#>   ~x,       ~d,
#>   "b",      4:6
#> )

^{Created on 2024-12-06 with reprex v2.1.1}

[IndrajeetPatil]

Wow, that's quite the slowdown. Need to figure out what's causing it once the scope is clear.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 27b5eca is merged into main:

• ✔️cache_applying: 164ms -> 162ms [-4.68%, +1.83%]
• ❗🐌cache_recording: 542ms -> 1.27s [+128.12%, +139.62%]
• ❗🐌without_cache: 1000ms -> 2.42s [+137.76%, +146.26%]

Further explanation regarding interpretation and methodology can be found in the documentation.

[IndrajeetPatil]

@lorenzwalthert Any thoughts?

I am blocked here without any further feedback from the team.

[lorenzwalthert]

Maybe we should talk about the abstract rule first. What is it exactly that we try to achieve?
As for the code, I just wanted to say that you can easily figure out if there are line-breaks in a nest by looking at the respective column. To figure out if there are line breaks hidden in one of the children, consier the is_multi_line column. That will save you a lot of computations I think.

Also important: in the example you gave, all line breaks should simply be removed if we had #247.