Ignore multi-line string contents optionally in line_length_linter

[MichaelChirico]

Closes #856. Progress on #2737

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.24%. Comparing base (a6ac90b) to head (69e1fd8).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2799   +/-   ##
=======================================
  Coverage   99.24%   99.24%           
=======================================
  Files         129      129           
  Lines        7282     7316   +34     
=======================================
+ Hits         7227     7261   +34     
  Misses         55       55           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[AshesITR]

Looking at the tests I'm wondering whether all you need is supress lints if the lint line satisfies line1 < line < line2 for some //STR_CONST?

[MichaelChirico]

Looking at the tests I'm wondering whether all you need is supress lints if the lint line satisfies line1 < line < line2 for some //STR_CONST?

there's also the issue of how to deal with lines like this:

x <- 'long string starting on a short
line which also has long strings in
the other parts of the body'

Before this rule, there's 3 lints, but under ignore_string_bodies=TRUE here, there are 0

[MichaelChirico]

Bumping for review :)

[Bisaloo]

I had to come back to this several time because the current version of is_in_string_body() is quite complex. I think part of it is the switch between vectorized and vapply logic and the easily missed changes between str_idx/long_idx as well as str_data/parse_data.

[Bisaloo]

Suggested change

	#' mainly affects wide multiline strings, e.g. SQL queries.
	#' only affects wide multiline strings, e.g. SQL queries.

?

Looking at the code, this seems clearer & more precise.

[MichaelChirico]

Added a new example related to this -- what I'm trying to convey is mainly that the closing quote counts towards the line length where it's found.

1234567890
----------
string <- "
 too long!
"

The less common case is that the opening " here will also trigger a lint (with max width 10)

lint("string <- '\n  too long\n'", line_length_linter(10, ignore_string_bodies=TRUE))
# <text>:1:11: style: [line_length_linter] Lines should not be more than 10 characters. This line is 11 characters.
# string <- '
# ~~~~~~~~~~^