Michael Chirico

CRAN sources won't have anything .Rbuildignore'd: https://github.com/r-lib/lintr/blob/19308c1d61dae2a9cc921a9cbd921c1df31c867e/.Rbuildignore#L13 I suggest we look for GitHub URLs in our reverse suggests and run compare_branches for all of them too to try and give...

oh that's right, `available.packages()` doesn't include the `URL` key of the package's DESCRIPTIONs -- @hadley is there a tool to retrieve those that you're aware of? if not we can...

right... does the script i just added to ` ` above look about right? basically (1) find rev-imports/suggests from `available.packages()` (2) `read_html()` the CRAN URL (3) scrape for a URL...

This one definitely seems susceptible to edge cases -- have you tried any .dev/compare_branches here to see what sort of patterns might be missed "in the wild"? (can only turn...

Thanks for the work so far! Heads up it may take me longer to review this one as I'm on vacation.

> Maybe we need an option to allow unnamed arguments on the same line as the function call, followed by only named arguments on subsequent lines, to be block-indented instead...

> Since that style is caught by brace_linter(), indentation linter should let it slide. Working in an exception for the hanging indent of function() s if they start with {....

I would lint the first one -- that style is only allowed for function definitions IINM. observe what styler does to it. See also https://github.com/tidyverse/style/pull/174

I think the style guide is pretty clear here, which is what we enforce: > If a function call is too long to fit on a single line, use one...

Bumping to [3.1.0](https://github.com/r-lib/lintr/milestone/4) following the most recent comment