feat(new linter): added a new linter

[kpagacz]

Hi all,

I have added a new linter that checks for a space after the closing parenthesis of the list of formal function arguments in cases where no braces are used and the function definition spans only a single line.

Examples:

function()print("test") # BAD

function() print("test") # OK

function()
  print("test")               # OK

• Added a new linter that checks for a space between the right parenthesis and the function body in function definitions that span one line and don't use braces

Closes #809

[kpagacz]

Locally, some of the checks don't pass for me. I have branched off master, is this intended?
The checks that don't pass:

[AshesITR]

Hi @kpagacz,

thanks for taking the time to write a PR!

Locally, some of the checks don't pass for me. I have branched off master, is this intended?
The checks that don't pass:

I had similar issues on my Windows machine.
One or more of the following may help:

1. Make sure your lintr Project directory is on your system (C:) drive and contains no spaces
2. Make sure to start the checks from a fresh R session that has no packages attached

[AshesITR]

Looks good so far, just a few things need changing.

[AshesITR]

Very nice, thank you.
I think this is ready to merge. Would you add a line to NEWS.md and update the roxygen documentation to reflect the extensions?

[kpagacz]

Done. Thanks for taking the time to review!

[AshesITR]

LGTM

[AshesITR]

@jimhester @MichaelChirico @renkun-ken @russHyde do we want this in the default linters?
It's mentioned in the tidyverse style guide.

[MichaelChirico]

yes agree. IINM styler::style_file also fixes this

[AshesITR]

The lambda syntax should only be checked in R >= 4.0.
Curiously, the test also fails for R 4.0.

Can you investigate?

[MichaelChirico]

for failure on 4.0 itself, wasn't there a bug with utils::getParseData() on that release?

[kpagacz]

I am going to look into running that particular test only for R>=4.0 and will check if the parsed tree is somehow wrong for version 4.0

[kpagacz]

Regarding the failure on the R version 4.0.5 (I assume that's the one running there). The new lambda syntax was introduced in the version 4.1.0 as per this NEWS document: https://cran.r-project.org/doc/manuals/r-release/NEWS.pdf (3rd point on the 3rd page). For obvious reasons parsing () lambdas returns an error on the version 4.0.x

[AshesITR]

My bad re the version number. The condition needs to be on R version >= 4.1.0, not 4.0.0
Do you need help with adding the linter to the defaults?

[kpagacz]

I have added the necessary skip condition.

[russHyde]

I don't think unbraced blocks of the form

f <- function(x)
    print(x)

are allowed in the tidyverse style (see section "inline statements" in https://style.tidyverse.org/syntax.html#control-flow); though they are allowed if on a single line.
So this linter doesn't seem to fit with the tidyverse style, and IMO shouldn't be in the defaults

[MichaelChirico]

Russ, AIUI that example is just to show another usage that wouldn't be flagged by this linter, rather than showing code that is "OK".

That is, the code maybe should be flagged, but it would be by a different linter, unless you are proposing to extend the scope of this linter.

[jimhester]

I think this could be added to the default linters.

[russHyde]

OK cool.