patterns() throws new informative error when cols is non-character or has NA

[tdhock]

before this PR

> melt(data.table(x1=1,x2=2,y1=3,y2=4),measure.vars=patterns("2",cols=NA_character_))
Error in (function (..., cols = character(0L), ignore.case = FALSE, perl = FALSE,  : 
  Pattern not found: [[2]]

above is not very informative for the user.

after this PR

> melt(data.table(x1=1,x2=2,y1=3,y2=4),measure.vars=patterns("2",cols=NA_character_))
Error in (function (..., cols = character(0L), ignore.case = FALSE, perl = FALSE,  : 
  cols must be a character vector of column names

above is more helpful for the user, and consistent with the error message in measure().

[github-actions]

Generated via commit eac1f8a

Download link for the artifact containing the test results: ↓ atime-results.zip

Time taken to finish the standard R installation steps: 3 minutes and 32 seconds

Time taken to run atime::atime_pkg on the tests: 8 minutes and 15 seconds

[MichaelChirico]

Great find & fix!

[tdhock]

thanks!