revdep dbi.table broken by with=FALSE

[tdhock]

revdep check machine
says #6487 broke dbi.table tests:

* checking tests ...
  Running 'testthat.R'
 ERROR
Running the tests in 'tests/testthat.R' failed.
Complete output:
  > # This file is part of the standard setup for testthat.
  > # It is recommended that you do not modify it.
  > #
  > # Where should you do additional test configuration?
  > # Learn more about the roles of various files in:
  > # * https://r-pkgs.org/testing-design.html#sec-tests-files-overview
  > # * https://testthat.r-lib.org/articles/special-files.html
  > 
  > library(testthat)
  > library(dbi.table)
  Loading required package: data.table
  > 
  > test_check("dbi.table")
  [ FAIL 2 | WARN 332 | SKIP 0 | PASS 264 ]
  
  == Failed tests ================================================================
  -- Failure ('test-dbi.table.R:77:5'): j syntax consistent with data.table ------
  reference.test(Track[, 1:MediaTypeId], verbose = FALSE) is not TRUE
  
  `actual`:   FALSE
  `expected`: TRUE 
  -- Failure ('test-dbi.table.R:77:5'): j syntax consistent with data.table ------
  reference.test(Track[, 1:MediaTypeId], verbose = FALSE) is not TRUE
  
  `actual`:   FALSE
  `expected`: TRUE 
  
  [ FAIL 2 | WARN 332 | SKIP 0 | PASS 264 ]
  Error: Test failures
  Execution halted
* checking for unstated dependencies in vignettes ... OK

@MichaelChirico @jangorecki you commented on that PR so can you please investigate?

[MichaelChirico]

OK, here's a reprex:

DT = data.table(a=1, b=2:3)
DT[, 1:b]

## master
# [1] 1 2
# Warning message:
# In 1:b : numerical expression has 2 elements: only the first used

## 1.16.4
#        a     b
#    <num> <num>
# 1:     1     2
# 2:     1     3

I think with=FALSE should be retained in this case, i.e. we should fix the regression.

[tdhock]

fixed using 6ad0524