https://github.com/linkrope/uncovered examines coverage listing files to identify the ones with the most uncovered lines. The results are sorted by absolute number, rather than percentage - the most actionable at the end. There is also a package summary.

For example, this is the output after a Phobos unittest run:

            ...
  257/802   std-socket.lst
  262/3568  std-range-package.lst
  265/1282  std-stdio.lst
  267/267   std-regex-internal-backtracking.lst
  282/323   std-regex-internal-kickstart.lst
  284/1084  std-net-curl.lst
  303/3176  std-datetime-interval.lst
  445/717   std-regex-internal-parser.lst
  777/3366  std-uni-package.lst
  854/3499  std-format.lst
-----------
    0/0     std-windows
    3/5     etc-c
  196/1940  std-digest
  211/7103  std-algorithm
  366/4044  std-range
  371/2480  std-container
  400/1840  std-net
  437/2592  std-internal
  514/7178  std-experimental
  658/14567 std-datetime
  777/3366  std-uni
 1122/2318  std-regex
 3960/36485 std
-----------
 9015/83918 lines uncovered

(quite impressive)

On Tuesday, 21 July 2020 at 20:39:54 UTC, Mario Kröplin wrote:
> https://github.com/linkrope/uncovered examines coverage listing files to identify the ones with the most uncovered lines.

Was this done with the new coverage check during CTFE?
I would expect the numbers to be better then...