On 02/07/2019 18:23, Junio C Hamano wrote: > Phillip Wood <phillip.wood123@xxxxxxxxx> writes: > >>> As long as sed implementation used here does not do anything funny >>> to CR, I think the approach to strip everything before the last CR >>> on the line is sensible. As I am not familiar with how Windows port >>> of sed wants to treat a CR byte in the pattern, I am not sure about >>> the precondition of the above statement, though. >> >> I wondered about that too, but it passes the CI tests under windows. > > Hopefully Git for Windows, MinGW, and CygWin would all behave > similarly. > >>> I also have to wonder if we can/want to do this without an extra >>> printf process every time we sanitize the output, though I do not >>> think I care too deeply about it. >> >> I could add 're="$(printf ...)"' to the setup at the top of the file >> if you want > > As I do not care too deeply about it, we recently saw a lot about > reducing number of processes in the tests, so apparently some folks > care and I presume they want to see something like that to happen. > I do not think $re is a good name for such a variable, though ;-) Yes, $re was just a place holder - naming is hard ... Best Wishes Phillip
