On 9/7/18 7:24 AM, Barath Aron wrote: > Any thoughts on posix_spawn()? It seems to be possible now with the ...addfchdir() function for find/exec.c. In xargs, the error detection is a bit more detailed, so this might become a bit more complicated.
In general, I think the code will have to be doubled, i.e., it seems we'll have to maintain both the fork+execv and the posix_spawn way in code. It should be configurable which of them has to be activated; however, for GNU/Linux, I'd stick to the well-working fork+exec per default. Anyway, I think this should go into a release following the imminent 4.6.1. Finally, if you contribute some code, then we need to roll the ball for the copyright assignment to the FSF. Have a nice day, Berny