On Tue, Feb 09, 2016 at 08:19:13PM +0000, Michael van Elst wrote: > [email protected] (Manuel Bouyer) writes: > > >It does retry because the same code is used after the pipe has been closed > >(the child exited or has been killed) in this case it's OK to wait for > >data or EOF which will happen very soon. > > Shouldn't it turn on blocking mode on the filedescriptor if it intends > to wait ?
it should probably but I don't know why the code is this way. Maybe there are systems where turning off O_NONBLOCK doens't work ... -- Manuel Bouyer <[email protected]> NetBSD: 26 ans d'experience feront toujours la difference --
