[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Wed, 3 Feb 2021 15:46:21 -0600
Ok, if that restriction is not going to be lifted, then the documentation
should reflect it.
On Tue, Feb 2, 2021 at 6:19 PM Chet Ramey <firstname.lastname@example.org> wrote:
> On 2/2/21 1:49 PM, Lawrence Velázquez wrote:
> >> Of course, removing this restriction opens up the possibility that a
> >> can shoot themselves by trying to use job control at the same time.
> Remember that `job control' isn't just suspending and restarting jobs. It's
> mostly about multiplexing jobs (identified by process groups) onto the
> terminal, which has a single foreground process group at a time. Your
> change leaves a pipeline with processes running in two separate process
> groups, and one of them isn't going to have access to the terminal. That
> means that any keyboard-generated signal, say SIGINT, won't get to all of
> the processes in the pipeline. Let's not even get into the signals that
> get sent when a process that isn't in the terminal's process group tries to
> read or write from the terminal.
> To make this work correctly requires a lot more work, and you might as well
> solve the SIGTSTP problem while you're at it.
> >> I think
> >> that is a pretty rare case where one would want to do that, like, close
> >> never.
> I think it's more common than you might think, and I think it's very common
> to want to interrupt a pipeline.
> > Here is Chet's response to a similar request nearly four years ago
> > (https://lists.gnu.org/archive/html/bug-bash/2017-03/msg00197.html):
> Thanks, this saved me from looking it up.
> ``The lyf so short, the craft so long to lerne.'' - Chaucer
> ``Ars longa, vita brevis'' - Hippocrates
> Chet Ramey, UTech, CWRU email@example.com http://tiswww.cwru.edu/~chet/
Description: Binary data
Description: Text Data