bug-bash
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] builtin: test


From: Pourko
Subject: Re: [PATCH] builtin: test
Date: Sat, 4 Oct 2025 03:19:23 +0200 (CEST)

On Fri, 3 Oct 2025 20:51:02 -0400, Grisha Levit wrote:
> On Thu, Oct 2, 2025 at 5:48 PM Mike Jonkmans <bashbug@jonkmans.nl> wrote:
> >
> > It is not a common case.
> >
> > Can't that check be done with something like:
> >         trap '' SIGTTIN
> >         if read -rd '' -n 0; then
> >                 echo Foreground
> >         else
> >                 echo Background
> >         fi
> >         trap SIGTTIN
> 
> Unfortunately, this relies on non-portable behavior of read(2) with a
> size argument of 0. Error detection is not required in this case per
> POSIX [1] and, indeed, on macOS `read -rd '' -n 0' returns 0 when called
> in a background job.
> 
> [1] https://pubs.opengroup.org/onlinepubs/9799919799/functions/read.html
> 
> > > On Tue, 23 Sep 2025 05:07:18 +0200 (CEST), Pourko wrote:
> > > > Now...
> > > > [ -t 0 ] && [ ! -T 0 ]
> > > > ...means we are running in the background.
> 
> I wonder if the `tcgetpgrp((fd) == getpgid(0)' test you propose wouldn't
> be more appropriate for `read -n0' (or some other option combination) to
> do rather than have it as a `test' operation.

I went with "test -T 0" because to me it seems the easiest, and the most 
intuitive. I mean, (judging from me personally), the average user won't know 
how to dive into the quirks of read -n0, let alone even think about SIGTTIN.





reply via email to

[Prev in Thread] Current Thread [Next in Thread]