Re: [PATCH v3] Allow trap to un-ignore SIGINT/SIGQUIT in async subshells

On Sun, Apr 07, 2024 at 07:18:52PM +0800, Herbert Xu wrote:
> On Fri, Mar 29, 2024 at 04:39:01PM +0100, Johannes Altmanninger wrote:
> >
> > diff --git a/src/trap.c b/src/trap.c
> > index cd84814..dbf81ea 100644
> > --- a/src/trap.c
> > +++ b/src/trap.c
> > @@ -272,7 +272,7 @@ ignoresig(int signo)
> >  		signal(signo, SIG_IGN);
> >  	}
> >  	if (!vforked)
> > -		sigmode[signo - 1] = S_HARD_IGN;
> > +		sigmode[signo - 1] = S_IGN;
> 
> This is buggy if sigmode is already S_HARD_IGN.  You can fix
> this by moving the if statement inside the previous one.
> 
> Please also add a comment stating that sigmode has already been
> initialised by setinteractive, as otherwise we may also lose a
> hard ignore.

I'm not really following the last part;
maybe it's no longer relevant with the bug fix.
Note that it works the same whether "set -i" or "trap - INT" is used,
also in noninteractive shells like

	dash -c '( trap - INT; sleep inf ) & read _'