Re: PATCH: traps, yet again

[Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx>]

On Mar 9,  1:11pm, Peter Stephenson wrote:
}
} `trapreturn' mechanism which is used to make TRAPNAL functions force
} the enclosing function to behave as if the signal wasn't handled.
} The actual use is a bit different, but the basic idea is similar:
} if we are in a trap, remember the return status and use it when
} we clear up after the trap has finished.
} 
} I've put back the original execsave/execrestore to avoid another
} headache.

I had been staring at this late last night and was about to suggest
something very close to what you did, but decided to sleep on it.  It
looks like you thought of more contingencies than I had, so it's just
as well.

I think this looks right.