On Thu, Dec 06, 2018 at 09:35:47PM +0000, Harald van Dijk wrote: > On 04/12/2018 23:57, Harald van Dijk wrote: > > This has the benefit of fixing one other test case, a small modification > > from one of Martijn Dekker's: > > > > $SHELL -c 'trap "set -o bad@option" INT; kill -s INT $$' && echo BUG > > Another test case, one that still fails: > > trap exit INT > trap 'true; kill -s INT $$' EXIT > false > > Here, inside the EXIT handler, "the command that executed immediately > preceding the trap action" is `false`, but inside the INT handler, it's > either `true` or `kill -s INT $$` (I think the latter, but it doesn't > matter). dash treats it as if it were still `false`. I think this makes sense. The EXIT trap trumps whatever happens inside it. FWIW ksh/mksh both do the same thing. Cheers, -- Email: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx> Home Page: http://gondor.apana.org.au/~herbert/ PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt