On Jun 6, 7:18pm, dholland-sourcechan...@netbsd.org (David Holland) wrote: -- Subject: Re: CVS commit: src/tests/util/sh
| On Sun, Jun 06, 2010 at 07:11:05PM +0000, David Holland wrote: | > On Sun, Jun 06, 2010 at 08:08:57AM -0400, Christos Zoulas wrote: | > > | If it's more important that the nightly test runs come up clean than | > > | that it actually check for correct behavior so sh gets fixed, the test | > > | script should at least document what the correct behavior is. I have | > > | done this much, although I'm inclined to think the changes should be | > > | reverted entirely. | > > | > > Please show me a shell implementation where those tests work, and then | > > I will accept that reverting them is TRT. I personally cannot find one. | > > Also all the tests now behave consistently in ksh, bash, and ash so I | > > am inclined to believe that they are correct. | > | > The last time I tried, the behavior of sh, ksh, and bash diverged | > substantially on some of them. Others, yes, it's posix on one side and | > all extant implementations on the other... this is why I was trying to | > secure a consensus, but that hasn't happened; it's too fiddly and | > consequently ~nobody cares enough to wade into the gory details. | | Ok, right now ksh diverges from sh on 32 of 75 cases, including | several of the ones you changed. bash diverges on 3. | | Anyway, the test script now at least documents what the deal is, so | I'm ok with leaving it until we can reach some conclusions on the | proper behavior. Our ksh and GNU bash, version 3.2.48(1)-release (x86_64--netbsd) behave exactly like our sh. Which bash and ksh did you use? christos