Darren J Moffat wrote: > > Roland Mainz wrote: > > Darren J Moffat wrote: > > [snip] > >> On the other hand given that you already have to modify the code to have > >> pfexec used it doesn't seem unreasonable to make it work as the user > >> expects. I'd be happy to work with you offline to investigate how much > >> code change it would be to have the appropriate subset of builtins > >> disabled when running as pfksh93. I'd be happy to contribute the code > >> changes. > > > > Yes, but disabling the builtins is not that easy. My main concern is > > that such a change must not - in any case - cause the test suite to > > fail, which quickly leads to the question/problem which builtins should > > be disabled and which not (for example the "test" builtin cannot be > > disabled unless we make the /usr/bin/test utility aware of the > > [snip] > > > Ahhggrll... ;-( > > I really like to avoid a "TCR" here and work in peace on a solution. > > This seems to be more complex and rushing any solution without propper > > investigation may lead to something which may even be worse than the > > current status (which is at least predictable and controlable for the > > developers of scripts). > > Another issue is that I don't know much about RBAC (and AFAIK neither > > David&Glenn do)... I first have to learn how it works in detail and > > which side-effects it has... > > Given this I HIGHLY recommend that this case does NOT introduce pfksh93. > It will be a requirement on the future case that makes the ksh93 code > base /usr/bin/ksh to have this issue resolved - how that should be done > can be taken offline and may not, depending on other OpenSolaris > projects, require any change to ksh93 code at all.
We've removed pfksh93 from this case, mainly because I feared the "rushing" of a solution which later may bite back in a very bad way. The POSIX shell spec explicitly allows and describes builtins and their usage and IMO we need a generic solution which doesn't require to turn off all builtins (and I am little bit stunned that such a change is requested at all - ksh93 gives users and script writers full control over the builtins and their invocation which is much better than the status of the old ksh or other shells). I am already unhappy with the removal of the other builtins which were excluded to simplify the initial putback - the price is a huge performance loss and this shouldn't be neccesary, not even for pfksh93 (as other said "... not this case (anymore) ..."). ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.mainz at nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 7950090 (;O/ \/ \O;)
