On Tue, May 21, 2019 at 07:34:05AM +0200, Martijn van Duren wrote: > Hello Jesper, > > The behaviour should be identical, the only difference would be that > pledge catches programming errors. So I see no particular reason to use > -S over "set secure" for normal users; even without pledge.
Yeah, vi will of course still tell the user that external commands are not allowed. My point was more the inconsistency in regards to pledge. > I choose not to fail hard on pledge, since that could loose peoples > work, which is most definitely not what we want. While here fix a > lie that secure has an off parameter and inform the user that it can't > be turned off again. Nice! Your patch is more elegant and good point about not failing hard. Jesper Wallin
