Dave Miner wrote: > Tim Knitter wrote: >> >> >> Dave Miner wrote: >>> Tim S. Knitter wrote: >>>> Can someone please review the following? >>>> >>>> 1868 beadm destroy creates an empty grub menu >>>> >>>> http://defect.opensolaris.org/bz/show_bug.cgi?id=1868 >>>> http://cr.opensolaris.org/~tsk/1868_slim/ >>>> >>> >>> The fix seems a little problematic yet. The sequencing you've chosen >>> here means that if we fail to activate the current menu item (which, >>> though fairly unlikely, is certainly possible), then we still end up >>> with a GRUB menu without an active entry. I'd rather we did things >>> in an order that made that not possible. >>> >> >> I fixed this in the latest webrev. If you could verify when you can >> find a spare moment, I'd appreciate it. >> > > Two things: > > - I found it odd that be_activate_current_be is off in the be_utils.c > rather than in be_activate.c. Any particular reason it's there? >
Only that I thought it might be called from other be_*.c files in the future and if so would be better served in be_utils.c. But since it is only called from be_create.c I'll move it to be_activate.c. > - It seems like beadm perhaps should print a message noting which be > will be active, just so the user realizes this and can correct if > desired. Or, perhaps have the confirmation prompt that's put up note > this case. Yeah I thought about that too but opted against it since 'beadm list' would show the state of the BE's. It does seem better to be explicit and put out some sort of message though. I'll add that in to this fix. Thanks Tim > > Dave
