On Tue, Jul 26, 2022 at 10:54 PM Rob Landley <[email protected]> wrote:
> On 7/26/22 17:59, enh wrote: > > This broke a lot of tests in postsubmit, and even seems to have had the > > effect of causing reboots (presumably by accidentally killing init) > > which made it quite hard to track down. > > Isn't this why init has signals blocked? > > > Sadly I'm too stupid by this point in the afternoon to come up with the > > missing test :-( > > Properly testing killall and rm goes under mkroot, precisely _because_ it > can do > bad things to the system if it goes wrong... > we do have _some_ killall tests, i just didn't get past the "how did we get here?" to even work out whether we could write a test for this case, not the "okay, how do i write that test?"... > Rob >
_______________________________________________ Toybox mailing list [email protected] http://lists.landley.net/listinfo.cgi/toybox-landley.net
