Il giorno lun 23 ago 2021 alle ore 10:16 Roberto A. Foglietta < roberto.foglie...@gmail.com> ha scritto:
> > For the moment we can still assume that in busybox all exceptions are > deadly. > > More specifically, we do not need to assume that every exception are deadly but a less imperative constraint. We need to assume that there does not exist any condition that goes in this branch and does not bring to an exit. if((exitstatus && eflag) || (e)) { /* memory leak is not a problem, we are exiting */ } else { For example (exitstatus && eflag) == true does not means in general that the ash will exit because false || true However, when it comes to this point it is always true that it will exit. So, a more general statement about (e) wil be (isdeadly(e)) so if e != 0 but it has a value for which we know it is not deadly, we can separate it. Unfortunately, at the moment it is not clear for me which kind of value could receive e and why. This is the reason for which I cannot grant the right condition to happen at 100%. Best regards, -- Roberto A. Foglietta +39.349.33.30.697
_______________________________________________ busybox mailing list busybox@busybox.net http://lists.busybox.net/mailman/listinfo/busybox