https://bugzilla.suse.com/show_bug.cgi?id=1191260 https://bugzilla.suse.com/show_bug.cgi?id=1191260#c8 --- Comment #8 from Martin Wilck <martin.wilck@suse.com> --- (In reply to Michal Suchanek from comment #7)
There are branches that do nothing and branch that complains about wrong scriptlet.
So what? What's the benefit of running the noop branch, or seeing that complaint on systems where the entire concept of certificates doesn't matter? By definition, cert-script is a noop on systems that don't support UEFI. Therefore exiting early (and without a warning, which would just confuse users without good reason) on such a system is the right thing to do. As a side effect, it makes the code more compact and easier to read.
Exiting early because the test is needed on all branches now is prone to the same problem as we had with IFS in find-provides.ksyms. The test is completely separated from the code in question.
No. The problem in find-provides.ksyms was not the early check, but the fact that IFS was mangled, which is about the dirtiest thing that you can do in shell programming. If a program can determine early on that it has no purpose, quitting immediately is reasonable and clean. The scriptlet is small enough that even in the extremely unlikely case that some time in the future we'll add some certificate handling on non-UEFI systems, we'll figure out how to skip the test. -- You are receiving this mail because: You are on the CC list for the bug.