Works here!
+1
robert: Can you move your incrontab from /etc/incron.d to a root user incrontab and see if the problem persists?
My theory is that this fixes only the user incrontab case, not the system wide /etc/incron.d/ case...
robert: are you sure the new one was used? can you duplicate it?
I did a test compose with this and it worked: https://koji.fedoraproject.org/koji/taskinfo?taskID=31528556
Works.
+1
+1
+1
Arrrgh. Looks like it had a stable request that I missed. :(
Sorry about that everyone...
I have disabled autokarma to stable, we can leave this in testing until folks have time to sort things out. Just let me know.
Well, not really. Upstream listed it, so I copied that, but looking at it, it's much more complicated than the fix upstream. See: https://bugzilla.redhat.com/show_bug.cgi?id=CVE-2016-1238 for discussion. Basically changing this behavior in perl would break things, but perl itself (and now spamassassin) don't rely on this behavior. I also didn't mark it fixed here because its already marked WONTFIX in the above bug. Anyhow, hope that helps.
Sorry about that, I thought it only affected f28+ Will get a new update out here in a few min...
Seems good here.
I thought software collections packages needed to be explicitly activated?
I can look at adding the requires to be sure tho... can you file a bug on it so I don't forget it? or a PR?