So nice to have zchunk working again!

Sorry, just realized the command should be: dnf --enablerepo=rawhide list updates; rm /var/cache/dnf/rawhide*.solv*; rm /var/cache/dnf/rawhide-*/repodata/repomd.xml; dnf --enablerepo=rawhide list updates

Tested with zchunked rawhide repo and everything works correctly. If you want to test that it fixes the segfault, run: dnf --enablerepo=rawhide list updates; rm /var/cache/dnf/rawhide*.solv*; /var/cache/dnf/rawhide-*/repodata/repomd.xml; dnf --enablerepo=rawhide list updates

The second time you run dnf, the Rawhide repo should be about 242KB instead of the usual 55MB, but, most importantly, you shouldn't get a segfault.

The librepo update segfaults when using zchunk metadata: https://bugzilla.redhat.com/show_bug.cgi?id=1694411

@pkratosh, it looks like this won't automatically be pushed, even though the karma has been reset and there isn't any remaining negative karma. Since we're at +6, could you please push manually?

Still works perfectly

Still works perfectly

Tested with dnf group list with a test repository with zchunked updateinfo.xml.zck, and everything worked perfectly.

With vlc-3.0.5-14 in RPM Fusion, this fixes the crash in #1642729.

Tested on a FreeIPA server, and named-pkcs11 starts and works perfectly

Unfortunately, it's only testable on rawhide for the moment. See https://www.jdieter.net/posts/2018/12/24/zchunk-repodata-in-rawhide

This update has been unpushed.

b09fe797cdbb4c8ff1e43a4be64bbaac1d1c5818 (Fix use after free)
d26a039f9f5b817f8bebddc7aa4f3f5a7a20cc69 (Fix uninitialized variable)
625335db13d62b80a5ab8cd374ce9d7c091d6e32 (Fix dereference before NULL)
62733a8606afa2139908577d3970cb04f1b16170 (Fix dereference before NULL and add extra checks)
af3d18b4fc492a83fdc47a4f980d9ab7c1990c0c (Set umask before running mkstemp because some UNIX flavors don't do this)

This update has been unpushed.

Ok, changing my karma to +1 because the problem lies in a F27->F28 upgrade. A new install doesn't see this problem.

To fix the bug if you've done a F27->F28 update, change all the files in /var/log/nginx to have owner.group of root.root.

Bug #1573942 is marked as private so I can't see it, but this doesn't fix bug #1567768 which is marked as a duplicate of the first one.

This also fixed the unreported bug where you are unable to get a certificate because Let’s Encrypt has permanently disabled the TLS-SNI-01 challenge. Thanks for the update!