Bug 230982

Summary: [GTK] Sign releases using a modern GPG key and publish it on webkitgtk.org
Product: WebKit Reporter: Michael Catanzaro <mcatanzaro>
Component: WebKitGTKAssignee: Nobody <webkit-unassigned>
Status: NEW    
Severity: Normal CC: bugs-noreply
Priority: P2    
Version: WebKit Nightly Build   
Hardware: PC   
OS: Linux   

Michael Catanzaro
Reported 2021-09-29 13:48:27 PDT
Currently WebKitGTK release tarballs are signed using Carlos Garcia's personal GPG key. In the Fedora spec file, I do this: # Created from http://hkps.pool.sks-keyservers.net/pks/lookup?op=get&search=0xF3D322D0EC4582C3 # $ gpg --import 0xF3D322D0EC4582C3.asc # $ gpg --export --export-options export-minimal D7FCF61CF9A2DEAB31D81BD3F3D322D0EC4582C3 > gpgkey-D7FCF61CF9A2DEAB31D81BD3F3D322D0EC4582C3.gpg Source2: gpg-key-D7FCF61CF9A2DEAB31D81BD3F3D322D0EC4582C3.gpg This would catch a supply chain attack where (a) distro has previously imported Carlos's key, then (b) attacker compromises webkitgtk.org and replaces tarballs, but (c) attacker does not have access to Carlos's GPG private key. Sadly, hkps.pool.sks-keyservers.net doesn't seem to exist anymore, and sks-keyservers.net is using an unacceptable TLS certificate, so my comment on how to create the .gpg keyring is obsolete. That could be fixed by publishing a keyring on webkitgtk.org, so I could link to that rather than these instructions for how to manually create the keyring. But the keyring should really use a new key, rather than Carlos's existing key, because the existing key is a DSA 1024 key that is weak compared to modern standards. There are probably other best practices to follow (project keys instead of individual developer keys?), but I don't pretend to understand them. GPG is too hard for me....
Attachments
Note You need to log in before you can comment on or make changes to this bug.