You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Compare & explore the differences between git-tagged versions (retrievable as generated archives from e.g. GitHub) and Release assets. The "Asset" tarball for a given version can differ from what was tagged for that release, for what are perceived as good reasons.
That discrepancy was what allowed xz-utils Release assets to have parts of the backdoor embedded that did not match what was in the Git repo.
So: how common is that? Can we bring some extra scrutiny to the differences?
The text was updated successfully, but these errors were encountered:
For a first pass, I think we can do something like:
Check if github.com is in SRC_URI (in ebuilds) or whatever the packaging download URL is
Query github's API for assets for a tag (they're 1:1), the gh tool may be able to do this too
If any exist, see if there's one with an obvious name (like ${P}.tar.xz/.gz etc), or could even grab them all and look for e.g. configure
Compare
We would need to try extend this to at least gitlab and then also maybe figure out a solution for general dist-tarballs where we know of a repo but a self-hosted tarball too, but that last part feels like a stretch goal really (fairly uncommon nowadays).
Compare & explore the differences between git-tagged versions (retrievable as generated archives from e.g. GitHub) and Release assets. The "Asset" tarball for a given version can differ from what was tagged for that release, for what are perceived as good reasons.
That discrepancy was what allowed
xz-utils
Release assets to have parts of the backdoor embedded that did not match what was in the Git repo.So: how common is that? Can we bring some extra scrutiny to the differences?
The text was updated successfully, but these errors were encountered: