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
Somebody might want to run a rebuilder that follows the package index with a delay. An entity or organization could decide to run both a regular and a delayed rebuilder that both have their own queue and scheduler. The second, delayed rebuilder would act as a check to ensure an external review would be able to replicate our verification work.
An operator is encouraged to also run a regular, non-delayed rebuilder since those results are going to be more useful if the build artifact is intended to be rolled out to users immediately.
A failed delayed rebuild does not imply bad faith or a compromise of the earlier rebuilder, but it would work as a "nothing up my sleeve" mechanism the earlier rebuilder runs in addition to provide transparency which of their results we could replicated in the future and which we won't.
Future replication might still break for other reasons though, like network resources disappearing.
The text was updated successfully, but these errors were encountered:
Somebody might want to run a rebuilder that follows the package index with a delay. An entity or organization could decide to run both a regular and a delayed rebuilder that both have their own queue and scheduler. The second, delayed rebuilder would act as a check to ensure an external review would be able to replicate our verification work.
An operator is encouraged to also run a regular, non-delayed rebuilder since those results are going to be more useful if the build artifact is intended to be rolled out to users immediately.
A failed delayed rebuild does not imply bad faith or a compromise of the earlier rebuilder, but it would work as a "nothing up my sleeve" mechanism the earlier rebuilder runs in addition to provide transparency which of their results we could replicated in the future and which we won't.
Future replication might still break for other reasons though, like network resources disappearing.
The text was updated successfully, but these errors were encountered: