Page MenuHomeSoftware Heritage

embrace repository snapshot object in the data model (meta task)
Closed, ResolvedPublic

Description

(Multiple) origins might expose the very same state of a repository over time.
To both avoid bloat and be more coherent with the "big Merkle DAG" approach of the data model, we want to introduce a new kind of node in our DAGs that capture the current state of a repository and associate to it a content-adressable identifier.

These new Merkle DAG nodes will be called repository snapshot objects.
A more formal description of repository snapshots is available in

and on the wiki.

This meta task tracks the various changes needed throughout Software Heritage to fully embrace repository snapshot objects.

Related Objects

Event Timeline

zack created this task.Sep 15 2016, 4:26 PM
zack triaged this task as Normal priority.
zack updated the task description. (Show Details)
ardumont added a subscriber: ardumont.EditedFeb 2 2018, 11:19 AM

Current status on the development migration towards snapshot (branch wip/snapshot(s)) as far as I know:

  • swh-storage (opened endpoints to store/retrieve)
  • swh-model (expose identifiers functions to compute hash)
  • swh-loader-dir
  • swh-loader-tar
  • swh-loader-deposit
  • swh-loader-svn
  • swh-loader-mercurial
  • swh-loader-git
  • swh-loader-debian
  • swh-web
olasd added a subscriber: olasd.Feb 6 2018, 4:30 PM

swh-loader-git and swh-loader-debian have now been migrated to snapshots as well, and restarted.

zack closed this task as Resolved.Mar 12 2019, 10:10 AM
zack claimed this task.

unless i'm missing something, this has been completed a while ago (if not, please reopen, ideally adding the relevant open sub-task)