Jump to content

Revision hashing in MediaWiki

From IdeaWazaWiki
Revision as of 05:30, 4 October 2026 by Idea (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Q: Does MediaWiki have an edit hash capability or any Git-like qualities with respect to tracking changes (and verifying changes) and so forth?

A: MediaWiki does calculate content hashes, but its architecture is fundamentally different from Git's cryptographic DAG (Directed Acyclic Graph).

Revision Content Hashing (rev_sha1)

MediaWiki calculates a SHA-1 hash for the content of every edit.

  • Storage: The hash is stored in the database under revision.rev_sha1 (encoded in base-36 format).
  • Multi-Content Revisions (MCR): In modern MediaWiki versions, a page can hold multiple "slots" (e.g., main wikitext, structured metadata). The rev_sha1 field acts as a composite hash combining the individual content_sha1 hashes of each slot.
  • API Exposure: You can query the SHA-1 directly via the MediaWiki Action API using:
action=query&prop=revisions&rvprop=sha1

Purpose: Unlike Git, where hashes form an immutable cryptographic chain, MediaWiki uses rev_sha1 primarily for:

  • Duplicate detection: Fast checks to see if an edit actually modified the text or reverted back to an identical previous version.
  • Integrity & corruption checks: Verifying client-transmitted payloads against database writes.
  • Revision deduplication: Avoiding redundant blob storage in backend object systems.

MediaWiki vs. Git: Architecture Differences

Feature MediaWiki Git
Identity Identifier Monotonic integer (rev_id) Cryptographic hash (SHA-1 / SHA-256) of object contents
History Structure Relational SQL table pointing to a parent ID (rev_parent_id) Cryptographic Merkle DAG (commits hash parents + tree states)
Tamper Evidence Low (database rows can be altered or pruned without invalidating future revisions) High (modifying any past commit breaks downstream hashes)
Branching & Merging Linear, single-branch trunk per page (conflicts must be resolved inline) Full distributed branching, three-way merging, and rebasing
Storage Model Centralized SQL database + external blob storage Distributed content-addressable object store

Git-Like Qualities That Do Exist

  • Line-by-Line Attribution: The action=query&prop=revisions API and on-wiki diff tools provide functionality analogous to git blame and git diff.
  • Parent References: Each revision explicitly records its predecessor in the database via rev_parent_id, allowing linear traversal of history.
  • Three-Way Merge on Conflict: When two users edit a page simultaneously, MediaWiki attempts an automated, diff3-style three-way merge. If changes do not overlap, it commits automatically; otherwise, it triggers an edit conflict resolution screen.

If you need true Git-level cryptographic immutability, audit verification, or offline branching for wiki content, common solutions include deploying file-backed wiki engines like Gitit or Gollum, or streaming MediaWiki revisions via XML dumps (dumpBackup.php) into a local Git repository for cryptographic verification.

Readings

Wikipedia

See also

Retrieved from "https://ideawaza.com/index.php?title=Revision_hashing_in_MediaWiki&oldid=86644"