Artifact c67cb24f30ba2f6deb1bd905100a266a62eb0017cdcfbed3d76093024d11ad05:
- File
internal/verify/rebuild.go
— part of check-in
[612f2e7acc]
at
2026-10-23 20:05:29
on branch main
— feat(signing): sign check-ins and horizon seeds with gpg or ssh
go-libfossil signed nothing and verified nothing, so a structural-horizon clone could not tell a seed its origin minted from one substituted on disk (draft-fossil-horizon-00 §5.2.4, HZ-5.2-5 and HZ-5.2-6), and a repository with fossil's clearsign setting on stored its check-ins unsigned.
Signing follows fossil's contract, the same one zit follows. With clearsign on (and the legacy omitsign off), every check-in a commit or branch creation stores is clear-signed by pgp-command: unset, gpg --clearsign -o, run as COMMAND OUTFILE INFILE; a command named ssh* runs as COMMAND INFILE and its detached signature is wrapped in fossil's SSH envelope. A signer that fails, or returns text that does not strip back to the manifest, abandons the commit. NoSign and --nosign store one check-in unsigned.
An origin with clearsign on signs each seed it mints. It signs only after the request is authorized, outside the request's transaction, keeps one signature per seed and command so the seed's name is stable, and re-signs a kept signature its own horizon-seed-key no longer verifies. Keeping is best effort, so a read-only mirror still serves.
A clone given its origin's key (CloneOpts.SeedKey, repo clone --seed-key, kept as horizon-seed-key) verifies every seed on clone, pull and rebuild, and falls back to the content horizon from one unsigned or signed by another key. An SSH key is verified in process: namespace fossilscm, the configured key, no SHA-1 ssh-rsa, and exactly fossil's envelope. An OpenPGP fingerprint is verified by gpg, which must report one good signature by that key over text byte-equal to what the parser reads, with nothing after the signature. A verifier that cannot run keeps the held seed and fails the sync instead of discarding it.
fossil 2.28 and zit read the signed check-ins; a check-in zit signs with the same SSH key verifies here; zit clones a signing origin and holds its seed. (user: [email protected], size: 5504) [annotate] [blame] [check-ins using] [more...]
package verify import ( "database/sql" "errors" "fmt" "time" "github.com/danmestas/go-libfossil/db" "github.com/danmestas/go-libfossil/internal/horizon/seed" "github.com/danmestas/go-libfossil/internal/manifest" "github.com/danmestas/go-libfossil/internal/repo" ) // bookkeepingTables are the sync-bookkeeping tables the crosslink sweep does not // own. fossil's rebuild repopulates them; Rebuild clears and refills them after // the sweep so a rebuilt repository offers every artifact to its next push. var bookkeepingTables = []string{"unclustered", "unsent"} // rebuildTableList reports every table Rebuild reconstructs, for Report. var rebuildTableList = append(manifest.RederivedTables(), bookkeepingTables...) // Rebuild reconstructs all derived tables from raw blob content. It first // verifies blob integrity (read-only), then empties the crosslink-derived // tables and runs the same crosslink sweep a clone runs, so derived state has a // single owner -- manifest.Crosslink -- rather than a second, partial // derivation that diverged from `fossil rebuild`. Finally it refills the // sync-bookkeeping tables the sweep does not own. It holds one write // transaction for the whole sweep, so a concurrent writer waits up to // busy_timeout and then gets SQLITE_BUSY, as with fossil's rebuild. func Rebuild(r *repo.Repo) (*Report, error) { if r == nil { panic("verify.Rebuild: nil *repo.Repo") } start := time.Now() report := &Report{} // Phase 1: verify blobs (read-only, outside any transaction). This is the // responsibility the crosslink sweep does not carry: every blob is expanded // and its content hashed against its stored UUID, so a corrupt or // delta-dangling repository is reported before any derived row is rewritten. if err := checkBlobs(r, report); err != nil { return nil, fmt.Errorf("verify.Rebuild: %w", err) } // Everything after the read-only check runs in one transaction, as // fossil's rebuild_db does, so a sweep that fails partway leaves the // derived state Rebuild found rather than emptied tables. if err := r.WithTx(rederive); err != nil { return nil, fmt.Errorf("verify.Rebuild: %w", err) } // The rebuilt mlink is what elision is read from. The store cone is not a // derived table and survives the rebuild, so elided phantoms stay elided. if err := countAbsences(r, report); err != nil { return nil, fmt.Errorf("verify.Rebuild: %w", err) } report.TablesRebuilt = rebuildTableList report.Duration = time.Since(start) // Postcondition: TablesRebuilt must be populated after successful rebuild. if len(report.TablesRebuilt) == 0 { panic("verify.Rebuild: postcondition: TablesRebuilt empty after successful rebuild") } return report, nil } // rederive re-derives the crosslink tables (manifest.Rederive), applies the // horizon seed a structural-horizon clone holds outside the artifact store, // as draft-fossil-horizon-00 §5.2.3 requires of every derivation, and refills // the sync-bookkeeping tables the sweep does not own. manifest.Rederive runs // CrosslinkRebuild, not Crosslink, so the whole rebuild runs in tx: a check-in // whose files or parent are absent is linked with phantoms, and a delta // manifest whose baseline is absent is refused and filed in orphan, exactly as // Crosslink and `fossil rebuild` do. func rederive(tx *db.Tx) error { if _, err := manifest.Rederive(tx); err != nil { return err } // A held seed is re-verified against the clone's key (§5.2.4: every seed // it holds), and discarded if it fails; the next sync falls back. verify, err := seed.VerifierFor(tx) if err != nil { return err } if _, err := seed.Apply(tx, verify); err != nil { return err } return rebuildBookkeeping(tx) } // clearTables deletes all rows from the named tables that exist. A table may be // absent: a repository created by fossil(1) without the feature that owns it // (forumpost until the first forum post, say) has no such table until the // crosslink sweep's prepareCrosslinkTables creates it, and there is nothing to // clear in that case. func clearTables(tx *db.Tx, tables []string) error { if tx == nil { panic("verify.clearTables: nil *db.Tx") } for _, table := range tables { exists, err := tableExists(tx, table) if err != nil { return err } if !exists { continue } if _, err := tx.Exec("DELETE FROM " + table); err != nil { return fmt.Errorf("clear %s: %w", table, err) } } return nil } // tableExists reports whether the named table is present in the schema. func tableExists(tx *db.Tx, table string) (bool, error) { var name string err := tx.QueryRow( "SELECT name FROM sqlite_master WHERE type='table' AND name=?", table, ).Scan(&name) if errors.Is(err, sql.ErrNoRows) { return false, nil } if err != nil { return false, fmt.Errorf("table exists %s: %w", table, err) } return true, nil } // rebuildBookkeeping repopulates unclustered and unsent. Every non-phantom blob // is marked unclustered and unsent so sync re-evaluates what to push. It clears // first so a repeated Rebuild stays idempotent. func rebuildBookkeeping(tx *db.Tx) error { if tx == nil { panic("verify.rebuildBookkeeping: nil *db.Tx") } if err := clearTables(tx, bookkeepingTables); err != nil { return err } for _, table := range bookkeepingTables { _, err := tx.Exec("INSERT INTO " + table + "(rid) SELECT rid FROM blob WHERE size >= 0") if err != nil { return fmt.Errorf("rebuildBookkeeping %s: %w", table, err) } } return nil }