Artifact 99a187563e8e85c26e3b034e998c2eb9ad3d8457a173312314b1f516fcb1a64a:
- File
internal/repo/repo.go
— part of check-in
[6f269e32a9]
at
2026-10-21 18:45:52
on branch feat/submanifest-reader
— fix(repo): verify with one sweep down the delta forest
Repo.Verify expanded every non-phantom blob on its own with content.Verify. Each expansion re-applies the blob's whole delta chain, so the cost grows with the square of chain length. On a clone of fossil-scm.org's repository (67,894 blobs, chains averaging 532 deltas and reaching 2,546) repo verify took 49 minutes of CPU on both main and this branch, against six and a half minutes for fossil test-integrity.
content.ExpandAll already walks the delta forest from its roots and applies each delta once, binding every result to its name; internal/verify uses it. Verify now goes through it, with the same blob selection (size >= 0), and names the failing rid in its error. The same clone verifies in 19 seconds. (user: [email protected], size: 11709) [annotate] [blame] [check-ins using] [more...]
package repo import ( "fmt" "strconv" "strings" "github.com/danmestas/go-libfossil/db" "github.com/danmestas/go-libfossil/internal/content" libfossil "github.com/danmestas/go-libfossil/internal/fsltype" "github.com/danmestas/go-libfossil/internal/hash" "github.com/danmestas/go-libfossil/simio" ) type Repo struct { db *db.DB path string // receipt is what content stored through this Repo is recorded under; // nil leaves each transaction its own (see WithReceipt). receipt *db.Receipt } // Create creates a new repo at path, seeded with the given user. // If projectCode is empty, a fresh project-code is generated; otherwise // projectCode must already be valid (40-char lowercase hex) — validation // is enforced at the libfossil.Create boundary. func Create(path string, user string, rng simio.Rand, projectCode string) (*Repo, error) { env := simio.RealEnv() env.Rand = rng return CreateWithEnv(path, user, env, projectCode) } func CreateWithEnv(path string, user string, env *simio.Env, projectCode string) (*Repo, error) { if path == "" { panic("repo.CreateWithEnv: path must not be empty") } if env == nil { env = simio.RealEnv() } if env.Rand == nil { panic("repo.CreateWithEnv: env.Rand must not be nil") } if env.Storage == nil { env.Storage = simio.OSStorage{} } if _, err := env.Storage.Stat(path); err == nil { return nil, fmt.Errorf("repo.CreateWithEnv: file already exists: %s", path) } d, err := db.Open(path) if err != nil { return nil, fmt.Errorf("repo.CreateWithEnv open: %w", err) } if err := db.CreateRepoSchema(d); err != nil { d.Close() if rmErr := env.Storage.Remove(path); rmErr != nil { return nil, fmt.Errorf("repo.CreateWithEnv: %w (cleanup failed: %v)", err, rmErr) } return nil, fmt.Errorf("repo.CreateWithEnv schema: %w", err) } if err := db.SeedUser(d, user); err != nil { d.Close() if rmErr := env.Storage.Remove(path); rmErr != nil { return nil, fmt.Errorf("repo.CreateWithEnv: %w (cleanup failed: %v)", err, rmErr) } return nil, fmt.Errorf("repo.CreateWithEnv seed user: %w", err) } if err := db.SeedConfig(d, env.Rand, projectCode); err != nil { d.Close() if rmErr := env.Storage.Remove(path); rmErr != nil { return nil, fmt.Errorf("repo.CreateWithEnv: %w (cleanup failed: %v)", err, rmErr) } return nil, fmt.Errorf("repo.CreateWithEnv seed config: %w", err) } // After SeedConfig, so a seeded Rand still yields the project and server // codes it did before these users were added. if err := db.SeedDefaultUsers(d, env.Rand); err != nil { d.Close() if rmErr := env.Storage.Remove(path); rmErr != nil { return nil, fmt.Errorf("repo.CreateWithEnv: %w (cleanup failed: %v)", err, rmErr) } return nil, fmt.Errorf("repo.CreateWithEnv seed default users: %w", err) } return &Repo{db: d, path: path}, nil } func Open(path string) (*Repo, error) { return OpenWithEnv(path, nil) } func OpenWithEnv(path string, env *simio.Env) (*Repo, error) { if path == "" { panic("repo.OpenWithEnv: path must not be empty") } if env == nil { env = simio.RealEnv() } if env.Storage == nil { env.Storage = simio.OSStorage{} } if err := checkExists(env, path); err != nil { return nil, fmt.Errorf("repo.OpenWithEnv: %w", err) } d, err := db.Open(path) if err != nil { return nil, fmt.Errorf("repo.OpenWithEnv: %w", err) } id, err := d.ApplicationID() if err != nil { d.Close() return nil, fmt.Errorf("repo.OpenWithEnv application_id: %w", err) } if id != libfossil.FossilApplicationID { d.Close() return nil, fmt.Errorf("repo.OpenWithEnv: not a fossil repo (application_id=%d, want %d)", id, libfossil.FossilApplicationID) } // The conversion is best-effort: while another connection holds the // write lock it is left for a later open, rather than failing this one. if err := upgradeJulianSeconds(d); err != nil && !isBusy(err) { d.Close() return nil, fmt.Errorf("repo.OpenWithEnv: %w", err) } return &Repo{db: d, path: path}, nil } // julianSecondsRows matches the config and user rows whose mtime is a Julian // day, as go-libfossil wrote it before writing seconds since 1970 as fossil // does. A Julian day now is about 2.46e6; as seconds that is January 1970, // a time no row fossil or go-libfossil writes today holds. const julianSecondsRows = "mtime BETWEEN 2000000 AND 3000000" // isBusy reports whether err is SQLite's SQLITE_BUSY or SQLITE_LOCKED, as // either driver words it. func isBusy(err error) bool { msg := err.Error() for _, s := range []string{"database is locked", "database table is locked", "SQLITE_BUSY", "SQLITE_LOCKED"} { if strings.Contains(msg, s) { return true } } return false } // upgradeJulianSeconds converts those rows to the seconds they stand for. // It reads first and writes only when a row needs it, so opening an // up-to-date repository takes no write lock. func upgradeJulianSeconds(d *db.DB) error { var n int err := d.QueryRow("SELECT (SELECT count(*) FROM config WHERE " + julianSecondsRows + ") + (SELECT count(*) FROM user WHERE " + julianSecondsRows + ")").Scan(&n) if err != nil { return fmt.Errorf("check julian mtimes: %w", err) } if n == 0 { return nil } return d.WithTx(func(tx *db.Tx) error { for _, table := range []string{"config", "user"} { if _, err := tx.Exec("UPDATE " + table + " SET mtime=CAST((mtime-2440587.5)*86400 AS INTEGER) WHERE " + julianSecondsRows); err != nil { return fmt.Errorf("convert %s julian mtimes: %w", table, err) } } return nil }) } func (r *Repo) Close() error { return r.db.Close() } // Checkpoint runs PRAGMA wal_checkpoint(<mode>) against the repo's database. // See db.DB.Checkpoint for semantics. func (r *Repo) Checkpoint(mode db.CheckpointMode) error { return r.db.Checkpoint(mode) } func (r *Repo) Path() string { return r.path } func (r *Repo) DB() *db.DB { return r.db } func (r *Repo) WithTx(fn func(tx *db.Tx) error) error { if r.receipt == nil { return r.db.WithTx(fn) } return r.db.WithTx(func(tx *db.Tx) error { prev := tx.UseReceipt(r.receipt) defer tx.UseReceipt(prev) return fn(tx) }) } // WithReceipt returns a Repo on the same database that records the content // its transactions store under rc: one receipt for a whole commit, sync, // clone or served request, as fossil keeps one g.rcvid for its process. It is // valid for as long as r is, and closing it closes r. func (r *Repo) WithReceipt(rc *db.Receipt) *Repo { if rc == nil { panic("repo.WithReceipt: rc must not be nil") } return &Repo{db: r.db, path: r.path, receipt: rc} } // Receipt returns the receipt given by WithReceipt, nil when there is none. func (r *Repo) Receipt() *db.Receipt { return r.receipt } // LocalUID returns the uid of login in the user table, the user content is // credited to. A login the repository lacks falls back to its first setup // user, as fossil's clone takes the user holding s when it has no other // (src/clone.c); a repository with neither credits uid 0. func (r *Repo) LocalUID(login string) (int64, error) { var uid int64 err := r.db.QueryRow("SELECT coalesce("+ "(SELECT uid FROM user WHERE login=?),"+ "(SELECT uid FROM user WHERE cap LIKE '%s%' ORDER BY uid LIMIT 1), 0)", login).Scan(&uid) if err != nil { return 0, fmt.Errorf("repo.LocalUID %q: %w", login, err) } return uid, nil } // WithBoundTx runs fn in one write transaction with a Repo bound to it, as // db.DB.WithBoundTx binds its DB: everything fn does through that Repo, // including the transactions code running on it opens, commits together when // fn returns nil and rolls back together otherwise. The bound Repo is valid // only until fn returns and must not be closed. func (r *Repo) WithBoundTx(fn func(bound *Repo) error) error { if fn == nil { panic("repo.WithBoundTx: fn must not be nil") } return r.db.WithBoundTx(func(b *db.DB) error { if r.receipt != nil { prev := b.UseReceipt(r.receipt) defer b.UseReceipt(prev) } return fn(&Repo{db: b, path: r.path, receipt: r.receipt}) }) } // Config reads a configuration value from the repo's config table. func (r *Repo) Config(key string) (string, error) { var value string err := r.db.QueryRow("SELECT value FROM config WHERE name=?", key).Scan(&value) if err != nil { return "", fmt.Errorf("repo.Config %q: %w", key, err) } return value, nil } // UUIDByRID returns the UUID (manifest hash) of the artifact with the given // rid by querying the blob table. Returns sql.ErrNoRows (wrapped) when no // blob row exists for the rid. func (r *Repo) UUIDByRID(rid int64) (string, error) { var uuid string err := r.db.QueryRow("SELECT uuid FROM blob WHERE rid=?", rid).Scan(&uuid) if err != nil { return "", fmt.Errorf("repo.UUIDByRID rid=%d: %w", rid, err) } return uuid, nil } // SetConfig writes a configuration value to the repo's config table, stamped // with the time in seconds since 1970, as fossil's db_set stamps it (now()). // // A hash-policy value must be one of the ordinals 0 through 4, the only form // the reader accepts (AF-7.1-2). Fossil makes an unknown policy fatal where // the value comes in (hname.c hash_policy_command), so a bad value is refused // here instead of poisoning every later store. func (r *Repo) SetConfig(key, value string) error { if key == "hash-policy" { if n, err := strconv.Atoi(value); err != nil || n < 0 || n > 4 { return fmt.Errorf("repo.SetConfig: hash-policy %q is not a policy ordinal (0-4)", value) } } _, err := r.db.Exec( "INSERT OR REPLACE INTO config(name, value, mtime) VALUES(?, ?, strftime('%s','now'))", key, value, ) if err != nil { return fmt.Errorf("repo.SetConfig %q: %w", key, err) } return nil } // SetHashPolicy assigns the hash policy by its value string, as fossil's // hash-policy command does: name must be one of sha1, auto, sha3, sha3-only // or shun-sha1, matched case-sensitively and exactly, and anything else is // an error that leaves the stored policy alone (artifact-format-00 7.1, // AF-7.1-2). The ordinal is what is stored, the form the reader takes. // // Asking for auto in a repo that already holds a SHA3 artifact stores sha3 // (7.3, set-time coercion): auto would be promoted by the first SHA3 store // anyway, and fossil's command coerces the same way. func (r *Repo) SetHashPolicy(name string) error { if r == nil { panic("repo.SetHashPolicy: receiver must not be nil") } policy, err := hash.PolicyByName(name) if err != nil { return fmt.Errorf("repo.SetHashPolicy: %w", err) } if policy != db.HashPolicyAuto { return r.SetConfig("hash-policy", strconv.Itoa(policy)) } // The SHA3 check and the write are one statement, so a SHA3 artifact // stored concurrently cannot slip between them and have the write undo // the ratchet (7.3). _, err = r.db.Exec( "INSERT OR REPLACE INTO config(name, value, mtime) VALUES('hash-policy', "+ "CASE WHEN EXISTS(SELECT 1 FROM blob WHERE length(uuid)>40) THEN ? ELSE ? END, "+ "strftime('%s','now'))", strconv.Itoa(db.HashPolicySHA3), strconv.Itoa(db.HashPolicyAuto), ) if err != nil { return fmt.Errorf("repo.SetHashPolicy: %w", err) } return nil } func (r *Repo) Verify() error { if r == nil { panic("repo.Verify: receiver must not be nil") } // One sweep down the delta forest, not an expansion per blob: expanding // each blob alone re-applies its whole chain, and on fossil-scm's own // repository (chains averaging 532 deltas) that took 46 CPU minutes // where the sweep takes under one. return content.ExpandAll(r.db, func(rid libfossil.FslID, _ string, _ []byte, err error) error { if err != nil { return fmt.Errorf("repo.Verify: rid %d: %w", rid, err) } return nil }) }