feat: github_deb metadata-only package type

Add the Debian/apt analog of github_rpm: a metadata-only remote that scans
a GitHub repo's releases for .deb assets, derives per-asset control metadata
via a ranged prefix fetch (never downloading whole packages), synthesizes a
flat apt repository, and redirects .deb downloads to a backend releases_remote.

- Add PackageGitHubDeb to the package-type enum + validity map.
- Add the github_deb provider (internal/provider/deb/github.go): ServeRemote
  serves Packages/Packages.gz/Release, 404s the signed index variants
  (consumed via [trusted=yes]), and 302-redirects *.deb to the releases_remote;
  deriveAsset ranged-GETs the ar prefix, locates control.tar.*, and parses the
  control paragraph, doubling the range on truncation; sha256 comes from the
  asset digest when present, else a one-time full stream.
- Add the github_deb background Syncer (internal/provider/deb/syncer.go): its
  own worker pool, shared rate limiter, deduped queue, and DB-lease-gated scans.
- Add github_deb_sync_state table plus ListGitHubDebRemotes/Claim/Release DB
  helpers (separate from the rpm ones).
- Prime github_deb remotes on create and run the deb syncer alongside the rpm
  one; route prime-on-create by package type.
- Reuse the deb apt-index generators and control parser; skip empty hash lines
  in the Packages index so a SHA256-only metadata entry is valid.
This commit is contained in:
2026-08-11 22:21:05 +10:00
parent 60f008debc
commit b2a6be8eb5
11 changed files with 1889 additions and 14 deletions
+70
View File
@@ -0,0 +1,70 @@
package database
import (
"context"
"errors"
"time"
"github.com/jackc/pgx/v5"
"git.unkin.net/unkin/artifactapi/pkg/models"
)
// ListGitHubDebRemotes returns every github_deb remote so the syncer can sweep
// them on each poll tick.
func (db *DB) ListGitHubDebRemotes(ctx context.Context) ([]models.Remote, error) {
rows, err := db.Pool.Query(ctx, `SELECT `+remoteCols+` FROM remotes WHERE package_type = $1 ORDER BY name`, models.PackageGitHubDeb)
if err != nil {
return nil, err
}
defer rows.Close()
var remotes []models.Remote
for rows.Next() {
var r models.Remote
if err := scanRemote(rows, &r); err != nil {
return nil, err
}
remotes = append(remotes, r)
}
return remotes, rows.Err()
}
// ClaimGitHubDebSyncLease atomically claims the per-remote sync lease. It
// succeeds only when the remote is due (never synced, or synced longer than
// freshness ago) and no live lease is held by another replica. A zero freshness
// (prime scans) ignores the recency gate. The returned etag is the stored
// releases-list ETag, shared across replicas.
func (db *DB) ClaimGitHubDebSyncLease(ctx context.Context, remoteName, owner string, freshness, lease time.Duration) (bool, string, error) {
row := db.Pool.QueryRow(ctx, `
INSERT INTO github_deb_sync_state AS s (remote_name, sync_lease_owner, sync_lease_expires)
VALUES ($1, $2, now() + make_interval(secs => $4))
ON CONFLICT (remote_name) DO UPDATE
SET sync_lease_owner = $2,
sync_lease_expires = now() + make_interval(secs => $4)
WHERE (s.last_synced_at IS NULL OR s.last_synced_at < now() - make_interval(secs => $3))
AND (s.sync_lease_expires IS NULL OR s.sync_lease_expires < now())
RETURNING s.etag
`, remoteName, owner, freshness.Seconds(), lease.Seconds())
var etag string
if err := row.Scan(&etag); err != nil {
if errors.Is(err, pgx.ErrNoRows) {
return false, "", nil
}
return false, "", err
}
return true, etag, nil
}
// ReleaseGitHubDebSyncLease records the completed scan and frees the lease. Only
// the owning replica may release; last_synced_at advances so the next poll waits
// a full freshness window, and etag is persisted for the next conditional request.
func (db *DB) ReleaseGitHubDebSyncLease(ctx context.Context, remoteName, owner, etag string, syncedAt time.Time) error {
_, err := db.Pool.Exec(ctx, `
UPDATE github_deb_sync_state
SET last_synced_at = $3, etag = $4, sync_lease_owner = '', sync_lease_expires = NULL
WHERE remote_name = $1 AND sync_lease_owner = $2
`, remoteName, owner, syncedAt, etag)
return err
}