Accounts people can create, and photos on a server with no bucket

A tenant had exactly the users somebody had created with a command on the
server. That is not a missing screen: a shop with an owner and four staff
either shared one password or raised a ticket per person, and a phone app
for the shop floor could not exist while there was one account to sign in
as.

Registration is by invitation, never open signup - the same line already
drawn around creating a company. The code carries the address and the role
and the request carries only a password, so a code that gets forwarded
cannot become somebody else's account, and a staff invitation cannot be
redeemed as an owner. Single use lives in the UPDATE and the account is
created in the same transaction.

Deactivating a member revokes their sessions in that transaction too. An
access token lives twelve hours, so without it "remove their access"
removed it sometime tomorrow. The session list and revoke that go with it
are the benefit of opaque tokens the product had been paying for and never
collecting: nothing could say what was signed in, let alone stop one.

Face images now work on a deployment with no object storage, which was
every local install and every self-hosted site - the arrivals feed said
"not storing customer photos" for every customer forever, on the screen
whose whole job is to show a face. Bounded to one row per visitor, so it
grows with the customer base and not with footfall; the bucket stays
primary wherever one exists.

Image.auth says whether a URL needs the session, because a browser img
cannot load one that does, a mobile image view can, and a webview can do
neither - the desktop client resolves those to a data URI in Go.

Found by running it, not by tests:

  * UPDATE ... RETURNING gives the value AFTER the update, so the prune
    read back empty keys, deleted nothing, and the table grew with
    footfall exactly as if it were not there. The fake agreed with either
    version; only the live Postgres test caught it.
  * Trusting only the auth flag broke every shop card, because Sites.jsx
    rebuilt a partial snapshot object and dropped it. A relative URL is
    now sufficient on its own.
  * ago() renders a future time as "just now", so a code valid for a week
    read "expires just now".

Verified live against real Postgres: invite, preview, escalation refused,
register into a session, replay 404, staff forbidden, device revoked and
401 at once, last owner refused, and a 92,405-byte camera JPEG stored,
served to its owner, 401 with no session, 404 to another tenant, and
rendered in a browser.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HViLj9gYNRtSr7YVZmW5sn

server/internal/api/handlers_images.go

@@ -91,31 +91,36 @@ func (s *Server) handleVisitorImage(w http.ResponseWriter, r *http.Request) {
 		writeErr(w, http.StatusNotFound, "not_found", "That customer no longer exists.")
 		return
 	}
-	if s.Blob == nil {
-		writeErr(w, http.StatusNotFound, "images_disabled",
-			"This server does not store images.")
-		return
-	}
 	key, err := s.Store.VisitorImageKey(r.Context(), p.ClientID, id)
-	if err != nil || key == "" {
-		writeErr(w, http.StatusNotFound, "no_image",
-			"There is no photo for this customer.")
-		return
-	}
-	url, err := s.Blob.PresignGet(key, viewTTL)
 	if err != nil {
-		s.serverError(w, "presign read", err)
+		key = ""
+	}
+	// The same function every other surface uses. Two ways to answer "where is
+	// this person's photo" would eventually answer differently, and the one
+	// that mattered would be whichever the customer was looking at.
+	img := s.imageFor(key)
+	if !img.Available {
+		// Absence, with the reason. `no_image` and `images_disabled` are
+		// separate codes because the desktop and mobile clients act on them
+		// differently: one is a customer with no picture yet, the other is a
+		// deployment that stores none and should stop asking.
+		code := "no_image"
+		if key == "" && s.Blob == nil {
+			code = "images_disabled"
+		}
+		writeErr(w, http.StatusNotFound, code, img.Reason)
 		return
 	}
 	// Every read of a face image is worth a row. If a client asks "who looked
 	// at my customers", an audit trail is the only answer that is not a guess.
+	// Recorded HERE, where the link is handed out, for both storage routes -
+	// the bucket's bytes never touch this server, so the fetch itself is not a
+	// place both paths could be counted.
 	s.Store.Audit(r.Context(), AuditEntry{
 		ClientID: p.ClientID, ActorID: p.UserID, ActorKind: "user",
 		Action: "image.view", Entity: "visitor", EntityID: id,
 	})
-	writeJSON(w, http.StatusOK, map[string]any{
-		"url": url, "expires_in": int(viewTTL.Seconds()),
-	})
+	writeJSON(w, http.StatusOK, img)
 }
 
 // handleForgetVisitor is the erasure path.
@@ -146,8 +151,21 @@ func (s *Server) handleForgetVisitor(w http.ResponseWriter, r *http.Request) {
 		s.serverError(w, "list images for erasure", err)
 		return
 	}
+	// Images this server holds itself. Deleted before the row, for the same
+	// reason the bucket objects are: if the database commits first and this
+	// fails, the keys are gone and nothing knows which images to remove.
+	if err := s.Store.DeleteVisitFaces(r.Context(), p.ClientID, keys); err != nil {
+		s.logf("ERROR erasure %s: cannot delete stored faces: %v", id, err)
+		writeErr(w, http.StatusBadGateway, "storage_error",
+			"The photo could not be deleted, so nothing was erased. "+
+				"Please try again.")
+		return
+	}
 	if s.Blob != nil {
 		for _, key := range keys {
+			if isDBKey(key) {
+				continue // already gone, above
+			}
 			if err := s.Blob.Delete(r.Context(), key); err != nil {
 				// Refuse the whole request. Reporting an erasure as done while
 				// a face image is still in the bucket is the one outcome this