don't diff against existing data if performing a resync

carstore/bs.go

@@ -875,7 +875,7 @@ func BlockDiff(ctx context.Context, bs blockstore.Blockstore, oldroot cid.Cid, n
 		keepset[c] = true
 		oblk, err := bs.Get(ctx, c)
 		if err != nil {
-			return nil, err
+			return nil, fmt.Errorf("get failed in new tree: %w", err)
 		}
 
 		if err := cbg.ScanForLinks(bytes.NewReader(oblk.RawData()), func(lnk cid.Cid) {
@@ -901,7 +901,7 @@ func BlockDiff(ctx context.Context, bs blockstore.Blockstore, oldroot cid.Cid, n
 
 		oblk, err := bs.Get(ctx, c)
 		if err != nil {
-			return nil, err
+			return nil, fmt.Errorf("get failed in old tree: %w", err)
 		}
 
 		if err := cbg.ScanForLinks(bytes.NewReader(oblk.RawData()), func(lnk cid.Cid) {

cmd/gosky/sync.go

@@ -27,6 +27,11 @@ var syncGetRepoCmd = &cli.Command{
 	Name:      "get-repo",
 	Usage:     "download repo from account's PDS to local file (or '-' for stdout). for hex combine with 'xxd -ps -u -c 0'",
 	ArgsUsage: `<at-identifier> [<car-file-path>]`,
+	Flags: []cli.Flag{
+		&cli.StringFlag{
+			Name: "host",
+		},
+	},
 	Action: func(cctx *cli.Context) error {
 		ctx := context.Background()
 		arg := cctx.Args().First()
@@ -57,6 +62,10 @@ var syncGetRepoCmd = &cli.Command{
 			return fmt.Errorf("no PDS endpoint for identity")
 		}
 
+		if h := cctx.String("host"); h != "" {
+			xrpcc.Host = h
+		}
+
 		log.Infof("downloading from %s to: %s", xrpcc.Host, carPath)
 		repoBytes, err := comatproto.SyncGetRepo(ctx, xrpcc, ident.DID.String(), "")
 		if err != nil {

repomgr/repomgr.go

@@ -769,6 +769,12 @@ func (rm *RepoManager) ImportNewRepo(ctx context.Context, user models.Uid, repoD
 		return err
 	}
 
+	if rev == nil {
+		// if 'rev' is nil, this implies a fresh sync.
+		// in this case, ignore any existing blocks we have and treat this like a clean import.
+		curhead = cid.Undef
+	}
+
 	if rev != nil && *rev != currev {
 		// TODO: we could probably just deal with this
 		return fmt.Errorf("ImportNewRepo called with incorrect base")