get rid of unnecessary bytes::length usage

[public-inbox.git] / lib / PublicInbox / Import.pm

diff --git a/lib/PublicInbox/Import.pm b/lib/PublicInbox/Import.pm
index 347382791c3c5bea4401aa8c17262d2296c8ca0d..362cdc47375db4b7947c41747c6d15bf91583d7d 100644 (file)
--- a/lib/PublicInbox/Import.pm
+++ b/lib/PublicInbox/Import.pm
@@ -25,7 +25,7 @@ sub default_branch () {
                                 { GIT_CONFIG => undef });
                chomp(my $h = <$r> // '');
                close $r;
-               $h eq '' ? 'refs/heads/master' : $h;
+               $h eq '' ? 'refs/heads/master' : "refs/heads/$h";
        }
 }
 
@@ -413,19 +413,19 @@ sub add {
                $smsg->{blob} = $self->get_mark(":$blob");
                $smsg->set_bytes($raw_email, $n);
                if (my $oidx = delete $smsg->{-oidx}) { # used by LeiStore
-                       my @docids = $oidx->blob_exists($smsg->{blob});
-                       my @vivify_xvmd;
-                       for my $id (@docids) {
-                               if (my $cur = $oidx->get_art($id)) {
-                                       # already imported if bytes > 0
-                                       return if $cur->{bytes} > 0;
-                                       push @vivify_xvmd, $id;
-                               } else {
-                                       warn "W: $smsg->{blob} ",
-                                               "#$id gone (bug?)\n";
-                               }
-                       }
-                       $smsg->{-vivify_xvmd} = \@vivify_xvmd;
+                       my $eidx_git = delete $smsg->{-eidx_git};
+
+                       # we need this sharedkv to dedupe blobs added in the
+                       # same fast-import transaction
+                       my $u = $self->{uniq_skv} //= do {
+                               require PublicInbox::SharedKV;
+                               my $x = PublicInbox::SharedKV->new;
+                               $x->dbh;
+                               $x;
+                       };
+                       return if !$u->set_maybe(pack('H*', $smsg->{blob}), 1);
+                       return if (!$oidx->vivify_xvmd($smsg) &&
+                                       $eidx_git->check($smsg->{blob}));
                }
        }
        my $ref = $self->{ref};
@@ -510,8 +510,8 @@ sub atfork_child {
        }
 }
 
-sub digest2mid ($$) {
-       my ($dig, $hdr) = @_;
+sub digest2mid ($$;$) {
+       my ($dig, $hdr, $fallback_time) = @_;
        my $b64 = $dig->clone->b64digest;
        # Make our own URLs nicer:
        # See "Base 64 Encoding with URL and Filename Safe Alphabet" in RFC4648
@@ -520,7 +520,7 @@ sub digest2mid ($$) {
        # Add a date prefix to prevent a leading '-' in case that trips
        # up some tools (e.g. if a Message-ID were a expected as a
        # command-line arg)
-       my $dt = msg_datestamp($hdr);
+       my $dt = msg_datestamp($hdr, $fallback_time);
        $dt = POSIX::strftime('%Y%m%d%H%M%S', gmtime($dt));
        "$dt.$b64" . '@z';
 }