clone|fetch|--mirror: cull manifest in partial mirrors

This makes it easier for users to enable fetching on a
previously read-only epoch.  Prior to this change, users were
required to delete manifest.js.gz in addition to adding the
writable bit.  Now, they just have to "chmod +w $EPOCH_DIR".

2 files changed, 35 insertions, 6 deletions

diff --git a/lib/PublicInbox/Fetch.pm b/lib/PublicInbox/Fetch.pm
index 464ffe12..7f60b619 100644
--- a/lib/PublicInbox/Fetch.pm
+++ b/lib/PublicInbox/Fetch.pm
@@ -12,6 +12,8 @@ use PublicInbox::LEI;
 use PublicInbox::LeiCurl;
 use PublicInbox::LeiMirror;
 use File::Temp ();
+use PublicInbox::Config;
+use IO::Compress::Gzip qw(gzip $GzipError);
 
 sub new { bless {}, __PACKAGE__ }
 
@@ -81,7 +83,7 @@ sub do_manifest ($$$) {
         }
         my (undef, $v1_path, @v2_epochs) =
                 PublicInbox::LeiMirror::deduce_epochs($mdiff, $ibx_uri->path);
-        [ 200, $v1_path, \@v2_epochs, $muri, $ft, $mf ];
+        [ 200, $v1_path, \@v2_epochs, $muri, $ft, $mf, $m1 ];
 }
 
 sub get_fingerprint2 {
@@ -133,7 +135,7 @@ EOM
         PublicInbox::LeiMirror::write_makefile($dir, $ibx_ver);
         $lei->qerr("# inbox URL: $ibx_uri/");
         my $res = do_manifest($lei, $dir, $ibx_uri) or return;
-        my ($code, $v1_path, $v2_epochs, $muri, $ft, $mf) = @$res;
+        my ($code, $v1_path, $v2_epochs, $muri, $ft, $mf, $m1) = @$res;
         if ($code == 404) {
                 # any pre-manifest.js.gz instances running? Just fetch all
                 # existing ones and unconditionally try cloning the next
@@ -145,6 +147,7 @@ EOM
         } else {
                 $code == 200 or die "BUG unexpected code $code\n";
         }
+        my $mculled;
         if ($ibx_ver == 2) {
                 defined($v1_path) and warn <<EOM;
 E: got v1 `$v1_path' when expecting v2 epoch(s) in <$muri>, WTF?
@@ -153,6 +156,12 @@ EOM
                                 my ($nr) = (m!/([0-9]+)\.git\z!g);
                                 $skip->{$nr} ? () : $nr;
                         } @$v2_epochs;
+                if ($m1 && scalar keys %$skip) {
+                        my $re = join('|', keys %$skip);
+                        my @del = grep(m!/git/$re\.git\z!, keys %$m1);
+                        delete @$m1{@del};
+                        $mculled = 1;
+                }
         } else {
                 $git_dir[0] = $dir;
         }
@@ -193,6 +202,10 @@ EOM
         for my $i (@new_epoch) { $mg->epoch_cfg_set($i) }
         if ($ft) {
                 my $fn = $ft->filename;
+                if ($mculled) {
+                        my $json = PublicInbox::Config->json->encode($m1);
+                        gzip(\$json => $fn) or die "gzip: $GzipError";
+                }
                 rename($fn, $mf) or die "E: rename($fn, $mf): $!\n";
                 $ft->unlink_on_destroy(0);
         }
diff --git a/lib/PublicInbox/LeiMirror.pm b/lib/PublicInbox/LeiMirror.pm
index fe81b967..1ab5e0d8 100644
--- a/lib/PublicInbox/LeiMirror.pm
+++ b/lib/PublicInbox/LeiMirror.pm
@@ -6,7 +6,9 @@ package PublicInbox::LeiMirror;
 use strict;
 use v5.10.1;
 use parent qw(PublicInbox::IPC);
+use PublicInbox::Config;
 use IO::Uncompress::Gunzip qw(gunzip $GunzipError);
+use IO::Compress::Gzip qw(gzip $GzipError);
 use PublicInbox::Spawn qw(popen_rd spawn run_die);
 use File::Temp ();
 use Fcntl qw(SEEK_SET O_CREAT O_EXCL O_WRONLY);
@@ -267,14 +269,14 @@ EOM
         close $fh or die "close:($f): $!";
 }
 
-sub clone_v2 ($$) {
-        my ($self, $v2_epochs) = @_;
+sub clone_v2 ($$;$) {
+        my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
         my $lei = $self->{lei};
         my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
         my $pfx = $curl->torsocks($lei, (values %$v2_epochs)[0]) or return;
         my $dst = $self->{dst};
         my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
-        my (@src_edst, @read_only);
+        my (@src_edst, @read_only, @skip_nr);
         for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
                 my $uri = $v2_epochs->{$nr};
                 my $src = $uri->as_string;
@@ -289,8 +291,15 @@ failed to extract epoch number from $src
                 } else { # create a placeholder so users only need to chmod +w
                         init_placeholder($src, $edst);
                         push @read_only, $edst;
+                        push @skip_nr, $nr;
                 }
         }
+        if (@skip_nr) { # filter out the epochs we skipped
+                my $re = join('|', @skip_nr);
+                my @del = grep(m!/git/$re\.git\z!, keys %$m);
+                delete @$m{@del};
+                $self->{-culled_manifest} = 1;
+        }
         my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
         _try_config($self);
         my $on_destroy = $lk->lock_for_scope($$);
@@ -379,13 +388,20 @@ EOM
                         my ($n) = ("$uri" =~ m!/([0-9]+)\.git\z!);
                         $n => $uri->clone
                 } @v2_epochs;
-                clone_v2($self, \%v2_epochs);
+                clone_v2($self, \%v2_epochs, $m);
         } elsif (defined $v1_path) {
                 clone_v1($self);
         } else {
                 die "E: confused by <$uri>, possible matches:\n\t",
                         join(', ', sort keys %$m), "\n";
         }
+        if (delete $self->{-culled_manifest}) { # set by clone_v2
+                # write the smaller manifest if epochs were skipped so
+                # users won't have to delete manifest if they +w an
+                # epoch they no longer want to skip
+                my $json = PublicInbox::Config->json->encode($m);
+                gzip(\$json => $fn) or die "gzip: $GzipError";
+        }
         my $fin = "$self->{dst}/manifest.js.gz";
         rename($fn, $fin) or die "E: rename($fn, $fin): $!";
         $ft->unlink_on_destroy(0);