fetch: support v2 w/o manifest on old WWW

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

diff --git a/lib/PublicInbox/Fetch.pm b/lib/PublicInbox/Fetch.pm
index 993e5b190818548d4bc07d0f8202ac7d7146561e..7881b402e3f6c2b7d870a115b50f30286a029ab4 100644 (file)
--- 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__ }
 
@@ -58,11 +60,13 @@ sub do_manifest ($$$) {
        $opt->{$_} = $lei->{$_} for (0..2);
        my $cerr = PublicInbox::LeiMirror::run_reap($lei, $curl_cmd, $opt);
        if ($cerr) {
-               return [ 404 ] if ($cerr >> 8) == 22; # 404 Missing
+               return [ 404, $muri ] if ($cerr >> 8) == 22; # 404 Missing
                $lei->child_error($cerr, "@$curl_cmd failed");
                return;
        }
-       my $m1 = PublicInbox::LeiMirror::decode_manifest($ft, $fn, $muri);
+       my $m1 = eval {
+               PublicInbox::LeiMirror::decode_manifest($ft, $fn, $muri);
+       } or return [ 404, $muri ];
        my $mdiff = { %$m1 };
 
        # filter out unchanged entries.  We check modified, too, since
@@ -81,7 +85,7 @@ 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, $muri, $v1_path, \@v2_epochs, $ft, $mf, $m1 ];
 }
 
 sub get_fingerprint2 {
@@ -96,7 +100,7 @@ sub do_fetch { # main entry point
        my $ibx_ver;
        $lei->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
        my $dir = PublicInbox::Admin::resolve_inboxdir($cd, \$ibx_ver);
-       my ($ibx_uri, @git_dir, @epochs, $mg, @new_epoch);
+       my ($ibx_uri, @git_dir, @epochs, $mg, @new_epoch, $skip);
        if ($ibx_ver == 1) {
                my $url = remote_url($lei, $dir) //
                        die "E: $dir missing remote.origin.url\n";
@@ -104,18 +108,24 @@ sub do_fetch { # main entry point
        } else { # v2:
                require PublicInbox::MultiGit;
                $mg = PublicInbox::MultiGit->new($dir, 'all.git', 'git');
-               my @epochs = $mg->git_epochs;
+               @epochs = $mg->git_epochs;
                my ($git_url, $epoch);
                for my $nr (@epochs) { # try newest epoch, first
                        my $edir = "$dir/git/$nr.git";
+                       unless (-d $edir && -w _) { # must be writable dir
+                               $skip->{$nr} = 1;
+                               next;
+                       }
+                       next if defined $git_url;
                        if (defined(my $url = remote_url($lei, $edir))) {
                                $git_url = $url;
                                $epoch = $nr;
-                               last;
                        } else {
                                warn "W: $edir missing remote.origin.url\n";
                        }
                }
+               @epochs = grep { !$skip->{$_} } @epochs if $skip;
+               $skip //= {}; # makes code below easier
                $git_url or die "Unable to determine git URL\n";
                my $inbox_url = $git_url;
                $inbox_url =~ s!/git/$epoch(?:\.git)?/?\z!! or
@@ -127,21 +137,33 @@ 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, $muri, $v1_path, $v2_epochs, $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
                $v2_epochs = [ map { "$dir/git/$_.git" } @epochs ];
-               push @$v2_epochs, "$dir/git/".($epochs[-1] + 1) if @epochs;
+               if (@epochs) {
+                       my $n = $epochs[-1] + 1;
+                       push @$v2_epochs, "$dir/git/$n.git" if !$skip->{$n};
+               }
        } 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 <$muri>, WTF?
 EOM
-               @git_dir = map { "$dir/git/$_.git" } sort { $a <=> $b }
-                       map { my ($nr) = (m!/([0-9]+)\.git\z!g) } @$v2_epochs;
+               @git_dir = map { "$dir/git/$_.git" } sort { $a <=> $b } map {
+                               my ($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;
        }
@@ -182,6 +204,10 @@ 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);
        }