about summary refs log tree commit homepage
path: root/t/v2mirror.t
diff options
context:
space:
mode:
Diffstat (limited to 't/v2mirror.t')
-rw-r--r--t/v2mirror.t224
1 files changed, 192 insertions, 32 deletions
diff --git a/t/v2mirror.t b/t/v2mirror.t
index 81b9544d..b8824182 100644
--- a/t/v2mirror.t
+++ b/t/v2mirror.t
@@ -1,19 +1,20 @@
-# Copyright (C) 2018-2020 all contributors <meta@public-inbox.org>
+# Copyright (C) all contributors <meta@public-inbox.org>
 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 use strict;
-use warnings;
-use Test::More;
+use v5.10.1;
 use PublicInbox::TestCommon;
-use File::Path qw(remove_tree);
+use File::Path qw(remove_tree make_path);
 use Cwd qw(abs_path);
+use Carp ();
+use PublicInbox::Spawn qw(which);
 require_git(2.6);
+require_cmd('curl');
 local $ENV{HOME} = abs_path('t');
+use IO::Uncompress::Gunzip qw(gunzip $GunzipError);
 
 # Integration tests for HTTP cloning + mirroring
 require_mods(qw(Plack::Util Plack::Builder
-                HTTP::Date HTTP::Status Search::Xapian DBD::SQLite));
-use IO::Socket;
-use POSIX qw(dup2);
+                HTTP::Date HTTP::Status Xapian DBD::SQLite));
 use_ok 'PublicInbox::V2Writable';
 use PublicInbox::InboxWritable;
 use PublicInbox::Eml;
@@ -25,7 +26,8 @@ my $pi_config = "$tmpdir/config";
         open my $fh, '>', $pi_config or die "open($pi_config): $!";
         print $fh <<"" or die "print $pi_config: $!";
 [publicinbox "v2"]
-        inboxdir = $tmpdir/in
+; using "mainrepo" rather than "inboxdir" for v1.1.0-pre1 WWW compat below
+        mainrepo = $tmpdir/in
         address = test\@example.com
 
         close $fh or die "close($pi_config): $!";
@@ -36,6 +38,7 @@ my $cfg = PublicInbox::Config->new($pi_config);
 my $ibx = $cfg->lookup('test@example.com');
 ok($ibx, 'inbox found');
 $ibx->{version} = 2;
+$ibx->{-no_fsync} = 1;
 my $v2w = PublicInbox::V2Writable->new($ibx, 1);
 ok $v2w, 'v2w loaded';
 $v2w->{parallel} = 0;
@@ -63,22 +66,17 @@ $v2w->done;
 }
 $ibx->cleanup;
 
-my $sock = tcp_server();
-ok($sock, 'sock created');
-my $cmd = [ '-httpd', '-W0', "--stdout=$tmpdir/out", "--stderr=$tmpdir/err" ];
-my $td = start_script($cmd, undef, { 3 => $sock });
-my ($host, $port) = ($sock->sockhost, $sock->sockport);
-$sock = undef;
+local $ENV{TEST_IPV4_ONLY} = 1; # plackup (below) doesn't do IPv6
+my $rdr = { 3 => tcp_server() };
+my @cmd = ('-httpd', '-W0', "--stdout=$tmpdir/out", "--stderr=$tmpdir/err");
+my $td = start_script(\@cmd, undef, $rdr);
+my ($host, $port) = tcp_host_port(delete $rdr->{3});
 
-my @cmd;
-foreach my $i (0..$epoch_max) {
-        my $sfx = $i == 0 ? '.git' : '';
-        @cmd = (qw(git clone --mirror -q),
-                "http://$host:$port/v2/$i$sfx",
-                "$tmpdir/m/git/$i.git");
+@cmd = (qw(-clone -q), "http://$host:$port/v2/", "$tmpdir/m");
+run_script(\@cmd) or xbail '-clone';
 
-        is(xsys(@cmd), 0, "cloned $i.git");
-        ok(-d "$tmpdir/m/git/$i.git", "mirror $i OK");
+for my $i (0..$epoch_max) {
+        ok(-d "$tmpdir/m/git/$i.git", "epoch $i cloned");
 }
 
 @cmd = ("-init", '-j1', '-V2', 'm', "$tmpdir/m", 'http://example.com/m',
@@ -93,7 +91,6 @@ my $mibx = { inboxdir => "$tmpdir/m", address => 'alt@example.com' };
 $mibx = PublicInbox::Inbox->new($mibx);
 is_deeply([$mibx->mm->minmax], [$ibx->mm->minmax], 'index synched minmax');
 
-$v2w->{rotate_bytes} = $old_rotate_bytes;
 for my $i (10..15) {
         $mime->header_set('Message-ID', "<$i\@example.com>");
         $mime->header_set('Subject', "subject = $i");
@@ -102,12 +99,16 @@ for my $i (10..15) {
 $v2w->done;
 $ibx->cleanup;
 
+my @new_epochs;
 my $fetch_each_epoch = sub {
-        foreach my $i (0..$epoch_max) {
-                my $dir = "$tmpdir/m/git/$i.git";
-                is(xsys('git', "--git-dir=$dir", 'fetch', '-q'), 0,
-                        'fetch successful');
-        }
+        my %before = map { $_ => 1 } glob("$tmpdir/m/git/*");
+        run_script([qw(-fetch --exit-code -q)], undef, {-C => "$tmpdir/m"}) or
+                xbail('-fetch fail ',
+                        [ xqx([which('find'), "$tmpdir/m", qw(-type f -ls) ]) ],
+                        Carp::longmess());
+        is($?, 0, '--exit-code 0 after fetch updated');
+        my @after = grep { !$before{$_} } glob("$tmpdir/m/git/*");
+        push @new_epochs, @after;
 };
 
 $fetch_each_epoch->();
@@ -233,10 +234,169 @@ EOF
         $mset = $mibx->search->reopen->mset('m:2big@a');
         is(scalar($mset->items), 0, 'large message not re-indexed');
 }
+ok(scalar(@new_epochs), 'new epochs were created and fetched');
+for my $d (@new_epochs) {
+        is(xqx(['git', "--git-dir=$d", 'config', qw(include.path)]),
+                "../../all.git/config\n",
+                'include.path set');
+}
+
+if ('test read-only epoch dirs') {
+        my @git = ('git', "--git-dir=$new_epochs[0]");
+        my $get_objs = [@git,
+                qw(cat-file --buffer --batch-check --batch-all-objects)];
+        my $before = [sort xqx($get_objs)];
+
+        remove_tree(map { "$new_epochs[0]/$_" } qw(objects refs/heads));
+        chmod(0555, $new_epochs[0]) or xbail "chmod: $!";
+
+        # force a refetch
+        unlink("$tmpdir/m/manifest.js.gz") or xbail "unlink: $!";
+
+        run_script([qw(-fetch -q)], undef, {-C => "$tmpdir/m"}) or
+                xbail '-fetch failed';
+
+        ok(!-d "$new_epochs[0]/objects", 'no objects after fetch to R/O dir');
+
+        chmod(0755, $new_epochs[0]) or xbail "chmod: $!";
+        mkdir("$new_epochs[0]/objects") or xbail "mkdir: $!";
+        mkdir("$new_epochs[0]/refs/heads") or xbail "mkdir: $!";
+
+        my $err = '';
+        run_script([qw(-fetch -q)], undef, {-C => "$tmpdir/m", 2 => \$err}) or
+                xbail '-fetch failed '.$err;
+        is_deeply([ sort xqx($get_objs) ], $before,
+                'fetch restored objects once GIT_DIR became writable');
+}
+
+{
+        my $dst = "$tmpdir/partial";
+        run_script([qw(-clone -q --epoch=~0), "http://$host:$port/v2/", $dst]);
+        is($?, 0, 'no error from partial clone');
+        my @g = glob("$dst/git/*.git");
+        my @w = grep { -w $_ } @g;
+        my @r = grep { ! -w $_ } @g;
+        if ($> == 0) {
+                @w = grep { (stat($_))[2] & 0200 } @g;
+                @r = grep { !((stat($_))[2] & 0200) } @g;
+        }
+        is(scalar(@w), 1, 'one writable directory');
+        my ($w) = ($w[0] =~ m!/([0-9]+)\.git\z!);
+        is((grep {
+                m!/([0-9]+)\.git\z! or xbail "no digit in $_";
+                $w > ($1 + 0)
+        } @r), scalar(@r), 'writable epoch # exceeds read-only ones');
+        run_script([qw(-fetch -q)], undef, { -C => $dst });
+        is($?, 0, 'no error from partial fetch');
+        remove_tree($dst);
 
-ok($td->kill, 'killed httpd');
-$td->join;
+        run_script([qw(-clone -q --epoch=~1..),
+                        "http://$host:$port/v2/", $dst]);
+        my @g2 = glob("$dst/git/*.git") ;
+        is_deeply(\@g2, \@g, 'cloned again');
+        is(scalar(grep { (stat($_))[2] & 0200 } @g2), scalar(@w) + 1,
+                'got one more cloned epoch');
 
-done_testing();
+        # make 0.git writable and fetch into it, relies on culled manifest
+        chmod(0755, $g2[0]) or xbail "chmod: $!";
+        my @before = glob("$g2[0]/objects/*/*");
+        run_script([qw(-fetch -q)], undef, { -C => $dst });
+        is($?, 0, 'no error from partial fetch');
+        my @after = glob("$g2[0]/objects/*/*");
+        ok(scalar(@before) < scalar(@after), 'fetched after chmod 0755 0.git');
+
+        # ensure culled manifest is maintained after fetch
+        gunzip("$dst/manifest.js.gz" => \(my $m), MultiStream => 1) or
+                xbail "gunzip: $GunzipError";
+        $m = PublicInbox::Config->json->decode($m);
+        for my $k (keys %$m) { # /$name/git/$N.git
+                my ($nr) = ($k =~ m!/git/([0-9]+)\.git\z!);
+                ok(-w "$dst/git/$nr.git", "writable $nr.git in manifest");
+        }
+        for my $ro (grep { !-w $_ } @g2) {
+                my ($nr) = ($ro =~ m!/git/([0-9]+)\.git\z!);
+                is(grep(m!/git/$nr\.git\z!, keys %$m), 0,
+                        "read-only $nr.git not in manifest")
+                        or xbail([sort keys %$m]);
+        }
+}
+
+my $err = '';
+my $oldrev = '0b3e19584c90d958a723ac2d3dec3f84f5513688~1';
+# 3e0e596105198cfa (wwwlisting: allow hiding entries from manifest, 2019-06-09)
+$oldrev = xqx([qw(git rev-parse), $oldrev], undef, { 2 => \$err });
+SKIP: {
+        skip("no detected public-inbox GIT_DIR ($err)", 1) if $?;
+        require_mods('Email::MIME', 1); # for legacy revision
+        # using plackup to test old PublicInbox::WWW since -httpd from
+        # back then relied on some packages we no longer depend on
+        my $plackup = require_cmd('plackup', 1) or skip('no plackup in path', 1);
+        require PublicInbox::Lock;
+        chomp $oldrev;
+        my ($base) = ($0 =~ m!\b([^/]+)\.[^\.]+\z!);
+        my $wt = "t/data-gen/$base.pre-manifest-$oldrev";
+        my $lk = PublicInbox::Lock->new(__FILE__);
+        $lk->lock_acquire;
+        my $psgi = "$wt/app.psgi";
+        if (!-f $psgi) { # checkout a pre-manifest.js.gz version
+                my $t = File::Temp->new(TEMPLATE => 'g-XXXX', TMPDIR => 1);
+                my $env = { GIT_INDEX_FILE => $t->filename };
+                xsys([qw(git read-tree), $oldrev], $env) and xbail 'read-tree';
+                xsys([qw(git checkout-index -a), "--prefix=$wt/"], $env)
+                        and xbail 'checkout-index';
+                my $f = "$wt/app.psgi.tmp.$$";
+                open my $fh, '>', $f or xbail $!;
+                print $fh <<'EOM' or xbail $!;
+use Plack::Builder;
+use PublicInbox::WWW;
+my $www = PublicInbox::WWW->new;
+builder { enable 'Head'; sub { $www->call(@_) } }
+EOM
+                close $fh or xbail $!;
+                rename($f, $psgi) or xbail $!;
+        }
+        $lk->lock_release;
+
+        $rdr->{run_mode} = 0;
+        $rdr->{-C} = $wt;
+        my $cmd = [$plackup, qw(-Enone -Ilib), "--host=$host", "--port=$port"];
+        $td->join('TERM');
+        open $rdr->{2}, '>>', "$tmpdir/plackup.err.log" or xbail "open: $!";
+        open $rdr->{1}, '>>&', $rdr->{2} or xbail "open: $!";
+        my $env = { PERL5LIB => 'lib', PERL_INLINE_DIRECTORY => undef };
+        $td = start_script($cmd, $env, $rdr);
+        # wait for plackup socket()+bind()+listen()
+        my %opt = ( Proto => 'tcp', Type => Socket::SOCK_STREAM(),
+                PeerAddr => "$host:$port" );
+        for (0..100) {
+                tick();
+                last if IO::Socket::INET->new(%opt);
+        }
+        IO::Socket::INET->new(%opt) or xbail "connect $host:$port: $!";
+        my $dst = "$tmpdir/scrape";
+        @cmd = (qw(-clone -q), "http://$host:$port/v2", $dst);
+        run_script(\@cmd, undef, { 2 => \($err = '') });
+        is($?, 0, 'scraping clone on old PublicInbox::WWW')
+                or diag $err;
+        my @g_all = glob("$dst/git/*.git");
+        ok(scalar(@g_all) > 1, 'cloned multiple epochs');
+
+        remove_tree($dst);
+        @cmd = (qw(-clone -q --epoch=~0), "http://$host:$port/v2", $dst);
+        run_script(\@cmd, undef, { 2 => \($err = '') });
+        is($?, 0, 'partial scraping clone on old PublicInbox::WWW');
+        my @g_last = grep { (stat($_))[2] & 0200 } glob("$dst/git/*.git");
+        is_deeply(\@g_last, [ $g_all[-1] ], 'partial clone of ~0 worked');
+
+        chmod(0755, $g_all[0]) or xbail "chmod $!";
+        my @before = glob("$g_all[0]/objects/*/*");
+        run_script([qw(-fetch -v)], undef, { -C => $dst, 2 => \($err = '') });
+        is($?, 0, 'scraping fetch on old PublicInbox::WWW') or diag $err;
+        my @after = glob("$g_all[0]/objects/*/*");
+        ok(scalar(@before) < scalar(@after),
+                'fetched 0.git after enabling write-bit');
+
+        $td->join('TERM');
+}
 
-1;
+done_testing;