diff options
-rw-r--r-- | lib/PublicInbox/ExtMsg.pm | 87 | ||||
-rw-r--r-- | lib/PublicInbox/Msgmap.pm | 18 | ||||
-rw-r--r-- | lib/PublicInbox/Search.pm | 1 | ||||
-rw-r--r-- | lib/PublicInbox/SearchIdx.pm | 7 | ||||
-rw-r--r-- | t/cgi.t | 6 | ||||
-rw-r--r-- | t/msgmap.t | 3 | ||||
-rw-r--r-- | t/psgi_search.t | 23 | ||||
-rw-r--r-- | t/search.t | 30 |
8 files changed, 113 insertions, 62 deletions
diff --git a/lib/PublicInbox/ExtMsg.pm b/lib/PublicInbox/ExtMsg.pm index 04cb4062..51e7799d 100644 --- a/lib/PublicInbox/ExtMsg.pm +++ b/lib/PublicInbox/ExtMsg.pm @@ -26,6 +26,52 @@ our @EXT_URL = ( 'doc-url=/lurker&format=en.html&query=id:%s' ); +sub PARTIAL_MAX () { 100 } + +sub search_partial ($$) { + my ($srch, $mid) = @_; + my $opt = { limit => PARTIAL_MAX, mset => 2 }; + my @try = ("m:$mid*"); + my $chop = $mid; + if ($chop =~ s/(\W+)(\w*)\z//) { + my ($delim, $word) = ($1, $2); + if (length($word)) { + push @try, "m:$chop$delim"; + push @try, "m:$chop$delim*"; + } + push @try, "m:$chop"; + push @try, "m:$chop*"; + } + + # break out long words individually to search for, because + # too many messages begin with "Pine.LNX." (or "alpine" or "nycvar") + if ($mid =~ /\w{9,}/) { + my @long = ($mid =~ m!(\w{3,})!g); + push(@try, join(' ', map { "m:$_" } @long)); + + # is the last element long enough to not trigger excessive + # wildcard matches? + if (length($long[-1]) > 8) { + $long[-1] .= '*'; + push(@try, join(' ', map { "m:$_" } @long)); + } + } + + foreach my $m (@try) { + my $mset = eval { $srch->query($m, $opt) }; + if (ref($@) eq 'Search::Xapian::QueryParserError') { + # If Xapian can't handle the wildcard since it + # has too many results. + next; + } + my @mids = map { + my $doc = $_->get_document; + PublicInbox::SearchMsg->load_doc($doc)->mid; + } $mset->items; + return \@mids if scalar(@mids); + } +} + sub ext_msg { my ($ctx) = @_; my $cur = $ctx->{-inbox}; @@ -56,41 +102,23 @@ sub ext_msg { return exact($ctx, \@found, $mid) if @found; # fall back to partial MID matching - my $n_partial = 0; my @partial; - - if (my $mm = $cur->mm) { - my $tmp_mid = $mid; - my $res = $mm->mid_prefixes($tmp_mid, 100); - if ($res && scalar(@$res)) { - $n_partial += scalar(@$res); - push @partial, [ $cur, $res ]; - # fixup common errors: - } elsif ($tmp_mid =~ s,/[tTf],,) { - $res = $mm->mid_prefixes($tmp_mid, 100); - if ($res && scalar(@$res)) { - $n_partial += scalar(@$res); - push @partial, [ $cur, $res ]; - } - } + my $n_partial = 0; + my $srch = $cur->search; + my $mids = search_partial($srch, $mid) if $srch; + if ($mids) { + $n_partial = scalar(@$mids); + push @partial, [ $cur, $mids ]; } # can't find a partial match in current inbox, try the others: if (!$n_partial && length($mid) >= 16) { - my $tmp_mid = $mid; -again: foreach my $ibx (@ibx) { - my $mm = $ibx->mm or next; - my $res = $mm->mid_prefixes($tmp_mid, 100); - if ($res && scalar(@$res)) { - $n_partial += scalar(@$res); - push @partial, [ $ibx, $res ]; - last if $n_partial >= 100; - } - } - # fixup common errors: - if (!$n_partial && $tmp_mid =~ s,/[tTf],,) { - goto again; + $srch = $ibx->search or next; + $mids = search_partial($srch, $mid) or next; + $n_partial += scalar(@$mids); + push @partial, [ $ibx, $mids]; + last if $n_partial >= PARTIAL_MAX; } } @@ -103,6 +131,7 @@ again: if ($n_partial) { $code = 300; my $es = $n_partial == 1 ? '' : 'es'; + $n_partial .= '+' if ($n_partial == PARTIAL_MAX); $s .= "\n$n_partial partial match$es found:\n\n"; my $cur_name = $cur->{name}; foreach my $pair (@partial) { diff --git a/lib/PublicInbox/Msgmap.pm b/lib/PublicInbox/Msgmap.pm index 6e758c1a..192e311a 100644 --- a/lib/PublicInbox/Msgmap.pm +++ b/lib/PublicInbox/Msgmap.pm @@ -36,7 +36,6 @@ sub dbh_new { ReadOnly => !$writable, sqlite_use_immediate_transaction => 1, }); - $dbh->do('PRAGMA case_sensitive_like = ON'); $dbh; } @@ -151,23 +150,6 @@ sub minmax { ($min, $sth->fetchrow_array); } -sub mid_prefixes { - my ($self, $pfx, $limit) = @_; - - die "No prefix given" unless (defined $pfx && $pfx ne ''); - $pfx =~ s/([%_])/\\$1/g; - $pfx .= '%'; - - $limit ||= 100; - $limit += 0; # force to integer - $limit ||= 100; - - $self->{dbh}->selectcol_arrayref('SELECT mid FROM msgmap ' . - 'WHERE mid LIKE ? ESCAPE ? ' . - "ORDER BY num DESC LIMIT $limit", - undef, $pfx, '\\'); -} - sub mid_delete { my ($self, $mid) = @_; my $dbh = $self->{dbh}; diff --git a/lib/PublicInbox/Search.pm b/lib/PublicInbox/Search.pm index 7175ddc5..5aabda02 100644 --- a/lib/PublicInbox/Search.pm +++ b/lib/PublicInbox/Search.pm @@ -260,6 +260,7 @@ sub qp { $qp->set_database($self->{xdb}); $qp->set_stemmer($self->stemmer); $qp->set_stemming_strategy(STEM_SOME); + $qp->set_max_wildcard_expansion(100); $qp->add_valuerangeprocessor( Search::Xapian::NumberValueRangeProcessor->new(YYYYMMDD, 'd:')); $qp->add_valuerangeprocessor( diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm index 4dc81352..aeb363e0 100644 --- a/lib/PublicInbox/SearchIdx.pm +++ b/lib/PublicInbox/SearchIdx.pm @@ -331,6 +331,13 @@ sub add_message { foreach my $mid (@$mids) { $tg->index_text($mid, 1, 'XM'); + + # because too many Message-IDs are prefixed with + # "Pine.LNX."... + if ($mid =~ /\w{12,}/) { + my @long = ($mid =~ /(\w{3,}+)/g); + $tg->index_text(join(' ', @long), 1, 'XM'); + } $tg->increase_termpos; } $smsg->{to} = $smsg->{cc} = ''; @@ -155,22 +155,16 @@ EOF $res = cgi_run("/test/blahblah\@example.com/raw"); like($res->{body}, qr/Message-Id: <blahblah\@example\.com>/, "mid raw hit"); - $res = cgi_run("/test/blahblah\@example.con/raw"); - like($res->{head}, qr/Status: 300 Multiple Choices/, "mid raw miss"); $res = cgi_run("/test/blahblah\@example.com/"); like($res->{body}, qr/\A<html>/, "mid html hit"); like($res->{head}, qr/Status: 200 OK/, "200 response"); - $res = cgi_run("/test/blahblah\@example.con/"); - like($res->{head}, qr/Status: 300 Multiple Choices/, "mid html miss"); $res = cgi_run("/test/blahblah\@example.com/f/"); like($res->{head}, qr/Status: 301 Moved/, "301 response"); like($res->{head}, qr!^Location: http://[^/]+/test/blahblah\@example\.com/\r\n!ms, '301 redirect location'); - $res = cgi_run("/test/blahblah\@example.con/"); - like($res->{head}, qr/Status: 300 Multiple Choices/, "mid html miss"); $res = cgi_run("/test/new.html"); like($res->{body}, qr/slashy%2Fasdf\@example\.com/, @@ -38,9 +38,6 @@ foreach my $mid (@mids) { is($d->num_for($mid), $mid2num{$mid}, "mid:$mid maps correctly"); } -is_deeply($d->mid_prefixes('a'), [qw(aa@cc aa@bb a@b)], "mid_prefixes match"); -is_deeply($d->mid_prefixes('A'), [], "mid_prefixes is case sensitive"); - is(undef, $d->last_commit, "last commit not set"); my $lc = 'deadbeef' x 5; is(undef, $d->last_commit($lc), 'previous last commit (undef) returned'); diff --git a/t/psgi_search.t b/t/psgi_search.t index 2f033016..a057a994 100644 --- a/t/psgi_search.t +++ b/t/psgi_search.t @@ -20,11 +20,14 @@ my $git_dir = "$tmpdir/a.git"; is(0, system(qw(git init -q --bare), $git_dir), "git init (main)"); my $rw = PublicInbox::SearchIdx->new($git_dir, 1); ok($rw, "search indexer created"); -my $data = <<'EOF'; +my $digits = '10010260936330'; +my $ua = 'Pine.LNX.4.10'; +my $mid = "$ua.$digits.2460-100000\@penguin.transmeta.com"; +my $data = <<"EOF"; Subject: test -Message-Id: <utf8@example> -From: Ævar Arnfjörð Bjarmason <avarab@example> -To: git@vger.kernel.org +Message-ID: <$mid> +From: Ævar Arnfjörð Bjarmason <avarab\@example> +To: git\@vger.kernel.org EOF @@ -37,8 +40,7 @@ foreach (reverse split(/\n\n/, $data)) { my $mime = Email::MIME->new(\$_); my $bytes = bytes::length($mime->as_string); my $doc_id = $rw->add_message($mime, $bytes, ++$num, 'ignored'); - my $mid = $mime->header('Message-Id'); - ok($doc_id, 'message added: '. $mid); + ok($doc_id, 'message added'); } $rw->commit_txn_lazy; @@ -72,6 +74,15 @@ test_psgi(sub { $www->call(@_) }, sub { $res = $cb->(POST('/test/?q=s:bogus&x=m')); is($res->code, 404, 'failed search result gives 404'); is_deeply([], $warn, 'no warnings'); + + my $mid_re = qr/\Q$mid\E/o; + while (length($digits) > 8) { + $res = $cb->(GET("/test/$ua.$digits/")); + is($res->code, 300, 'partial match found while truncated'); + like($res->content, qr/\b1 partial match found\b/); + like($res->content, $mid_re, 'found mid in response'); + chop($digits); + } }); done_testing(); @@ -430,6 +430,36 @@ foreach my $f ("$git_dir/public-inbox/msgmap.sqlite3", "sharedRepository respected for $bn"); } +$ibx->with_umask(sub { + $rw_commit->(); + my $digits = '10010260936330'; + my $ua = 'Pine.LNX.4.10'; + my $mid = "$ua.$digits.2460-100000\@penguin.transmeta.com"; + is($ro->reopen->query("m:$digits", { mset => 1})->size, 0, + 'no results yet'); + my $pine = Email::MIME->create( + header_str => [ + Subject => 'blah', + 'Message-ID' => "<$mid>", + From => 'torvalds@transmeta', + To => 'list@example.com', + ], + body => "" + ); + my $x = $rw->add_message($pine); + $rw->commit_txn_lazy; + is($ro->reopen->query("m:$digits", { mset => 1})->size, 1, + 'searching only digit yielded result'); + + my $wild = $digits; + for my $i (1..6) { + chop($wild); + is($ro->query("m:$wild*", { mset => 1})->size, 1, + "searching chopped($i) digit yielded result $wild "); + } + is($ro->query("m:Pine m:LNX m:10010260936330", {mset=>1})->size, 1); +}); + done_testing(); 1; |