X-Git-Url: http://www.git.stargrave.org/?p=public-inbox.git;a=blobdiff_plain;f=lib%2FPublicInbox%2FSearchView.pm;h=777710e3649148bcfec7f2e552ba9bb1416bdca9;hp=ab3dddb48252ea62197fd79d8fbb2d746175825b;hb=ef9ce6f2dfc9fcbb91e3cec0b7a2b7f7e359be5d;hpb=57dbfc8fca52b4c986ad53f40301c7ff7a464f78

diff --git a/lib/PublicInbox/SearchView.pm b/lib/PublicInbox/SearchView.pm
index ab3dddb4..777710e3 100644
--- a/lib/PublicInbox/SearchView.pm
+++ b/lib/PublicInbox/SearchView.pm
@@ -5,17 +5,22 @@
 package PublicInbox::SearchView;
 use strict;
 use warnings;
+use URI::Escape qw(uri_unescape uri_escape);
 use PublicInbox::SearchMsg;
-use PublicInbox::Hval;
+use PublicInbox::Hval qw/ascii_html obfuscate_addrs/;
 use PublicInbox::View;
-use PublicInbox::MID qw(mid2path mid_clean);
-use Email::MIME;
+use PublicInbox::WwwAtomStream;
+use PublicInbox::MID qw(mid2path mid_mime mid_clean mid_escape MID_ESC);
+use PublicInbox::MIME;
 require PublicInbox::Git;
-our $LIM = 50;
+require PublicInbox::SearchThread;
+our $LIM = 200;
+
+sub noop {}
 
 sub sres_top_html {
 	my ($ctx) = @_;
-	my $q = PublicInbox::SearchQuery->new($ctx->{cgi});
+	my $q = PublicInbox::SearchQuery->new($ctx->{qp});
 	my $code = 200;
 
 	# double the limit for expanded views:
@@ -25,71 +30,119 @@ sub sres_top_html {
 		mset => 1,
 		relevance => $q->{r},
 	};
-	my ($mset, $total);
-
+	my ($mset, $total, $err, $cb);
+retry:
 	eval {
-		$mset = $ctx->{srch}->query($q->{q}, $opts);
+		$mset = $ctx->{srch}->query($q->{'q'}, $opts);
 		$total = $mset->get_matches_estimated;
 	};
-	my $err = $@;
-	my $res = html_start($q, $ctx) . '<pre>';
+	$err = $@;
+	ctx_prepare($q, $ctx);
 	if ($err) {
 		$code = 400;
-		$res .= err_txt($err) . "</pre><hr /><pre>" . foot($ctx);
+		$ctx->{-html_tip} = '<pre>'.err_txt($ctx, $err).'</pre><hr>';
+		$cb = *noop;
 	} elsif ($total == 0) {
+		if (defined($ctx->{-uxs_retried})) {
+			# undo retry damage:
+			$q->{'q'} = $ctx->{-uxs_retried};
+		} elsif (index($q->{'q'}, '%') >= 0) {
+			$ctx->{-uxs_retried} = $q->{'q'};
+			$q->{'q'} = uri_unescape($q->{'q'});
+			goto retry;
+		}
 		$code = 404;
-		$res .= "\n\n[No results found]</pre><hr /><pre>".foot($ctx);
+		$ctx->{-html_tip} = "<pre>\n[No results found]</pre><hr>";
+		$cb = *noop;
 	} else {
 		my $x = $q->{x};
-		return sub { adump($_[0], $mset, $q, $ctx) } if ($x eq 'A');
+		return adump($_[0], $mset, $q, $ctx) if $x eq 'A';
 
-		$res .= search_nav_top($mset, $q) . "\n\n";
+		$ctx->{-html_tip} = search_nav_top($mset, $q, $ctx) . "\n\n";
 		if ($x eq 't') {
-			return sub { tdump($_[0], $res, $mset, $q, $ctx) };
+			$cb = mset_thread($ctx, $mset, $q);
+		} else {
+			$cb = mset_summary($ctx, $mset, $q);
 		}
-		dump_mset(\$res, $mset);
-		$res .= '</pre>' . search_nav_bot($mset, $q) .
-			"\n\n" . foot($ctx);
 	}
+	PublicInbox::WwwStream->response($ctx, $code, $cb);
+}
+
+# allow undef for individual doc loads...
+sub load_doc_retry {
+	my ($srch, $mitem) = @_;
 
-	$res .= "</pre></body></html>";
-	[$code, ['Content-Type'=>'text/html; charset=UTF-8'], [$res]];
+	eval {
+		$srch->retry_reopen(sub {
+			PublicInbox::SearchMsg->load_doc($mitem->get_document)
+		});
+	}
 }
 
-sub dump_mset {
-	my ($res, $mset) = @_;
+# display non-threaded search results similar to what users expect from
+# regular WWW search engines:
+sub mset_summary {
+	my ($ctx, $mset, $q) = @_;
 
 	my $total = $mset->get_matches_estimated;
 	my $pad = length("$total");
 	my $pfx = ' ' x $pad;
+	my $res = \($ctx->{-html_tip});
+	my $srch = $ctx->{srch};
+	my $obfs = $ctx->{-inbox}->{obfuscate};
 	foreach my $m ($mset->items) {
 		my $rank = sprintf("%${pad}d", $m->get_rank + 1);
 		my $pct = $m->get_percent;
-		my $smsg = PublicInbox::SearchMsg->load_doc($m->get_document);
-		my $s = PublicInbox::Hval->new_oneline($smsg->subject);
-		my $f = $smsg->from_name;
-		$f = PublicInbox::Hval->new_oneline($f)->as_html;
+		my $smsg = load_doc_retry($srch, $m);
+		unless ($smsg) {
+			eval {
+				$m = "$m ".$m->get_docid . " expired\n";
+				$ctx->{env}->{'psgi.errors'}->print($m);
+			};
+			next;
+		}
+		my $s = ascii_html($smsg->subject);
+		my $f = ascii_html($smsg->from_name);
+		if ($obfs) {
+			obfuscate_addrs($s);
+			obfuscate_addrs($f);
+		}
 		my $ts = PublicInbox::View::fmt_ts($smsg->ts);
-		my $mid = PublicInbox::Hval->new_msgid($smsg->mid)->as_href;
+		my $mid = PublicInbox::Hval->new_msgid($smsg->mid)->{href};
 		$$res .= qq{$rank. <b><a\nhref="$mid/">}.
-			$s->as_html . "</a></b>\n";
+			$s . "</a></b>\n";
 		$$res .= "$pfx  - by $f @ $ts UTC [$pct%]\n\n";
 	}
+	$$res .= search_nav_bot($mset, $q);
+	*noop;
 }
 
 sub err_txt {
-	my ($err) = @_;
-	my $u = 'http://xapian.org/docs/queryparser.html';
+	my ($ctx, $err) = @_;
+	my $u = $ctx->{-inbox}->base_url($ctx->{env}) . '_/text/help/';
 	$err =~ s/^\s*Exception:\s*//; # bad word to show users :P
-	$err = PublicInbox::Hval->new_oneline($err)->as_html;
-	"\n\nBad query: <b>$err</b>\n" .
-		qq{See <a\nhref="$u">$u</a> for Xapian query syntax};
+	$err = ascii_html($err);
+	"\nBad query: <b>$err</b>\n" .
+		qq{See <a\nhref="$u">$u</a> for help on using search};
 }
 
 sub search_nav_top {
-	my ($mset, $q) = @_;
+	my ($mset, $q, $ctx) = @_;
+
+	my $rv = '<pre>';
+	my $initial_q = $ctx->{-uxs_retried};
+	if (defined $initial_q) {
+		my $rewritten = $q->{'q'};
+		utf8::decode($initial_q);
+		utf8::decode($rewritten);
+		$initial_q = ascii_html($initial_q);
+		$rewritten = ascii_html($rewritten);
+		$rv .= " Warning: Initial query:\n <b>$initial_q</b>\n";
+		$rv .= " returned no results, used:\n";
+		$rv .= " <b>$rewritten</b>\n instead\n\n";
+	}
 
-	my $rv = "Search results ordered by [";
+	$rv .= 'Search results ordered by [';
 	if ($q->{r}) {
 		my $d = $q->qs_html(r => 0);
 		$rv .= qq{<a\nhref="?$d">date</a>|<b>relevance</b>};
@@ -119,148 +172,132 @@ sub search_nav_bot {
 	my $o = $q->{o};
 	my $end = $o + $nr;
 	my $beg = $o + 1;
-	my $rv = "<hr /><pre>Results $beg-$end of $total";
+	my $rv = '</pre><hr><pre id=t>';
+	if ($beg <= $end) {
+		$rv .= "Results $beg-$end of $total";
+		$rv .= ' (estimated)' if $end != $total;
+	} else {
+		$rv .= "No more results, only $total";
+	}
 	my $n = $o + $LIM;
 
 	if ($n < $total) {
 		my $qs = $q->qs_html(o => $n);
-		$rv .= qq{, <a\nhref="?$qs">next</a>}
+		$rv .= qq{  <a\nhref="?$qs"\nrel=next>next</a>}
 	}
 	if ($o > 0) {
-		$rv .= $n < $total ? '/' : ',      ';
+		$rv .= $n < $total ? '/' : '       ';
 		my $p = $o - $LIM;
 		my $qs = $q->qs_html(o => ($p > 0 ? $p : 0));
-		$rv .= qq{<a\nhref="?$qs">prev</a>};
-	}
-	$rv;
-}
-
-sub tdump {
-	my ($cb, $res, $mset, $q, $ctx) = @_;
-	my $fh = $cb->([200, ['Content-Type'=>'text/html; charset=UTF-8']]);
-	$fh->write($res);
-	my %pct;
-	my @m = map {
-		my $i = $_;
-		my $m = PublicInbox::SearchMsg->load_doc($i->get_document);
-		$pct{$m->mid} = $i->get_percent;
-		$m = $m->mini_mime;
-		$m;
-	} ($mset->items);
-
-	require PublicInbox::Thread;
-	my $th = PublicInbox::Thread->new(@m);
-	{
-		no warnings 'once';
-		$Mail::Thread::nosubject = 0;
-	}
-	$th->thread;
-	if ($q->{r}) {
-		$th->order(sub {
-			sort { (eval { $pct{$b->topmost->messageid} } || 0)
-					<=>
-				(eval { $pct{$a->topmost->messageid} } || 0)
-			} @_;
-		});
-	} else {
-		no warnings 'once';
-		$th->order(*PublicInbox::View::rsort_ts);
+		$rv .= qq{<a\nhref="?$qs"\nrel=prev>prev</a>};
 	}
-
-	my $git = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir});
-	my $state = { ctx => $ctx, anchor_idx => 0, pct => \%pct };
-	$ctx->{searchview} = 1;
-	tdump_ent($fh, $git, $state, $_, 0) for $th->rootset;
-	Email::Address->purge_cache;
-
-	$fh->write(search_nav_bot($mset, $q). "\n\n" .
-			foot($ctx). '</pre></body></html>');
-
-	$fh->close;
+	$rv .= '</pre>';
 }
 
-sub tdump_ent {
-	my ($fh, $git, $state, $node, $level) = @_;
-	return unless $node;
-	my $mime = $node->message;
-
-	if ($mime) {
-		# lazy load the full message from mini_mime:
-		my $mid = $mime->header('Message-ID');
-		$mime = eval {
-			my $path = mid2path(mid_clean($mid));
-			Email::MIME->new($git->cat_file('HEAD:'.$path));
-		};
-	}
-	if ($mime) {
-		PublicInbox::View::index_entry($fh, $mime, $level, $state);
-	} else {
-		my $mid = $node->messageid;
-		$fh->write(PublicInbox::View::ghost_table('', $mid, $level));
-	}
-	tdump_ent($fh, $git, $state, $node->child, $level + 1);
-	tdump_ent($fh, $git, $state, $node->next, $level);
+sub sort_relevance {
+	my ($pct) = @_;
+	sub {
+		[ sort { (eval { $pct->{$b->topmost->{id}} } || 0)
+				<=>
+			(eval { $pct->{$a->topmost->{id}} } || 0)
+	} @{$_[0]} ] };
 }
 
-sub foot {
-	my ($ctx) = @_;
-	my $foot = $ctx->{footer} || '';
-	qq{Back to <a\nhref=".">index</a>.\n$foot};
+sub mset_thread {
+	my ($ctx, $mset, $q) = @_;
+	my %pct;
+	my $msgs = $ctx->{srch}->retry_reopen(sub { [ map {
+		my $i = $_;
+		my $smsg = PublicInbox::SearchMsg->load_doc($i->get_document);
+		$pct{$smsg->mid} = $i->get_percent;
+		$smsg;
+	} ($mset->items) ]});
+	my $r = $q->{r};
+	my $rootset = PublicInbox::SearchThread::thread($msgs,
+		$r ? sort_relevance(\%pct) : *PublicInbox::View::sort_ts);
+	my $skel = search_nav_bot($mset, $q). "<pre>";
+	my $inbox = $ctx->{-inbox};
+	$ctx->{-upfx} = '';
+	$ctx->{anchor_idx} = 1;
+	$ctx->{cur_level} = 0;
+	$ctx->{dst} = \$skel;
+	$ctx->{mapping} = {};
+	$ctx->{pct} = \%pct;
+	$ctx->{prev_attr} = '';
+	$ctx->{prev_level} = 0;
+	$ctx->{seen} = {};
+	$ctx->{s_nr} = scalar(@$msgs).'+ results';
+
+	# reduce hash lookups in skel_dump
+	$ctx->{-obfuscate} = $ctx->{-inbox}->{obfuscate};
+	PublicInbox::View::walk_thread($rootset, $ctx,
+		*PublicInbox::View::pre_thread);
+
+	@$msgs = reverse @$msgs if $r;
+	my $mime;
+	sub {
+		return unless $msgs;
+		while ($mime = pop @$msgs) {
+			$mime = $inbox->msg_by_smsg($mime) and last;
+		}
+		if ($mime) {
+			$mime = PublicInbox::MIME->new($mime);
+			return PublicInbox::View::index_entry($mime, $ctx,
+				scalar @$msgs);
+		}
+		$msgs = undef;
+		$skel .= "\n</pre>";
+	};
 }
 
-sub html_start {
+sub ctx_prepare {
 	my ($q, $ctx) = @_;
-	my $query = PublicInbox::Hval->new_oneline($q->{q});
-
-	my $qh = $query->as_html;
-	my $A = $q->qs_html(x => 'A', r => undef);
-	my $res = "<html><head><title>$qh - search results</title>" .
-		qq{<link\nrel=alternate\ntitle="Atom feed"\n} .
-		qq!href="?$A"\ntype="application/atom+xml"/></head>! .
-		qq{<body><form\naction="">} .
-		qq{<input\nname=q\nvalue="$qh"\ntype=text />};
-
-	$res .= qq{<input\ntype=hidden\nname=r />} if $q->{r};
+	my $qh = $q->{'q'};
+	utf8::decode($qh);
+	$qh = ascii_html($qh);
+	$ctx->{-q_value_html} = $qh;
+	$ctx->{-atom} = '?'.$q->qs_html(x => 'A', r => undef);
+	$ctx->{-title_html} = "$qh - search results";
+	my $extra = '';
+	$extra .= qq{<input\ntype=hidden\nname=r />} if $q->{r};
 	if (my $x = $q->{x}) {
-		my $xh = PublicInbox::Hval->new_oneline($x)->as_html;
-		$res .= qq{<input\ntype=hidden\nname=x\nvalue="$xh" />};
+		$x = ascii_html($x);
+		$extra .= qq{<input\ntype=hidden\nname=x\nvalue="$x" />};
 	}
-
-	$res .= qq{<input\ntype=submit\nvalue=search /></form>};
+	$ctx->{-extra_form_html} = $extra;
 }
 
 sub adump {
 	my ($cb, $mset, $q, $ctx) = @_;
-	my $fh = $cb->([ 200, ['Content-Type' => 'application/atom+xml']]);
-	my $git = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir});
-	my $feed_opts = PublicInbox::Feed::get_feedopts($ctx);
-	my $x = PublicInbox::Hval->new_oneline($q->{q})->as_html;
-	$x = qq{$x - search results};
-	$feed_opts->{atomurl} = $feed_opts->{url} . '?'. $q->qs_html;
-	$feed_opts->{url} .= '?'. $q->qs_html(x => undef);
-	$x = PublicInbox::Feed::atom_header($feed_opts, $x);
-	$fh->write($x. PublicInbox::Feed::feed_updated());
-
-	for ($mset->items) {
-		$x = PublicInbox::SearchMsg->load_doc($_->get_document)->mid;
-		$x = mid2path($x);
-		PublicInbox::Feed::add_to_feed($feed_opts, $fh, $x, $git);
-	}
-	PublicInbox::Feed::end_feed($fh);
+	my $ibx = $ctx->{-inbox};
+	my @items = $mset->items;
+	$ctx->{search_query} = $q;
+	my $srch = $ctx->{srch};
+	PublicInbox::WwwAtomStream->response($ctx, 200, sub {
+		while (my $x = shift @items) {
+			$x = load_doc_retry($srch, $x);
+			$x = $ibx->msg_by_smsg($x) and
+					return PublicInbox::MIME->new($x);
+		}
+		return undef;
+	});
 }
 
 package PublicInbox::SearchQuery;
 use strict;
 use warnings;
+use URI::Escape qw(uri_escape);
 use PublicInbox::Hval;
+use PublicInbox::MID qw(MID_ESC);
 
 sub new {
-	my ($class, $cgi) = @_;
-	my $r = $cgi->param('r');
+	my ($class, $qp) = @_;
+
+	my $r = $qp->{r};
 	bless {
-		q => $cgi->param('q'),
-		x => $cgi->param('x') || '',
-		o => int($cgi->param('o') || 0) || 0,
+		q => $qp->{'q'},
+		x => $qp->{x} || '',
+		o => (($qp->{o} || '0') =~ /(\d+)/),
 		r => (defined $r && $r ne '0'),
 	}, $class;
 }
@@ -276,7 +313,7 @@ sub qs_html {
 		$self = $tmp;
 	}
 
-	my $q = PublicInbox::Hval->new($self->{q})->as_href;
+	my $q = uri_escape($self->{'q'}, MID_ESC);
 	$q =~ s/%20/+/g; # improve URL readability
 	my $qs = "q=$q";