X-Git-Url: http://www.git.stargrave.org/?p=public-inbox.git;a=blobdiff_plain;f=lib%2FPublicInbox%2FView.pm;h=687a0acb9367737410d5ff4aae641513fd49b629;hp=44130b904d94c32fc3356268f6265e39f1ee02c5;hb=ef9ce6f2dfc9fcbb91e3cec0b7a2b7f7e359be5d;hpb=e29518088b3f4a4759280f76dd5416c376ee683e diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm index 44130b90..687a0acb 100644 --- a/lib/PublicInbox/View.pm +++ b/lib/PublicInbox/View.pm @@ -6,16 +6,14 @@ package PublicInbox::View; use strict; use warnings; -use URI::Escape qw/uri_escape_utf8/; use Date::Parse qw/str2time/; -use Encode::MIME::Header; -use Plack::Util; -use PublicInbox::Hval qw/ascii_html/; +use PublicInbox::Hval qw/ascii_html obfuscate_addrs/; use PublicInbox::Linkify; -use PublicInbox::MID qw/mid_clean id_compress mid_mime/; +use PublicInbox::MID qw/mid_clean id_compress mid_mime mid_escape/; use PublicInbox::MsgIter; use PublicInbox::Address; use PublicInbox::WwwStream; +use PublicInbox::Reply; require POSIX; use constant INDENT => ' '; @@ -24,14 +22,15 @@ sub th_pfx ($) { $_[0] == 0 ? '' : TCHILD }; # public functions: (unstable) sub msg_html { - my ($ctx, $mime, $footer) = @_; + my ($ctx, $mime) = @_; my $hdr = $mime->header_obj; - my $tip = _msg_html_prepare($hdr, $ctx); - PublicInbox::WwwStream->new($ctx, sub { + my $obfs = $ctx->{-inbox}->{obfuscate}; + my $tip = _msg_html_prepare($hdr, $ctx, $obfs); + PublicInbox::WwwStream->response($ctx, 200, sub { my ($nr, undef) = @_; if ($nr == 1) { - $tip . multipart_text_as_html($mime, '') . - '
' + $tip . multipart_text_as_html($mime, '', $obfs) . + '
' } elsif ($nr == 2) { # fake an EOF if generating the footer fails; # we want to at least show the message if something @@ -49,41 +48,65 @@ sub msg_html { # /$INBOX/$MESSAGE_ID/#R sub msg_reply { my ($ctx, $hdr) = @_; - my $se_url = 'https://git-htmldocs.bogomips.org/git-send-email.html'; + my $se_url = + 'https://kernel.org/pub/software/scm/git/docs/git-send-email.html'; + my $p_url = + 'https://en.wikipedia.org/wiki/Posting_style#Interleaved_style'; - my ($arg, $link) = mailto_arg_link($hdr); + my $info = ''; + my $ibx = $ctx->{-inbox}; + if (my $url = $ibx->{infourl}) { + $url = PublicInbox::Hval::prurl($ctx->{env}, $url); + $info = qq(\n List information: $url\n); + } + + my ($arg, $link) = PublicInbox::Reply::mailto_arg_link($ibx, $hdr); push @$arg, '/path/to/YOUR_REPLY'; + $arg = ascii_html(join(" \\\n ", '', @$arg)); + <
Reply instructions:
+
+You may reply publically to this message via plain-text email
+using any one of the following methods:
+
+* Save the following mbox file, import it into your mail client,
+  and reply-to-all from there: mbox
+
+  Avoid top-posting and favor interleaved quoting:
+  $p_url
+$info
+* Reply to all the recipients using the --to, --cc,
+  and --in-reply-to switches of git-send-email(1):
+
+  git send-email$arg
+
+  $se_url
 
-	"".
-	"You may reply publically to this message via\n".
-	"plain-text email using any one of the following methods:\n\n" .
-	"* Save the following mbox file, import it into your mail client,\n" .
-	"  and reply-to-all from there: mbox\n\n" .
-	"* Reply to all the recipients using the --to, --cc,\n" .
-	"  and --in-reply-to switches of git-send-email(1):\n\n" .
-	"  git send-email \\\n    " .
-	join(" \\\n    ", @$arg ). "\n\n" .
-	qq(  $se_url\n\n) .
-	"* If your mail client supports setting the In-Reply-To" .
-	" header\n  via mailto: links, try the " .
-	qq(mailto: link\n) .
-	'
'; +* If your mail client supports setting the In-Reply-To header + via mailto: links, try the mailto: link +EOF } sub in_reply_to { my ($hdr) = @_; - my $irt = $hdr->header_raw('In-Reply-To'); - - return mid_clean($irt) if (defined $irt); - - my $refs = $hdr->header_raw('References'); - if ($refs && $refs =~ /<([^>]+)>\s*\z/s) { - return $1; + my %mid = map { $_ => 1 } $hdr->header_raw('Message-ID'); + my @refs = ($hdr->header_raw('References'), + $hdr->header_raw('In-Reply-To')); + @refs = ((join(' ', @refs)) =~ /<([^>]+)>/g); + while (defined(my $irt = pop @refs)) { + next if $mid{"<$irt>"}; + return $irt; } undef; } -sub _hdr_names ($$) { +sub _hdr_names_html ($$) { my ($hdr, $field) = @_; my $val = $hdr->header($field) or return ''; ascii_html(join(', ', PublicInbox::Address::names($val))); @@ -105,60 +128,75 @@ sub index_entry { my $mid_raw = mid_clean(mid_mime($mime)); my $id = id_compress($mid_raw, 1); my $id_m = 'm'.$id; - my $mid = PublicInbox::Hval->new_msgid($mid_raw); my $root_anchor = $ctx->{root_anchor} || ''; - my $irt = in_reply_to($hdr); - - my $rv = ''.ascii_html($subj).''; - $rv = "$rv" if $root_anchor eq $id_m; - $rv .= "\n"; - $rv .= _th_index_lite($mid_raw, $irt, $id, $ctx); + my $irt; + my $obfs = $ctx->{-obfuscate}; + + my $rv = "* "; + $subj = ''.ascii_html($subj).''; + obfuscate_addrs($subj) if $obfs; + $subj = "$subj" if $root_anchor eq $id_m; + $rv .= $subj . "\n"; + $rv .= _th_index_lite($mid_raw, \$irt, $id, $ctx); my @tocc; foreach my $f (qw(To Cc)) { - my $dst = _hdr_names($hdr, $f); - push @tocc, "$f: $dst" if $dst ne ''; + my $dst = _hdr_names_html($hdr, $f); + if ($dst ne '') { + obfuscate_addrs($dst) if $obfs; + push @tocc, "$f: $dst"; + } } - $rv .= "From: "._hdr_names($hdr, 'From').' @ '._msg_date($hdr)." UTC"; + my $from = _hdr_names_html($hdr, 'From'); + obfuscate_addrs($from) if $obfs; + $rv .= "From: $from @ "._msg_date($hdr)." UTC"; my $upfx = $ctx->{-upfx}; - my $mhref = $upfx . $mid->as_href . '/'; - $rv .= qq{ (permalink / }; - $rv .= qq{raw)\n}; + my $mhref = $upfx . mid_escape($mid_raw) . '/'; + $rv .= qq{ (permalink / }; + $rv .= qq{raw)\n}; $rv .= ' '.join('; +', @tocc) . "\n" if @tocc; my $mapping = $ctx->{mapping}; - if (!$mapping && $irt) { - my $mirt = PublicInbox::Hval->msgid($irt); - my $href = $upfx . $mirt->as_href . '/'; + if (!$mapping && (defined($irt) || defined($irt = in_reply_to($hdr)))) { + my $mirt = PublicInbox::Hval->new_msgid($irt); + my $href = $upfx . $mirt->{href}. '/'; my $html = $mirt->as_html; - $rv .= qq(In-Reply-To: <$html>\n) + $rv .= qq(In-Reply-To: <$html>\n) } $rv .= "\n"; # scan through all parts, looking for displayable text - msg_iter($mime, sub { $rv .= add_text_body($mhref, $_[0]) }); + msg_iter($mime, sub { $rv .= add_text_body($mhref, $obfs, $_[0]) }); # add the footer $rv .= "\n^ ". "permalink" . - " / raw" . - " / reply"; + " raw" . + " reply"; + + my $hr; if (my $pct = $ctx->{pct}) { # used by SearchView.pm - $rv .= " [relevance $pct->{$mid_raw}%]"; + $rv .= "\t[relevance $pct->{$mid_raw}%]"; + $hr = 1; } elsif ($mapping) { my $threaded = 'threaded'; my $flat = 'flat'; my $end = ''; if ($ctx->{flat}) { + $hr = 1; $flat = "$flat"; } else { $threaded = "$threaded"; } - $rv .= " [$threaded"; - $rv .= "|$flat]"; + $rv .= "\t[$flat"; + $rv .= "|$threaded]"; + $rv .= " $ctx->{s_nr}"; + } else { + $hr = $ctx->{-hr}; } - $rv .= $more ? "\n\n" : "\n"; + $rv .= $more ? '
' : '
' if $hr; + $rv; } sub pad_link ($$;$) { @@ -173,82 +211,83 @@ sub _th_index_lite { my $rv = ''; my $mapping = $ctx->{mapping} or return $rv; my $pad = ' '; - # map = [children, attr, node, idx, level] - my $map = $mapping->{$mid_raw}; - my $nr_c = scalar @{$map->[0]}; + my $mid_map = $mapping->{$mid_raw}; + defined $mid_map or + return 'public-inbox BUG: '.ascii_html($mid_raw).' not mapped'; + my ($attr, $node, $idx, $level) = @$mid_map; + my $children = $node->{children}; + my $nr_c = scalar @$children; my $nr_s = 0; - my $level = $map->[4]; - my $idx = $map->[3]; - if (defined $irt) { - my $irt_map = $mapping->{$irt}; - my $siblings = $irt_map->[0]; + my $siblings; + if (my $smsg = $node->{smsg}) { + ($$irt) = (($smsg->{references} || '') =~ m/<([^>]+)>\z/); + } + my $irt_map = $mapping->{$$irt} if defined $$irt; + if (defined $irt_map) { + $siblings = $irt_map->[1]->{children}; $nr_s = scalar(@$siblings) - 1; - $rv .= $pad . $irt_map->[1]; + $rv .= $pad . $irt_map->[0]; if ($idx > 0) { my $prev = $siblings->[$idx - 1]; - my $pmid = $prev->messageid; + my $pmid = $prev->{id}; if ($idx > 2) { my $s = ($idx - 1). ' preceding siblings ...'; $rv .= pad_link($pmid, $level, $s); } elsif ($idx == 2) { - my $ppmid = $siblings->[0]->messageid; - $rv .= $pad . $mapping->{$ppmid}->[1]; + my $ppmid = $siblings->[0]->{id}; + $rv .= $pad . $mapping->{$ppmid}->[0]; } - $rv .= $pad . $mapping->{$pmid}->[1]; + $rv .= $pad . $mapping->{$pmid}->[0]; } } my $s_s = nr_to_s($nr_s, 'sibling', 'siblings'); my $s_c = nr_to_s($nr_c, 'reply', 'replies'); - my $this = $map->[1]; - $this =~ s!\n\z!\n!s; - $this =~ s! !!s; # no point in duplicating subject - $rv .= "@ $this"; - my $node = $map->[2]; - if (my $child = $node->child) { - my $cmid = $child->messageid; - $rv .= $pad . $mapping->{$cmid}->[1]; + $attr =~ s!\n\z!\n!s; + $attr =~ s! !!s; # no point in duplicating subject + $attr =~ s!]+>([^<]+)!$1!s; # no point linking to self + $rv .= "@ $attr"; + if ($nr_c) { + my $cmid = $children->[0]->{id}; + $rv .= $pad . $mapping->{$cmid}->[0]; if ($nr_c > 2) { my $s = ($nr_c - 1). ' more replies'; $rv .= pad_link($cmid, $level + 1, $s); - } elsif (my $cn = $child->next) { - $rv .= $pad . $mapping->{$cn->messageid}->[1]; + } elsif (my $cn = $children->[1]) { + $rv .= $pad . $mapping->{$cn->{id}}->[0]; } } - if (my $next = $node->next) { - my $nmid = $next->messageid; - $rv .= $pad . $mapping->{$nmid}->[1]; + + my $next = $siblings->[$idx+1] if $siblings && $idx >= 0; + if ($next) { + my $nmid = $next->{id}; + $rv .= $pad . $mapping->{$nmid}->[0]; my $nnext = $nr_s - $idx; if ($nnext > 2) { my $s = ($nnext - 1).' subsequent siblings'; $rv .= pad_link($nmid, $level, $s); - } elsif (my $nn = $next->next) { - $rv .= $pad . $mapping->{$nn->messageid}->[1]; + } elsif (my $nn = $siblings->[$idx + 2]) { + $rv .= $pad . $mapping->{$nn->{id}}->[0]; } } - $rv .= "_ $s_s, $s_c\n"; + $rv .= $pad ."$s_s, $s_c; $ctx->{s_nr}\n"; } sub walk_thread { - my ($th, $ctx, $cb) = @_; - my @q = map { (0, $_) } $th->rootset; + my ($rootset, $ctx, $cb) = @_; + my @q = map { (0, $_, -1) } @$rootset; while (@q) { - my $level = shift @q; - my $node = shift @q or next; - $cb->($ctx, $level, $node); - unshift @q, $level+1, $node->child, $level, $node->next; + my ($level, $node, $i) = splice(@q, 0, 3); + defined $node or next; + $cb->($ctx, $level, $node, $i); + ++$level; + $i = 0; + unshift @q, map { ($level, $_, $i++) } @{$node->{children}}; } } sub pre_thread { - my ($ctx, $level, $node) = @_; - my $mapping = $ctx->{mapping}; - my $idx = -1; - if (my $parent = $node->parent) { - my $m = $mapping->{$parent->messageid}->[0]; - $idx = scalar @$m; - push @$m, $node; - } - $mapping->{$node->messageid} = [ [], '', $node, $idx, $level ]; + my ($ctx, $level, $node, $idx) = @_; + $ctx->{mapping}->{$node->{id}} = [ '', $node, $idx, $level ]; skel_dump($ctx, $level, $node); } @@ -259,34 +298,37 @@ sub thread_index_entry { } sub stream_thread ($$) { - my ($th, $ctx) = @_; + my ($rootset, $ctx) = @_; my $inbox = $ctx->{-inbox}; my $mime; - my @q = map { (0, $_) } $th->rootset; + my @q = map { (0, $_) } @$rootset; my $level; while (@q) { $level = shift @q; my $node = shift @q or next; - unshift @q, $level+1, $node->child, $level, $node->next; - $mime = $inbox->msg_by_mid($node->messageid) and last; + my $cl = $level + 1; + unshift @q, map { ($cl, $_) } @{$node->{children}}; + $mime = $inbox->msg_by_smsg($node->{smsg}) and last; } return missing_thread($ctx) unless $mime; - $mime = Email::MIME->new($mime); + $ctx->{-obfuscate} = $ctx->{-inbox}->{obfuscate}; + $mime = PublicInbox::MIME->new($mime); $ctx->{-title_html} = ascii_html($mime->header('Subject')); $ctx->{-html_tip} = thread_index_entry($ctx, $level, $mime); - my $body = PublicInbox::WwwStream->new($ctx, sub { + PublicInbox::WwwStream->response($ctx, 200, sub { return unless $ctx; while (@q) { $level = shift @q; my $node = shift @q or next; - unshift @q, $level+1, $node->child, $level, $node->next; - my $mid = $node->messageid; - if ($mime = $inbox->msg_by_mid($mid)) { - $mime = Email::MIME->new($mime); + my $cl = $level + 1; + unshift @q, map { ($cl, $_) } @{$node->{children}}; + my $mid = $node->{id}; + if ($mime = $inbox->msg_by_smsg($node->{smsg})) { + $mime = PublicInbox::MIME->new($mime); return thread_index_entry($ctx, $level, $mime); } else { - return ghost_index_entry($ctx, $level, $mid); + return ghost_index_entry($ctx, $level, $node); } } my $ret = join('', thread_adj_level($ctx, 0)); @@ -294,22 +336,24 @@ sub stream_thread ($$) { $ctx = undef; $ret; }); - [ 200, ['Content-Type', 'text/html; charset=UTF-8'], $body ]; } sub thread_html { my ($ctx) = @_; my $mid = $ctx->{mid}; - my $sres = $ctx->{srch}->get_thread($mid, { asc => 1 }); - my $msgs = load_results($sres); + my $srch = $ctx->{srch}; + my $sres = $srch->get_thread($mid); + my $msgs = load_results($srch, $sres); my $nr = $sres->{total}; return missing_thread($ctx) if $nr == 0; - my $skel = '
';
+	my $skel = '
';
 	$skel .= $nr == 1 ? 'only message in thread' : 'end of thread';
-	$skel .= ", back to index";
-	$skel .= "\n$nr+ messages in thread: (download: ";
-	$skel .= "mbox.gz";
+	$skel .= ", back to index\n\n";
+	$skel .= "Thread overview: ";
+	$skel .= $nr == 1 ? '(only message)' : "$nr+ messages";
+	$skel .= " (download: mbox.gz";
 	$skel .= " / follow: Atom feed)\n";
+	$skel .= "-- links below jump to the message on this page --\n";
 	$ctx->{-upfx} = '../../';
 	$ctx->{cur_level} = 0;
 	$ctx->{dst} = \$skel;
@@ -318,48 +362,48 @@ sub thread_html {
 	$ctx->{root_anchor} = anchor_for($mid);
 	$ctx->{seen} = {};
 	$ctx->{mapping} = {};
+	$ctx->{s_nr} = "$nr+ messages in thread";
+
+	my $rootset = thread_results($msgs);
+
+	# reduce hash lookups in pre_thread->skel_dump
+	$ctx->{-obfuscate} = $ctx->{-inbox}->{obfuscate};
+	walk_thread($rootset, $ctx, *pre_thread);
 
-	my $th = thread_results($msgs);
-	walk_thread($th, $ctx, *pre_thread);
 	$skel .= '
'; - return stream_thread($th, $ctx) unless $ctx->{flat}; + return stream_thread($rootset, $ctx) unless $ctx->{flat}; - # flat display: lazy load the full message from mini_mime: + # flat display: lazy load the full message from smsg my $inbox = $ctx->{-inbox}; my $mime; while ($mime = shift @$msgs) { - $mime = $inbox->msg_by_mid(mid_clean(mid_mime($mime))) and last; + $mime = $inbox->msg_by_smsg($mime) and last; } return missing_thread($ctx) unless $mime; - $mime = Email::MIME->new($mime); + $mime = PublicInbox::MIME->new($mime); $ctx->{-title_html} = ascii_html($mime->header('Subject')); $ctx->{-html_tip} = '
'.index_entry($mime, $ctx, scalar @$msgs);
 	$mime = undef;
-	my $body = PublicInbox::WwwStream->new($ctx, sub {
+	PublicInbox::WwwStream->response($ctx, 200, sub {
 		return unless $msgs;
 		while ($mime = shift @$msgs) {
-			$mid = mid_clean(mid_mime($mime));
-			$mime = $inbox->msg_by_mid($mid) and last;
+			$mime = $inbox->msg_by_smsg($mime) and last;
 		}
 		if ($mime) {
-			$mime = Email::MIME->new($mime);
+			$mime = PublicInbox::MIME->new($mime);
 			return index_entry($mime, $ctx, scalar @$msgs);
 		}
 		$msgs = undef;
-		'
'.$skel; + $skel; }); - [ 200, ['Content-Type', 'text/html; charset=UTF-8'], $body ]; } sub multipart_text_as_html { - my ($mime, $upfx) = @_; + my ($mime, $upfx, $obfs) = @_; my $rv = ""; # scan through all parts, looking for displayable text - msg_iter($mime, sub { - my ($p) = @_; - $rv .= add_text_body($upfx, $p); - }); + msg_iter($mime, sub { $rv .= add_text_body($upfx, $obfs, $_[0]) }); $rv; } @@ -377,14 +421,16 @@ sub flush_quote { $$s .= qq() . $rv . '' } -sub attach_link ($$$$) { - my ($upfx, $ct, $p, $fn) = @_; +sub attach_link ($$$$;$) { + my ($upfx, $ct, $p, $fn, $err) = @_; my ($part, $depth, @idx) = @$p; my $nl = $idx[-1] > 1 ? "\n" : ''; my $idx = join('.', @idx); my $size = bytes::length($part->body); - $ct ||= 'text/plain'; - $ct =~ s/;.*//; # no attributes + + # hide attributes normally, unless we want to aid users in + # spotting MUA problems: + $ct =~ s/;.*// unless $err; $ct = ascii_html($ct); my $desc = $part->header('Content-Description'); $desc = $fn unless defined $desc; @@ -397,62 +443,85 @@ sub attach_link ($$$$) { } else { $sfn = 'a.bin'; } - my @ret = qq($nl[-- Attachment #$idx: ); + my $ret = qq($nl); + if ($err) { + $ret .= +"[-- Warning: decoded text below may be mangled --]\n"; + } + $ret .= "[-- Attachment #$idx: "; my $ts = "Type: $ct, Size: $size bytes"; - push(@ret, ($desc eq '') ? "$ts --]" : "$desc --]\n[-- $ts --]"); - join('', @ret, "\n"); + $desc = ascii_html($desc); + $ret .= ($desc eq '') ? "$ts --]" : "$desc --]\n[-- $ts --]"; + $ret .= "\n"; } sub add_text_body { - my ($upfx, $p) = @_; # from msg_iter: [ Email::MIME, depth, @idx ] - my ($part, $depth, @idx) = @$p; - my $ct = $part->content_type; + my ($upfx, $obfs, $p) = @_; + # $p - from msg_iter: [ Email::MIME, depth, @idx ] + my ($part, $depth) = @$p; # attachment @idx is unused + my $ct = $part->content_type || 'text/plain'; my $fn = $part->filename; - if (defined $ct && $ct =~ m!\btext/x?html\b!i) { + if ($ct =~ m!\btext/x?html\b!i) { return attach_link($upfx, $ct, $p, $fn); } my $s = eval { $part->body_str }; # badly-encoded message? tell the world about it! - return attach_link($upfx, $ct, $p, $fn) if $@; + my $err = $@; + if ($err) { + if ($ct =~ m!\btext/plain\b!i) { + # Try to assume UTF-8 because Alpine seems to + # do wacky things and set charset=X-UNKNOWN + $part->charset_set('UTF-8'); + $s = eval { $part->body_str }; + + # If forcing charset=UTF-8 failed, + # attach_link will warn further down... + $s = $part->body if $@; + } else { + return attach_link($upfx, $ct, $p, $fn); + } + } my @lines = split(/^/m, $s); $s = ''; - if (defined($fn) || $depth > 0) { - $s .= attach_link($upfx, $ct, $p, $fn); + if (defined($fn) || $depth > 0 || $err) { + $s .= attach_link($upfx, $ct, $p, $fn, $err); $s .= "\n"; } my @quot; my $l = PublicInbox::Linkify->new; - while (defined(my $cur = shift @lines)) { + foreach my $cur (@lines) { if ($cur !~ /^>/) { # show the previously buffered quote inline flush_quote(\$s, $l, \@quot) if @quot; # regular line, OK - $cur = $l->linkify_1($cur); - $cur = ascii_html($cur); - $s .= $l->linkify_2($cur); + $l->linkify_1($cur); + $s .= $l->linkify_2(ascii_html($cur)); } else { push @quot, $cur; } } - my $end = "\n"; - if (@quot) { - $end = ''; + if (@quot) { # ugh, top posted flush_quote(\$s, $l, \@quot); + obfuscate_addrs($s) if $obfs; + $s; + } else { + obfuscate_addrs($s) if $obfs; + if ($s =~ /\n\z/s) { # common, last line ends with a newline + $s; + } else { # some editors don't do newlines... + $s .= "\n"; + } } - $s =~ s/[ \t]+$//sgm; # kill per-line trailing whitespace - $s =~ s/\A\n+//s; # kill leading blank lines - $s =~ s/\s+\z//s; # kill all trailing spaces - $s .= $end; } sub _msg_html_prepare { - my ($hdr, $ctx) = @_; + my ($hdr, $ctx, $obfs) = @_; my $srch = $ctx->{srch} if $ctx; my $atom = ''; my $rv = ""; # anchor for body start @@ -461,7 +530,7 @@ sub _msg_html_prepare { $ctx->{-upfx} = '../'; } my @title; - my $mid = $hdr->header_raw('Message-ID'); + my $mid = mid_clean($hdr->header_raw('Message-ID')); $mid = PublicInbox::Hval->new_msgid($mid); foreach my $h (qw(From To Cc Subject Date)) { my $v = $hdr->header($h); @@ -471,6 +540,7 @@ sub _msg_html_prepare { if ($h eq 'From') { my @n = PublicInbox::Address::names($v->raw); $title[1] = ascii_html(join(', ', @n)); + obfuscate_addrs($title[1]) if $obfs; } elsif ($h eq 'Subject') { $title[0] = $v->as_html; if ($srch) { @@ -480,7 +550,7 @@ sub _msg_html_prepare { } } $v = $v->as_html; - $v =~ s/(\@[^,]+,) /$1\n\t/g if ($h eq 'Cc' || $h eq 'To'); + obfuscate_addrs($v) if $obfs; $rv .= "$h: $v\n"; } @@ -498,11 +568,12 @@ sub thread_skel { my $mid = mid_clean($hdr->header_raw('Message-ID')); my $sres = $srch->get_thread($mid); my $nr = $sres->{total}; - my $expand = qq(expand ) . + my $expand = qq(expand ) . qq(/ mbox.gz ) . qq(/ Atom feed); my $parent = in_reply_to($hdr); + $$dst .= "\nThread overview: "; if ($nr <= 1) { if (defined $parent) { $$dst .= "($expand)\n "; @@ -524,7 +595,12 @@ sub thread_skel { $ctx->{prev_attr} = ''; $ctx->{prev_level} = 0; $ctx->{dst} = $dst; - walk_thread(thread_results(load_results($sres)), $ctx, *skel_dump); + $sres = load_results($srch, $sres); + + # reduce hash lookups in skel_dump + $ctx->{-obfuscate} = $ctx->{-inbox}->{obfuscate}; + walk_thread(thread_results($sres), $ctx, *skel_dump); + $ctx->{parent_msg} = $parent; } @@ -534,9 +610,9 @@ sub _parent_headers { my $irt = in_reply_to($hdr); if (defined $irt) { - my $v = PublicInbox::Hval->new_msgid($irt, 1); + my $v = PublicInbox::Hval->new_msgid($irt); my $html = $v->as_html; - my $href = $v->as_href; + my $href = $v->{href}; $rv .= "In-Reply-To: <"; $rv .= "$html>\n"; } @@ -565,50 +641,6 @@ sub _parent_headers { $rv; } -sub squote_maybe ($) { - my ($val) = @_; - if ($val =~ m{([^\w@\./,\%\+\-])}) { - $val =~ s/(['!])/'\\$1'/g; # '!' for csh - return "'$val'"; - } - $val; -} - -sub mailto_arg_link { - my ($hdr) = @_; - my %cc; # everyone else - my $to; # this is the From address - - foreach my $h (qw(From To Cc)) { - my $v = $hdr->header($h); - defined($v) && ($v ne '') or next; - my @addrs = PublicInbox::Address::emails($v); - foreach my $address (@addrs) { - my $dst = lc($address); - $cc{$dst} ||= $address; - $to ||= $dst; - } - } - my @arg; - - my $subj = $hdr->header('Subject') || ''; - $subj = "Re: $subj" unless $subj =~ /\bRe:/i; - my $mid = $hdr->header_raw('Message-ID'); - push @arg, '--in-reply-to='.ascii_html(squote_maybe(mid_clean($mid))); - my $irt = uri_escape_utf8($mid); - delete $cc{$to}; - push @arg, '--to=' . ascii_html($to); - $to = uri_escape_utf8($to); - $subj = uri_escape_utf8($subj); - my $cc = join(',', sort values %cc); - push @arg, '--cc=' . ascii_html($cc); - $cc = uri_escape_utf8($cc); - my $href = "mailto:$to?In-Reply-To=$irt&Cc=${cc}&Subject=$subj"; - $href =~ s/%20/+/g; - - (\@arg, ascii_html($href)); -} - sub html_footer { my ($hdr, $standalone, $ctx, $rhref) = @_; @@ -620,21 +652,35 @@ sub html_footer { if ($idx && $srch) { $idx .= "\n"; thread_skel(\$idx, $ctx, $hdr, $tpfx); - my $p = $ctx->{parent_msg}; - my $next = $ctx->{next_msg}; - if ($p) { - $p = PublicInbox::Hval->new_msgid($p); - $p = $p->as_href; - $irt = "parent "; - } else { - $irt = ' ' x length('parent '); + my ($next, $prev); + my $parent = ' '; + $next = $prev = ' '; + + if (my $n = $ctx->{next_msg}) { + $n = PublicInbox::Hval->new_msgid($n)->{href}; + $next = "next"; } - if ($next) { - my $n = PublicInbox::Hval->new_msgid($next)->as_href; - $irt .= "next "; - } else { - $irt .= ' ' x length('next '); + my $u; + my $par = $ctx->{parent_msg}; + if ($par) { + $u = PublicInbox::Hval->new_msgid($par)->{href}; + $u = "$upfx$u/"; + } + if (my $p = $ctx->{prev_msg}) { + $prev = PublicInbox::Hval->new_msgid($p)->{href}; + if ($p && $par && $p eq $par) { + $prev = "prev parent'; + $parent = ''; + } else { + $prev = "prev'; + $parent = " parent" if $u; + } + } elsif ($u) { # unlikely + $parent = " parent"; } + $irt = "$next $prev$parent "; } else { $irt = ''; } @@ -644,9 +690,9 @@ sub html_footer { } sub linkify_ref_nosrch { - my $v = PublicInbox::Hval->new_msgid($_[0], 1); + my $v = PublicInbox::Hval->new_msgid($_[0]); my $html = $v->as_html; - my $href = $v->as_href; + my $href = $v->{href}; "<$html>"; } @@ -657,11 +703,9 @@ sub anchor_for { sub ghost_parent { my ($upfx, $mid) = @_; - # 'subject dummy' is used internally by Mail::Thread - return '[no common parent]' if ($mid eq 'subject dummy'); $mid = PublicInbox::Hval->new_msgid($mid); - my $href = $mid->as_href; + my $href = $mid->{href}; my $html = $mid->as_html; qq{[parent not found: <$html>]}; } @@ -672,9 +716,9 @@ sub indent_for { } sub load_results { - my ($sres) = @_; - - [ map { $_->mini_mime } @{delete $sres->{msgs}} ]; + my ($srch, $sres) = @_; + my $msgs = delete $sres->{msgs}; + $srch->retry_reopen(sub { [ map { $_->mid; $_ } @$msgs ] }); } sub msg_timestamp { @@ -685,11 +729,8 @@ sub msg_timestamp { sub thread_results { my ($msgs) = @_; - require PublicInbox::Thread; - my $th = PublicInbox::Thread->new(@$msgs); - $th->thread; - $th->order(*sort_ts); - $th + require PublicInbox::SearchThread; + PublicInbox::SearchThread::thread($msgs, *sort_ts); } sub missing_thread { @@ -706,31 +747,35 @@ sub _msg_date { sub fmt_ts { POSIX::strftime('%Y-%m-%d %k:%M', gmtime($_[0])) } -sub _skel_header { - my ($ctx, $hdr, $level) = @_; +sub skel_dump { + my ($ctx, $level, $node) = @_; + my $smsg = $node->{smsg} or return _skel_ghost($ctx, $level, $node); my $dst = $ctx->{dst}; my $cur = $ctx->{cur}; - my $mid = mid_clean($hdr->header_raw('Message-ID')); - my $f = ascii_html($hdr->header('X-PI-From')); - my $d = _msg_date($hdr) . ' ' . indent_for($level) . th_pfx($level); + my $mid = $smsg->{mid}; + + my $f = ascii_html($smsg->from_name); + my $obfs = $ctx->{-obfuscate}; + obfuscate_addrs($f) if $obfs; + + my $d = fmt_ts($smsg->{ts}) . ' ' . indent_for($level) . th_pfx($level); my $attr = $f; $ctx->{first_level} ||= $level; if ($attr ne $ctx->{prev_attr} || $ctx->{prev_level} > $level) { $ctx->{prev_attr} = $attr; - } else { - $attr = ''; } $ctx->{prev_level} = $level; if ($cur) { if ($cur eq $mid) { delete $ctx->{cur}; - $$dst .= $d; - $$dst .= "". + $$dst .= "$d". "$attr [this message]\n"; return; + } else { + $ctx->{prev_msg} = $mid; } } else { $ctx->{next_msg} ||= $mid; @@ -739,175 +784,194 @@ sub _skel_header { # Subject is never undef, this mail was loaded from # our Xapian which would've resulted in '' if it were # really missing (and Filter rejects empty subjects) - my $s = $hdr->header('Subject'); - my $h = $ctx->{srch}->subject_path($s); + my $subj = $smsg->subject; + my $h = $ctx->{srch}->subject_path($subj); if ($ctx->{seen}->{$h}) { - $s = undef; + $subj = undef; } else { $ctx->{seen}->{$h} = 1; - $s = PublicInbox::Hval->new($s); - $s = $s->as_html; + $subj = PublicInbox::Hval->new($subj); + $subj = $subj->as_html; + obfuscate_addrs($subj) if $obfs; } - my $m = PublicInbox::Hval->new_msgid($mid); + my $m; my $id = ''; my $mapping = $ctx->{mapping}; - my $end = defined($s) ? "$s $f\n" : "$f\n"; + my $end = defined($subj) ? "$subj $f\n" : "$f\n"; if ($mapping) { my $map = $mapping->{$mid}; $id = id_compress($mid, 1); $m = '#m'.$id; - $map->[1] = "$d$end"; + $map->[0] = "$d$end"; $id = "\nid=r".$id; } else { - $m = $ctx->{-upfx}.$m->as_href.'/'; + $m = $ctx->{-upfx}.mid_escape($mid).'/'; } $$dst .= $d . "" . $end; } -sub skel_dump { +sub _skel_ghost { my ($ctx, $level, $node) = @_; - if (my $mime = $node->message) { - _skel_header($ctx, $mime->header_obj, $level); + + my $mid = $node->{id}; + my $d = $ctx->{pct} ? ' [irrelevant] ' # search result + : ' [not found] '; + $d .= indent_for($level) . th_pfx($level); + my $upfx = $ctx->{-upfx}; + my $m = PublicInbox::Hval->new_msgid($mid); + my $href = $upfx . $m->{href} . '/'; + my $html = $m->as_html; + + my $mapping = $ctx->{mapping}; + my $map = $mapping->{$mid} if $mapping; + if ($map) { + my $id = id_compress($mid, 1); + $map->[0] = $d . qq{<$html>\n}; + $d .= qq{<$html>\n}; } else { - my $mid = $node->messageid; - my $dst = $ctx->{dst}; - my $mapping = $ctx->{mapping}; - my $map = $mapping->{$mid} if $mapping; - if ($mid eq 'subject dummy') { - my $ncp = "\t[no common parent]\n"; - $map->[1] = $ncp if $map; - $$dst .= $ncp; - return; - } - my $d = $ctx->{pct} ? ' [irrelevant] ' # search result - : ' [not found] '; - $d .= indent_for($level) . th_pfx($level); - my $upfx = $ctx->{-upfx}; - my $m = PublicInbox::Hval->new_msgid($mid); - my $href = $upfx . $m->as_href . '/'; - my $html = $m->as_html; - - if ($map) { - my $id = id_compress($mid, 1); - $map->[1] = $d . qq{<$html>\n}; - $d .= qq{<$html>\n}; - } else { - $d .= qq{<$html>\n}; - } - $$dst .= $d; + $d .= qq{<$html>\n}; } + my $dst = $ctx->{dst}; + $$dst .= $d; } sub sort_ts { - sort { - (eval { $a->topmost->message->header('X-PI-TS') } || 0) <=> - (eval { $b->topmost->message->header('X-PI-TS') } || 0) - } @_; -} - -sub _tryload_ghost ($$) { - my ($srch, $mid) = @_; - my $smsg = $srch->lookup_mail($mid) or return; - $smsg->mini_mime; + [ sort { + (eval { $a->topmost->{smsg}->ts } || 0) <=> + (eval { $b->topmost->{smsg}->ts } || 0) + } @{$_[0]} ]; } # accumulate recent topics if search is supported -# returns 1 if done, undef if not -sub add_topic { +# returns 200 if done, 404 if not +sub acc_topic { my ($ctx, $level, $node) = @_; my $srch = $ctx->{srch}; - my $mid = $node->messageid; - my $x = $node->message || _tryload_ghost($srch, $mid); + my $mid = $node->{id}; + my $x = $node->{smsg} || $srch->lookup_mail($mid); my ($subj, $ts); + my $topic; if ($x) { - $x = $x->header_obj; - $subj = $x->header('Subject'); + $subj = $x->subject; $subj = $srch->subject_normalized($subj); - $ts = $x->header('X-PI-TS'); - } else { # ghost message, do not bump level - $ts = -666; - $subj = "<$mid>"; - } - if (++$ctx->{subjs}->{$subj} == 1) { - push @{$ctx->{order}}, [ $level, $subj ]; - } - my $exist = $ctx->{latest}->{$subj}; - if (!$exist || $exist->[1] < $ts) { - $ctx->{latest}->{$subj} = [ $mid, $ts ]; + $ts = $x->ts; + if ($level == 0) { + $topic = [ $ts, 1, { $subj => $mid }, $subj ]; + $ctx->{-cur_topic} = $topic; + push @{$ctx->{order}}, $topic; + return; + } + + $topic = $ctx->{-cur_topic}; # should never be undef + $topic->[0] = $ts if $ts > $topic->[0]; + $topic->[1]++; + my $seen = $topic->[2]; + if (scalar(@$topic) == 3) { # parent was a ghost + push @$topic, $subj; + } elsif (!$seen->{$subj}) { + push @$topic, $level, $subj; + } + $seen->{$subj} = $mid; # latest for subject + } else { # ghost message + return if $level != 0; # ignore child ghosts + $topic = [ -666, 0, {} ]; + $ctx->{-cur_topic} = $topic; + push @{$ctx->{order}}, $topic; } } -sub emit_topics { +sub dump_topics { my ($ctx) = @_; - my $order = $ctx->{order}; - my $subjs = $ctx->{subjs}; - my $latest = $ctx->{latest}; - my $fh = $ctx->{fh}; - return $fh->write("\n[No topics in range]
") unless scalar @$order; - my $pfx; - my $prev = 0; - my $prev_attr = ''; - my $cur; - my @recent; - while (defined(my $info = shift @$order)) { - my ($level, $subj) = @$info; - my $n = delete $subjs->{$subj}; - my ($mid, $ts) = @{delete $latest->{$subj}}; - $mid = PublicInbox::Hval->new_msgid($mid)->as_href; - $pfx = indent_for($level); - my $nl = $level == $prev ? "\n" : ''; - if ($nl && $cur) { - push @recent, $cur; - $cur = undef; - } - $cur ||= [ $ts, '' ]; - $cur->[0] = $ts if $ts > $cur->[0]; - $cur->[1] .= $nl . $pfx . th_pfx($level); - if ($ts == -666) { # ghost - $cur->[1] .= ghost_parent('', $mid) . "\n"; - next; # child will have mbox / atom link - } - - $subj = PublicInbox::Hval->new($subj)->as_html; - $cur->[1] .= "$subj\n"; + my $order = delete $ctx->{order}; # [ ts, subj1, subj2, subj3, ... ] + if (!@$order) { + $ctx->{-html_tip} = '
[No topics in range]
'; + return 404; + } + + my @out; + my $obfs = $ctx->{-inbox}->{obfuscate}; + + # sort by recency, this allows new posts to "bump" old topics... + foreach my $topic (sort { $b->[0] <=> $a->[0] } @$order) { + my ($ts, $n, $seen, $top, @ex) = @$topic; + @$topic = (); + next unless defined $top; # ghost topic + my $mid = delete $seen->{$top}; + my $href = mid_escape($mid); + $top = PublicInbox::Hval->new($top)->as_html; $ts = fmt_ts($ts); - my $attr = " $ts UTC"; # $n isn't the total number of posts on the topic, # just the number of posts in the current results window - $n = $n == 1 ? '' : " ($n+ messages)"; - - if ($level == 0 || $attr ne $prev_attr) { - my $mbox = qq(mbox.gz); - my $atom = qq(Atom); - $pfx .= INDENT if $level > 0; - $cur->[1] .= $pfx . $attr . $n . " - $mbox / $atom\n"; - $prev_attr = $attr; + my $anchor; + if ($n == 1) { + $n = ''; + $anchor = '#u'; # top of only message + } else { + $n = " ($n+ messages)"; + $anchor = '#t'; # thread skeleton } + + my $mbox = qq(mbox.gz); + my $atom = qq(Atom); + my $s = "$top\n" . + " $ts UTC $n - $mbox / $atom\n"; + for (my $i = 0; $i < scalar(@ex); $i += 2) { + my $level = $ex[$i]; + my $subj = $ex[$i + 1]; + $mid = delete $seen->{$subj}; + $subj = ascii_html($subj); + obfuscate_addrs($subj) if $obfs; + $href = mid_escape($mid); + $s .= indent_for($level) . TCHILD; + $s .= "$subj\n"; + } + push @out, $s; } - push @recent, $cur if $cur; - @recent = map { $_->[1] } sort { $b->[0] <=> $a->[0] } @recent; - $fh->write(join('', @recent) . ''); + $ctx->{-html_tip} = '
' . join("\n", @out) . '
'; + 200; } -sub emit_index_topics { +sub index_nav { # callback for WwwStream + my (undef, $ctx) = @_; + delete $ctx->{qp} or return; + my ($next, $prev); + $next = $prev = ' '; + my $latest = ''; + + my $next_o = $ctx->{-next_o}; + if ($next_o) { + $next = qq!next!; + } + if (my $cur_o = $ctx->{-cur_o}) { + $latest = qq! latest!; + + my $o = $cur_o - ($next_o - $cur_o); + if ($o > 0) { + $prev = qq!prev!; + } elsif ($o == 0) { + $prev = qq!prev!; + } + } + "
page: $next $prev$latest
"; +} + +sub index_topics { my ($ctx) = @_; - my ($off) = (($ctx->{cgi}->param('o') || '0') =~ /(\d+)/); + my ($off) = (($ctx->{qp}->{o} || '0') =~ /(\d+)/); + my $opts = { offset => $off, limit => 200 }; + $ctx->{order} = []; - $ctx->{subjs} = {}; - $ctx->{latest} = {}; - my $max = 25; - my %opts = ( offset => $off, limit => $max * 4 ); - while (scalar @{$ctx->{order}} < $max) { - my $sres = $ctx->{srch}->query('', \%opts); - my $nr = scalar @{$sres->{msgs}} or last; - $sres = load_results($sres); - walk_thread(thread_results($sres), $ctx, *add_topic); - $opts{offset} += $nr; + my $srch = $ctx->{srch}; + my $sres = $srch->query('', $opts); + my $nr = scalar @{$sres->{msgs}}; + if ($nr) { + $sres = load_results($srch, $sres); + walk_thread(thread_results($sres), $ctx, *acc_topic); } - - emit_topics($ctx); - $opts{offset}; + $ctx->{-next_o} = $off+ $nr; + $ctx->{-cur_o} = $off; + PublicInbox::WwwStream->response($ctx, dump_topics($ctx), *index_nav); } sub thread_adj_level { @@ -935,9 +999,10 @@ sub thread_adj_level { } sub ghost_index_entry { - my ($ctx, $level, $mid) = @_; + my ($ctx, $level, $node) = @_; my ($beg, $end) = thread_adj_level($ctx, $level); - $beg . '
'. ghost_parent($ctx->{-upfx}, $mid) . '
' . $end; + $beg . '
'. ghost_parent($ctx->{-upfx}, $node->{id})
+		. '
' . $end; } 1;