-# Copyright (C) 2014, Eric Wong <normalperson@yhbt.net> and all contributors
+# Copyright (C) 2014-2015 all contributors <meta@public-inbox.org>
# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
+#
+# Used for displaying the HTML web interface.
+# See Documentation/design_www.txt for this.
package PublicInbox::View;
use strict;
use warnings;
use URI::Escape qw/uri_escape_utf8/;
use Date::Parse qw/str2time/;
-use Encode qw/find_encoding/;
use Encode::MIME::Header;
-use Email::MIME::ContentType qw/parse_content_type/;
-use PublicInbox::Hval;
-use PublicInbox::MID qw/mid_clean mid_compressed mid2path/;
-use Digest::SHA;
+use Plack::Util;
+use PublicInbox::Hval qw/ascii_html/;
+use PublicInbox::Linkify;
+use PublicInbox::MID qw/mid_clean id_compress mid2path mid_mime/;
+use PublicInbox::MsgIter;
+use PublicInbox::Address;
require POSIX;
-# TODO: make these constants tunable
-use constant MAX_INLINE_QUOTED => 12; # half an 80x24 terminal
-use constant MAX_TRUNC_LEN => 72;
-use constant PRE_WRAP => "<pre\nstyle=\"white-space:pre-wrap\">";
-use constant T_ANCHOR => '#u';
+use constant INDENT => ' ';
+use constant TCHILD => '` ';
+sub th_pfx ($) { $_[0] == 0 ? '' : TCHILD };
-*ascii_html = *PublicInbox::Hval::ascii_html;
-
-my $enc_utf8 = find_encoding('UTF-8');
-
-# public functions:
+# public functions: (unstable)
+# TODO: stream this, since threading is expensive but also oh-so-important
sub msg_html {
- my ($class, $mime, $full_pfx, $footer, $srch) = @_;
- if (defined $footer) {
- $footer = "\n" . $footer;
- } else {
- $footer = '';
- }
- headers_to_html_header($mime, $full_pfx, $srch) .
- multipart_text_as_html($mime, $full_pfx) .
- '</pre><hr />' . PRE_WRAP .
- html_footer($mime, 1, $full_pfx, $srch) . $footer .
- '</pre></body></html>';
+ my ($ctx, $mime, $footer) = @_;
+ $footer = defined($footer) ? "\n$footer" : '';
+ my $hdr = $mime->header_obj;
+ my $n = 0;
+ Plack::Util::inline_object(
+ close => sub {}, # noop
+ getline => sub {
+ my $nr = $n++;
+ if ($nr == 0) {
+ headers_to_html_header($hdr, $ctx) .
+ multipart_text_as_html($mime, '') .
+ '</pre><hr />'
+ } elsif ($nr == 1) {
+ '<pre>' .
+ html_footer($hdr, 1, $ctx) .
+ '</pre>' . msg_reply($ctx, $hdr) .
+ '<hr /><pre>'. $footer .
+ '</pre></body></html>'
+ } else {
+ undef
+ }
+ }
+ )
+}
+
+# /$INBOX/$MESSAGE_ID/#R
+sub msg_reply {
+ my ($ctx, $hdr) = @_;
+ my $se_url =
+ 'https://kernel.org/pub/software/scm/git/docs/git-send-email.html';
+
+ my ($arg, $link) = mailto_arg_link($hdr);
+ push @$arg, '/path/to/YOUR_REPLY';
+
+ "<hr /><pre\nid=R>".
+ "You may reply publically to <a\nhref=#t>this message</a> via\n".
+ "plain-text email using any one of the following methods:\n\n" .
+ "* Save the following mbox file, import it into your mail client,\n" .
+ " and reply-to-all from there: <a\nhref=raw>mbox</a>\n\n" .
+ "* Reply to all the recipients using the <b>--to</b>, <b>--cc</b>,\n" .
+ " and <b>--in-reply-to</b> switches of git-send-email(1):\n\n" .
+ "\tgit send-email \\\n\t\t" .
+ join(" \\\n\t\t", @$arg ). "\n\n" .
+ qq( <a\nhref="$se_url">$se_url</a>\n\n) .
+ "* If your mail client supports setting the <b>In-Reply-To</b>" .
+ " header\n via mailto: links, try the " .
+ qq(<a\nhref="$link">mailto: link</a>\n) .
+ '</pre>';
}
-sub feed_entry {
- my ($class, $mime, $full_pfx) = @_;
+sub in_reply_to {
+ my ($hdr) = @_;
+ my $irt = $hdr->header_raw('In-Reply-To');
+
+ return mid_clean($irt) if (defined $irt);
- PRE_WRAP . multipart_text_as_html($mime, $full_pfx) . '</pre>';
+ my $refs = $hdr->header_raw('References');
+ if ($refs && $refs =~ /<([^>]+)>\s*\z/s) {
+ return $1;
+ }
+ undef;
}
# this is already inside a <pre>
-# state = [ time, seen = {}, first_commit, page_nr = 0 ]
sub index_entry {
- my (undef, $mime, $level, $state) = @_;
- my (undef, $seen, $first_commit) = @$state;
- my $midx = $state->[3]++;
- my ($prev, $next) = ($midx - 1, $midx + 1);
- my $rv = '';
- my $part_nr = 0;
- my $enc_msg = enc_for($mime->header("Content-Type"));
- my $subj = $mime->header('Subject');
- my $header_obj = $mime->header_obj;
-
- my $mid_raw = $header_obj->header_raw('Message-ID');
+ my ($mime, $level, $state) = @_;
+ my $midx = $state->{anchor_idx}++;
+ my $ctx = $state->{ctx};
+ my $srch = $ctx->{srch};
+ my $hdr = $mime->header_obj;
+ my $subj = $hdr->header('Subject');
+
+ my $mid_raw = mid_clean(mid_mime($mime));
my $id = anchor_for($mid_raw);
- $seen->{$id} = "#$id"; # save the anchor for later
+ my $seen = $state->{seen};
+ $seen->{$id} = "#$id"; # save the anchor for children, later
my $mid = PublicInbox::Hval->new_msgid($mid_raw);
- my $from = PublicInbox::Hval->new_oneline($mime->header('From'))->raw;
- my @from = Email::Address->parse($from);
- $from = $from[0]->name;
- (defined($from) && length($from)) or $from = $from[0]->address;
-
- $from = PublicInbox::Hval->new_oneline($from)->as_html;
- $subj = PublicInbox::Hval->new_oneline($subj)->as_html;
- my $pfx = (' ' x $level);
- my $root_anchor = $seen->{root_anchor};
- my $path;
- my $more = 'permalink';
- if ($root_anchor) {
- $path = '../';
- $subj = "<u\nid=\"u\">$subj</u>" if $root_anchor eq $id;
- } else {
- $path = '';
- }
+ my $from = PublicInbox::Address::from_name($hdr->header('From'));
- my $ts = $mime->header('X-PI-TS');
- my $fmt = '%Y-%m-%d %H:%M UTC';
- $ts = POSIX::strftime($fmt, gmtime($ts));
+ my $root_anchor = $state->{root_anchor} || '';
+ my $path = $root_anchor ? '../../' : '';
+ my $href = $mid->as_href;
+ my $irt = in_reply_to($hdr);
+ my $parent_anchor = $seen->{anchor_for($irt)} if defined $irt;
- $rv .= "$pfx<b\nid=\"$id\">$subj</b>\n$pfx";
- $rv .= "- by $from @ $ts - ";
- $rv .= "<a\nid=\"s$midx\"\nhref=\"#s$next\">next</a>";
- if ($prev >= 0) {
- $rv .= "/<a\nhref=\"#s$prev\">prev</a>";
- }
- $rv .= "\n\n";
+ $from = ascii_html($from);
+ $subj = ascii_html($subj);
+ $subj = "<a\nhref=\"${path}$href/\">$subj</a>";
+ $subj = "<u\nid=u>$subj</u>" if $root_anchor eq $id;
- my $irt = $header_obj->header_raw('In-Reply-To');
- my ($anchor_idx, $anchor, $t_anchor);
- if (defined $irt) {
- $anchor_idx = anchor_for($irt);
- $anchor = $seen->{$anchor_idx};
- $t_anchor = T_ANCHOR;
- } else {
- $t_anchor = '';
- }
- my $href = $mid->as_href;
- my $mhref = "${path}m/$href.html";
- my $fhref = "${path}f/$href.html";
- # scan through all parts, looking for displayable text
- $mime->walk_parts(sub {
- $rv .= index_walk($_[0], $pfx, $enc_msg, $part_nr, $fhref,
- \$more);
- $part_nr++;
- });
+ my $ts = _msg_date($hdr);
+ my $rv = "<pre\nid=s$midx>";
+ $rv .= "<b\nid=$id>$subj</b>\n";
+ my $txt = "${path}$href/raw";
+ my $fh = $state->{fh};
+ $fh->write($rv .= "- $from @ $ts UTC (<a\nhref=\"$txt\">raw</a>)\n\n");
- $rv .= "\n$pfx<a\nhref=\"$mhref\">$more</a> ";
- my $txt = "${path}m/$href.txt";
- $rv .= "<a\nhref=\"$txt\">raw</a> ";
- $rv .= html_footer($mime, 0);
+ my $mhref = "${path}$href/";
+
+ # scan through all parts, looking for displayable text
+ msg_iter($mime, sub { index_walk($fh, $mhref, $_[0]) });
+ $rv = "\n" . html_footer($hdr, 0, $ctx, "$path$href/#R");
if (defined $irt) {
- unless (defined $anchor) {
- my $v = PublicInbox::Hval->new_msgid($irt);
+ unless (defined $parent_anchor) {
+ my $v = PublicInbox::Hval->new_msgid($irt, 1);
$v = $v->as_href;
- $anchor = "${path}m/$v.html";
- $seen->{$anchor_idx} = $anchor;
+ $parent_anchor = "${path}$v/";
}
- $rv .= " <a\nhref=\"$anchor\">parent</a>";
+ $rv .= " <a\nhref=\"$parent_anchor\">parent</a>";
}
-
- if ($first_commit) {
- $rv .= " <a\nhref=\"t/$href.html$t_anchor\">thread</a>";
+ if (my $pct = $state->{pct}) { # used by SearchView.pm
+ $rv .= " [relevance $pct->{$mid_raw}%]";
+ } elsif ($srch) {
+ my $threaded = 'threaded';
+ my $flat = 'flat';
+ my $end = '';
+ if ($ctx->{flat}) {
+ $flat = "<b>$flat</b>";
+ $end = "\n"; # for lynx
+ } else {
+ $threaded = "<b>$threaded</b>";
+ }
+ $rv .= " [<a\nhref=\"${path}$href/t/#u\">$threaded</a>";
+ $rv .= "|<a\nhref=\"${path}$href/T/#u\">$flat</a>]$end";
}
-
- $rv . "\n\n";
+ $fh->write($rv .= '</pre>');
}
sub thread_html {
- my (undef, $ctx, $foot, $srch) = @_;
- my $mid = mid_compressed($ctx->{mid});
- my $res = $srch->get_thread($mid);
- my $rv = '';
- my $msgs = load_results($ctx, $res);
- my $nr = scalar @$msgs;
- return $rv if $nr == 0;
- require PublicInbox::Thread;
- my $th = PublicInbox::Thread->new(@$msgs);
- $th->thread;
- $th->order(*PublicInbox::Thread::sort_ts);
- my $state = [ undef, { root_anchor => anchor_for($mid) }, undef, 0 ];
- thread_entry(\$rv, $state, $_, 0) for $th->rootset;
- my $final_anchor = $state->[3];
- my $next = "<a\nid=\"s$final_anchor\">end of thread</a>\n";
-
- $rv .= "</pre><hr />" . PRE_WRAP . $next . $foot . "</pre>";
-}
-
-sub subject_path_html {
- my (undef, $ctx, $foot, $srch) = @_;
- my $path = $ctx->{subject_path};
- my $res = $srch->get_subject_path($path);
- my $rv = '';
- my $msgs = load_results($ctx, $res);
- my $nr = scalar @$msgs;
- return $rv if $nr == 0;
- require PublicInbox::Thread;
- my $th = PublicInbox::Thread->new(@$msgs);
- $th->thread;
- $th->order(*PublicInbox::Thread::sort_ts);
- my $state = [ undef, { root_anchor => 'dummy' }, undef, 0 ];
- thread_entry(\$rv, $state, $_, 0) for $th->rootset;
- my $final_anchor = $state->[3];
- my $next = "<a\nid=\"s$final_anchor\">end of thread</a>\n";
-
- $rv .= "</pre><hr />" . PRE_WRAP . $next . $foot . "</pre>";
+ my ($ctx, $foot, $srch) = @_;
+ # $_[0] in sub is the Plack callback
+ sub { emit_thread_html($_[0], $ctx, $foot, $srch) }
}
# only private functions below.
-sub index_walk {
- my ($part, $pfx, $enc_msg, $part_nr, $fhref, $more) = @_;
- my $rv = '';
- return $rv if $part->subparts; # walk_parts already recurses
- my $ct = $part->content_type;
-
- # account for filter bugs...
- return if defined $ct && $ct =~ m!\btext/[xh]+tml\b!i;
-
- my $enc = enc_for($ct, $enc_msg);
-
- if ($part_nr > 0) {
- my $fn = $part->filename;
- defined($fn) or $fn = "part #" . ($part_nr + 1);
- $rv .= $pfx . add_filename_line($enc->decode($fn));
+sub emit_thread_html {
+ my ($res, $ctx, $foot, $srch) = @_;
+ my $mid = $ctx->{mid};
+ my $flat = $ctx->{flat};
+ my $msgs = load_results($srch->get_thread($mid, { asc => $flat }));
+ my $nr = scalar @$msgs;
+ return missing_thread($res, $ctx) if $nr == 0;
+ my $seen = {};
+ my $state = {
+ res => $res,
+ ctx => $ctx,
+ seen => $seen,
+ root_anchor => anchor_for($mid),
+ anchor_idx => 0,
+ cur_level => 0,
+ };
+
+ require PublicInbox::Git;
+ $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir});
+ if ($flat) {
+ pre_anchor_entry($seen, $_) for (@$msgs);
+ __thread_entry($state, $_, 0) for (@$msgs);
+ } else {
+ my $th = thread_results($msgs);
+ thread_entry($state, $_, 0) for $th->rootset;
+ if (my $max = $state->{cur_level}) {
+ $state->{fh}->write(
+ ('</ul></li>' x ($max - 1)) . '</ul>');
+ }
}
- my $s = add_text_body_short($enc, $part, $part_nr, $fhref);
-
- # drop the remainder of git patches, they're usually better
- # to review when the full message is viewed
- $s =~ s!^---+\n.*\z!!ms and $$more = 'more...';
+ # there could be a race due to a message being deleted in git
+ # but still being in the Xapian index:
+ my $fh = delete $state->{fh} or return missing_thread($res, $ctx);
+
+ my $final_anchor = $state->{anchor_idx};
+ my $next = "<a\nid=s$final_anchor>";
+ $next .= $final_anchor == 1 ? 'only message in' : 'end of';
+ $next .= " thread</a>, back to <a\nhref=\"../../\">index</a>";
+ $next .= "\ndownload thread: ";
+ $next .= "<a\nhref=\"../t.mbox.gz\">mbox.gz</a>";
+ $next .= " / follow: <a\nhref=\"../t.atom\">Atom feed</a>";
+ $fh->write('<hr /><pre>' . $next . "\n\n".
+ $foot . '</pre></body></html>');
+ $fh->close;
+}
- # Drop signatures
- $s =~ s/^-- \n.*\z//ms and $$more = 'more...';
+sub index_walk {
+ my ($fh, $upfx, $p) = @_;
+ my $s = add_text_body($upfx, $p);
- # kill any leading or trailing whitespace
- $s =~ s/\A\s+//s;
- $s =~ s/\s+\z//s;
+ return if $s eq '';
- if (length $s) {
- # add prefix:
- $s =~ s/^/$pfx/sgm;
+ $s .= "\n"; # ensure there's a trailing newline
- $rv .= $s . "\n";
- }
- $rv;
-}
-
-sub enc_for {
- my ($ct, $default) = @_;
- $default ||= $enc_utf8;
- defined $ct or return $default;
- my $ct_parsed = parse_content_type($ct);
- if ($ct_parsed) {
- if (my $charset = $ct_parsed->{attributes}->{charset}) {
- my $enc = find_encoding($charset);
- return $enc if $enc;
- }
- }
- $default;
+ $fh->write($s);
}
sub multipart_text_as_html {
- my ($mime, $full_pfx, $srch) = @_;
+ my ($mime, $upfx) = @_;
my $rv = "";
- my $part_nr = 0;
- my $enc_msg = enc_for($mime->header("Content-Type"));
# scan through all parts, looking for displayable text
- $mime->walk_parts(sub {
- my ($part) = @_;
- return if $part->subparts; # walk_parts already recurses
- my $ct = $part->content_type;
+ msg_iter($mime, sub {
+ my ($p) = @_;
+ $p = add_text_body($upfx, $p);
+ $rv .= $p;
+ $rv .= "\n" if $p ne '';
+ });
+ $rv;
+}
- # account for filter bugs...
- return if defined $ct && $ct =~ m!\btext/[xh]+tml\b!i;
+sub flush_quote {
+ my ($s, $l, $quot) = @_;
- my $enc = enc_for($ct, $enc_msg);
+ # show everything in the full version with anchor from
+ # short version (see above)
+ my $rv = $l->linkify_1(join('', @$quot));
+ @$quot = ();
- if ($part_nr > 0) {
- my $fn = $part->filename;
- defined($fn) or $fn = "part #" . ($part_nr + 1);
- $rv .= add_filename_line($enc->decode($fn));
- }
+ # we use a <div> here to allow users to specify their own
+ # color for quoted text
+ $rv = $l->linkify_2(ascii_html($rv));
+ $$s .= qq(<span\nclass="q">) . $rv . '</span>'
+}
- if (defined $full_pfx) {
- $rv .= add_text_body_short($enc, $part, $part_nr,
- $full_pfx);
- } else {
- $rv .= add_text_body_full($enc, $part, $part_nr);
- }
- $rv .= "\n" unless $rv =~ /\n\z/s;
- ++$part_nr;
- });
- $rv;
+sub attach_link ($$$$) {
+ my ($upfx, $ct, $p, $fn) = @_;
+ 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
+ $ct = ascii_html($ct);
+ my $desc = $part->header('Content-Description');
+ $desc = $fn unless defined $desc;
+ $desc = '' unless defined $desc;
+ my $sfn;
+ if (defined $fn && $fn =~ /\A[[:alnum:]][\w\.-]+[[:alnum:]]\z/) {
+ $sfn = $fn;
+ } elsif ($ct eq 'text/plain') {
+ $sfn = 'a.txt';
+ } else {
+ $sfn = 'a.bin';
+ }
+ my @ret = qq($nl<a\nhref="$upfx$idx-$sfn">[-- Attachment #$idx: );
+ my $ts = "Type: $ct, Size: $size bytes";
+ push(@ret, ($desc eq '') ? "$ts --]" : "$desc --]\n[-- $ts --]");
+ join('', @ret, '</a>');
}
-sub add_filename_line {
- my ($fn) = @_;
- my $len = 72;
- my $pad = "-";
+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 $fn = $part->filename;
- $len -= length($fn);
- $pad x= ($len/2) if ($len > 0);
- "$pad " . ascii_html($fn) . " $pad\n";
-}
+ if (defined $ct && $ct =~ m!\btext/x?html\b!i) {
+ return attach_link($upfx, $ct, $p, $fn);
+ }
-my $LINK_RE = qr!\b((?:ftp|https?|nntp)://[@\w\+\&\?\.\%\;/#=-]+)!;
+ my $s = eval { $part->body_str };
-sub linkify {
- # no newlines added here since it'd break the splitting we do
- # to fold quotes
- $_[0] =~ s!$LINK_RE!<a href="$1">$1</a>!g;
-}
+ # badly-encoded message? tell the world about it!
+ return attach_link($upfx, $ct, $p, $fn) if $@;
-sub add_text_body_short {
- my ($enc, $part, $part_nr, $full_pfx) = @_;
- my $n = 0;
- my $s = ascii_html($enc->decode($part->body));
- linkify($s);
- $s =~ s!^((?:(?:>[^\n]*)\n)+)!
- my $cur = $1;
- my @lines = split(/\n/, $cur);
- if (@lines > MAX_INLINE_QUOTED) {
- # show a short snippet of quoted text
- $cur = join(' ', @lines);
- $cur =~ s/^>\s*//;
-
- my @sum = split(/\s+/, $cur);
- $cur = '';
- do {
- my $tmp = shift(@sum);
- my $len = length($tmp) + length($cur);
- if ($len > MAX_TRUNC_LEN) {
- @sum = ();
- } else {
- $cur .= $tmp . ' ';
- }
- } while (@sum && length($cur) < MAX_TRUNC_LEN);
- $cur =~ s/ \z/ .../;
- "> <<a\nhref=\"${full_pfx}#q${part_nr}_" . $n++ .
- "\">$cur<\/a>>\n";
+ my @lines = split(/^/m, $s);
+ $s = '';
+ if (defined($fn) || $depth > 0) {
+ $s .= attach_link($upfx, $ct, $p, $fn);
+ $s .= "\n\n";
+ }
+ my @quot;
+ my $l = PublicInbox::Linkify->new;
+ while (defined(my $cur = shift @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);
} else {
- $cur;
+ push @quot, $cur;
}
- !emg;
- $s;
-}
+ }
-sub add_text_body_full {
- my ($enc, $part, $part_nr) = @_;
- my $n = 0;
- my $s = ascii_html($enc->decode($part->body));
- linkify($s);
- $s =~ s!^((?:(?:>[^\n]*)\n)+)!
- my $cur = $1;
- my @lines = split(/\n/, $cur);
- if (@lines > MAX_INLINE_QUOTED) {
- "<a\nid=q${part_nr}_" . $n++ . ">$cur</a>";
- } else {
- $cur;
- }
- !emg;
+ flush_quote(\$s, $l, \@quot) if @quot;
+ $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 (final "\n" added if ne '')
$s;
}
sub headers_to_html_header {
- my ($mime, $full_pfx, $srch) = @_;
+ my ($hdr, $ctx) = @_;
+ my $srch = $ctx->{srch} if $ctx;
+ my $atom = '';
+ my $rv = '';
+ my $upfx = '';
+
+ if ($srch) {
+ $atom = qq{<link\nrel=alternate\ntitle="Atom feed"\n} .
+ qq!href="${upfx}t.atom"\ntype="application/atom+xml"/>!;
+ }
- my $rv = "";
my @title;
+ my $mid = $hdr->header_raw('Message-ID');
+ $mid = PublicInbox::Hval->new_msgid($mid);
foreach my $h (qw(From To Cc Subject Date)) {
- my $v = $mime->header($h);
- defined($v) && length($v) or next;
- $v = PublicInbox::Hval->new_oneline($v);
+ my $v = $hdr->header($h);
+ defined($v) && ($v ne '') or next;
+ $v = PublicInbox::Hval->new($v);
if ($h eq 'From') {
- my @from = Email::Address->parse($v->raw);
- $title[1] = ascii_html($from[0]->name);
+ my $n = PublicInbox::Address::from_name($v->raw);
+ $title[1] = ascii_html($n);
} elsif ($h eq 'Subject') {
$title[0] = $v->as_html;
if ($srch) {
- my $path = $srch->subject_path($v->raw);
- $rv .= "$h: <a\nhref=\"../s/$path.html\">";
+ $rv .= qq($h: <a\nhref="#r"\nid=t>);
$rv .= $v->as_html . "</a>\n";
next;
}
$rv .= "$h: " . $v->as_html . "\n";
}
-
- my $header_obj = $mime->header_obj;
- my $mid = $header_obj->header_raw('Message-ID');
- $mid = PublicInbox::Hval->new_msgid($mid);
$rv .= 'Message-ID: <' . $mid->as_html . '> ';
- my $href = $mid->as_href;
- $href = "../m/$href" unless $full_pfx;
- $rv .= "(<a\nhref=\"$href.txt\">raw</a>)\n";
+ $rv .= "(<a\nhref=\"${upfx}raw\">raw</a>)\n";
+ $rv .= _parent_headers($hdr, $srch);
+ $rv .= "\n";
+
+ ("<html><head><title>". join(' - ', @title) . "</title>$atom".
+ PublicInbox::Hval::STYLE .
+ "</head><body><pre\nid=b>" . # anchor for body start
+ $rv);
+}
+
+sub thread_skel {
+ my ($dst, $ctx, $hdr, $tpfx) = @_;
+ my $srch = $ctx->{srch};
+ my $mid = mid_clean($hdr->header_raw('Message-ID'));
+ my $sres = $srch->get_thread($mid);
+ my $nr = $sres->{total};
+ my $expand = qq(<a\nhref="${tpfx}t/#u">expand</a> ) .
+ qq(/ <a\nhref="${tpfx}t.mbox.gz">mbox.gz</a> ) .
+ qq(/ <a\nhref="${tpfx}t.atom">Atom feed</a>);
+
+ my $parent = in_reply_to($hdr);
+ if ($nr <= 1) {
+ if (defined $parent) {
+ $$dst .= "($expand)\n ";
+ $$dst .= ghost_parent("$tpfx../", $parent) . "\n";
+ } else {
+ $$dst .= "[no followups, yet] ($expand)\n";
+ }
+ $ctx->{next_msg} = undef;
+ $ctx->{parent_msg} = $parent;
+ return;
+ }
+
+ $$dst .= "$nr+ messages in thread ($expand";
+ $$dst .= qq! / <a\nhref="#b">[top]</a>)\n!;
+
+ my $subj = $srch->subject_path($hdr->header('Subject'));
+ my $state = {
+ seen => { $subj => 1 },
+ srch => $srch,
+ cur => $mid,
+ prev_attr => '',
+ prev_level => 0,
+ };
+ for (thread_results(load_results($sres))->rootset) {
+ skel_dump($dst, $state, $tpfx, $_, 0);
+ }
+ $ctx->{next_msg} = $state->{next_msg};
+ $ctx->{parent_msg} = $parent;
+}
+
+sub _parent_headers {
+ my ($hdr, $srch) = @_;
+ my $rv = '';
- my $irt = $header_obj->header_raw('In-Reply-To');
+ my $irt = in_reply_to($hdr);
if (defined $irt) {
- my $v = PublicInbox::Hval->new_msgid($irt);
+ my $v = PublicInbox::Hval->new_msgid($irt, 1);
my $html = $v->as_html;
my $href = $v->as_href;
$rv .= "In-Reply-To: <";
- $rv .= "<a\nhref=\"$href.html\">$html</a>>\n";
+ $rv .= "<a\nhref=\"../$href/\">$html</a>>\n";
}
- my $refs = $header_obj->header_raw('References');
+ # do not display References: if search is present,
+ # we show the thread skeleton at the bottom, instead.
+ return $rv if $srch;
+
+ my $refs = $hdr->header_raw('References');
if ($refs) {
- $refs =~ s/\s*\Q$irt\E\s*// if (defined $irt);
- my @refs = ($refs =~ /<([^>]+)>/g);
+ # avoid redundant URLs wasting bandwidth
+ my %seen;
+ $seen{$irt} = 1 if defined $irt;
+ my @refs;
+ my @raw_refs = ($refs =~ /<([^>]+)>/g);
+ foreach my $ref (@raw_refs) {
+ next if $seen{$ref};
+ $seen{$ref} = 1;
+ push @refs, linkify_ref_nosrch($ref);
+ }
+
if (@refs) {
- $rv .= 'References: '. linkify_refs(@refs) . "\n";
+ $rv .= 'References: '. join(' ', @refs) . "\n";
}
}
-
- $rv .= "\n";
-
- ("<html><head><title>". join(' - ', @title) .
- '</title></head><body>' . PRE_WRAP . $rv);
+ $rv;
}
-sub html_footer {
- my ($mime, $standalone, $full_pfx, $srch) = @_;
+sub mailto_arg_link {
+ my ($hdr) = @_;
my %cc; # everyone else
my $to; # this is the From address
foreach my $h (qw(From To Cc)) {
- my $v = $mime->header($h);
- defined($v) && length($v) or next;
- my @addrs = Email::Address->parse($v);
- foreach my $recip (@addrs) {
- my $address = $recip->address;
+ 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;
}
}
- Email::Address->purge_cache if $standalone;
+ my @arg;
- my $subj = $mime->header('Subject') || '';
- $subj = "Re: $subj" unless $subj =~ /\bRe:/;
- my $mid = $mime->header_obj->header_raw('Message-ID');
+ my $subj = $hdr->header('Subject') || '';
+ $subj = "Re: $subj" unless $subj =~ /\bRe:/i;
+ my $mid = $hdr->header_raw('Message-ID');
+ push @arg, "--in-reply-to='" . ascii_html($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 = uri_escape_utf8(join(',', sort values %cc));
+ 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, $href);
+}
+
+sub html_footer {
+ my ($hdr, $standalone, $ctx, $rhref) = @_;
- my $idx = $standalone ? " <a\nhref=\"../\">index</a>" : '';
+ my $srch = $ctx->{srch} if $ctx;
+ my $upfx = '../';
+ my $tpfx = '';
+ my $idx = $standalone ? " <a\nhref=\"$upfx\">index</a>" : '';
+ my $irt = '';
if ($idx && $srch) {
- $irt = $mime->header_obj->header_raw('In-Reply-To') || '';
- $mid = mid_compressed(mid_clean($mid));
- my $t_anchor = length $irt ? T_ANCHOR : '';
- $idx = " <a\nhref=\"../t/$mid.html$t_anchor\">thread</a>$idx";
- my $res = $srch->get_replies($mid);
- if (my $c = $res->{count}) {
- $c = $c == 1 ? '1 reply' : "$c replies";
- $idx .= "\n$c:\n";
- thread_replies(\$idx, $mime, $res);
+ $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;
+ $irt = "<a\nhref=\"$upfx$p/\"rel=prev>parent</a> ";
} else {
- $idx .= "\n(no replies yet)\n";
+ $irt = ' ' x length('parent ');
}
- if ($irt) {
- $irt = PublicInbox::Hval->new_msgid($irt);
- $irt = $irt->as_href;
- $irt = "<a\nhref=\"$irt\">parent</a> ";
+ if ($next) {
+ my $n = PublicInbox::Hval->new_msgid($next)->as_href;
+ $irt .= "<a\nhref=\"$upfx$n/\"\nrel=next>next</a> ";
} else {
- $irt = ' ' x length('parent ');
+ $irt .= ' ' x length('next ');
}
} else {
$irt = '';
}
-
- "$irt<a\nhref=\"" . ascii_html($href) . '">reply</a>' . $idx;
+ $rhref ||= '#R';
+ $irt .= qq(<a\nhref="$rhref">reply</a>);
+ $irt .= $idx;
}
-sub linkify_refs {
- join(' ', map {
- my $v = PublicInbox::Hval->new_msgid($_);
- my $html = $v->as_html;
- my $href = $v->as_href;
- "<<a\nhref=\"$href.html\">$html</a>>";
- } @_);
+sub linkify_ref_nosrch {
+ my $v = PublicInbox::Hval->new_msgid($_[0], 1);
+ my $html = $v->as_html;
+ my $href = $v->as_href;
+ "<<a\nhref=\"../$href/\">$html</a>>";
}
sub anchor_for {
my ($msgid) = @_;
- 'm' . mid_compressed(mid_clean($msgid));
+ my $id = $msgid;
+ if ($id !~ /\A[a-f0-9]{40}\z/) {
+ $id = id_compress(mid_clean($id), 1);
+ }
+ 'm' . $id;
}
-sub simple_dump {
- my ($dst, $root, $node, $level) = @_;
- my $pfx = ' ' x $level;
- $$dst .= $pfx;
- if (my $x = $node->message) {
- my $mid = $x->header('Message-ID');
- if ($root->[0] ne $mid) {
- my $s = $x->header('Subject');
- my $h = hash_subj($s);
- if ($root->[1]->{$h}) {
- $s = '';
- } else {
- $root->[1]->{$h} = 1;
- $s = PublicInbox::Hval->new($s);
- $s = $s->as_html;
- }
- my $m = PublicInbox::Hval->new_msgid($mid);
- my $f = PublicInbox::Hval->new($x->header('X-PI-From'));
- my $d = PublicInbox::Hval->new($x->header('X-PI-Date'));
- $m = $m->as_href . '.html';
- $f = $f->as_html;
- $d = $d->as_html . ' UTC';
- if (length($s) == 0) {
- $$dst .= "` <a\nhref=\"$m\">$f @ $d</a>\n";
- } else {
- $$dst .= "` <a\nhref=\"$m\">$s</a>\n" .
- "$pfx by $f @ $d\n";
- }
+sub thread_html_head {
+ my ($hdr, $state) = @_;
+ my $res = delete $state->{res} or die "BUG: no Plack callback in {res}";
+ my $fh = $res->([200, ['Content-Type'=> 'text/html; charset=UTF-8']]);
+ $state->{fh} = $fh;
+
+ my $s = ascii_html($hdr->header('Subject'));
+ $fh->write("<html><head><title>$s</title>".
+ qq{<link\nrel=alternate\ntitle="Atom feed"\n} .
+ qq!href="../t.atom"\ntype="application/atom+xml"/>! .
+ PublicInbox::Hval::STYLE .
+ "</head><body>");
+}
+
+sub pre_anchor_entry {
+ my ($seen, $mime) = @_;
+ my $id = anchor_for(mid_mime($mime));
+ $seen->{$id} = "#$id"; # save the anchor for children, later
+}
+
+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 $html = $mid->as_html;
+ qq{[parent not found: <<a\nhref="$upfx$href/">$html</a>>]};
+}
+
+sub thread_adj_level {
+ my ($state, $level) = @_;
+
+ my $max = $state->{cur_level};
+ if ($level <= 0) {
+ return '' if $max == 0; # flat output
+
+ # reset existing lists
+ my $x = $max > 1 ? ('</ul></li>' x ($max - 1)) : '';
+ $state->{fh}->write($x . '</ul>');
+ $state->{cur_level} = 0;
+ return '';
+ }
+ if ($level == $max) { # continue existing list
+ $state->{fh}->write('<li>');
+ } elsif ($level < $max) {
+ my $x = $max > 1 ? ('</ul></li>' x ($max - $level)) : '';
+ $state->{fh}->write($x .= '<li>');
+ $state->{cur_level} = $level;
+ } else { # ($level > $max) # start a new level
+ $state->{cur_level} = $level;
+ $state->{fh}->write(($max ? '<li>' : '') . '<ul><li>');
+ }
+ '</li>';
+}
+
+sub ghost_flush {
+ my ($state, $upfx, $mid, $level) = @_;
+ my $end = '<pre>'. ghost_parent($upfx, $mid) . '</pre>';
+ $state->{fh}->write($end .= thread_adj_level($state, $level));
+}
+
+sub __thread_entry {
+ my ($state, $mime, $level) = @_;
+
+ # lazy load the full message from mini_mime:
+ $mime = eval {
+ my $path = mid2path(mid_clean(mid_mime($mime)));
+ Email::MIME->new($state->{ctx}->{git}->cat_file('HEAD:'.$path));
+ } or return;
+
+ thread_html_head($mime, $state) if $state->{anchor_idx} == 0;
+ if (my $ghost = delete $state->{ghost}) {
+ # n.b. ghost messages may only be parents, not children
+ foreach my $g (@$ghost) {
+ ghost_flush($state, '../../', @$g);
}
}
- simple_dump($dst, $root, $node->child, $level + 1) if $node->child;
- simple_dump($dst, $root, $node->next, $level) if $node->next;
+ my $end = thread_adj_level($state, $level);
+ index_entry($mime, $level, $state);
+ $state->{fh}->write($end) if $end;
+
+ 1;
}
-sub hash_subj {
- my ($subj) = @_;
- $subj =~ s/\A\s+//;
- $subj =~ s/\s+\z//;
- $subj =~ s/^(?:re|aw):\s*//i; # remove reply prefix (aw: German)
- $subj =~ s/\s+/ /;
- Digest::SHA::sha1($subj);
+sub indent_for {
+ my ($level) = @_;
+ INDENT x ($level - 1);
}
-sub thread_replies {
- my ($dst, $root, $res) = @_;
- my @msgs = map { $_->mini_mime } @{$res->{msgs}};
- foreach (@{$res->{msgs}}) {
- print STDERR "smsg->path: <", $_->path, ">\n";
+sub __ghost_prepare {
+ my ($state, $node, $level) = @_;
+ my $ghost = $state->{ghost} ||= [];
+ push @$ghost, [ $node->messageid, $level ];
+}
+
+sub thread_entry {
+ my ($state, $node, $level) = @_;
+ return unless $node;
+ if (my $mime = $node->message) {
+ unless (__thread_entry($state, $mime, $level)) {
+ __ghost_prepare($state, $node, $level);
+ }
+ } else {
+ __ghost_prepare($state, $node, $level);
}
+
+ thread_entry($state, $node->child, $level + 1);
+ thread_entry($state, $node->next, $level);
+}
+
+sub load_results {
+ my ($sres) = @_;
+
+ [ map { $_->mini_mime } @{delete $sres->{msgs}} ];
+}
+
+sub msg_timestamp {
+ my ($hdr) = @_;
+ my $ts = eval { str2time($hdr->header('Date')) };
+ defined($ts) ? $ts : 0;
+}
+
+sub thread_results {
+ my ($msgs) = @_;
require PublicInbox::Thread;
- $root->header_set('X-PI-TS', '0');
- my $th = PublicInbox::Thread->new($root, @msgs);
+ my $th = PublicInbox::Thread->new(@$msgs);
$th->thread;
- $th->order(*PublicInbox::Thread::sort_ts);
- $root = [ $root->header('Message-ID'),
- { hash_subj($root->header('Subject')) => 1 } ];
- simple_dump($dst, $root, $_, 0) for $th->rootset;
+ $th->order(*sort_ts);
+ $th
}
-sub thread_html_head {
- my ($mime) = @_;
- my $s = PublicInbox::Hval->new_oneline($mime->header('Subject'));
- $s = $s->as_html;
- "<html><head><title>$s</title></head><body>" . PRE_WRAP
+sub missing_thread {
+ my ($res, $ctx) = @_;
+ require PublicInbox::ExtMsg;
+ $res->(PublicInbox::ExtMsg::ext_msg($ctx))
}
-sub thread_entry {
- my ($dst, $state, $node, $level) = @_;
- # $state = [ $search_res, $seen, undef, 0 (msg_nr) ];
- # $seen is overloaded with 3 types of fields:
- # 1) "root_anchor" => anchor_for(Message-ID),
- # 2) seen subject hashes: sha1(subject) => 1
- # 3) anchors hashes: "#$sha1_hex" (same as $seen in index_entry)
+sub _msg_date {
+ my ($hdr) = @_;
+ my $ts = $hdr->header('X-PI-TS') || msg_timestamp($hdr);
+ fmt_ts($ts);
+}
+
+sub fmt_ts { POSIX::strftime('%Y-%m-%d %k:%M', gmtime($_[0])) }
+
+sub _skel_header {
+ my ($dst, $state, $upfx, $hdr, $level) = @_;
+
+ my $cur = $state->{cur};
+ my $mid = mid_clean($hdr->header_raw('Message-ID'));
+ my $f = ascii_html($hdr->header('X-PI-From'));
+ my $d = _msg_date($hdr);
+ my $pfx = "$d " . indent_for($level) . th_pfx($level);
+ my $attr = $f;
+ $state->{first_level} ||= $level;
+
+ if ($attr ne $state->{prev_attr} || $state->{prev_level} > $level) {
+ $state->{prev_attr} = $attr;
+ } else {
+ $attr = '';
+ }
+ $state->{prev_level} = $level;
+
+ if ($cur) {
+ if ($cur eq $mid) {
+ delete $state->{cur};
+ $$dst .= "$pfx<b><a\nid=r\nhref=\"#t\">".
+ "$attr [this message]</a></b>\n";
+
+ return;
+ }
+ } else {
+ $state->{next_msg} ||= $mid;
+ }
+
+ # 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 = $state->{srch}->subject_path($s);
+ if ($state->{seen}->{$h}) {
+ $s = undef;
+ } else {
+ $state->{seen}->{$h} = 1;
+ $s = PublicInbox::Hval->new($s);
+ $s = $s->as_html;
+ }
+ my $m = PublicInbox::Hval->new_msgid($mid);
+ $m = $upfx . '../' . $m->as_href . '/';
+ $$dst .= "$pfx<a\nhref=\"$m\">";
+ $$dst .= defined($s) ? "$s</a> $f\n" : "$f</a>\n";
+}
+
+sub skel_dump {
+ my ($dst, $state, $upfx, $node, $level) = @_;
+ return unless $node;
if (my $mime = $node->message) {
- if (length($$dst) == 0) {
- $$dst .= thread_html_head($mime);
+ my $hdr = $mime->header_obj;
+ my $mid = mid_clean($hdr->header_raw('Message-ID'));
+ _skel_header($dst, $state, $upfx, $hdr, $level);
+ } else {
+ my $mid = $node->messageid;
+ if ($mid eq 'subject dummy') {
+ $$dst .= "\t[no common parent]\n";
+ } else {
+ $$dst .= ' [not found] ';
+ $$dst .= indent_for($level) . th_pfx($level);
+ $mid = PublicInbox::Hval->new_msgid($mid);
+ my $href = "$upfx../" . $mid->as_href . '/';
+ my $html = $mid->as_html;
+ $$dst .= qq{<<a\nhref="$href">$html</a>>\n};
}
- $$dst .= index_entry(undef, $mime, $level, $state);
}
- thread_entry($dst, $state, $node->child, $level + 1) if $node->child;
- thread_entry($dst, $state, $node->next, $level) if $node->next;
+ skel_dump($dst, $state, $upfx, $node->child, $level+1);
+ skel_dump($dst, $state, $upfx, $node->next, $level);
}
-sub load_results {
- my ($ctx, $res) = @_;
-
- require PublicInbox::GitCatFile;
- my $git = PublicInbox::GitCatFile->new($ctx->{git_dir});
- my @msgs;
- while (my $smsg = shift @{$res->{msgs}}) {
- my $m = $smsg->mid;
- my $path = mid2path($m);
-
- # FIXME: duplicated code from Feed.pm
- my $mime = eval {
- my $str = $git->cat_file("HEAD:$path");
- Email::MIME->new($str);
- };
- unless ($@) {
- my $t = eval { str2time($mime->header('Date')) };
- defined($t) or $t = 0;
- $mime->header_set('X-PI-TS', $t);
- push @msgs, $mime;
+sub sort_ts {
+ sort {
+ (eval { $a->topmost->message->header('X-PI-TS') } || 0) <=>
+ (eval { $b->topmost->message->header('X-PI-TS') } || 0)
+ } @_;
+}
+
+# accumulate recent topics if search is supported
+# returns 1 if done, undef if not
+sub add_topic {
+ my ($state, $node, $level) = @_;
+ return unless $node;
+ my $child_adjust = 1;
+
+ if (my $x = $node->message) {
+ $x = $x->header_obj;
+ my $subj;
+
+ $subj = $x->header('Subject');
+ $subj = $state->{srch}->subject_normalized($subj);
+
+ if (++$state->{subjs}->{$subj} == 1) {
+ push @{$state->{order}}, [ $level, $subj ];
+ }
+
+ my $mid = mid_clean($x->header_raw('Message-ID'));
+
+ my $ts = $x->header('X-PI-TS');
+ my $exist = $state->{latest}->{$subj};
+ if (!$exist || $exist->[1] < $ts) {
+ $state->{latest}->{$subj} = [ $mid, $ts ];
}
+ } else {
+ # ghost message, do not bump level
+ $child_adjust = 0;
}
- \@msgs;
+
+ add_topic($state, $node->child, $level + $child_adjust);
+ add_topic($state, $node->next, $level);
+}
+
+sub emit_topics {
+ my ($state) = @_;
+ my $order = $state->{order};
+ my $subjs = $state->{subjs};
+ my $latest = $state->{latest};
+ my $fh = $state->{fh};
+ return $fh->write("\n[No topics in range]</pre>") unless scalar @$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;
+ $subj = PublicInbox::Hval->new($subj)->as_html;
+ $pfx = indent_for($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) .
+ "<a\nhref=\"$mid/t/#u\"><b>" .
+ $subj . "</b></a>\n";
+
+ $ts = fmt_ts($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 $mbox = qq(<a\nhref="$mid/t.mbox.gz">mbox.gz</a>);
+ my $atom = qq(<a\nhref="$mid/t.atom">Atom</a>);
+ $pfx .= INDENT if $level > 0;
+ $cur->[1] .= $pfx . $attr . $n . " - $mbox / $atom\n";
+ $prev_attr = $attr;
+ }
+ }
+ push @recent, $cur if $cur;
+ @recent = map { $_->[1] } sort { $b->[0] <=> $a->[0] } @recent;
+ $fh->write(join('', @recent) . '</pre>');
+}
+
+sub emit_index_topics {
+ my ($state) = @_;
+ my ($off) = (($state->{ctx}->{cgi}->param('o') || '0') =~ /(\d+)/);
+ $state->{order} = [];
+ $state->{subjs} = {};
+ $state->{latest} = {};
+ my $max = 25;
+ my %opts = ( offset => $off, limit => $max * 4 );
+ while (scalar @{$state->{order}} < $max) {
+ my $sres = $state->{srch}->query('', \%opts);
+ my $nr = scalar @{$sres->{msgs}} or last;
+
+ for (thread_results(load_results($sres))->rootset) {
+ add_topic($state, $_, 0);
+ }
+ $opts{offset} += $nr;
+ }
+
+ emit_topics($state);
+ $opts{offset};
}
1;