}
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) . $footer .
+ '</pre><hr /><pre>' .
+ html_footer($mime, 1, $full_pfx, $srch) .
+ $footer .
'</pre></body></html>';
}
my ($srch, $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');
$from = PublicInbox::Hval->new_oneline($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 = '../';
- } else {
- $path = '';
- }
+ my $path = $root_anchor ? '../' : '';
my $href = $mid->as_href;
my $irt = $header_obj->header_raw('In-Reply-To');
my ($anchor_idx, $anchor, $t_anchor);
} else {
$t_anchor = '';
}
-
if (defined $srch) {
$subj = "<a\nhref=\"${path}t/$href.html#u\">$subj</a>";
}
}
my $ts = $mime->header('X-PI-TS');
- my $fmt = '%Y-%m-%d %H:%M UTC';
+ unless (defined $ts) {
+ $ts = msg_timestamp($mime);
+ }
+ my $fmt = '%Y-%m-%d %H:%M';
$ts = POSIX::strftime($fmt, gmtime($ts));
- $rv .= "$pfx<b\nid=\"$id\">$subj</b>\n$pfx";
- $rv .= "- by $from @ $ts - ";
+ my $rv = "<table\nsummary=l$level><tr>";
+ if ($level) {
+ $rv .= '<td><pre>' . (' ' x $level) . '</pre></td>';
+ }
+ $rv .= '<td>' . PRE_WRAP;
+ $rv .= "<b\nid=\"$id\">$subj</b>\n";
+ $rv .= "- by $from @ $ts UTC - ";
$rv .= "<a\nid=\"s$midx\"\nhref=\"#s$next\">next</a>";
if ($prev >= 0) {
$rv .= "/<a\nhref=\"#s$prev\">prev</a>";
}
$rv .= "\n\n";
+ my ($fhref, $more_ref);
my $mhref = "${path}m/$href.html";
- my $fhref = "${path}f/$href.html";
+ if ($level > 0) {
+ $fhref = "${path}f/$href.html";
+ $more_ref = \$more;
+ }
# scan through all parts, looking for displayable text
$mime->walk_parts(sub {
- $rv .= index_walk($_[0], $pfx, $enc_msg, $part_nr, $fhref,
- \$more);
+ $rv .= index_walk($_[0], $enc_msg, $part_nr, $fhref, $more_ref);
$part_nr++;
});
+ $mime->body_set('');
- $rv .= "\n
$pfx<a\nhref=\"$mhref\">$more</a> ";
+ $rv .= "\n<a\nhref=\"$mhref\">$more</a> ";
my $txt = "${path}m/$href.txt";
$rv .= "<a\nhref=\"$txt\">raw</a> ";
$rv .= html_footer($mime, 0);
"threadlink</a>";
}
- $rv . "\n\n";
+ $rv .= '</pre></td></tr></table>';
}
sub thread_html {
my $mid = mid_compressed($ctx->{mid});
my $res = $srch->get_thread($mid);
my $rv = '';
- my $msgs = load_results($ctx, $res);
+ my $msgs = load_results($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 $th = thread_results($msgs);
my $state = [ $srch, { root_anchor => anchor_for($mid) }, undef, 0 ];
- thread_entry(\$rv, $state, $_, 0) for $th->rootset;
+ {
+ require PublicInbox::GitCatFile;
+ my $git = PublicInbox::GitCatFile->new($ctx->{git_dir});
+ thread_entry(\$rv, $git, $state, $_, 0) for $th->rootset;
+ }
my $final_anchor = $state->[3];
my $next = "<a\nid=\"s$final_anchor\">";
} else {
$next .= 'end of thread';
}
- $next .= ", back to <a\nhref=\"../\">index</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 = [ $srch, { 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";
+ $next .= "</a>, back to <a\nhref=\"../\">index</a>\n";
- $rv .= "</pre><hr />" . PRE_WRAP . $next . $foot . "</pre>";
+ $rv .= "<hr />" . PRE_WRAP . $next . $foot . "</pre>";
}
# only private functions below.
sub index_walk {
- my ($part, $pfx, $enc_msg, $part_nr, $fhref, $more) = @_;
+ my ($part, $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;
+ if (defined $ct && $ct =~ m!\btext/[xh]+tml\b!i) {
+ $part->body_set('');
+ return '';
+ }
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));
+ $rv .= add_filename_line($enc->decode($fn));
}
- my $s = add_text_body_short($enc, $part, $part_nr, $fhref);
+ my $s = add_text_body($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 $$more = 'more...';
+ if ($more) {
+ # drop the remainder of git patches, they're usually better
+ # to review when the full message is viewed
+ $s =~ s!^---+\n.*\z!!ms and $$more = 'more...';
- # Drop signatures
- $s =~ s/^-- \n.*\z//ms and $$more = 'more...';
+ # Drop signatures
+ $s =~ s/^-- \n.*\z//ms and $$more = 'more...';
+ }
# kill any leading or trailing whitespace lines
$s =~ s/^\s*$//sgm;
$s =~ s/\s+\z//s;
if (length $s) {
- # add prefix:
- $s =~ s/^/$pfx/sgm;
-
# kill per-line trailing whitespace
$s =~ s/[ \t]+$//sgm;
- $rv .= $s . "\n";
+ $rv .= $s;
+ $s = undef;
+ $rv .= "\n";
}
$rv;
}
$rv .= add_filename_line($enc->decode($fn));
}
- 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);
- }
+ $rv .= add_text_body($enc, $part, $part_nr, $full_pfx);
$rv .= "\n" unless $rv =~ /\n\z/s;
++$part_nr;
});
+ $mime->body_set('');
$rv;
}
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;
+ $_[0] =~ s!$LINK_RE!<a\nhref="$1">$1</a>!g;
}
-sub add_text_body_short {
- my ($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";
- } else {
- $cur;
+sub flush_quote {
+ my ($quot, $n, $part_nr, $full_pfx, $final) = @_;
+
+ if ($full_pfx) {
+ if (!$final && scalar(@$quot) <= MAX_INLINE_QUOTED) {
+ # show quote inline
+ my $rv = join("\n", map { linkify($_); $_ } @$quot);
+ @$quot = ();
+ return $rv . "\n";
}
- !emg;
- $s;
+
+ # show a short snippet of quoted text and link to full version:
+ @$quot = map { s/^(?:>\s*)+//gm; $_ } @$quot;
+ my $cur = join(' ', @$quot);
+ @$quot = split(/\s+/, $cur);
+ $cur = '';
+ do {
+ my $tmp = shift(@$quot);
+ my $len = length($tmp) + length($cur);
+ if ($len > MAX_TRUNC_LEN) {
+ @$quot = ();
+ } else {
+ $cur .= $tmp . ' ';
+ }
+ } while (@$quot && length($cur) < MAX_TRUNC_LEN);
+ @$quot = ();
+ $cur =~ s/ \z/ .../s;
+ my $nr = ++$$n;
+ "> [<a\nhref=\"$full_pfx#q${part_nr}_$nr\">$cur</a>]\n";
+ } else {
+ # show everything in the full version with anchor from
+ # short version (see above)
+ my $nr = ++$$n;
+ my $rv = "<a\nid=q${part_nr}_$nr></a>";
+ $rv .= join("\n", map { linkify($_); $_ } @$quot) . "\n";
+ @$quot = ();
+ $rv;
+ }
}
-sub add_text_body_full {
- my ($enc, $part, $part_nr) = @_;
+sub add_text_body {
+ my ($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/, $cur);
- if (@lines > MAX_INLINE_QUOTED) {
- "<a\nid=q${part_nr}_" . $n++ . ">$cur</a>";
+ my $nr = 0;
+ my $s = $part->body;
+ $part->body_set('');
+ $s = $enc->decode($s);
+ $s = ascii_html($s);
+ my @lines = split(/\n/, $s);
+ $s = '';
+ my @quot;
+ while (defined(my $cur = shift @lines)) {
+ if ($cur !~ /^>/) {
+ # show the previously buffered quote inline
+ if (scalar @quot) {
+ $s .= flush_quote(\@quot, \$n, $part_nr,
+ $full_pfx, 0);
+ }
+
+ # regular line, OK
+ linkify($cur);
+ $s .= $cur;
+ $s .= "\n";
} else {
- $cur;
+ push @quot, $cur;
}
- !emg;
+ }
+ $s .= flush_quote(\@quot, \$n, $part_nr, $full_pfx, 1) if scalar @quot;
$s;
}
my $refs = $header_obj->header_raw('References');
if ($refs) {
- $refs =~ s/\s*\Q$irt\E\s*// if (defined $irt);
- my @refs = ($refs =~ /<([^>]+)>/g);
+ # avoid redundant URLs wasting bandwidth
+ my %seen;
+ $seen{mid_clean($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($ref);
+ }
+
if (@refs) {
- $rv .= 'References: '. linkify_refs(@refs) . "\n";
+ $rv .= 'References: '. join(' ', @refs) . "\n";
}
}
if (my $c = $res->{count}) {
$c = $c == 1 ? '1 followup' : "$c followups";
$idx .= "\n$c:\n";
+ $res->{srch} = $srch;
thread_followups(\$idx, $mime, $res);
} else {
$idx .= "\n(no followups, yet)\n";
"$irt<a\nhref=\"" . ascii_html($href) . '">reply</a>' . $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 {
+ my $v = PublicInbox::Hval->new_msgid($_[0]);
+ my $html = $v->as_html;
+ my $href = $v->as_href;
+ "<<a\nhref=\"$href.html\">$html</a>>";
}
sub anchor_for {
my ($msgid) = @_;
- 'm' . mid_compressed(mid_clean($msgid));
+ my $id = $msgid;
+ if ($id !~ /\A[a-f0-9]{40}\z/) {
+ $id = mid_compressed(mid_clean($id), 1);
+ }
+ 'm' . $id;
}
sub simple_dump {
my ($dst, $root, $node, $level) = @_;
- my $pfx = ' ' x $level;
- $$dst .= $pfx;
+ # $root = [ Root Message-ID, \%seen, $srch ];
if (my $x = $node->message) {
my $mid = $x->header('Message-ID');
if ($root->[0] ne $mid) {
+ my $pfx = ' ' x $level;
+ $$dst .= $pfx;
my $s = $x->header('Subject');
- my $h = hash_subj($s);
+ my $h = $root->[2]->subject_path($s);
if ($root->[1]->{$h}) {
$s = '';
} else {
simple_dump($dst, $root, $node->next, $level) if $node->next;
}
-sub hash_subj {
- my ($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 thread_followups {
my ($dst, $root, $res) = @_;
- my @msgs = map { $_->mini_mime } @{$res->{msgs}};
- require PublicInbox::Thread;
$root->header_set('X-PI-TS', '0');
- my $th = PublicInbox::Thread->new($root, @msgs);
- $th->thread;
- $th->order(*PublicInbox::Thread::sort_ts);
- $root = [ $root->header('Message-ID'),
- { hash_subj($root->header('Subject')) => 1 } ];
+ my $msgs = load_results($res);
+ push @$msgs, $root;
+ my $th = thread_results($msgs);
+ my $srch = $res->{srch};
+ my $subj = $srch->subject_path($root->header('Subject'));
+ my %seen = ($subj => 1);
+ $root = [ $root->header('Message-ID'), \%seen, $srch ];
simple_dump($dst, $root, $_, 0) for $th->rootset;
}
my ($mime) = @_;
my $s = PublicInbox::Hval->new_oneline($mime->header('Subject'));
$s = $s->as_html;
- "<html><head><title>$s</title></head><body>" . PRE_WRAP
-
+ "<html><head><title>$s</title></head><body>";
}
sub thread_entry {
- my ($dst, $state, $node, $level) = @_;
+ my ($dst, $git, $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)
if (my $mime = $node->message) {
- if (length($$dst) == 0) {
- $$dst .= thread_html_head($mime);
+
+ # lazy load the full message from mini_mime:
+ my $path = mid2path(mid_clean($mime->header('Message-ID')));
+ $mime = eval { Email::MIME->new($git->cat_file("HEAD:$path")) };
+ if ($mime) {
+ if (length($$dst) == 0) {
+ $$dst .= thread_html_head($mime);
+ }
+ $$dst .= index_entry(undef, $mime, $level, $state);
}
- $$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;
+ my $cur;
+ $cur = $node->child and
+ thread_entry($dst, $git, $state, $cur, $level + 1);
+ $cur = $node->next and
+ thread_entry($dst, $git, $state, $cur, $level);
}
sub load_results {
- my ($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, $mime;
- }
- }
- \@msgs;
+ my ($res) = @_;
+
+ [ map { $_->mini_mime } @{delete $res->{msgs}} ];
+}
+
+sub msg_timestamp {
+ my ($mime) = @_;
+ my $ts = eval { str2time($mime->header('Date')) };
+ defined($ts) ? $ts : 0;
+}
+
+sub thread_results {
+ my ($msgs) = @_;
+ require PublicInbox::Thread;
+ my $th = PublicInbox::Thread->new(@$msgs);
+ $th->thread;
+ $th->order(*PublicInbox::Thread::sort_ts);
+ $th
}
1;