From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <e@80x24.org>
X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on dcvr.yhbt.net
X-Spam-Level: 
X-Spam-ASN:  
X-Spam-Status: No, score=-4.0 required=3.0 tests=ALL_TRUSTED,BAYES_00,
	T_SCC_BODY_TEXT_LINE shortcircuit=no autolearn=ham autolearn_force=no
	version=3.4.2
Received: from localhost (dcvr.yhbt.net [127.0.0.1])
	by dcvr.yhbt.net (Postfix) with ESMTP id 5F8F71F8C4;
	Tue, 15 Mar 2022 20:45:02 +0000 (UTC)
Date: Tue, 15 Mar 2022 20:45:02 +0000
From: Eric Wong <e@80x24.org>
To: Vlastimil Babka <vbabka@suse.cz>
Cc: meta@public-inbox.org,
	Konstantin Ryabitsev <konstantin@linuxfoundation.org>
Subject: [PATCH] www: loosen deep-linking prevention
Message-ID: <20220315204502.GA2275@dcvr>
References: <93ebfbd1-9924-481c-4edc-9b232d1e995c@suse.cz>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <93ebfbd1-9924-481c-4edc-9b232d1e995c@suse.cz>
List-Id: <meta.public-inbox.org>

Vlastimil Babka <vbabka@suse.cz> wrote:
> Hi,
> 
> When opening:
> https://lore.kernel.org/all/20220309021531.GA22223@xsang-OptiPlex-9020/
> 
> And clicking at [-- Attachment #4: dmesg.xz --]
> Which links to
> https://lore.kernel.org/all/20220309021531.GA22223@xsang-OptiPlex-9020/4-dmesg.xz
> 
> I get with a high probability a "Deep-linking prevented" response instead of
> the attachment. In Firefox and Chrome. Sometimes it does succeed and
> provides the attachment. When opening the attachment URL directly so that
> there's no Referer, it succeeds reliably. Other people confirmed this too.
> Reporting here per Konstantin's advise.

Not sure exactly why this is on the browser side, but I think
the patch below fixes it.  I've deployed to
<https://yhbt.net/lore/>, and tested going through
<https://80x24.org/deep_link.html> via lynx and dillo (w3m
doesn't send Referer in this case)

------------8<---------
Subject: [PATCH] www: loosen deep-linking prevention

Apparently some browsers can set a Referer: header which fails
to match.  I'm not certain why, but making "$schema://$HOST_PORT"
matches case-insensitive seems more correct regardless.

In case that doesn't work, we'll also allow bypassing deep-link
prevention via a POST form button.

Reported-by: Vlastimil Babka <vbabka@suse.cz>
Link: https://public-inbox.org/meta/93ebfbd1-9924-481c-4edc-9b232d1e995c@suse.cz/
---
 lib/PublicInbox/WWW.pm       |  6 +++++-
 lib/PublicInbox/WwwAttach.pm | 18 ++++++++++++------
 2 files changed, 17 insertions(+), 7 deletions(-)

diff --git a/lib/PublicInbox/WWW.pm b/lib/PublicInbox/WWW.pm
index a282784a..755d7558 100644
--- a/lib/PublicInbox/WWW.pm
+++ b/lib/PublicInbox/WWW.pm
@@ -1,4 +1,4 @@
-# Copyright (C) 2014-2021 all contributors <meta@public-inbox.org>
+# Copyright (C) all contributors <meta@public-inbox.org>
 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 #
 # Main web interface for mailing list archives
@@ -64,6 +64,10 @@ sub call {
 				serve_git($ctx, $epoch, $path);
 		} elsif ($path_info =~ m!$INBOX_RE/(\w+)\.sql\.gz\z!o) {
 			return get_altid_dump($ctx, $1, $2);
+		} elsif ($path_info =~ m!$INBOX_RE/$MID_RE/$ATTACH_RE\z!o) {
+			my ($idx, $fn) = ($3, $4);
+			return invalid_inbox_mid($ctx, $1, $2) ||
+				get_attach($ctx, $idx, $fn);
 		} elsif ($path_info =~ m!$INBOX_RE/!o) {
 			return invalid_inbox($ctx, $1) || mbox_results($ctx);
 		}
diff --git a/lib/PublicInbox/WwwAttach.pm b/lib/PublicInbox/WwwAttach.pm
index c17394af..87844bf3 100644
--- a/lib/PublicInbox/WwwAttach.pm
+++ b/lib/PublicInbox/WwwAttach.pm
@@ -1,4 +1,4 @@
-# Copyright (C) 2016-2021 all contributors <meta@public-inbox.org>
+# Copyright (C) all contributors <meta@public-inbox.org>
 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 
 # For retrieving attachments from messages in the WWW interface
@@ -11,16 +11,17 @@ use PublicInbox::Eml;
 sub referer_match ($) {
 	my ($ctx) = @_;
 	my $env = $ctx->{env};
-	my $referer = $env->{HTTP_REFERER} // '';
+	return 1 if $env->{REQUEST_METHOD} eq 'POST';
+	my $referer = lc($env->{HTTP_REFERER} // '');
 	return 1 if $referer eq ''; # no referer is always OK for wget/curl
 
 	# prevent deep-linking from other domains on some browsers (Firefox)
 	# n.b.: $ctx->{ibx}->base_url($env) with INBOX_URL won't work
 	# with dillo, we can only match "$url_scheme://$HTTP_HOST/" without
 	# path components
-	my $base_url = $env->{'psgi.url_scheme'} . '://' .
+	my $base_url = lc($env->{'psgi.url_scheme'} . '://' .
 			($env->{HTTP_HOST} //
-			 "$env->{SERVER_NAME}:$env->{SERVER_PORT}") . '/';
+			 "$env->{SERVER_NAME}:$env->{SERVER_PORT}") . '/');
 	index($referer, $base_url) == 0;
 }
 
@@ -46,8 +47,13 @@ sub get_attach_i { # ->each_part callback
 			$part = $part->body;
 		} else {
 			$res->[0] = 403;
-			$res->[1]->[1] = 'text/plain';
-			$part = "Deep-linking prevented\n";
+			$res->[1]->[1] = 'text/html';
+			$part = <<"";
+<html><head><title>download
+attachment</title><body><pre>Deep-linking prevented</pre><form
+method=post\naction=""><input type=submit value="Download attachment"
+/></form></body></html>
+
 		}
 	}
 	push @{$res->[1]}, 'Content-Length', length($part);