www: attempt to handle Message-IDs with slashes

Unfortunately, some HTTP servers will try to be clever
with %2F and escape it to '/', making life difficult for
us.  Fortunately, not many Message-IDs have slashes in
them.

1 files changed, 30 insertions, 27 deletions

diff --git a/lib/PublicInbox/WWW.pm b/lib/PublicInbox/WWW.pm
index 11b54027..6eebf628 100644
--- a/lib/PublicInbox/WWW.pm
+++ b/lib/PublicInbox/WWW.pm
@@ -17,12 +17,13 @@ use constant SSOMA_URL => 'http://ssoma.public-inbox.org/';
 use constant PI_URL => 'http://public-inbox.org/';
 our $LISTNAME_RE = qr!\A/([\w\.\-]+)!;
 our $MID_RE = qr!([^/]+)!;
+our $END_RE = qr!(f/|T/|t/|t\.mbox(?:\.gz)?|t\.atom|raw|)!;
 our $pi_config;
 
 sub run {
         my ($cgi, $method) = @_;
         $pi_config ||= PublicInbox::Config->new;
-        my %ctx = (cgi => $cgi, pi_config => $pi_config);
+        my $ctx = { cgi => $cgi, pi_config => $pi_config };
         if ($method !~ /\AGET|HEAD\z/) {
                 return r(405, 'Method Not Allowed');
         }
@@ -32,40 +33,26 @@ sub run {
         if ($path_info eq '/') {
                 r404();
         } elsif ($path_info =~ m!$LISTNAME_RE\z!o) {
-                invalid_list(\%ctx, $1) || r301(\%ctx, $1);
+                invalid_list($ctx, $1) || r301($ctx, $1);
         } elsif ($path_info =~ m!$LISTNAME_RE(?:/|/index\.html)?\z!o) {
-                invalid_list(\%ctx, $1) || get_index(\%ctx);
+                invalid_list($ctx, $1) || get_index($ctx);
         } elsif ($path_info =~ m!$LISTNAME_RE/(?:atom\.xml|new\.atom)\z!o) {
-                invalid_list(\%ctx, $1) || get_atom(\%ctx);
+                invalid_list($ctx, $1) || get_atom($ctx);
 
-        # thread display
-        } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/t/\z!o) {
-                invalid_list_mid(\%ctx, $1, $2) || get_thread(\%ctx);
-        } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/t\.mbox(\.gz)?\z!o) {
-                my $sfx = $3;
-                invalid_list_mid(\%ctx, $1, $2) || get_thread_mbox(\%ctx, $sfx);
-        } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/t\.atom\z!o) {
-                invalid_list_mid(\%ctx, $1, $2) || get_thread_atom(\%ctx);
-        } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/T/\z!o) {
-                $ctx{flat} = 1;
-                invalid_list_mid(\%ctx, $1, $2) || get_thread(\%ctx);
-
-        # single-message pages
-        } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/\z!o) {
-                invalid_list_mid(\%ctx, $1, $2) || get_mid_html(\%ctx);
-        } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/raw\z!o) {
-                invalid_list_mid(\%ctx, $1, $2) || get_mid_txt(\%ctx);
+        } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/$END_RE\z!o) {
+                msg_page($ctx, $1, $2, $3);
 
-        # full-message page
-        } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/f/\z!o) {
-                invalid_list_mid(\%ctx, $1, $2) || get_full_html(\%ctx);
+        # some Message-IDs have slashes in them and the HTTP server
+        # may try to be clever and unescape them :<
+        } elsif ($path_info =~ m!$LISTNAME_RE/(\S+/\S+)/$END_RE\z!o) {
+                msg_page($ctx, $1, $2, $3);
 
         # convenience redirects order matters
         } elsif ($path_info =~ m!$LISTNAME_RE/([^/]{2,})\z!o) {
-                r301(\%ctx, $1, $2);
+                r301($ctx, $1, $2);
 
         } else {
-                legacy_redirects(\%ctx, $path_info);
+                legacy_redirects($ctx, $path_info);
         }
 }
 
@@ -203,10 +190,11 @@ sub get_full_html {
 
 # /$LISTNAME/$MESSAGE_ID/t/
 sub get_thread {
-        my ($ctx) = @_;
+        my ($ctx, $flat) = @_;
         my $srch = searcher($ctx) or return need_search($ctx);
         require PublicInbox::View;
         my $foot = footer($ctx);
+        $ctx->{flat} = $flat;
         PublicInbox::View::thread_html($ctx, $foot, $srch);
 }
 
@@ -390,4 +378,19 @@ sub r301 {
           [ "Redirecting to $url\n" ] ]
 }
 
+sub msg_page {
+        my ($ctx, $list, $mid, $e) = @_;
+        unless (invalid_list_mid($ctx, $list, $mid)) {
+                '' eq $e and return get_mid_html($ctx);
+                't/' eq $e and return get_thread($ctx);
+                't.atom' eq $e and return get_thread_atom($ctx);
+                't.mbox' eq $e and return get_thread_mbox($ctx);
+                't.mbox.gz' eq $e and return get_thread_mbox($ctx, '.gz');
+                'T/' eq $e and return get_thread($ctx, 1);
+                'raw' eq $e and return get_mid_txt($ctx);
+                'f/' eq $e and return get_full_html($ctx);
+        }
+        r404($ctx);
+}
+
 1;