From: Micah Cowan <micah@cowan.name>
Date: Thu, 2 Jul 2009 06:17:33 +0000 (-0700)
Subject: Referer/IRI fix.
X-Git-Tag: v1.13~327
X-Git-Url: http://sjero.net/git/?p=wget;a=commitdiff_plain;h=a00b834bb39fd2982ebb247d52e576208d4934ab

Referer/IRI fix.
---

diff --git a/src/ChangeLog b/src/ChangeLog
index 2d54cb34..fca73ac2 100644
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,14 @@
+2009-07-01  Micah Cowan  <micah@cowan.name>
+
+	* retr.c (retrieve_url): Use the existing "redirect" label,
+	instead of superfluous "second_try". Removed no-longer-accurate
+	debug statement. Use the "newloc" parameter to store the fallback
+	URL, when IRI version was rejected.
+
+	* recur.c (retrieve_tree): Always use the parsed URL for tracking
+	the Referer, since that's the one we actually requested (if
+	there's a difference in terms of percent-encodings and such).
+
 2009-07-01  Steven Schubiger  <stsc@member.fsf.org>
 
 	* Makefile.am: Add a rule to generate build_info.c and list
diff --git a/src/recur.c b/src/recur.c
index e4ffa424..b0042dff 100644
--- a/src/recur.c
+++ b/src/recur.c
@@ -320,6 +320,11 @@ retrieve_tree (struct url *start_url_parsed, struct iri *pi)
               xfree (url);
               url = redirected;
             }
+          else
+            {
+              xfree (url);
+              url = xstrdup (url_parsed->url);
+            }
           url_free(url_parsed);
         }
 
diff --git a/src/retr.c b/src/retr.c
index 8752dce7..a1f045be 100644
--- a/src/retr.c
+++ b/src/retr.c
@@ -605,6 +605,7 @@ retrieve_url (struct url * orig_parsed, const char *origurl, char **file,
   uerr_t result;
   char *url;
   bool location_changed;
+  bool iri_fallbacked = 0;
   int dummy;
   char *mynewloc, *proxy;
   struct url *u = orig_parsed, *proxy_url;
@@ -628,15 +629,11 @@ retrieve_url (struct url * orig_parsed, const char *origurl, char **file,
   if (file)
     *file = NULL;
 
- second_try:
-  DEBUGP (("[IRI Retrieving %s with %s (UTF-8=%d)\n", quote_n (0, url),
-           iri->uri_encoding ? quote_n (1, iri->uri_encoding) : "None",
-           iri->utf8_encode));
-
   if (!refurl)
     refurl = opt.referer;
 
  redirected:
+  /* (also for IRI fallbacking) */
 
   result = NOCONERROR;
   mynewloc = NULL;
@@ -805,7 +802,9 @@ retrieve_url (struct url * orig_parsed, const char *origurl, char **file,
       if (u)
         {
           DEBUGP (("[IRI fallbacking to non-utf8 for %s\n", quote (url)));
-          goto second_try;
+          url = xstrdup (u->url);
+          iri_fallbacked = 1;
+          goto redirected;
         }
       else
           DEBUGP (("[Couldn't fallback to non-utf8 for %s\n", quote (url)));
@@ -840,7 +839,7 @@ retrieve_url (struct url * orig_parsed, const char *origurl, char **file,
       url_free (u);
     }
 
-  if (redirection_count)
+  if (redirection_count || iri_fallbacked)
     {
       if (newloc)
         *newloc = url;
diff --git a/tests/ChangeLog b/tests/ChangeLog
index 6fd859c0..8e209251 100644
--- a/tests/ChangeLog
+++ b/tests/ChangeLog
@@ -1,3 +1,12 @@
+2009-07-01  Micah Cowan  <micah@cowan.name>
+
+	* HTTPServer.pm (send_response): Invocation of
+	verify_request_headers, to support testing of Wget-sent header
+	values.
+	(verify_request_headers): Added.
+
+	* Test-iri.px: Added verification checks for Referer values.
+
 2009-06-29  Micah Cowan  <micah@cowan.name>
 
 	* WgetTest.pm.in (_cleanup): Allow cleanup of test directories to
diff --git a/tests/HTTPServer.pm b/tests/HTTPServer.pm
index 5252b5b8..58b1a363 100644
--- a/tests/HTTPServer.pm
+++ b/tests/HTTPServer.pm
@@ -68,6 +68,9 @@ sub send_response {
     if (exists $url_rec->{'auth_method'}) {
         ($send_content, $code, $msg, $headers) =
             $self->handle_auth($req, $url_rec);
+    } elsif (!$self->verify_request_headers ($req, $url_rec)) {
+        ($send_content, $code, $msg, $headers) =
+            ('', 400, 'Mismatch on expected headers', {});
     } else {
         ($code, $msg) = @{$url_rec}{'code', 'msg'};
         $headers = $url_rec->{headers};
@@ -210,6 +213,22 @@ sub verify_auth_basic {
     }
 }
 
+sub verify_request_headers {
+    my ($self, $req, $url_rec) = @_;
+
+    return 1 unless exists $url_rec->{'request_headers'};
+    for my $hdrname (keys %{$url_rec->{'request_headers'}}) {
+        my $rhdr = $req->header ($hdrname);
+        my $ehdr = $url_rec->{'request_headers'}{$hdrname};
+        unless (defined $rhdr && $rhdr =~ $ehdr) {
+            print STDERR "\n*** Mismatch on $hdrname: $rhdr =~ $ehdr\n";
+            return undef;
+        }
+    }
+
+    return 1;
+}
+
 sub _substitute_port {
     my $self = shift;
     my $ret = shift;
diff --git a/tests/Test-iri.px b/tests/Test-iri.px
index 738c304a..01e1c50a 100755
--- a/tests/Test-iri.px
+++ b/tests/Test-iri.px
@@ -148,14 +148,9 @@ my %urls = (
     '/p2_%C3%A9%C3%A9n.html' => {	# UTF-8 encoded
         code => "200",
         msg => "Ok",
-        headers => {
-            "Content-type" => "text/html; charset=ISO-8859-1",
+        request_headers => {
+            "Referer" => qr|http://localhost:[0-9]+/p1_fran%E7ais.html|,
         },
-        content => $pageeen,
-    },
-    '/p2_%E9%E9n.html' => {
-        code => "200",
-        msg => "Ok",
         headers => {
             "Content-type" => "text/html; charset=ISO-8859-1",
         },
@@ -180,6 +175,9 @@ my %urls = (
     '/p4_m%C3%A9%C3%A9r.html' => {
         code => "200",
         msg => "Ok",
+        request_headers => {
+            "Referer" => qr|http://localhost:[0-9]+/p2_%C3%A9%C3%A9n.html|,
+        },
         headers => {
             "Content-type" => "text/plain; charset=UTF-8",
         },
@@ -187,7 +185,7 @@ my %urls = (
     },
 );
 
-my $cmdline = $WgetTest::WGETPATH . " --iri --restrict-file-names=nocontrol -nH -r http://localhost:{{port}}/";
+my $cmdline = $WgetTest::WGETPATH . " -d --iri --restrict-file-names=nocontrol -nH -r http://localhost:{{port}}/";
 
 my $expected_error_code = 0;