What does Safari do when it gets an RSS feed?
Sorry if this is not the right forum for a question like this. It *could* be a bug in Safari, but our production RSS feed seems fine (see http://keepmedia.com/). We're experiencing a situation where Safari begins to load our RSS feed, and then suddenly redirects to http://rss/. Obviously, this is not correct. I've included a snippet of the feed below, withe bulk of the items removed. Does anyone see anything that would cause Safari to redirect. ---------- HTTP/1.1 200 OK Server: Resin/3.0.14 Cache-Control: private X-Actual-Host: twinaero.corp.keepmedia.int (192.168.1.143) Set-Cookie: customer_id=636377; path=/; expires=Fri, 11-Aug-2006 01:14:25 GMT Set-Cookie: customer_id=636377; path=/; expires=Fri, 11-Aug-2006 01:14:25 GMT Set-Cookie: customer_id=636377; path=/; expires=Fri, 11-Aug-2006 01:14:25 GMT Set-Cookie: customer_id=636377; path=/; expires=Fri, 11-Aug-2006 01:14:25 GMT Set-Cookie: JSESSIONID=c2p1lnq6sscMVitGXq; path=/ Content-Type: text/xml; charset=utf-8 Transfer-Encoding: chunked Date: Wed, 12 Jul 2006 01:14:25 GMT 1722 <?xml version="1.0" encoding="UTF-8"?> <rss version="2.0"> <channel> <title>KeepMedia: Home</title> <link>http://localhost/topics/</link> <description> Home materials available through KeepMedia. </description> <generator>http://www.keepmedia.com/</generator> <language>en</language> <item> <title>10 Things You Don't Know About Women</title> <link>http://localhost/pubs/Esquire/ 2003/11/01/279725?extId=10029</link> <pubDate>Sat, 01 Nov 2003 00:00:00 PST</ pubDate> <guid isPermaLink="true">http://localhost/ pubs/Esquire/2003/11/01/279725</guid> <description><![CDATA[ <p>1. Women are innately self- conscious. This is not a choice; it's a genderwide condition. On a bad day, I look in the mirror and see my ten-pound-heavier alter ego. Her name is Bertha. On a really bad day, Bertha sees her two- hundred-pound-heavier alter ego. Her name is Brian Dennehy.</p> ]]></description> </item> <item> <title>Remembrance of Books Past</title> <link>http://localhost/pubs/ OpinionJournal.com/2004/04/23/356242?extId=10029</link> <pubDate>Fri, 23 Apr 2004 14:57:38 PDT</ pubDate> <guid isPermaLink="true">http://localhost/ pubs/OpinionJournal.com/2004/04/23/356242</guid> <description><![CDATA[ <p>You could write your own sequel to "Fahrenheit 451."</p> ]]></description> </item> </channel> </rss> 0 ----------
It seems a bit odd that the same cookie is set four times. What's the "X-Actual-Host" header for? What did the original request headers from Safari look like? Have you used a packet sniffer like Ethereal or tcpdump to find out what the traffic looks like between your development server and Safari? Have you compared them to traffic to your production server? Do you have proper DNS entries set up (both forward and reverse) for your internal server on your internal network? Dave On Jul 11, 2006, at 8:24 PM, Rick Mann wrote:
Sorry if this is not the right forum for a question like this. It *could* be a bug in Safari, but our production RSS feed seems fine (see http://keepmedia.com/).
We're experiencing a situation where Safari begins to load our RSS feed, and then suddenly redirects to http://rss/. Obviously, this is not correct. I've included a snippet of the feed below, withe bulk of the items removed. Does anyone see anything that would cause Safari to redirect.
---------- HTTP/1.1 200 OK Server: Resin/3.0.14 Cache-Control: private X-Actual-Host: twinaero.corp.keepmedia.int (192.168.1.143) Set-Cookie: customer_id=636377; path=/; expires=Fri, 11-Aug-2006 01:14:25 GMT Set-Cookie: customer_id=636377; path=/; expires=Fri, 11-Aug-2006 01:14:25 GMT Set-Cookie: customer_id=636377; path=/; expires=Fri, 11-Aug-2006 01:14:25 GMT Set-Cookie: customer_id=636377; path=/; expires=Fri, 11-Aug-2006 01:14:25 GMT Set-Cookie: JSESSIONID=c2p1lnq6sscMVitGXq; path=/ Content-Type: text/xml; charset=utf-8 Transfer-Encoding: chunked Date: Wed, 12 Jul 2006 01:14:25 GMT
1722 <?xml version="1.0" encoding="UTF-8"?> <rss version="2.0">
<channel> <title>KeepMedia: Home</title> <link>http://localhost/topics/</link> <description> Home materials available through KeepMedia. </description>
<generator>http://www.keepmedia.com/</generator> <language>en</language>
<item> <title>10 Things You Don't Know About Women</title> <link>http://localhost/pubs/Esquire/ 2003/11/01/279725?extId=10029</link> <pubDate>Sat, 01 Nov 2003 00:00:00 PST</ pubDate>
<guid isPermaLink="true">http://localhost/ pubs/Esquire/2003/11/01/279725</guid> <description><![CDATA[ <p>1. Women are innately self- conscious. This is not a choice; it's a genderwide condition. On a bad day, I look in the mirror and see my ten-pound-heavier alter ego. Her name is Bertha. On a really bad day, Bertha sees her two-hundred-pound-heavier alter ego. Her name is Brian Dennehy.</p>
]]></description>
</item>
<item> <title>Remembrance of Books Past</title> <link>http://localhost/pubs/ OpinionJournal.com/2004/04/23/356242?extId=10029</link> <pubDate>Fri, 23 Apr 2004 14:57:38 PDT</ pubDate>
<guid isPermaLink="true">http://localhost/ pubs/OpinionJournal.com/2004/04/23/356242</guid> <description><![CDATA[ <p>You could write your own sequel to "Fahrenheit 451."</p>
]]></description>
</item>
</channel> </rss>
0 ----------
_______________________________________________ webkit-dev mailing list webkit-dev@opendarwin.org http://www.opendarwin.org/mailman/listinfo/webkit-dev
On Jul 11, 2006, at 19:15 , David D. Kilzer wrote:
It seems a bit odd that the same cookie is set four times.
Yeah, it's one of the things that needs cleaning up, but I don't think that's causing the problem.
What's the "X-Actual-Host" header for?
X-Actual-Host is used in debugging problems. It lets us know which server actually fielded the request when hitting the load balancer.
What did the original request headers from Safari look like? Have you used a packet sniffer like Ethereal or tcpdump to find out what the traffic looks like between your development server and Safari? Have you compared them to traffic to your production server?
I tried to use tcpdump to see this, but got useless results. Clearly I don't know how to call it. sudo tcpdump -i lo0 -qA didn't really work for me. I got mostly periods, and what little recognizable text did appear didn't seem to be complete (in this case I'm running the server and Safari on the same machine).
Do you have proper DNS entries set up (both forward and reverse) for your internal server on your internal network?
Fairly sure, but it's not clear to me why this would make a difference. Other browsers/feed readers seem to have no problem. Thanks! -- Rick
Add an -X parameter in tcpdump and you'll get a hex/ascii output. On Jul 12, 2006, at 12:39 PM, Rick Mann wrote:
On Jul 11, 2006, at 19:15 , David D. Kilzer wrote:
It seems a bit odd that the same cookie is set four times.
Yeah, it's one of the things that needs cleaning up, but I don't think that's causing the problem.
What's the "X-Actual-Host" header for?
X-Actual-Host is used in debugging problems. It lets us know which server actually fielded the request when hitting the load balancer.
What did the original request headers from Safari look like? Have you used a packet sniffer like Ethereal or tcpdump to find out what the traffic looks like between your development server and Safari? Have you compared them to traffic to your production server?
I tried to use tcpdump to see this, but got useless results. Clearly I don't know how to call it.
sudo tcpdump -i lo0 -qA
didn't really work for me. I got mostly periods, and what little recognizable text did appear didn't seem to be complete (in this case I'm running the server and Safari on the same machine).
Do you have proper DNS entries set up (both forward and reverse) for your internal server on your internal network?
Fairly sure, but it's not clear to me why this would make a difference. Other browsers/feed readers seem to have no problem.
Thanks!
-- Rick
_______________________________________________ webkit-dev mailing list webkit-dev@opendarwin.org http://www.opendarwin.org/mailman/listinfo/webkit-dev
Actually, it occurs to me that packet sniffing is not going to reveal much. I need to know what Safari does in terms of processing an RSS feed. When does it do the change to feed:, what part of the feed does it look at when it decides to redirect? It's making the first request fine, and it's getting the feed. I need to know what it does next when preparing to render the feed. -- Rick
On 15.07.2006 01:08, "Rick Mann" <rmann@latencyzero.com> wrote:
I need to know what Safari does in terms of processing an RSS feed. When does it do the change to feed:, what part of the feed does it look at when it decides to redirect?
I believe most of this code is not even in Safari, but in system frameworks (developed by separate teams at Apple). Getting such an overview could be quite tricky. - WBR, Alexey Proskuryakov
On Jul 16, 2006, at 23:47 , Alexey Proskuryakov wrote:
I need to know what Safari does in terms of processing an RSS feed. When does it do the change to feed:, what part of the feed does it look at when it decides to redirect?
I believe most of this code is not even in Safari, but in system frameworks (developed by separate teams at Apple). Getting such an overview could be quite tricky.
Oh, wow. Figures. This is going to make things difficult. Can anyone verify that this is indeed the case, that the code to download a feed is outside of Safari proper? Thanks! -- Rick
On Jul 12, 2006, at 2:39 PM, Rick Mann wrote:
On Jul 11, 2006, at 19:15 , David D. Kilzer wrote:
What did the original request headers from Safari look like? Have you used a packet sniffer like Ethereal or tcpdump to find out what the traffic looks like between your development server and Safari? Have you compared them to traffic to your production server?
I tried to use tcpdump to see this, but got useless results. Clearly I don't know how to call it.
sudo tcpdump -i lo0 -qA
didn't really work for me. I got mostly periods, and what little recognizable text did appear didn't seem to be complete (in this case I'm running the server and Safari on the same machine).
Ethereal decodes the packets for you fully, but you'll either have to install Fink or Darwin Ports plus the X11 environment to use it. I know, it's a lot of work just to get a packet sniffer running. You could tell tcpdump to output to a file, then open the file on a Linux or Windows system with Ethereal, too. (Ethereal could use a nice Cocoa interface port!) Does production Safari behave the same way as a WebKit nightly (http://nightly.webkit.org/)? Dave
participants (5)
-
Alexey Proskuryakov
-
Darin Adler
-
David D. Kilzer
-
Rick Mann
-
Rudi Sherry