MIME-Version: 1.0
Received: by 10.150.57.17 with SMTP id f17mr1117996yba.9.1230901246222; Fri, 
	02 Jan 2009 05:00:46 -0800 (PST)
Date: Fri, 2 Jan 2009 05:00:46 -0800 (PST)
In-Reply-To: <b3a47e21-3546-4bb4-89b1-8b2a80f6aa98@o4g2000pra.googlegroups.com>
X-IP: 76.15.231.44
References: <cbabe4cd-4e2d-449f-aac9-557e584f126e@r15g2000prh.googlegroups.com> 
	<b493c375-0087-441e-8f56-b6b97849c913@o4g2000pra.googlegroups.com> 
	<70d32762-6ac8-4707-a1be-68eac63b427b@b41g2000pra.googlegroups.com> 
	<f98eb678-f0e0-4796-9eeb-b9b7e02fd04c@g39g2000pri.googlegroups.com> 
	<50175a1c-3c98-4958-a0d8-3e70aed1ba80@r10g2000prf.googlegroups.com> 
	<d3833782-9fbf-4f8a-b180-a96abbc44133@r36g2000prf.googlegroups.com> 
	<140202d6-7915-4242-bda0-6b9d2d9c1215@z6g2000pre.googlegroups.com> 
	<984c8e16-32a3-44f4-8e3a-9a4a21cac538@t39g2000prh.googlegroups.com> 
	<b3a47e21-3546-4bb4-89b1-8b2a80f6aa98@o4g2000pra.googlegroups.com>
User-Agent: G2/1.0
X-HTTP-UserAgent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; 
	rv:1.9.0.5) Gecko/2008120121 Firefox/3.0.5,gzip(gfe),gzip(gfe)
Message-ID: <b3e1e486-0913-41f9-b826-beac406da39d@b38g2000prf.googlegroups.com>
Subject: Re: Getting logged out of applications using Fluid (SSBs)/Webkit
From: Benjamin Stein <benst...@gmail.com>
To: fluidapp <fluidapp@googlegroups.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

http://discussions.apple.com/message.jspa?messageID=3D8680137#8680137

On Dec 31 2008, 8:02=A0pm, todd.ditchend...@gmail.com wrote:
> i've done some testing on a machine running OS X 10.5.5 with Safari
> 3.2.1 (the exact same build of safari i have tested on 10.5.6 which i
> believe is the very latest public release of safari).
>
> I *cannot* reproduce the strange behavior on 10.5.5 using either
> Shiira or a Fluid 0.9.5 SSB.
>
> If I had to guess, I would say this is a new problem in CFNetwork
> (which webkit uses for http networking/cookies) shipping with 10.5.6.
> AFAICT, the issue is not present in 10.5.5. However, I do not have
> conclusive evidence of that. (one should always keep in mind that most
> systems have many extensions and applications installed, and should
> never be too quick to point fingers... different apps/plugins/
> extensions can interact in surprising ways).
>
> now I'm going to work on a bug report for apple including all this
> info. i kinda doubt this will be news to them tho.
>
> td
>
> On Dec 31, 11:54=A0am, todd.ditchend...@gmail.com wrote:
>
> > thanks for the details ben. i rarely use safari, so other than the
> > notes from you guys here, I cant tell when this began, but it sounds
> > like its prolly a 10.5.6/3.2.1 issue.
>
> > i'm working on a bug for apple (i like to be thorough).
>
> > td
>
> > On Dec 30, 9:03=A0pm, Benjamin Stein <benst...@gmail.com> wrote:
>
> > > Yes, it started happening to me as soon as I upgraded to 10.5.6. =A0I
> > > also upgraded to Safari 3.2.1 (5525.27.1) in the same system update,
> > > so I can't confirm if it's one or the other.
>
> > > Just some more notes, in case it's helpful:
>
> > > I do get the logout problem in Facebook, but I'm also getting strange
> > > Gmail functionality. =A0After some period of time, it appears to thin=
k
> > > I'm offline (I get a Google's javascript "not connected, retrying in
> > > 60 seconds" message). =A0And then other times Gmail.app just displays
> > > "Bad Request Error 400". =A0I need to restart the app (reload or new
> > > window doesn't help). =A0This started happening right after my update=
,
> > > so I can only conclude it's just a different manifestation of the sam=
e
> > > problem.
>
> > > HTH,
> > > --Ben
>
> > > On Dec 30, 11:03=A0pm, todd.ditchend...@gmail.com wrote:
>
> > > > Update folks:
>
> > > > I have been able to reproduce this same behavior using other webkit=
-
> > > > based applications on two separate macs. Specifically NetNewsWire a=
nd
> > > > Shiira. Both NNW and Shiira use the WebKit located on your system.
> > > > just like Fluid. I *cannot* reproduce this in OmniWeb which ships w=
ith
> > > > its *own*, private (presumably custom) WebKit framework.
>
> > > > In other words, this behavior is *not* specific to Fluid SSBs, but
> > > > rather seems to be affecting any WebKit-based application that uses
> > > > the WebKit framework already located on your system.
>
> > > > As soon as you perform some network activity in one of these apps, =
it
> > > > seems to clobber cookie changes you've made in safari. (note launch=
ing
> > > > the app is not enough. you have to do something that performs a
> > > > network request... looks like ajax or traditional http requests bot=
h
> > > > do the trick). I even used HTTPScoop.app to confirm this. It happen=
s
> > > > immediately after a silent/background ajax request.
>
> > > > So if this is a bug (and it seems like it probably is), its looking
> > > > like a bug in WebKit. I'm not sure if this behavior is new to 10.5.=
6,
> > > > or some specific version of WebKit, but I dont recall seeing or
> > > > hearing about this before... plz correct me on that if i'm wrong.
>