gnash-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnash-dev] youtube cookie bug


From: strk
Subject: Re: [Gnash-dev] youtube cookie bug
Date: Wed, 26 May 2010 15:04:13 +0200

On Tue, May 25, 2010 at 05:48:22PM -0600, Rob Savoye wrote:
> On 05/25/10 15:51, Rob Savoye wrote:
> 
> >> The curl adaptar code is libbase/curl_adapter.cpp, function importCookies.
> 
>   Thanks for the pointer, I just fixed this bug in my branch. I wound up
> writing the cookie file like it currently works. The only thing
> different is instead of piles of ugly XPCOM code, it's just an NPAPI
> call. GetValueFromURL() is supported as of Firefox 3.1, and 3.5 is
> what's currently in most distributions, so it should be safe to depend
> on it.

Good job !
Does that fix the youtube cookies problem ? (fingers crossed).

It'd be interesting to hear from other browsers (or pre-3.1 users)
to see what happens in that case. I might have one (ubuntu 8.10).

>   We could get rid of the file and pass it as a command line parameter,
> but then we'd have to pass it down through some layers before it got to
> importCookie(). Or set the actual cookie value in the environment
> variable instead of the file name, and use CURLOPT_COOKIE. That would
> require the least changes.

Yeah, the file isn't very good.
One good thing about it is that if the responses
Gnash get also set cookies, the hosting app can read them
back on gnash exit using the GNASH_COOKIES_OUT env variable.

Anyway, a limitation of the file-based approach is that gnash
only sees cookies which were set in the browser at time of
invocation so any newly injected cookie gnash won't know about.

A possible improvement would be adding a get-cookie kind of
request to the current gnash-host protocol which the gnash
curl adapter might use whenever it's asked to fetch a resource.

--strk;

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html



reply via email to

[Prev in Thread] Current Thread [Next in Thread]