Re: [Gnash-dev] using gnash plugin to playing embedded flash video

[Torsten Schlabach]

Hi Nevo!

> Anyway, having no
> good knowledge about that ,so just a curious wonder. :)

It's never wrong to wonder and to ask intelligent questions!

> You mean FLV file is a binary executable ?

No, I mean the SWF file is a binary executable. All the rest you wrote 
is exactly to the point.

> I could see such a Flash method as "get_videoUrl", which I think is
> what I want to acquire

What you want to acquire is the return value of that method, i.e. the 
actual URL of the video. (You are still assuming there is only one!) The 
only way way to predict the return value form a method that I know if is 
to actually run the code in it. And this method may be named entirely 
different in any other video player.

> We cannot tell how the ActionScript code will be like for all video
> host sites, right ?

Right. It there was a limited number of standard "video player" SWF 
files, you might have a way to predict this, but there aren't. Basically 
everyone is writing his / her own. Of course, algorithms for popular 
sites may get detected and published on the web, but this is not a 
generic solution and by the way has little to do with the goals of the 
Gnash project, AFAIK.

If you come across a very straightforward implementation where the URL 
to be played is one of the parameters passed to the player, you could 
find your way, but this isn't what you really want.

> Have you studied the reason ?  I have encountered such a case last
> week, it just return "Unauthorized" though the URL looked okay (I
> extracted the URL from some website).

I can only guess it: The content owners don't want you to watch their 
content using your own player which is not embedded into their website 
because if you play the URL directly you will not see the advertising 
that they try to make some money of. That would apply to any site which 
plays the content without a login. Those sites create a session key for 
you somewhere in the browser when you visit the site, then pass this on 
to the ActionScript which will pass this on to the request for the 
actual content. That way they make sure that the request for the video 
comes from a player which is embedded in their web page. (And not for 
example someone else's web page.)

> Just a thought.  I saw by using Adobe Flex3 SDK, we could create swf
> file , and Flex3 SDK is open sourced ( havn't finished downloading) ,
> is that helpful to implement the unimplemented parts ?

I am not sure, but I don't actually think so. IIUC, the SDK is open 
sources, not the VM on which the result of what you built with the SDK ( 
the actual Flash movie) will run.

Compare this to having a FOSS SDK for Windows apps (MinGW). That does 
not mean that Windows itself (on which software built with MinGW will 
run) is open source.

Would MinGW help you if you tried to create a clean-room implementation 
of the Win32 API? It would for sure provide some input to your 
requirements, but it's not an implementation you can learn / steal from.

Regards,
Torsten


Nevo schrieb:
> hi, Strk:
> > On Tue, Feb 24, 2009 at 02:17:10PM +0100, Torsten Schlabach wrote:
> >
> >  
> > > What I mean is the AVM2 branch. I have no idea yet how this might be
> > > different from the "normal" development version (trunk) but what I found
> > > is: It has different (not less, but different) unimplemented warnings in
> > > my case.
> > >     
> >
> > The avm2 branch is likely a bit out-dated.
> > Should get a merge *from* trunk to be in sync.
> > The avm2 code should be pretty separated from the
> > old vm, so theoretically it could be merged *into*
> > trunk, except that the parsers for  avm2-specific tags
> > (DoABC, DoABCDefine) are still bogus thus potentially
> > making gnash unstable.
> >
> > Anyway, I suggest you merge trunk into the avm2
> > branch as one of the early stages if you plan
> > to work there.
> >
> >   
> Just a thought.  I saw by using Adobe Flex3 SDK, we could create swf 
> file , and Flex3 SDK is open sourced ( havn't finished downloading) , is 
> that helpful to implement the unimplemented parts ? Anyway, having no 
> good knowledge about that ,so just a curious wonder. :)
>
> Nevo
>
>
>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Gnash-dev mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/gnash-dev