[Opera-Linux] video playback using the mplayerplugin succesor "Gecko media player" is broken due to a bug in Opera

[Eirik Byrkjeflot Anonsen]

Daniel Eckl <daniel.eckl at gmx.de> writes:

> I understand it's difficult and specs might be unclear. Perhaps it's not
> possible at all.
>
> It was just that Eiriks words below sounded as if he just was tired of
> this plugin subject in general.

I see what you mean.  It could definitely be read that way.  That was
not my intention, though.

>
> Thank you very much for your nice words and your clarification.
>
> Best,
> Daniel
>
>>>> There's way too much black magic in plug-in-land.  The whole plug-in
>>>> specification is really broken when it comes to interoperability.  And
>>>> nobody follows the spec anyway (which just makes it worse).
>>>>
>>>> eirik
>>>>
>>>>       
>>> Hi Eirik!
>>>
>>> How can I read your writing? Is this a confirmation of the supposed bug
>>> in opera? I assume so.

Actually, no.  I haven't looked at it.  It was entirely meant as some
observations on the general problems.  I could have said it better,
though...  I know that there has been some weirdness in the setwindow
calls though, so I assume there's some merit to the complaint.

Actually, I'm no longer working on the plug-in code, and haven't been
for quite a while.  Which is why I'm not being any more specific about
the problems that are found.  I quite simply don't know all that much
about it any more.

>>>
>>> So in other words you mean:
>>> "Everyone writes buggy plugins, so we can legitimately write a buggy
>>> plugin runner. Above that I cannot catch on with this stuff anymore,
>>> because it's a mess."
>>>
>>> ???
>>>
>>> _In case I read your writing correct_: Sorry, but this saying is not
>>> acceptable. If someone writes a buggy plugin: his fault. If opera writes
>>> a buggy implementation: Operas fault!

Completely agreed.

Unfortunately, in some cases we're not following the spec.  Typically
this is because some or other plug-in won't work if we follow the
spec.  That's one of the things I mean with "nobody follows the spec
anyway".

There's also the problem that the spec is designed to make it very
hard to make interoperable plug-ins.  Various internal design choices
of the browser does affect the context the plug-in is running in.
Thus the problem where mozilla was compiled with gcc-2.95 for a long
time after nobody else used it:  Flash was compiled with gcc 2.95, and
couldn't be run by a mozilla compiled with gcc 3.

The plug-in spec is broken.  Unfortunately, there's not much we can do
about that.  We just have to try to make plug-ins work anyway.

>>> With your point of view you make the story go round and round. Soon a
>>> plugin dev will say "Why should I write spec conform plugins? The
>>> browser developers don't follow the spec anyway!"

Which I'm sure is the case today, unfortunately.  I seem to remember
that mozilla used to recommend writing plug-ins that would only work
in mozilla (they were advocating using the gtk main loop from the
plug-in).

There are so many things that can go wrong with plug-ins that I'm sure
both browser devs and plug-in devs just make their browser/plug-in
work with a suitable set of plug-ins/browsers.


Anyway, I'm sure the current owner of the plug-in code will get it to
work with the main plug-ins, and some mplayer plug-in is surely on
that list.


eirik