[Gnash-dev] architecture question

[Charles Crisler]

I am curious about the architecture plan for gnash relative to GUIs and
rendering. Currently you support several GUIs and 3 rendering backends.
I also haven't found that you can download Linux binaries. Maybe there
is a correlation? :-)

While working through the rendering code I saw that it doesn't support
alpha blending. Is that because you support 3 backends and they are all
different? If so, how do you intend to address this for eventual
release? If a flash file uses alpha blending, will the current (0.8.4)
version properly decode it to the intermediate RGB image? I am
considering 'grabbing' that RGB image to render it myself and really
need the alpha blending info. If you haven't included it, how would I go
about adding it? I guess that the first step would be to download (and
read) the Flash file format spec.

BTW - I saw mention by someone of suggesting the Flash ActiveX control
with gnash on Windows. That doesn't work so well relative to the alpha
blending. It seems that you have to blit to black, blit to white,
examine both and adjust/set the alpha blending, all on a pixel by pixel
basis because it seems random. I wasn't involved on that project and
don't know the minute details. However, I know someone who has
implemented this and the performance sucks (rendering twice
unnecessarily never helps). If anyone knows something that I don't about
how to do this correctly (rendering Flash using the ActiveX control on
Windows and handling alpha blending), please let me know. Otherwise,
keeping your own flash decoder may be the smartest approach by far.

Thank you,
Chuck Crisler