I am very happy. Today I managed to write the GstDasfSink. It is a GStreamer element which detects what kind of component is on his left. If it have a Goo component inside, then the GstDasfSink will activate de DASF mode on it et voila, the speakers reproduce the sound without further processing.
I set the filter component to send empty GstBuffers to the Sink if it have activated de DASF mode, but I do not know if it is necessary. I am going to try tomorrow if the pipeline do not collapse without sending GstBuffers to the next element. Because in this way the performance will be improved.
Ivan just amazed me again. He add the setup_tunnel method in the GooComponent after writing the classes for the MPEG4 decoder and the Video PostProcessor. They are not clean, and randomly breaks, but the tunneling is working! There is too much janitor work to do, but the overall advances are very exciting.
The other thing missing is the documentation. The design specification for Goo is the biggest missing parts in the project.
TODO: clean the current code; correct the dist target in autoconf; design spec; be showy with management about the results; clock issue; and a lot of things more. And bug fixing something utterly broke (the current gstelements for OpenMAX).
En otro orden de ideas, sigo leyendo Arte, mente y cerebro de Howard Garnder. Impresionantes ideas sobre el arte, la psicologia y la epistemologia.