We need to keep the sources around until all sources are dispatched. If
not: source A and B are active after poll, A is dispatched first and
removes B, when B is then dispatched, the memory is gone. We don't
free the source but simply mark the fd invalid and move it do a
destroy list. After all sources are dispatched we destroy.
This is safe because removing a source is either done from the poll
context (with invoke) or when holding the right locks.
Implement audioconvert as a complex element of fmtconver,
channelmix and resample.
Make copying resample just to test.
Plug the converter into pw_stream.
Separate buffer sizes in the info. Make it possible to not set
data pointers for when we use memfd for the memory.
Improve format conversion.
Work on in-place conversion in test-convert2
Make explicit links between elements that are used to activate the
next element in the graph.
Make subgraphs a special regular node. Make a link from the
subgraph children to the parent so that the subgraph completes when
all the children completed.
Implement a single process function in plugins
Remove many messages in the client node
A device that supports continuous/step-wise frame interval will
report the correct frame intervals for index 0 and EINVAL for
every other index.
(https://linuxtv.org/downloads/v4l-dvb-apis/uapi/v4l/vidioc-enum-frameintervals.html)
Previously, the code would exit from the loop after successfully
reading index 0, without marking that the iteration of frame
interval completed. Therefore, the next time the function was
called the same frame intervals would be read, instead of advancing
to the next frame size or format type.
Instead, mark that we need to try and read the next format size.
Fixes#56
Don't use a destroy_list anymore, it's not safe anyway. Instead require
that sources be removed with (a blocking) invoke. Make sure we don't
dispatch sources that were removed from the loop.
