membrane_core
membrane_core copied to clipboard
membraneframework
The core of the Membrane Framework, advanced multimedia processing framework
The demand unit and other properties of pads are now (after #313) ignored as the pads are linked directly. We should either disallow passing them to avoid confusion or use...
We found out that there's no generic enough way to handle `{:error, reason}` tuples returned from callbacks. Currently, errors that can be recovered from are now handled with use of...
- [ ] Make sure other links are successful, there's no memory leak and a parent doesn't fail when a child fails between a link request and actual linking -...
1. The name of this metric isn't descriptive enough, it should indicate what happened to the buffer (i.e. whether it was sent or received), especially since we still have conflicting...
Since the separation of Start/End-of-Stream events, we allow to forward `:end_of_stream` as well (or at least the type says so). The docs, however: * [ ] Do not mention `handle_end_of_stream`...
Action redemand with a list of pads is supported but not documented in https://hexdocs.pm/membrane_core/Membrane.Element.Action.html#t:redemand_t/0
When Bin has dynamic input pads, linking them in handle_stopped_to_prepared instead of handle_pad_added doesn't work as expected. Expected behaviour: Dynamic input pads of the Bin should be linked correctly in...
We should revisit the shape of playback states, considering the following: - [ ] Playback states currently are not really related to the actual playbacks, but rather the readiness of...
I'm not sure if you want to include it here, but it's high time we removed those ugly "common" types 1. It adds useless jump when trying to find the...
@mat-hek wrote: I'm going to rename this and more functions in element internals to reflect better if they're handling incoming or outgoing data in a separate PR _Originally posted by...
Metadata
Owner
Metadata
The core of the Membrane Framework, advanced multimedia processing framework