This parameter is a weird one. With the abstraction of requests from
the underlying websocket protocol, there theoretically should be no
need to ignore invalid messages, because the implementation of the
low-level protocol on clients should be solid, with the requests
themselves not being fatal to the session.
As such, I consider this to be feature bloat, with lots of messy code
attributed to it.
We were previously storing config values in the WebSocketServer class
itself, in a weak attempt at modularity. It's better to give up on that
idea and just get it from the Config object itself.
- Removes the `isGroup` boolean field from the scene object, and does
not include any groups in the returned array.
- Reintroduces ordered results. Previous versions used a method which
did not return the scene list in the same order as the UI. This change
also means that this request is more susceptible to crashing OBS if
called during a scene collection change.
- Adds the `sceneIndex` number to the scene object. 0 being the bottom
of the scene list, just like in other requests like `GetSceneItemList`.
More docs-related commits will follow, but this needs to be merged in order to continue with other development.
* Docs: Overhaul docs generator (beginning)
* docs: Rename comments file
* docs: Move comments gitignore
* docs: Initial request documentation
* docs: Improvements to comment processing
* docs: More improvements
* docs: Add enum functionality for protocol.json
* WebSocketServer: Document enums
* RequestHandler: Document RequestStatus enum
* Base: Move ObsWebSocketRequestBatchExecutionType to its own file
Moves it to its own file, renaming it to `RequestBatchExecutionType`.
Changes the RPC to use integer values for selecting execution type
instead of strings.
* docs: Update introduction header
Removes the enum section, and documents RequestBatchExecutionType.
* WebSocketCloseCode: Shuffle a bit
* Base: Use `field` instead of `key` or `parameter` in most places
* RequestStatus: Mild shuffle
It was really bothering me that OutputPaused and OutputNotPaused
had to be separated, so we're breaking it while we're breaking
other stuff.
* docs: Delete old files
They may be added back in some form, but for now I'm getting them
out of the way.
* docs: Add enum identifier value
Forgot to add this before, oops
* docs: Document more enums
* docs: Add basic protocol.md generator
* docs: More work on MD generator
* docs: MD generator should be finished now
* docs: More fixes
* docs: More fixes
* docs: More tweaks + add readme
* docs: Update readme and add inputs docs
* docs: More documentation
This is probably one of the most requested features for obs-websocket.
This currently works by firing an event to all explicit subscribers
with an array of all active audio sources every **60 milliseconds.**
The `inputLevelsMul` field follows this data format:
Base: [Channel, Channel]
Channel: [magnitude (mul), peak (mul), input_peak (mul)]
*Not Muted* *Muted*
Example: [[0.3, 0.5, 0.9], [0.0, 0.0, 0.0]]
(input_peak is the actual peak value, before volume adjustment.)
You may notice that the values are only in mul. This is because we are
trying to cut down on bandwidth. dB values can be calculated using this
formula:
`dB = 20.0 * log10(mul)`
obs-studio's LOG_DEBUG setting only works in very specific
circumstances, which is why we implement our own debug logging. This
will help a lot of code cleanup.
