0.2.10

Fix explode in connect when trying to open a port that's not there

0.2.9

Add support for multiple universes via multiple dongles

0.2.8

Add line to readme about linux

0.2.7

Do not explode in firefox - just bail out (more specifically, check if navigator.serial exists)

0.2.6

Generalize the backends into buffered and direct, as Enttec pro seems to be happy running the same instruction set that eurolite pro is

0.2.5

On failing open port still start ticking as we are relying on it. The error is shown in console. Potentially might want to catch it, but this is fine for now.

0.2.4

Simplify the api so that we don't have a mysterious init and then a connect on top of that. Also make sure that the ticker gets called regardless of whether we are talking to something or not. This detaches any painting logic from physical connection. Might later flip on this but sort of makes sense, as we can't rely on apps clock as that one might overwhelm the DMX widget

0.2.3

Failed miserably with the sniffing of model because serial numbers for these dongle's aren't specific to the repackager but rather they are just unique; so no way for now to tell an enttec apart from a eurolite. oh well. back to feeding in a model in constructor

0.2.2

Still battling timing woes; The dongles do not want to adhere to exact 40fps rather have to do a blocking send and then give a bit of time for the thing to breathe before sending the next package

0.2.0

Instead of requesting that a backend class is passed in, we ask user to grant us permission to the USB device and figure out the backend from the serial number. Cons: asking permission twice. Pros: don't have to think about backend model;
The change breaks API - now you just need to call .connect() without specifying the backend

0.1.9

Sort out timing so that we come back after 23ms and not wait for the operation to finish the 23ms correspond to the 44Hz that the DMX512 is capable of (effectively 44 fps)
Eurolite - when new data stops coming in, we still send another 4 frames to make sure the dongle hasn't stuck on an unfinished state (seems to happen at times)

0.1.8

Put setTimeout back as it was inexplicably driving eurolite's dongle nuts on non-linux systems or somesuch (still figuring it out proper)

0.1.7

Fix a requestPermission mixup (it is supposed to be a static method)

0.1.6

Lighten up the send loop by relying on setInterval rather than setTimeout
Fix init explodes when connecting happens on the fly