Chris Smith

Hi @baalexander - sorry to be a little late getting back to you. We should be good for the transfer - I'll take over control of the NPM package -...

Thanks! I believe I just need to publish to npm and that should handle it... @IanTheEngineer and I are planning to update the packages this week and then start pushing...

I believe you've stumbled on a case where we didn't update the on-the-fly message code to work the same way as the pre-generated messages. The generated messages specifically deal with...

As far as I'm aware. The uint8 arrays for pre-generated messages are already specially deserialized so we'd just be extending the case.

A few things I was trying to determine (and hadn't managed to yet)... - are TypedArrays strictly a superset of normal Arrays? Basically, that they officially supported all the same...

I started looking into doing this but discovered that the TypedArray views require that the byte offset be a multiple of the type size (see: [StackOverflow](https://stackoverflow.com/questions/15417310/why-typed-array-constructors-require-offset-to-be-multiple-of-underlying-type-si)). In specific circumstances (e.g...

Excluding on-the-fly messages, rosnodejs already does not allocate additional memory for uint8 arrays. For most other types of message, I think the current api does what users expect, though point...

Unfortunately I don't have any recommendations if these errors are specific to use with electron. I've used the two together but it's been a few years and I don't have...

In general during deserialization, no. It does happen when deserializing uint8 arrays unless you are using on-the-fly messages (#117). As I mentioned in #118 just now, the easiest way around...

`onTheFly` would mostly be useful if you are not in a ROS environment or are pre-kinetic (message generation via catkin was not released in earlier distributions). We aren't re-using some...