Issue 51 in vrjuggler: VRPN Text Gadgeteer Driver (Pull Request)

[vrju...@googlecode.com]

Status: New
Owner: ----
Labels: Type-Defect Priority-Medium

New issue 51 by firefly2...@gmail.com: VRPN Text Gadgeteer Driver (Pull
Request)
http://code.google.com/p/vrjuggler/issues/detail?id=51

I wrote a gadgeteer driver to support VRPN text. We're using this in
conjunction with a Kinect to pass text information back and forth. I made
a Git branch off of the current vrjuggler master branch and committed the
changes[1]. Note that since we use a heavily modified version of Vrjuggler
with the CMake build system[2] that I am unable to test these changes
against the VRJuggler master branch. I just wanted to make sure that they
get in the upstream version. Thanks.

https://github.com/carlsonp/vrjuggler/tree/VRPNText-squashed
[1]branch: VRPNText-squashed


https://github.com/rpavlik/vrjuggler
[2] We use the revised-cmake-3.0 branch from Ryan Pavlik.

--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings

[vrju...@googlecode.com]

Updates:
Labels: -Type-Defect Type-Enhancement

Comment #1 on issue 51 by patrick....@gmail.com: VRPN Text Gadgeteer Driver
(Pull Request)
http://code.google.com/p/vrjuggler/issues/detail?id=51

Is there a reason that this could not have been incorporated into the
existing gadget::Vrpn class?

[vrju...@googlecode.com]

Comment #2 on issue 51 by firefly2...@gmail.com: VRPN Text Gadgeteer Driver
(Pull Request)
http://code.google.com/p/vrjuggler/issues/detail?id=51

If I remember correctly, the way Gadgeteer is setup, you can have either a
tracked object or text, but not both. Thus the separation.

[vrju...@googlecode.com]

Comment #3 on issue 51 by patrick....@gmail.com: VRPN Text Gadgeteer Driver
(Pull Request)
http://code.google.com/p/vrjuggler/issues/detail?id=51

That is not correct. A single Gadgeeer device driver can provide multiple
types of input data. The gadget::Vrpn class is an example of that. It
supplies positional, digital, and analog data. It could also provide text
data.

[vrju...@googlecode.com]

Comment #4 on issue 51 by ryan.pav...@gmail.com: VRPN Text Gadgeteer Driver
(Pull Request)
http://code.google.com/p/vrjuggler/issues/detail?id=51

Actually, I'm pretty sure it was for 2 reasons:
1 - the existing "all in one" VRPN driver makes flawed assumptions, and
would ideally actually be split up, or at least made threadsafe
2 - It was difficult to come up with a suitable model for receiving voice
recognition results when servicing a VRPN text device asynchronously at a
high rate, and mainlooping any device on a connection may result in
delivery of messages to any device sharing that connection.

[vrju...@googlecode.com]

Comment #5 on issue 51 by patrick....@gmail.com: VRPN Text Gadgeteer Driver
(Pull Request)
http://code.google.com/p/vrjuggler/issues/detail?id=51

The VRPN driver has been through a lot of iterations in its lifetime, and
it makes me unhappy to hear that it is still deemed to be flawed. If a
rewrite is necessary, then so be it. It would be nice, however, if you
would elaborate on the problems in more detail--perhaps in a separate issue
related to refactoring or rewriting the VRPN driver class(es).

[vrju...@googlecode.com]

Comment #6 on issue 51 by firefly2...@gmail.com: VRPN Text Gadgeteer Driver
(Pull Request)
http://code.google.com/p/vrjuggler/issues/detail?id=51

There's a slightly cleaner branch with squashed commits in Ryan's Github
branch:

https://github.com/rpavlik/vrjuggler/tree/3.0-improvements