Paul,

Great, thanks so much.

Don't worry, I get the "no new features right before a stable release" -
I've done this "open-source thing" before so I understand the tricky
position I put you in.  I really thought this one was going to end up as a
bug fix, which is why I had mentioned it before it was ready.  It took
longer than expected because we wanted to review and test it well here
before sending in a patch.

We aren't actually using trunk yet here, which is why these patches I submit
are first against 2.8 - I backport them to trunk with a git rebase. I
haven't gotten a chance to set up a good trunk+external osgIntrospection and
wrappers build on all our platforms, and since our group uses osg almost
exclusively in conjunction with the osgIntrospection-based osgLua bindings,
that's kept us from working on trunk.

I'm looking to get a continuous integration build server set up here sooner
or later, so while I'm at it I'll get the full stack (osg/osgIntrospection)
building on there in both branches so that I can have a bit more confidence
before updating on my (slow...) main dev machine.

Thanks again!

Ryan

On Fri, May 27, 2011 at 9:22 AM, Paul Martz <[email protected]> wrote:

> Hi Ryan -- After performing some testing, I've committed this as r12463,
> with a tweak similar to the one made by Robert when he committed this to
> trunk.
>
> But at this point I'm going to have to say "no" to any new features for
> 2.8.5 that haven't been in a trunk dev release for at least a month. Just
> too much risk. The 2.8.5 stable release is not an appropriate place for
> untested modifications.
>   -Paul
>
>
>
>
> On 5/25/2011 8:28 PM, Paul Martz wrote:
>
>> My main concern is that we're very close to a 2.8.5 release, and your
>> change is
>> completely untested.
>>
>> My reason for pursuing 2.8.5 all along has been to get some of the changes
>> that
>> have been waiting on trunk into a stable release. Putting a change into
>> 2.8.5
>> that hasn't been tested yet, not even on trunk, when we're this close to a
>> release, doesn't seem appropriate for a stable release.
>>
>> Things to consider:
>>
>> * Can you get Robert to review this change and add it to trunk? If so,
>> that
>> would give your change a good vote of confidence, and make me more
>> receptive to it.
>>
>> * Does it have to be in 2.8.5? You seem to have been getting along just
>> fine
>> without it to date (didn't even submit it to trunk until now). How about
>> it
>> waits for a 2.8.6 release? Or a 3.0 release?
>>
>> I'm not trying to shaft you -- you've been very helpful with the wrappers.
>> But
>> the job of a release manager is to stop this kind of feature creep and get
>> the
>> release out the door.
>>
>> So please consider what I've said above. In the meantime, I've added it to
>> my
>> local build and will take a look.
>> -Paul
>>
>>
>>
>> On 5/25/2011 5:01 PM, Ryan Pavlik wrote:
>>
>>> Yes, this is the only osgconv change, and the one I was referring to
>>> earlier. We
>>> developed it here so that our plugin for SketchUp to export to OSG ( [1]
>>> which
>>> exports to collada then converts with osgconv) would produce the expected
>>> results, keeping the world axes where they are shown. We had earlier seen
>>> unexpected behavior when scaling and rotating using osgconv, so I had
>>> anticipated that I would have a bug fix to submit. However, inspection of
>>> the
>>> code revealed that the behavior was just different (differently-useful)
>>> and not
>>> immediately obvious, rather than necessarily wrong. Thus, the "fix" comes
>>> in the
>>> form of a small new feature to enable the behavior we expected on
>>> request. With
>>> this change included, 2.8.5 will be fully functional with the SketchUp
>>> export
>>> plugin for use in VR environments using either the binary or text
>>> formats.
>>>
>>> If it's easier to visualize, the diff can be seen here:
>>>
>>> https://github.com/rpavlik/osg/compare/OpenSceneGraph-2.8...osgconv-worldframe
>>> It's pretty straightforward, limited to only the osgconv binary, and
>>> doesn't
>>> result in changing existing behavior, which is why I was hoping it would
>>> be
>>> suitable for the 2.8.5 release.
>>>
>>> Thanks!
>>>
>>> Ryan
>>>
>>> [1] https://github.com/rpavlik/sketchupToOSG
>>>
>>> On Wed, May 25, 2011 at 4:58 PM, Paul Martz <[email protected]
>>> <mailto:[email protected]>> wrote:
>>>
>>> Ryan, is this the only change to osgconv you have? Are there any more
>>> changes coming for the 2.8.5 release?
>>>
>>> I know you had previously mentioned a change to osgconv, but I did not
>>> realize you were talking about an entirely new untested feature, as
>>> opposed
>>> to a backport of something already on trunk.
>>> -Paul
>>>
>>>
>>>
>>> On 5/25/2011 3:38 PM, Ryan Pavlik wrote:
>>>
>>> All,
>>>
>>> I've attached zipped modified full files for trunk and the 2.8 branch of
>>> the
>>> promised improvement to the osgconv tool.
>>>
>>> Existing osgconv behavior is to transform the model bounding sphere
>>> center to
>>> the world origin before performing transformations specified on the
>>> command
>>> line, and translating back after rotation and scaling unless an alternate
>>> translation is specified. This patch adds a setting to the
>>> OrientationConverter
>>> class in osgconv to disable this extra transformation, which has the
>>> effect of
>>> applying specified transforms with respect to the input world coordinate
>>> system,
>>> rather than to the center of the bounding sphere. It also adds a command
>>> line
>>> argument "--use-world-frame" to enable this behavior. When this command
>>> line
>>> argument is not passed, behavior is unchanged from before the patch. The
>>> usage
>>> text has been updated to reflect this additional option, and the comments
>>> in
>>> OrientationConverter are also updated.
>>>
>>> Thanks!
>>>
>>> Ryan
>>>
>>> --
>>> Ryan Pavlik
>>> HCI Graduate Student
>>> Virtual Reality Applications Center
>>> Iowa State University
>>>
>>> [email protected] <mailto:[email protected]>
>>> <mailto:[email protected] <mailto:[email protected]>>
>>> http://academic.cleardefinition.com
>>>
>>>
>>>
>>> _______________________________________________
>>> osg-submissions mailing list
>>> [email protected]
>>> <mailto:[email protected]>
>>>
>>> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
>>>
>>>
>>>
>>> --
>>> -Paul Martz Skew Matrix Software
>>> http://www.skew-matrix.com/
>>> _______________________________________________
>>> osg-submissions mailing list
>>> [email protected]
>>> <mailto:[email protected]>
>>>
>>> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
>>>
>>>
>>>
>>>
>>> --
>>> Ryan Pavlik
>>> HCI Graduate Student
>>> Virtual Reality Applications Center
>>> Iowa State University
>>>
>>> [email protected] <mailto:[email protected]>
>>> http://academic.cleardefinition.com
>>>
>>>
>>>
>>> _______________________________________________
>>> osg-submissions mailing list
>>> [email protected]
>>>
>>> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
>>>
>>
>>
>>
>
> --
>  -Paul Martz      Skew Matrix Software
>                   http://www.skew-matrix.com/
> _______________________________________________
> osg-submissions mailing list
> [email protected]
>
> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
>



-- 
Ryan Pavlik
HCI Graduate Student
Virtual Reality Applications Center
Iowa State University

[email protected]
http://academic.cleardefinition.com
_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org

Reply via email to