Export lights to dae #228
Export lights to dae #228
Conversation
Signed-off-by: ddengster <ed.fan@osrfoundation.org>
Signed-off-by: ddengster <ed.fan@osrfoundation.org>
github-actions
bot
added
🏢 edifice
graphics/src/ColladaExporter_TEST.cc
Outdated
| EXPECT_TRUE(technique); | ||
| auto spot = technique->FirstChildElement("spot"); | ||
| EXPECT_TRUE(spot); | ||
| auto color = spot->FirstChildElement("color"); |
There was a problem hiding this comment.
Do you need these variables?
Signed-off-by: ddengster <ed.fan@osrfoundation.org>
Codecov Report
@@ Coverage Diff @@
## ign-common4 #228 +/- ##
===============================================
+ Coverage 76.83% 76.99% +0.15%
===============================================
Files 74 74
Lines 10491 10563 +72
===============================================
+ Hits 8061 8133 +72
Misses 2430 2430
Continue to review full report at Codecov.
|
Signed-off-by: ddengster <ed.fan@osrfoundation.org>
Signed-off-by: ddengster <ed.fan@osrfoundation.org>
| math::Vector3d direction; | ||
|
|
||
| /// \brief Light position (non directional lights only) | ||
| math::Vector3d position; |
There was a problem hiding this comment.
As I commented on gazebosim/gz-sim#912 (comment), we should use the light's full math::Pose3d and apply a direction on top of that.
Now that this has been released, we can't extend this struct though. That's why we prefer full classes with getters and setters using the PIMPL pattern; they involve a lot more boilerplate, but at least they can be extended without breaking ABI.
I think the solution will have to be calculating the direction field based on the light's pose orientation together with the light's direction.
/// \brief Light position (non directional lights only)
For directional lights, the position doesn't affect how they light up the scene, but it's often useful to set so we choose where the light's visual representation shows up on the GUI.
There was a problem hiding this comment.
Sorry for the delayed reply, was pushing some deadlines.
Now that this has been released, we can't extend this struct though.
If I may ask, why so? Is there some technical reason?
This struct is specifically used for collada exporting - I tried to make it stick as close as possible to collada specifications. Most notably their <light> tag is omitting information on all orientations and direction - meaning that I have to push the transform data to it's parent node that instantiates the light under <node>.
We can use math::Pose3d but it's orientation component has it's members set to private as seen here, and in addition there's another problem where the collada specification's <rotate> tag is not be the same as a quaternion.
Reference https://www.khronos.org/files/collada_spec_1_4.pdf, search for <light>
This does mean if anyone decides to use it for something else they are taking on certain risks. It's something I would expect not many users to be using.
There was a problem hiding this comment.
If I may ask, why so? Is there some technical reason?
ABI compatibility 😕
I think you may have solved the issue in gazebosim/gz-sim#912 (comment)
🎉 New feature
Closes #901
Downstream PR
Summary
Test it
run
./build/ignition-common4/bin/UNIT_ColladaExporter_TESTChecklist
codecheckpassed (See contributing)Note to maintainers: Remember to use Squash-Merge