Robotics

Gazebo PMC postpones Ogre 3.0 upgrade, eyes Bevy integration

New committer sponsorship vote; texture loading gets 5-10% speed boost

Deep Dive

The Gazebo Project Management Committee (PMC) held its weekly meeting on June 8, 2026, covering rendering strategy shifts, package optimizations, and infrastructure enhancements. A formal vote was initiated to sponsor a new committer. The primary discussion centered on upgrading Gazebo's rendering from Ogre 2.3 to Ogre 3.0 (referenced in issue gz-rendering#1289). The team concluded they are not ready to proceed due to non-trivial review requirements. Instead, they are weighing whether to focus exclusively on Bevy integration, a long-term roadmap goal. Community members were reminded to contribute to the Gazebo Rotary artwork brainstorming document by June 12.

Several package updates were discussed. In gz-math, a pull request to cache vertices and edges was merged (gz-math#771), though backporting was skipped because upcoming Entity Component Manager improvements will supersede it. A gz-common optimization to speed up texture image decoding and channel conversion (gz-common#817) is under review, expected to improve loading times by 5% to 10%. A gz-physics PR implements AttachFixedJointFeature and DetachJointFeature in MuJoCo (gz-physics#1001). A fix for a Grid3D CMake warning on non-Linux OS was implemented in gz-gui (gz-gui#770). Memory allocation improvements were discussed: passing a unique name or hash from gz-common to gz-rendering to avoid redundant allocations for identical textures. Infrastructure work includes removing the flakyTest plugin to resolve high RAM usage, automating branch protection with Terraform, and using Anubis to block scraper attacks. The Rotary documentation PR (docs#685) is nearly ready, adding a separate directory; feedback is welcomed.

Key Points
  • Rendering upgrade from Ogre 2.3 to 3.0 postponed; Bevy integration under evaluation.
  • gz-common texture decoding optimization expected to improve loading times by 5–10%.
  • Memory allocation mechanism proposed to avoid redundant texture copies by passing unique hashes.

Why It Matters

Gazebo's rendering roadmap shift and performance tweaks directly impact simulation realism and developer efficiency.