Closed rapid-images-tore-levenstam closed 3 months ago
bjsplat
commented
3 months ago Please make sure to label your PR with "bug", "new feature" or "breaking change" label(s). To prevent this PR from going to the changelog marked it with the "skip changelog" label.
rapid-images-tore-levenstam
commented
3 months ago I want to note here that this fix seemed to work in our project which runs an earlier version of Babylon (6.31.0). I'm not sure if this change has other implications, but I couldn't find any direct issues from my limited testing.
bjsplat
commented
3 months ago Snapshot stored with reference name: refs/pull/15184/merge
Test environment: https://babylonsnapshots.z22.web.core.windows.net/refs/pull/15184/merge/index.html
To test a playground add it to the URL, for example:
https://babylonsnapshots.z22.web.core.windows.net/refs/pull/15184/merge/index.html#WGZLGJ#4600
Links to test babylon tools with this snapshot:
https://playground.babylonjs.com/?snapshot=refs/pull/15184/merge https://sandbox.babylonjs.com/?snapshot=refs/pull/15184/merge https://gui.babylonjs.com/?snapshot=refs/pull/15184/merge https://nme.babylonjs.com/?snapshot=refs/pull/15184/merge
To test the snapshot in the playground with a playground ID add it after the snapshot query string:
https://playground.babylonjs.com/?snapshot=refs/pull/15184/merge#BCU1XR#0
bjsplat
commented
3 months ago Visualization tests for WebGPU (Experimental)
bjsplat
commented
3 months ago WebGL2 visualization test reporter:
bjsplat
commented
3 months ago WebGL2 visualization test reporter:
bjsplat
commented
3 months ago Visualization tests for WebGPU (Experimental)
bjsplat
commented
3 months ago Visualization tests for WebGL 1 have failed. If some tests failed because the snapshots do not match, the report can be found at
If tests were successful afterwards, this report might not be available anymore.
rapid-images-tore-levenstam
commented
3 months ago Not sure why the formatting fails, or the WebGL1 visualization tests fail. Any ideas?
RaananW
commented
3 months ago Not sure why the formatting fails, or the WebGL1 visualization tests fail. Any ideas?
Formatting fails because something is off :-)
Run npm run format:fix, it will fix everything for you. Probably some whitespace issue.
About WebGL1 - I am running the tests again just to be sure it is related to this PR.
bjsplat
commented
3 months ago Visualization tests for WebGL 1 have failed. If some tests failed because the snapshots do not match, the report can be found at
If tests were successful afterwards, this report might not be available anymore.
rapid-images-tore-levenstam
commented
3 months ago Not sure why the formatting fails, or the WebGL1 visualization tests fail. Any ideas?
Formatting fails because something is off :-)
Run
npm run format:fix, it will fix everything for you. Probably some whitespace issue.About WebGL1 - I am running the tests again just to be sure it is related to this PR.
Alright, thanks for the tip, trying it :)
bjsplat
commented
3 months ago Visualization tests for WebGL 1 have failed. If some tests failed because the snapshots do not match, the report can be found at
If tests were successful afterwards, this report might not be available anymore.
RaananW
commented
3 months ago You can ignore the webgl1 error for now, it is related to some other PR (and webgl1 tests are optional)
Forum thread with more info: https://forum.babylonjs.com/t/babylon-leaks-msaa-renderbuffers-on-resize-dispose-with-webgl/51228
The short version is that renderbuffers created by MSAA RenderTargetTextures were never deleted, causing a (sometimes big) memory leak. In our project it could be several gigabytes of leaked memory in a few seconds. The proposed solution here is that thinEngine should use the
webGLHardwareTexture's functionreleaseto let it delete its resources, instead of directly deleting (only) the WebGLTexture from the context directly.