fix: memory issues

[kuruk-mm]

Fix #415

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 70.07042% with 85 lines in your changes are missing coverage. Please review.

Project coverage is 47.74%. Comparing base (0df27ef) to head (4d77015).
Report is 14 commits behind head on main.

Files	Patch %	Lines
...ntraland-godot-lib/src/content/content_provider.rs	70.09%	32 Missing ⚠️
...ntraland-godot-lib/src/godot_classes/dcl_config.rs	80.00%	11 Missing ⚠️
...and-godot-lib/src/godot_classes/resource_locker.rs	86.66%	4 Missing ⚠️
rust/decentraland-godot-lib/src/content/gltf.rs	84.21%	3 Missing ⚠️
...t-lib/src/scene_runner/components/avatar_attach.rs	0.00%	3 Missing ⚠️
...ot-lib/src/scene_runner/components/avatar_shape.rs	0.00%	3 Missing ⚠️
...ib/src/scene_runner/components/ui/ui_background.rs	0.00%	3 Missing ⚠️
...-lib/src/scene_runner/components/ui/ui_dropdown.rs	0.00%	3 Missing ⚠️
...dot-lib/src/scene_runner/components/ui/ui_input.rs	0.00%	3 Missing ⚠️
...odot-lib/src/scene_runner/components/ui/ui_text.rs	0.00%	3 Missing ⚠️
... and 10 more

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #414      +/-   ##
==========================================
+ Coverage   47.61%   47.74%   +0.12%     
==========================================
  Files         155      157       +2     
  Lines       19376    19599     +223     
==========================================
+ Hits         9226     9357     +131     
- Misses      10150    10242      +92     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[leanmendoza]

Great work! It looks good.
Something I'm not really sure, using queue_free instead of remove_child it's ok in terms of memory management, but is it desirable to keep it parented to the given node until the end of the frame?

You also have commented code with remove_child

[leanmendoza]

Wouldn't it be better to use self.cached.retain() and mutate it instead of creating a new one each time? Even with this rate 1/sec, it makes more sense