Update wgpu to 27 #11419
Update wgpu to 27 #11419
Conversation
Wumpf
added
🔺 re_renderer
|
Web viewer built successfully.
Note: This comment is updated whenever you push a commit. |
Wumpf
force-pushed
the
andreas/update-wgpu
branch
from
ec2576b to
7d010aa
Compare
There was a problem hiding this comment.
Pull Request Overview
This PR updates the wgpu graphics library from version 26 to 27, enabling the removal of unsafe code related to buffer mapping. The update includes API changes for error handling, device polling, and buffer management, along with additional documentation explaining limitations with the new map_buffer_on_submit feature.
Key changes:
- Updated wgpu dependency from 26.0.1 to 27.0.0
- Removed unsafe lifetime transmutation in CPU write GPU read buffer implementation
- Updated API calls to match wgpu 27's new interface patterns
Reviewed Changes
Copilot reviewed 5 out of 6 changed files in this pull request and generated 1 comment.
Show a summary per file
| File | Description |
|---|---|
| Cargo.toml | Updated wgpu version and removed egui "log" feature |
| .cargo/config.toml | Added rust-lld linker configuration for Windows |
| crates/viewer/re_renderer/src/device_caps.rs | Simplified device descriptor creation using default values |
| crates/viewer/re_renderer/src/context.rs | Updated error callback and polling API to new wgpu 27 format |
| crates/viewer/re_renderer/src/allocator/cpu_write_gpu_read_belt.rs | Removed unsafe lifetime transmutation and improved documentation |
| crates/viewer/re_renderer/src/allocator/gpu_readback_belt.rs | Added documentation about map_buffer_on_submit limitations |
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
| let write_view = buffer_slice.get_mapped_range_mut(); | ||
| self.unused_offset = end_offset; | ||
|
|
||
| #[allow(unsafe_code)] |
There was a problem hiding this comment.
sooo happy to have this thing gone
Wumpf
force-pushed
the
andreas/update-wgpu
branch
from
7d010aa to
3fa7196
Compare
… use `map_buffer_on_submit`
Wumpf
force-pushed
the
andreas/update-wgpu
branch
from
3fa7196 to
0fb42a1
Compare
Co-authored-by: Copilot <[email protected]>
… into andreas/update-wgpu
|
Timeout was too aggressive. Need |
|
@rerun-bot full-check |
|
Started a full build: https://github.com/rerun-io/rerun/actions/runs/18274497663 |
| /// Prepare currently mapped buffers for use in a submission. | ||
| /// | ||
| /// This must be called before the command encoder(s) used in [`CpuWriteGpuReadBuffer`] copy operations are submitted. | ||
| /// All existing [`CpuWriteGpuReadBuffer`] MUST be dropped before calling this function. |
There was a problem hiding this comment.
[VOICE OF JULES FROM PULP FICTION]
"He should have a borrow checker for this kind o' deal."
Allows us to remove some very dreaded
unsafecode \o/Went on a rabbit-hole trip to use the fancy new
map_buffer_on_submitmethods. But it's actually really hard to do so because by design, our read/write belts can use a single chunk (==buffer) on multiple command encoders. Therefore, it's really hard to know which command encoder is the one that you callmap_buffer_on_submiton.I noted down my learnings in the code comments, hopefully that prevents future generations to get confused by this.
Testing