In wgpu 0.7, this was a valid operation, the zero padding was done by hand.
However in #872 this was removed as an optimization since buffers are zero padded, however the slice end was taken from buffer.slice which has to be aligned when mapped.
894: Skip mapping of zero sized buffers in create_buffer_init to avoid a panic and avoid creating an unused padded buffer r=kvark a=Imberflur
Co-authored-by: Imbris <imbrisf@gmail.com>
1358: Big refactor of the profiling usage r=kvark a=kvark
**Connections**
**Description**
Improves the profiling markers in many ways:
1. remove unnecessary leftovers from the `tracing` era, such as querying device/adapter features. These are one-time and light-weight. It's unlikely we are going to be interested in them.
2. add markers to where it really matters, such as `maintain` pieces, `submit` phases, render pass preparation, etc.
3. shorten the markers to be strictly the function names, passing the class name as a tag
**Testing**
Tested on vange-rs
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1354: Update profiling to 1.0 r=kvark a=kvark
**Connections**
https://github.com/gfx-rs/gfx/pull/3734
**Description**
Some crates like to release 1.0
**Testing**
Should work
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1353: Update naga to gfx-23 r=kvark a=kvark
**Connections**
Picks up https://github.com/gfx-rs/gfx/pull/3733
**Description**
Just a bunch of Naga fixes
**Testing**
To be tested in wgpu-rs
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1351: Hard error on SPIRV parsing failure when VALIDATION is requested r=cwfitzgerald a=kvark
**Connections**
Follow-up from https://github.com/gfx-rs/wgpu/issues/1350#issuecomment-825186159
**Description**
If the user provided SPIR-V, and they requested VALIDATION, and we fail to parse the file - there is no point to proceed.
VALIDATION flag should be a gaurantee of safe operation, so we can't just throw the same SPIR-V at SPIRV-Cross and pretend that it's ok.
**Testing**
Tested on noclip shaders
Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
873: Boids example: cleaner rand code r=kvark a=dhardy
... for a certain definition of cleaner. Is this clearer or just more confusing?
Co-authored-by: Diggory Hardy <git@dhardy.name>
1348: Change error type for the backend shader r=kvark a=kvark
**Connections**
Fixes#1347
**Description**
Issues a better error for this case
**Testing**
Should work
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1346: Make implicit ids actually errors to start with r=kvark a=kvark
**Connections**
Follow-up to #1341
**Description**
Turns out, leaving the indices `Vacant` is also not an option, since the `hub.rs` code considers this an internal error (unexpected), and panics.
**Testing**
Tested on Austin's examples in Gecko
Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
1332: Validate min/max blending factors r=kvark a=kvark
**Connections**
Based on #1325
Matches https://github.com/gpuweb/gpuweb/pull/1635
**Description**
This PR refactors the color state errors, giving them more detailed context, and having a dedicated enum for them.
It also adds the check for the blend factors to be (1, 1) for the min/max operations.
**Testing**
Untested
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1341: Mark unused implicit BGL ids as vacant r=kvark a=kvark
**Connections**
Fixes https://github.com/gfx-rs/wgpu/issues/1340
**Description**
This is not an ideal solution. We aren't marking the IDs as errors. But it's an improvement, and it resolves the crash in FF, supposedly.
**Testing**
Tested on wgpu-rs examples with implicit layouts
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>