Changelog

The latest published Vello release is 0.3.0 which was released on 2024-10-04. You can find its changes documented below.

Unreleased

This release has an MSRV of 1.75.

0.3.0 - 2024-10-04

This release has an MSRV of 1.75.

Highlights

Support for most Emoji (#615, #641 by @DJMcNab)
GPU Friendly Stroke Expansion, which documents key parts of how Vello works, was released (by @raphlinus, @armansito)
Blurred rounded rectangles are now supported, which can be used for box shadows (#665 by @msiglreith)
Vello is no longer considered experimental 🎉 (#691 by @waywardmonkeys) Note that Vello is still an alpha, but we believe that the direction has been proven.

Added

Access to the Adapter from the utils DeviceHandle (#634 by @cfagot)
Support for compositing existing wgpu::Textures into a Vello scene (#636, #655 by @DJMcNab, @TrueDoctor)
Utilities for constructing an AaSupport from a set of AaConfigs (#654 by @simbleau)
An example which uses sdl2 (#671 by @TheNachoBIT)
The underlying Encoding for a scene can now be modified, circumventing guardrails for advanced use-cases (#701 by @timtom-dev)

Changed

Breaking: Updated wgpu to 22.1.0 (#635 by @waywardmonkeys)
Clipping more than four layers deep is now supported (#657 by @DJMcNab)
Significantly improved automated testing (#610, #643 by @DJMcNab)
Preliminary debug layers for Vello's internal development (#416 by @armansito)
Examples now use the run_app API from Winit (#626, #628 by @yutannihilation)
Labels on GPU objects are now prefixed with vello. (#677 by @waywardmonkeys)
The async render implementations are deprecated and unstable; you should use the non-async versions as they enable greater throughput (#706 by @DJMcNab)

Fixed

Example code in the repository README (#627 by @kmoon2437)
A possible crash on iOS working around an invariant undocumented by Apple (#639 by @DJMcNab)
Large number of clips now work (#659 by @raphlinus)
Empty clips now no longer cause artifacts (#651 by @raphlinus)
A potential panic in the presence of a weaker than default allocator (#675 by @timtom-dev)
Watertightness breaks causing artifacts with some rounded rectangles (#695 by @raphlinus)

Removed

Breaking: Pipelines API from vello_shaders (#612 by @DJMcNab)
The wgpu_profiler profiler feature is no longer stable (#694 by @DJMcNab)
Breaking: Moved the Recording abstraction into a low_level module, as almost all users should prefer the higher-level Renderer (#711 by @DJMcNab)

0.2.1 - 2024-07-16

This release has an MSRV of 1.75.

Fixed

Crash when there is no scene contents (#630by @DJMcNab)

Changed

Updated wgpu to 0.20.1. (#631by @waywardmonkeys)
Document the MSRV of releases in the changelog (#619 by @DJMcNab)

0.2.0 - 2024-06-08

This release has an MSRV of 1.75.

Added

Euler spiral based stroke expansion. (#496 by @raphlinus)
Sweep gradients. (#435 by @dfrg)
Bump allocation estimation. (#436, #454, #522 by @armansito)
Impl From<Encoding> for Scene. (#538 by @waywardmonkeys)
Glyph hinting support. (#544 by @dfrg)
Better glyph caching. (#555 by @dfrg)
vello_shaders crate to load and preprocess WGSL. (#563 by @armansito)
Coverage-mask specialization. (#540 by @armansito)
Support for the #enable post-process directive. (#550 by @armansito)

Changed

Better error types. (#516 by @DasLixou)
RenderContext::new() no longer returns a Result. (#547 by @waywardmonkeys)
Updated wgpu to 0.20. (#560 by @waywardmonkeys)

Removed

force_rw_storage feature. (#540 by @armansito)

Fixed

64k draw object limit. (#526 by @raphlinus)
Increased robustness of cubic params. (#521 by @raphlinus)
Increased robustness of GPU shaders. (#537 by @raphlinus)
draw_leaf uniformity. (#535 by @raphlinus)
Bug in join estimates in vello_encoding. (#573 by @armansito)
Incorrect use of numerical operators on atomics in binning. (#539 by @armansito)
path_reduced_scan buffer size. (#551 by @armansito)
Handling of upstream pipeline failure. (#553 by @armansito)
Very slow shader compilation. (#575 by @DJMcNab, @waywardmonkeys)
Full system hang on Apple systems. (#589 by @raphlinus)

0.1.0 - 2024-03-04

Initial release