CPU-side dashing for GPU strokes & encoding-time filtering of zero-length segments #412
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
armansito
force-pushed
the
gpu-strokes-1-cpu-dashing
branch
from
c9171b0
to
5c82590
Compare
dfrg
approved these changes
Nov 21, 2023
crates/encoding/src/path.rs
Outdated
Comment on lines
772
to
777
| fn start_tangent_for_curve( | ||
| p0: (f32, f32), | ||
| p1: (f32, f32), | ||
| p2: (f32, f32), | ||
| p3: (f32, f32), | ||
| p2: Option<(f32, f32)>, | ||
| p3: Option<(f32, f32)>, | ||
| ) -> Option<(f32, f32)> { |
There was a problem hiding this comment.
Small style suggestion: making this a method of PathEncoder and dropping the p0 parameter (since it always appears to be PathEncoder::first_point?) should simplify the call sites a bit.
Comment on lines
72
to
73
| let missing_movetos = [ | ||
| MoveTo((0., 0.).into()), |
There was a problem hiding this comment.
I find this more amusing than is reasonable :)
There was a problem hiding this comment.
🙃 It's probably a good idea to fix these up and remove the MoveTo(0, 0)s when we fix the path encoding.
When GPU-side stroking is enbled, a stroke with a dash style now gets converted to a series of undashed strokes which then get encoded separately and in sequence. To support this, `Encoding` now accepts an iterator over `PathEl` as an alternative to `Shape`.
Zero-length segments lead to numerical errors and lead to extra complexity and wasted work (in the form of thread allocation) for GPU flattening. We now detect and drop them at encoding time. Currently, a path segment is considered zero-length if all of its control points (in local coordinates) are within EPSILON (1e-12) of each other. This may not be the desired behavior under transform. Also since the length here isn't in terms of the actual arc length, this won't detect all curves that are within an EPSILON sized bounding box (which may have a larger distance between their control points). For stroking purposes, the distance between the control points is what matters most as that's what's used to compute a curve's tangent.
armansito
force-pushed
the
gpu-strokes-1-cpu-dashing
branch
from
5c82590
to
8001151
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When GPU-side stroking is enabled, a stroke with a dash style now gets converted to a series of regular strokes which then get encoded separately and in sequence.
Zero-length segments lead to numerical errors and lead to extra complexity and wasted work (in the form of thread
allocation) for GPU flattening. We now detect and drop them at encoding time.
Currently, a path segment is considered zero-length if all of its control points (in local coordinates) are within EPSILON (1e-12) of each other. This may not be the desired behavior under transform.
Also since the length here isn't in terms of the actual arc length, this won't detect all curves that are within an EPSILON sized bounding box (which may have a larger distance between their control points).
For stroking purposes, the distance between the control points is what matters most as that's what's used to compute a curve's tangent.
This PR is based on #410
#303