Fix feOffset/feDropShadow under rotated or skewed transforms#1063
Open
StefanoD wants to merge 1 commit into
Open
Fix feOffset/feDropShadow under rotated or skewed transforms#1063StefanoD wants to merge 1 commit into
StefanoD wants to merge 1 commit into
Conversation
feOffset's dx/dy were mapped through `scale_coordinates`, which only multiplies by the transform's scale factors and discards rotation and skew. As a result, an offset inside a rotated or skewed group was applied axis-aligned in device space instead of being rotated together with the filtered content, producing output inconsistent with browsers. Unlike blur radii, an offset is a vector and can be represented exactly, so map dx/dy through the full linear part of the transform via a new `transform_coordinates` helper, used by both feOffset and feDropShadow. Adds unit tests for the coordinate mapping and regenerates the feOffset/feMerge/feTile complex-transform references, which previously encoded the buggy behavior. Regenerated references show a shift: in these tests the filter output is the offset result itself (e.g. feOffset's complex-transform has only an feOffset), so correcting the offset moves the whole shape. For feOffset's complex-transform (dx=20 dy=40, transform skewX(30) translate(-50), plus 1.5x viewport scale), skewX adds tan(30)*dy to the x offset: the device offset goes from ~(34.6, 60)px to ~(64.6, 60)px, i.e. ~30px further right, y unchanged. This matches the SVG spec and Chrome. Known limitation (pre-existing, not introduced here): the filter region is still clipped using the axis-aligned bounding box of the (sheared or rotated) region rather than the exact transformed rectangle. With this fix the offset is correctly sheared further out, so it can now reach the default filter region boundary and get clipped there. The clip edge is therefore axis-aligned in device space instead of following the skew as in Chrome. Because the AABB is larger than the true region, resvg never clips more than a spec-perfect renderer would. Making the region clip exact would require reworking filter-region handling and is out of scope for this fix. Fixes linebender#949 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
feOffset's dx/dy were mapped through
scale_coordinates, which only multiplies by the transform's scale factors and discards rotation and skew. As a result, an offset inside a rotated or skewed group was applied axis-aligned in device space instead of being rotated together with the filtered content, producing output inconsistent with browsers.Unlike blur radii, an offset is a vector and can be represented exactly, so map dx/dy through the full linear part of the transform via a new
transform_coordinateshelper, used by both feOffset and feDropShadow.Adds unit tests for the coordinate mapping and regenerates the feOffset/feMerge/feTile complex-transform references, which previously encoded the buggy behavior.
Regenerated references show a shift: in these tests the filter output is the offset result itself (e.g. feOffset's complex-transform has only an feOffset), so correcting the offset moves the whole shape. For feOffset's complex-transform (dx=20 dy=40, transform skewX(30) translate(-50), plus 1.5x viewport scale), skewX adds tan(30)*dy to the x offset: the device offset goes from ~(34.6, 60)px to ~(64.6, 60)px, i.e. ~30px further right, y unchanged. This matches the SVG spec and Chrome.
Known limitation (pre-existing, not introduced here): the filter region is still clipped using the axis-aligned bounding box of the (sheared or rotated) region rather than the exact transformed rectangle. With this fix the offset is correctly sheared further out, so it can now reach the default filter region boundary and get clipped there. The clip edge is therefore axis-aligned in device space instead of following the skew as in Chrome. Because the AABB is larger than the true region, resvg never clips more than a spec-perfect renderer would. Making the region clip exact would require reworking filter-region handling and is out of scope for this fix.
Fixes #949
Generated by Claude