The five widgets the audit could not judge were not broken (#212) - #215
Merged
Conversation
Scale, RangeSlider, ColorChooser, ColorPicker and Agenda were left out of the
metric-scale catalogue because their runs did not double. I wrote that up as "a
scaled metric composed with an unscaled one-pixel something" and said it needed
reading their Draw. It did, and the answer is that there is nothing wrong with
them.
The edge of a rounded shape is one pixel of anti-aliased coverage at ANY scale,
plus a border that does scale. So the overhead around a feature's interior is
partly constant, and interior = size - 2*(AA + border) is 12 at one scale and 26
to 28 at twice it -- never 24. What doubles is the WHOLE feature:
1x runs: [1 71 1 1 12 1 1 71 1] thumb = 1+1+12+1+1 = 16
2x runs: [1 143 2 28 2 143 1] thumb = 2+28+2 = 32
Every merge policy that would let the run-length audit see this also hides the
defect it exists for: merging the slivers into their neighbour turns an unscaled
one-pixel border into a rounding difference of two pixels in a run of three
hundred. I tried four of them.
So the audit stays strict and blind to this one shape, and the five get the
assertion that suits them: the thumb's whole extent, edges included, doubles.
An exemption nobody checks would have been the wrong answer to the same
question.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This was referenced Aug 16, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #212.
Scale,RangeSlider,ColorChooser,ColorPickerandAgendawere left out of the metric-scale catalogue because their runs did not double. I wrote that up as "a scaled metric composed with an unscaled one-pixel something" and said it needed reading theirDraw. It did — and the answer is that there is nothing wrong with them.The edge of a rounded shape is one pixel of anti-aliased coverage at any scale, plus a border that does scale. So the overhead around a feature's interior is partly constant, and
interior = size − 2*(AA + border)is 12 at one scale and 26–28 at twice it — never 24. What doubles is the whole feature:Why the audit is not "fixed" to see this
Every merge policy that would let a run-length comparison see it also hides the defect it exists for. Merging slivers into their neighbour turns an unscaled one-pixel border into a rounding difference of two pixels in a run of three hundred — which is exactly the
Card/Buttondefect this whole audit started from. I tried four policies; each traded one class for the other.So the audit stays strict and blind to this one shape, and the five get the assertion that suits them:
TestRoundedThumbDoublesmeasures the thumb's whole extent, edges included, and requires it to double. An exemption nobody checks would have been the wrong answer to the same question.🤖 Generated with Claude Code