Skip to content

0.3.0 follow-up: junction case reads the visible corner; Windows contract hashes; landing page - #32

Merged
Stefatorus merged 3 commits into
mainfrom
ccr-73a29992-0f58d2
Oct 11, 2026
Merged

Stefatorus merged 3 commits into
mainfrom
ccr-73a29992-0f58d2

Conversation

@Stefatorus

Copy link
Copy Markdown
Member

What changed

Three follow-ups to the 0.3.0 merge (#30). The v0.3.0 tag waits on this PR.

  • bench/cases/junctions.py: junction_tri reads the junction where it shows.
    • Failure: main's regression-gate job failed its hard cases, Quality step, with junction_tri reading 0.48 px against a 0.2 px limit.
    • Cause: the trace's junction is exactly where 0.2.7 put it. The check measured how near each region's path comes to the true junction. Since 0.3.0, Quality completes a face beneath the faces painted over it. The lower wedge of the Y junction is now a rectangle under the other two. The visible corner is where the covering faces' edges meet. The rectangle's own edge is 0.48 px away, under paint that hides it.
    • Fix:
      • The check now renders the 4×4 source pixels around the true junction at 64 samples per pixel.
      • Each of the artist's regions is matched to the traced area of its ink that overlaps it most.
      • The error is the furthest any of them comes to the junction, minus the truth's own reach.
    • Calibration: I moved the true geometry by known amounts. Shifts of 0.05, 0.1, 0.2 and 0.5 px read 0.06, 0.11, 0.20 and 0.43–0.50 on both junction cases.
  • bindings/contract/cases.json: the x86_64-windows-msvc hashes, recorded by CI (facade job on main, 3d1841c).
    • The release's Linux re-bless had dropped them.
    • The file is taken unchanged from the artifact: 48 lines added, nothing else differs.
  • web/index.html: the landing page's "What's new" is now 0.3.0's.
    • Six cards, each figure taken from the 0.3.0 changelog.
    • The Space build used for the deploy carries this page.

No engine code changes, so output is identical to main.

How verified

  • cargo test --release --workspace: not re-run, no Rust changes. cargo test --release -p inkvec --test contract passes with the new hashes.
  • cargo fmt / clippy: no Rust changes.
  • Hard cases: bench/cases.py --ratchet quality 42/42 and --ratchet fast (28 recorded) pass.
    • junction_tri reads 0.080 px on this build and on 0.2.7 alike.
    • junction_quad reads 0.035 px here and 0.014 px on 0.2.7.
  • Output byte-identical to main: there are no engine changes.
  • docs/THIRD_PARTY.md: no dependency changes.
  • CHANGELOG.md: nothing user-visible beyond the website.
  • The landing page was checked in headless Chromium inside the full Space build (studio/scripts/build-web.mjs). The Studio Lite smoke test passes (studio/web/smoke.py, engine 0.3.0, 4 threads, cross-origin isolated).

🤖 Generated with Claude Code

https://claude.ai/code/session_016AZWMqqHhRQ337N5EWhNC6


Generated by Claude Code

Six cards from the 0.3.0 changelog, with its figures: fidelity against 0.2.7, JPEG halos
and parameters on the noisy tier, faces completed behind their covers and counters cut
out, strokes by default, nodes at extrema and --editability, and the Lean theory with its
generated checkers. The 0.2.0 cards are in that release's changelog.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016AZWMqqHhRQ337N5EWhNC6
`junction_tri` failed on main after the 0.3.0 merge (junction_err_px 0.48 against 0.2),
with the junction exactly where 0.2.7 put it. The check measured how near each large
region's path comes to the true junction, which assumed a path is what shows. Since
0.3.0, Quality completes a face beneath the faces painted over it: the Y junction's
lower wedge comes back as a rectangle under the other two (`<rect y="62.75" ...>`), the
visible corner is where its covers' edges meet, and the rectangle's own edge is 0.48 px
from the junction under paint that hides it.

The junction is now read from the render. The four source pixels about the true junction
are rendered at 64 samples per pixel (the document nested in a window viewBox, so both
frames stay the document's own), each artist region is matched to the traced area of its
ink that overlaps it most, and the furthest any of them comes to the junction, less the
truth's own reach (0.01-0.03 px of blended samples), is the error. Moving the true
geometry by 0.05, 0.1, 0.2, 0.3 and 0.5 px reads 0.06, 0.11, 0.20, 0.27-0.31 and
0.43-0.50 on both cases, so the 0.2 px limit means what it said. At the whole-image
SCALE the nearest sample inside a 120-degree corner can sit 0.2 px from its apex, which
is why the window.

junction_tri reads 0.080 on this build and on 0.2.7 alike (the visible corner, 0.09 px
off, has not moved); junction_quad 0.035 and 0.014. Quality 42/42 and Fast's 28
recorded cases pass their ratchets.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016AZWMqqHhRQ337N5EWhNC6
…output

Taken unchanged from the facade (windows-latest) job's INKVEC_BLESS=add artifact on main
at 3d1841c (bindings run 38135195486): 48 lines added, nothing else differs from the
committed file. The release's re-bless on Linux had dropped every Windows entry, so the
Windows contract ran on everything but the exact bytes until now.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016AZWMqqHhRQ337N5EWhNC6
@Stefatorus
Stefatorus merged commit a1e7be1 into main Oct 11, 2026
45 of 46 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants