Repository navigation
Rebuild the viewer as canvases (Phase 31 slice I) - #253
Merged
Merged
Conversation
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Phase 31 slice I, the last slice. The phase began when framesmith judged its own viewer and gave every page a perfect usability score while the viewer had a 17px Back link and 21 click targets built from plain spans. This PR rebuilds the viewer as canvases on the finished phase, through the real tools, and keeps the rebuild as a test. The run surfaced three framesmith defects, fixed here.
The rebuild
test-viewer-rebuild.tsdrives the MCP server over stdio, astest-ridgeline.tsdoes for Phase 30:project_create, then one design system for the whole app at project scope:generate_design_systemwith seed#B45309(amber, the viewer's accent), thetechnicalpersonality and the 14px desktop-app body.canvas_create({ type: "desktop-app" }), thenapply_structure, thenbatch_design:browser-app.editor-inspector-app.split-workbench-app.canvas_add_variant:comment-open: a popover under the Comment button, placed from the button'ssnapshot_layoutbox.confirm-delete: the delete dialog.emptystate.No autofix, no token overrides and no rule changes. Every canvas resolves its tokens from the project, and every colour on every canvas is a
$token.What every one of the six shows, in both modes:
desktop-app-handbuilt-targetfindings.surface.rulesOnholds all 14desktop-app-*rules, andrulesOffholdstype-scale-ratio.rulesOnlists the rules that are enabled.desktop-app-row-rhythmis enabled but not exercised, because none of the six canvases has a table.big-numbershas nothing to widen on the empty archive, which shows no figures.The test also proves its own checks are live: ten twins that must fail, nine of them on throwaway clones of the detail page and one a check of the test's own duplicate-card predicate. They cover:
The six baseline defects, by name
$text-secondarymeasures 5.58 / 6.00 / 5.58:1 on the three surfaces in light and 4.64 / 5.38 / 4.64:1 in darklink(iconarrow-left, label "Back"), 24px tall, with nodesktop-app-targetfindingtabsnode: Quality, Design system, Feedback, one activelink("Show 4 nodes")destructivebutton behind a divider; Archive is secondaryThe full screens, in the dark theme, are in
docs/viewer-rebuild/:Where the rebuild differs from the live viewer:
list-rowhas no selected state and the detail panel shows one selected canvas.Three fixes the rebuild needed
Detailed evaluate now measures layout in the design system's fonts (approved before building).
canvas_evaluatelaid the canvas out in the system fallback font.screenshot,canvas_stressandsnapshot_layoutrender in the design system's own faces. So a detailed CLEAN was judged on different glyph widths than the screen shows.Font warningsitem, as it does forscreenshot.test-evaluate-fonts.tspins this. It sets amonospacebody, which flips two detailed verdicts, and runs asans-seriftwin where neither fires.A drawer and a scrolling popover or menu read as scrolling under stress.
overflow: "auto"were already read as scrolling, but overlays build their own markup and missed the marker. So the comment popover (a fixed height with a scrolling body) and every drawer were reported as clipped under long text.overflow: "hidden"still clips.Fonts used only inside a component are loaded.
A smaller correction came with these:
Docs
canvas_evaluatedocstring, README and GUIDELINES say detailed mode lays out in the design system's fonts.canvas_stressdocstring, README and GUIDELINES name the drawer and the scrolling popover or menu.batch_designdocstring and GUIDELINES recommendoverflow: "auto"on a fixed-height popover or menu.nav-itemcounts are not a stress slot.Testing
npm run test:full: 138/138 on the final revision.test-ridgeline.tsand the desktop pattern gates still pass with detailed evaluate in the real fonts.