docs(website): re-shoot the stale stills on the 0.67+ UI - #829
Open
edspencer wants to merge 5 commits into
Open
Conversation
added 3 commits
August 10, 2026 14:09
Every UI-bearing image on the site predated the redesign, including the seven committed by the previous pass — those were merged 41 minutes before the first design commit, so age was never the criterion; what landed 41 minutes after them was. Foundation dark at the out-of-box default (DEFAULT_APPEARANCE = foundation, hue null, tint 0, and paddock:theme defaults to dark) — what a reader sees on first boot. The theme quartet is the deliberate exception: there the choice itself is the subject, so all four appear. Shot against a rig rebuilt at main and verified post-redesign from inside the browser: the Appearance section with four theme cards cannot exist on any build before 75c3935. Per-frame provenance (theme, accent, app version) is recorded in tools/docs-media/shots.manifest.json.
An asset and the reference to it belong in the same PR. These five illustrate the 0.67 entry, which lives in whats-new.mdx on docs/whats-new-v069 (#815) — merged from here they would be assets referenced by nothing, and referenced from there they would be a 404 the Astro build does not reliably catch. Files stay on disk for #815 to pull across.
The alt text said four chats are unread; the frame shows nine. Alt text is the only description a screen-reader user gets, so a stale one is not cosmetic.
edspencer
pushed a commit
that referenced
this pull request
Aug 10, 2026
The stills were pushed onto this branch by a concurrent worker. This PR is the rig and the tooling, which is code-shaped and independently reviewable; the frames it produces are a separate review with separate criteria.
An asset and the reference to it belong in one PR. These five illustrate the 0.67 entry in whats-new.mdx, which lives on #815 — merged from here they are assets referenced by nothing; referenced from there without being there they are a 404 the Astro build does not reliably catch. They remain committed on docs/media-v069 for #815 to take across. NB an earlier attempt used `git rm --cached`, which left them in the working tree where the next blanket `git add -A` silently re-staged them.
Deploying paddock with
|
| Latest commit: |
399d7ac
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://37d9fbb6.paddock-7u2.pages.dev |
| Branch Preview URL: | https://docs-media-captures.paddock-7u2.pages.dev |
…ated row Two unread rows both read "Compare the two survey methods" — genuinely two chats in two workspaces, but in a published frame it reads as the list rendering a duplicate, in the one image that introduces Home to a new reader. The root chat is now "Pick a survey method for the tide pools". The unread arithmetic is unchanged and still reconciles: UNREAD 9 = 9 rows = badges (Home 2 + Lanternfish 2 + Tidepool 4 + Harbour Notes 1), with root CHATS 2 matching the two untagged rows.
edspencer
pushed a commit
that referenced
this pull request
Aug 10, 2026
docs-discover.png was captured and deliberately not committed: the rig has no discoverable Claude Code history, so the route renders its empty state. A manifest entry for a frame that exists in no PR invites someone to go looking for it. The twelve remaining entries all ship — seven in #829 and five in #815. Those five are NOT pruned despite having moved out of #829: they still ship, and the manifest is the only record of which theme and solved accent each was shot at, which is exactly what a reviewer of #815 needs.
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.
Every UI-bearing image on the site was pre-redesign — including the seven committed by the previous pass, which merged 41 minutes before the first design commit. Age was never the criterion; what landed after them was.
What changed
12 stills, all shot against a rig rebuilt at
main:root-home,project-settings,adopt-row,adopt-modal,adopted-badge,promote-to-project,trigger-tool-picker-bash.appearance-panelplus the four-theme quartet. These are not re-shoots — 0.67 has no What's New entry at all, so there is no historical record to protect.Theme: Foundation dark at the out-of-box default (
DEFAULT_APPEARANCE = { foundation, hue: null, tint: 0 };paddock:themedefaults to dark), because that is what a reader sees on first boot. The quartet is the deliberate exception — there the choice itself is the subject.Archive stills under
whats-new/are untouched on purpose: a What's New entry records what shipped on a date, so replacing its frame would falsify it.How I know these are not the old UI again
The single biggest risk here was re-shooting against a stale build — it fails silently, because the rig comes up, the seed passes, the leak scan passes and every shot succeeds. Guarded three ways:
b2503ec, not the one that predated the redesign./configrenders the Appearance section with four theme cards, which cannot exist before75c3935.--accent, app version) intools/docs-media/shots.manifest.json. The quartet records four distinct accent triples, which is what proves the themes really applied rather than four labels being written over one appearance.Verification
cd website && npm install && npm run build→ exit 0, 52 pages.md5sum | uniq -D -w32across every committed still → no duplicates. (Two unframed shots of one URL at one viewport land byte-identical; that has produced a phantom capture before.)assertCleanthrows rather than writing), and I looked at the frames.127.0.0.1count is 46, matching the baseline exactly — and the pattern was positive-controlled, so the clean result is real rather than a broken regex./home/demopaths, stock branding.Findings, not fixed here
getting-started.mdshows the wrong screen. The prose is about Discover ("Each row is one directory… press Import") but the image is Home. The image needs to become a Discover shot; that is a prose decision, so I only corrected the alt text, which claimed four unread against a frame showing nine.Docs-only: no changeset, no version bump.
Interaction with #835 (
fix/prose-code-quote-tokens) — checked, no re-shoot needed#835 changes how inline
codeand blockquotes paint in chat prose, acrossindex.cssand all four theme files. I checked whether that would re-stale anything here.No frame in this PR renders inline code or a blockquote as a subject. Verified per frame, using the
routerecorded in each sidecar to find the candidates rather than guessing:adopt-row,adopted-badge/projects/tidepool/chatadopt-modal/projects/tidepool/chatpromote-to-project/chatroot-home/project-settings,trigger-tool-picker-bashSo #835 can land before or after this PR without invalidating a frame.
Where the real exposure is, for whoever merges #835: the frames that do show rendered transcript prose are
using/sent-attachments.pngandusing/chat-streaming-queued.png, plus several underwhats-new/. None of those are in this PR — they were not re-shot in this pass. That is where to look if #835 lands.Three of the rig's fixture replies do contain the affected constructs (two with inline code, one with a blockquote), so a future transcript shot against this rig will exercise #835's change — which is why the check above is per-frame rather than per-fixture.