[work-73gq] feat(studio): the egg catalogue — all 72, across twelve mechanisms #72
No reviewers
Labels
No labels
review:human
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
ember/mossfire!72
Loading…
Reference in a new issue
No description provided.
Delete branch "eggs-integration"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Every egg in
Mossfire/Design/Ornamenting the Studio.md§4, plusstudio/EGGS.md, the ledger that records each one with its law-checks so 72 of them cannot quietly drift from canon.Eight workstreams, one per mechanism group, each in its own worktree and branch — the eight merge commits are in this branch's history, one per bead:
.1data/hall.js, 120 authored#mossfirelines interleaved with the bash.org corpus, and the typed commands (/consult,/awaken,/trace,/read the omens,/tend the fire,/whois,/quit,/tally, the seven marks, law 11, the Beast) +EGGS.md.2#000000gives three seconds of the Hush.3js/cycle.js.4#seer).5TAKESROOT), the aspect sign column, realgit hash-objectcommits, the true test count, and the tokens that light up in the sample when you click a swatch.6CHRONICLE.txt.7?beast.8What the per-branch suites could not see, and this branch fixes. Every test loads its modules in isolation, so the integration bugs were invisible until the browser:
MF.codex.The token table was clobbered by the projection of the book, which loads later;app.jsthen threw insiderenderThemeListduring init, soMF.appnever appeared — the whole studio was dead while 33 tests stayed green. The book keepsMF.codex; the token module is nowMF.vat.codex.sealedvscodex.seal), because my two task specs disagreed. Merged as-is, the vat lock would have set a field the export path ignored: shipped-looking and doing nothing. One shape now, with the importer normalising the old name instead of dropping it.js/cycle.jsnested the clock inside.s-right, andinsertBeforeneeds a direct child — soviewMarkthrew and abortedupdate()before the clock and the Cycle painted.Verified, not assumed
build_data.test.pygreen on the integrated branch, andbuild_data.pyleavesstudio/datawith no diff/whois everesolves through the projection — proof the two modules are wired together), the vat lock seals to{seal:true,lot:1}, 27 presets with element marks, the register flip repaints and labels itself, the acrostic readsTAKESROOT, htop reports the institutions,22 passedand9.81 7.42 5.77= 23, and both gates' footer logic holdsNot verified: the time-of-day Cycle branches (real clock), a real download of a sealed ZIP, and the gate-conditional pages on their actual hosts. Those rest on the tests.
Tests were updated in three places where they encoded pre-merge expectations — the lynx page set, the tally's shape, and the seal field name. None was weakened; each now says why.
Beads: work-73gq.1 … work-73gq.8
Eight parallel branches, merged by hand. What the unit suites could not see, because every test loads its modules in isolation: - two modules claimed MF.codex. js/codex.js (the token table, sigil helper and seal field) was clobbered by data/codex.js (the projection of the book), which loads later; app.js then threw inside renderThemeList during init, so MF.app never appeared and the studio was dead while 33 tests stayed green. The book keeps MF.codex; the token module is MF.vat. - two branches built the same seal with different field names — codex.sealed and codex.seal. Merged as-is the vat lock would have set a field the export path ignored. One shape now: { seal: true, lot: n }, with the importer normalising the old name rather than dropping it. - js/cycle.js nested the clock inside .s-right, so the register flip's viewMark threw (insertBefore needs a direct child) and aborted update() before the clock and the Cycle painted. Also: /tally now runs the tally module's own line instead of the Hall's stale 'assumption about its shape; the test that encoded the old expectations says so; the discovery line for the tend-the-fire command no longer uses a word that reads as a slur on a public page; and the Atlas says the count aloud (egg 56). Tests updated where they encoded pre-merge expectations: gates (the lynx page set), irc-commands (the tally's shape and words), vatlock (the seal shape). Beads: work-73gq.1 through work-73gq.8