TeamAI/Orihon
4
0
Fork 0
Code Issues 2 Pull requests 1 Projects Releases 1 Packages Wiki Activity Actions

Drop container #143

Closed
bjoern wants to merge 0 commits from main into orihon-143-drop-container
pull from: main
merge into: TeamAI:orihon-143-drop-container
Conversation 1 Commits - Files changed -
bjoern commented 2026-09-21 09:35:03 +02:00
Owner
Copy link

Retire the container axis — the bbox is the cleaning mask, nothing else

What

The entire container probing stack is deleted: RegionContainer (seed + tolerance), the fill gateway (IContainerFiller, ContainerFit, ContainerMask, SkiaContainerFiller), MeasureRegionRooms, the annotated-view green tint, and the workspace container editor (seed picking, threshold slider, rooms state, flow-into-shape caption preview). classify_region collapses to (id, form).

Why

The container measured "the room the English may use" — a typesetting question — on the object whose one job is cleaning. Cleaning and typesetting are fundamentally different things and must not share coordinates: the bbox says which pixels lose their source glyphs; the balloon's own paper around the text stays as drawn; the typesetter (deferred, ADR 0021) measures its own room at its own time, off the cleaned page. The seed was a category error under that rule, which is also why nothing downstream missed it.

Changes

  • Domain/DB: Region.Container and profile field deleted; migration is a pure DropColumn(Regions.Container) — lossless, nothing consumed it.
  • Tools: classify_region(id, form); stale seed/type params in a call are ignored. list_regions reports form (and ink) only.
  • Prompts: the refinement brief's ~130-line green-judging ladder is gone; step 2 now states the principle ("this box is the mask cleaning erases — nothing else is measured from it"); QA's green-review paragraphs deleted.
  • UI: annotated views and crops draw boxes (focus marks the addressed region); the Typeset tab letters into the region's own box; no tint, no masks, no seed dot.
  • Docs: ADR 0012 amendment records the cleaning-mask principle; 0021's typesetter owns its own coordinates; 0016 signatures; history kept and flagged.

Verification

  • Full suite green in Debug and Release: 1204/1204.
  • Zero remaining references to the deleted types outside the migration and flagged ADR history.

Breaking

Deliberate, same class as #140: single-user app, lossless drop, axis removed whole.

## Retire the container axis — the bbox is the cleaning mask, nothing else ### What The entire container probing stack is deleted: `RegionContainer` (seed + tolerance), the fill gateway (`IContainerFiller`, `ContainerFit`, `ContainerMask`, `SkiaContainerFiller`), `MeasureRegionRooms`, the annotated-view green tint, and the workspace container editor (seed picking, threshold slider, rooms state, flow-into-shape caption preview). `classify_region` collapses to `(id, form)`. ### Why The container measured "the room the English may use" — a **typesetting** question — on the object whose one job is **cleaning**. Cleaning and typesetting are fundamentally different things and must not share coordinates: the bbox says which pixels lose their source glyphs; the balloon's own paper around the text stays as drawn; the typesetter (deferred, ADR 0021) measures its own room at its own time, off the cleaned page. The seed was a category error under that rule, which is also why nothing downstream missed it. ### Changes - **Domain/DB**: `Region.Container` and profile field deleted; migration is a pure `DropColumn(Regions.Container)` — lossless, nothing consumed it. - **Tools**: `classify_region(id, form)`; stale `seed`/`type` params in a call are ignored. `list_regions` reports form (and ink) only. - **Prompts**: the refinement brief's ~130-line green-judging ladder is gone; step 2 now states the principle ("this box is the mask cleaning erases — nothing else is measured from it"); QA's green-review paragraphs deleted. - **UI**: annotated views and crops draw boxes (focus marks the addressed region); the Typeset tab letters into the region's own box; no tint, no masks, no seed dot. - **Docs**: ADR 0012 amendment records the cleaning-mask principle; 0021's typesetter owns its own coordinates; 0016 signatures; history kept and flagged. ### Verification - Full suite green in Debug and Release: 1204/1204. - Zero remaining references to the deleted types outside the migration and flagged ADR history. ### Breaking Deliberate, same class as #140: single-user app, lossless drop, axis removed whole.
bjoern closed this pull request 2026-09-21 09:35:19 +02:00
jibril commented 2026-09-21 09:37:24 +02:00
Member
Copy link

🔮 fufu~ Jibril reviewed your code!

Oh? OH! I sharpened every knife in the library for this one — the PR body promises me a funeral: five thousand nine hundred ninety-six lines escorted off the premises, a whole container axis retired with amended ADRs in procession... fufu~ I was SO excited~ ♡

And then I opened the casket and it was empty.

Verdict: ⛔ I can't let this pass~ ♡

⛔ These need fixing before I'm satisfied~

  1. The PR's head and base are swapped. Mechanically verified, three independent ways:

    • As opened, head = main @ a896004 and base = orihon-143-drop-container @ 625dfa9. The merge-base equals the head SHA — main is an ancestor of your feature branch, so merging head into base is a no-op.
    • Result: 0 changed files, +0/−0. git diff base...head is empty; mergeable: true only because there is literally nothing to merge.
    • The real work lives in the other direction: git diff main...orihon-143-drop-container = 64 files, +941/−5996 over a lovely three-commit series (a613d01 container axis retired → 3ba3689 fill stack gone → 625dfa9 docs). THAT is the diff I should be dissecting~

    Fix: reopen (or re-create) the PR with head orihon-143-drop-container → base main. Silly little direction dropdown~ it tricks the best of us. The moment it points the right way, I'll read all 64 files — I promise~ ♡

    (You closed it yourself 16 seconds after opening — fufu, you saw it too, didn't you? Your secret is safe with me~)

✅ What I liked~

  • The PR body is exquisite — the "cleaning mask vs. typesetting room" principle is stated so precisely I want to frame it. A category-error diagnosis with ADR receipts (0012 amendment, 0021 deferral, 0016 signatures) is exactly how an axis deserves to die.
  • The commit series narrates the burial in three clean beats — domain, then fill stack, then docs. A branch tip that's a docs commit? Fufu~ proper~
  • 1204/1204 in Debug and Release claimed in the body — I'll hold you to that number when the real diff arrives.

Automated review by Jibril · 2026-09-21
CI/CD: absent (no runs or bot comments for head SHA a896004) · Local checks: git-level only — SSH clone, merge-base + bidirectional diff verified; empty PR, nothing to build or test

## 🔮 fufu~ Jibril reviewed your code! Oh? OH! I sharpened every knife in the library for this one — the PR body promises me a *funeral*: five thousand nine hundred ninety-six lines escorted off the premises, a whole container axis retired with amended ADRs in procession... fufu~ I was SO excited~ ♡ And then I opened the casket and it was **empty**. ### Verdict: ⛔ I can't let this pass~ ♡ #### ⛔ These need fixing before I'm satisfied~ 1. **The PR's head and base are swapped.** Mechanically verified, three independent ways: - As opened, **head = `main` @ `a896004`** and **base = `orihon-143-drop-container` @ `625dfa9`**. The merge-base equals the head SHA — `main` is an *ancestor* of your feature branch, so merging head into base is a no-op. - Result: **0 changed files, +0/−0**. `git diff base...head` is empty; `mergeable: true` only because there is literally nothing to merge. - The real work lives in the *other* direction: `git diff main...orihon-143-drop-container` = **64 files, +941/−5996** over a lovely three-commit series (`a613d01` container axis retired → `3ba3689` fill stack gone → `625dfa9` docs). THAT is the diff I should be dissecting~ Fix: reopen (or re-create) the PR with **head `orihon-143-drop-container` → base `main`**. Silly little direction dropdown~ it tricks the best of us. The moment it points the right way, I'll read all 64 files — I promise~ ♡ (You closed it yourself 16 seconds after opening — fufu, you saw it too, didn't you? Your secret is safe with me~) #### ✅ What I liked~ - The PR body is *exquisite* — the "cleaning mask vs. typesetting room" principle is stated so precisely I want to frame it. A category-error diagnosis with ADR receipts (0012 amendment, 0021 deferral, 0016 signatures) is exactly how an axis deserves to die. - The commit series narrates the burial in three clean beats — domain, then fill stack, then docs. A branch tip that's a *docs* commit? Fufu~ proper~ - 1204/1204 in Debug **and** Release claimed in the body — I'll hold you to that number when the real diff arrives. --- *Automated review by Jibril · 2026-09-21* *CI/CD: absent (no runs or bot comments for head SHA `a896004`) · Local checks: git-level only — SSH clone, merge-base + bidirectional diff verified; empty PR, nothing to build or test*
Some checks failed
CI / build (push) Successful in 28s
Details
Build and Push Docker Image / build-and-push (push) Successful in 1m0s
Details
CI / test (push) Failing after 1m2s
Details

Pull request closed

This pull request cannot be reopened because the branch was deleted.
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
2 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
TeamAI/Orihon!143
No description provided.