Artstr-Studio/TODO.md

# Artstr Studio — TODO

Running list of deferred work. Detailed specs for the larger items live in
`docs/`. Each entry notes scope and where the spec is (if any).

## Adam's working list

- **Click-canvas-background-to-deselect.** In the vector canvas editor,
  clicking on empty canvas (outside any layer) should clear the layer
  selection. Today the selection persists, which makes it hard to "exit"
  a selected layer without picking another one.
- **Toggle crop marks for export previews.** A new toggle in the
  vector designer to hide crop marks from the PDF Print/Save and
  JPEG Export flows. Today crop marks bleed into clean exports.
- **Custom-art canvas size from imported background URL.** When a
  user pastes a background image URL on a Custom Art canvas, offer
  a one-click "Match canvas to image size" button (or apply
  automatically on first import) so the canvas dims = image dims.
  Edge cases: huge images (clamp to CUSTOM_ART_MAX), aspect-locked
  layouts (Slide is locked 16:9 — should be skipped or warn).
- **JPEG export: hide layer-selection outlines.** The Save-as-JPEG
  flow currently renders the selected-layer dashed outline in the
  output. Add `body.exporting` (or similar) state that strips
  selection chrome before the raster snapshot.
- **One-click JPEG export (skip the PDF round-trip).** Today the
  flow is: print → user saves PDF → user re-picks PDF → PDF.js
  converts to JPEG. We can't cache the PDF in between — `window
  .print()` doesn't expose the generated bytes to JS. Instead,
  write a direct canvas rasterizer that walks the layer model
  (image / text / shape / color / qr / pixelart) and draws each
  layer onto an off-screen canvas at the artboard's design dims,
  then `toBlob('image/jpeg')`. Same approach the pixel-art PNG
  export already uses, generalised. CORS-tainted remote images
  (cross-origin host without `Access-Control-Allow-Origin`) would
  break the direct path — fall back to the existing PDF round-trip
  for those with a one-line "this image needs the PDF route" toast.
  Most modern image hosts (CDNs, Blossom blobs) send CORS so the
  fallback should be rare. ~300–500 LoC of layer-by-layer drawing
  code; reuses `renderShapeIntoElement` etc. as references for
  geometry but writes to canvas instead of DOM.
- **Independent-axis corner-node resize.** Corner resize handles
  currently lock the aspect ratio (or scale both axes together).
  Behaviour should follow the actual drag vector: dragging mostly
  horizontally with a small vertical move resizes width a lot,
  height a little — both axes scale proportional to their own
  movement, not to each other. Applies to every layer type's
  resize handles, not just text.

## CORS for remote-image raster export

Quick reply, no action yet:

- The PDF Print/Save path renders via `window.print()` → the browser's
  native print/PDF engine, which handles cross-origin images fine. No
  CORS issue there.
- The JPEG export path already round-trips through PDF.js (PDF → JPEG),
  so it inherits the PDF path's CORS tolerance. If it's failing for
  some images, the root cause is likely the print path itself, not a
  canvas taint.
- The canvas-tainted situation only bites if we ever do a direct
  `<canvas>.toBlob('image/jpeg')` against a layout that contains a
  cross-origin remote `<img>` without CORS headers. Options if it
  becomes a real problem: (a) keep using the PDF.js round-trip;
  (b) recommend CORS-friendly hosts (Blossom blobs, ImgBB, most
  CDNs send `Access-Control-Allow-Origin: *`); (c) add an
  optional same-origin proxy endpoint — needs a server, which the
  app currently doesn't have.

## QR code layer — remaining phases

Phases A and B shipped (the `qr` layer type: data, error-correction level,
quiet zone, module / background colors, transparent background, renders
on every surface + canvas export, JSON round-trip; plus "Insert npub" and
"Insert share link" buttons under the QR data field).

- **Phase C — visual polish.** Optional center logo overlay and rounded
  modules. Riskier: a logo eats into the error-correction budget, so it
  should nudge the ECC level up. Skippable.

## Editor / tool UX

Small editor-usability improvements — no spec doc; scoped here.

Shipped: on-canvas drag-resize handles for every layer type, text-box
resize that never rescales the point size, the expanded Text tool panel
(font / size / color / alignment / bold / italic), a collapsible sidebar
toggle that fits the early-web aesthetic. Add new editor-UX items below.

## Profile pages — deferred polish

(Core profile pages at `/u/<npub>` already shipped.)

- Make author chips clickable in the **community thread view** (comments and
  reposts) — currently only feed-card and preview-meta author chips route to
  `openProfilePage`.
- **About-text clamp** — long bios overflow the profile header; clamp to ~4
  lines with a "Show more" toggle.
- **Profile-grid pagination** — an author with 100+ designs renders them all
  at once; cap the initial render and add a "Show more" pager.

## Edit-published-designs — deferred polish

(Core edit-in-place via NIP-33 replaceable semantics already shipped.)

- Replace the remaining `confirm()` calls in the edit / publish flow with the
  styled in-app modal pattern, for consistency with the close-project and
  publish-confirm modals.
- "Editing a previously-deleted design" warning — if the user opens Edit on a
  design they had tombstoned with a kind-5, warn that publishing will
  republish it.

## Pixel art — future polish

- **`pixelsB64` size-fallback encoder.** The spec calls for each frame
  to store whichever is smaller: `pixels` (RLE pairs) or `pixelsB64`
  (base64 of the raw `Uint8Array` of indices). Only RLE is implemented
  today. For noisy / photo-pixelized content with many palette colours,
  base64 is ~5× smaller — a 24-frame 64×64 sprite tested at 785 KB
  RLE would drop to ~150 KB. Worth shipping alongside a live event-size
  readout in the editor (soft warn at 50% of relay budget, hard block
  at 90%) so users see the budget before publish.

## Drafted-but-unbuilt features

Each has a full spec in `docs/`.

- **Private (encrypted) publishing** — `docs/PRIVATE_PUBLISHING_FEATURE.md`.
  Publish any design encrypted, readable only by the author (Phase 1
  "private to me"). The agreed next build; Phase 0 is a validation spike.
- **Artstr Stacks** — `docs/STACKS_FEATURE.md`. Interactive fullscreen
  page/card stacks (HyperCard for Nostr) — the Slide Deck extended with
  layer actions and an interactive viewer. Queued after private publishing.
- **Badges (NIP-58)** — `docs/BADGES_FEATURE.md`. Award / display Nostr
  badges on profile pages.
- **Collections (NIP-51 sets)** — `docs/COLLECTIONS_FEATURE.md`. Let users
  organize favorite community templates into named sets.
- **Zap-gated premium templates** — `docs/ZAP_GATED_TEMPLATES_FEATURE.md`.
  Creators mark a template premium with a minimum-zap threshold.
- **Zaps via Nostr Wallet Connect** — `docs/ZAPS_NWC_FEATURE.md`. In-app
  zapping without the wallet handoff. (Tracked LNURL-pay tips already ship;
  this would be the NWC path.)

## Shipped (recent)

- **Slide deck PDF export** — multi-page PDF, one page per slide. The
  existing Print / Save PDF button is context-aware: in the Deck Builder
  it builds a hidden print-only container with one page per slide (themed
  via `composeSlideForDeck`, each page sized to the preview's native
  dimensions so the slide fills the page exactly) and emits one PDF via
  `window.print()`. Inside a single-slide editor, the button still prints
  just that slide.
- **Copy / paste transform consistency** — paste now computes a single
  group bounding box across all clipboard items, applies one shared scale
  (only shrinks, never enlarges) and one shared translation, so cross-
  target paste keeps every layer's relative position and size to the
  others intact.
- **Pixel-editor colour picker** — ported the OS native `<input type="color">`
  over from the vector tools (same swatch styling; opens the OS picker with
  its built-in eyedropper on click). Removed the Screen-Capture-API screen-
  eyedropper workaround (~170 lines of dead code). Palette swatches and
  the in-canvas eyedropper tool now sync the picker too, so grabbing a
  palette colour as the starting point for a tweaked variant just works.

- **Pixel Art** — `docs/PIXEL_ART_FEATURE.md`. Phases A–D all shipped:
  embedded `pixelart` layer + standalone `casewrap-pixelart` design,
  PNG / sprite-sheet / GIF export, drawing tools (line / rect / ellipse,
  flip / rotate, grid resize), animation (multi-frame timeline, playback,
  onion-skin), image → pixelize, and animated previews in every preview
  surface. Undo / redo integrated with the top-level toolbar.
- **Pixel editor layout expansion** — addressed by the collapsible
  sidebar (gives both the pixel and vector editors a lot more breathing
  room when needed).
- **Print / Save event ID wrapping** — the event-id `<dd>` already uses
  `overflow-wrap: anywhere; word-break: break-word`, so no horizontal
  scroll.
- **TMDB auto-fill cleanup** — `docs/TMDB_AUTOFILL_FEATURE.md` removed.