idml 0.2.6 → 0.2.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 02ab670c36492471dd8dd986380fd180826e9ce250d5e22dde6fb5fda18b9bd3
-  data.tar.gz: 73d2c392dee732f12b1551a445f5e1a866a551885742adf3552d81abe0c3e9d4
+  metadata.gz: e9833da9d4a8ddda74890980eab76f2a47b03cdb99977730fb1dee5d8809dd68
+  data.tar.gz: 1d771cc2775f1071329d2b2dbab681cc46197546be27681fcba39ee9313f9a2b
 SHA512:
-  metadata.gz: b5c44e670ba295c964c6b58f6cecab0dcde43b579d4e4eb282e2d5b66cf567df34f943377a84510df7edd84d831892b3f3d6e13d7a4b13eb5311058377b94494
-  data.tar.gz: d1692d7d0ed411358d2ed877497abe2838fe25d2a1aa6a9fc534561b116743fd0a4e279819ea0206b8dfb5ecaf29c8f806e4341cca941267a2c2dc71193dbd9e
+  metadata.gz: e3dfbfc1c8ccdccdcd53070ff9aaf8de3f150699e524c3b6b4668c9a2d266cfb530ed5d8fb1e5b515be6da145df17e60832d9f9f60c0c9b2c106092adf716dfd
+  data.tar.gz: 437d8bcdc82e87f999b916ea9fc10486f2f4b8ae57965b9753fae465ab8f570684f4b2ea7eb8ae81fbc40fdf6a66288387429d265053562cf2dc4581101473be

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    idml (0.2.6)
+    idml (0.2.8)
       bigdecimal
       fontisan
       lutaml-model (~> 0.8.18)
@@ -198,7 +198,7 @@ CHECKSUMS
   ffi (1.17.4-x86_64-linux-gnu) sha256=9d3db14c2eae074b382fa9c083fe95aec6e0a1451da249eab096c34002bc752d
   ffi (1.17.4-x86_64-linux-musl) sha256=3fdf9888483de005f8ef8d1cf2d3b20d86626af206cbf780f6a6a12439a9c49e
   fontisan (0.4.45) sha256=2337ca0205f806647b24b8cac57aa34f3769473026ae6f52dc36bde0feed6c62
-  idml (0.2.6)
+  idml (0.2.8)
   json (2.21.2) sha256=1f1d3b7cf2b3ba1a69beca0bb6db13d5438b80bff3cd54cdaaa620b9b07c1c6a
   language_server-protocol (3.17.0.6) sha256=5ef2c0c138f8267e1bc631d3328347d354f96724b0af22f2c79516120443b7f0
   lint_roller (1.1.0) sha256=2c0c845b632a7d172cb849cc90c1bce937a28c5c8ccccb50dfd46a485003cc87

data/TODO.pdf/52-font-subsetting.md

@@ -1,52 +1,45 @@
 # TODO PDF 52: Font subsetting via pdfrb
 
-## Status: BLOCKED (pdfrb has no subsetting API)
-
-## Goal
-
-Subset embedded TrueType fonts to include only the glyphs actually used
-in the document. This dramatically reduces PDF file size (from full font
-~500KB to subset ~5-20KB per font).
-
-## Blocker
-
-pdfrb 0.4.0 exposes no subsetting API:
-
-- `Pdfrb::Document::Fonts#add` accepts `**opts` but ignores `subset:`.
-- No `used_codepoints` collector.
-- `glyph_width` returns stub 500 for TrueType — the per-glyph data
-  needed to build a subset is unavailable.
-
-Without per-glyph metrics, subsetting cannot be implemented in the
-caller either. This is also the same blocker that blocks TODO 63
-(replace FontMetrics with pdfrb).
-
-## Path forward
-
-Either:
-
-1. pdfrb adds a subsetting API that consumes a `used_codepoints:` set
-   and emits a subsetted FontFile2 at write time, OR
-2. pdfrb grows real TTF table parsing (head/hmtx/cmap) so the caller
-   can collect used codepoints, build the subset, and pass the bytes
-   back to `Fonts#add` as pre-subsetted data.
-
-Once pdfrb unblocks, the idml side needs to:
-
-1. Pipeline collects all Unicode codepoints across all stories.
-2. Pipeline passes `used_codepoints: set` to `Fonts#add(subset: true, ...)`.
-3. Spec renders text with known characters and verifies the embedded
-   font only contains those glyphs.
-
-## Acceptance criteria (after pdfrb unblocks)
-
-- [ ] Pipeline collects all Unicode codepoints used across all stories.
-- [ ] For each embedded font, only the used glyphs are included.
-- [ ] FontFile2 contains a subsetted font table (not the full font).
-- [ ] PDF file size significantly reduced vs full-font embedding.
-- [ ] Spec: render text with known characters, verify embedded font
-      contains only those glyphs.
-
-## Dependencies
-
-- pdfrb subsetting API (NOT YET AVAILABLE in 0.4.0).
+## Status: DONE
+
+## What was implemented
+
+Pipeline now subsets embedded TrueType fonts via pdfrb 0.4.0's
+`Fonts#subset_fonts!`. The flow:
+
+1. Every `canvas.text` / `canvas.text_lines` call auto-populates
+   `document.fonts.used_codepoints(resource)` per font.
+2. `Pipeline#call` invokes `writer.subset_fonts!` after rendering all
+   spreads and before `writer.write`.
+3. `PdfrbWriter#subset_fonts!` delegates to
+   `document.fonts.subset_fonts!`, which rewrites each registered
+   font's FontFile2 with a subset containing only the used glyphs
+   (plus glyph 0 / notdef).
+4. Subsetting uses `Pdfrb::Font::TrueType::Subsetter`, which rebuilds
+   glyf/loca/cmap/hmtx/hhea/maxp/head tables and resolves composite
+   glyph references.
+
+## Opting out
+
+`Idml::Render.render(..., subset_fonts: false)` skips the call.
+The CLI exposes `--no-subset` for the same effect. Useful when a
+downstream consumer needs the full font table (e.g. searchable PDFs
+with a specific glyph set).
+
+## Verification
+
+- `lib/idml/render/pdfrb_writer.rb:48` — `subset_fonts!` delegation.
+- `lib/idml/render/pipeline.rb:38` — `subset_fonts!` call site.
+- `spec/idml/render/render_pdfrb_writer_spec.rb:120` — subsetting
+  produces a sub-15KB PDF for "Hello" in Arial.
+
+## Acceptance criteria
+
+- [x] Pipeline collects used codepoints automatically via pdfrb's
+      `encode_text` hook.
+- [x] Each embedded font's FontFile2 is rewritten with a subset
+      containing only used glyphs.
+- [x] PDF file size significantly reduced vs full-font embedding.
+- [x] Spec renders text with known characters and verifies the
+      output PDF is small.
+- [x] `subset_fonts: false` skips subsetting.

data/TODO.pdf/63-replace-fontmetrics-with-pdfrb.md

@@ -1,73 +1,60 @@
 # TODO PDF 63: Replace FontMetrics with pdfrb measurement API
 
-## Status: BLOCKED (pdfrb measurement APIs are stubs)
+## Status: PARTIALLY UNBLOCKED (subsetting works; measurement still AFM-only)
 
 ## Goal
 
 Replace `Idml::TextEngine::FontMetrics` (200+ lines of TTF binary parsing
 via Fontisan) with pdfrb's native `Fonts#measure_text`, `#glyph_width`,
-and `#metrics_for`. Eliminates the Fontisan dependency for text layout.
+and `#metrics_for`.
 
-## Blocker
+## What pdfrb 0.4.0 provides
 
-pdfrb 0.4.0 ships the API surface but the implementations are stubs
-that return constant values, not real per-glyph widths:
+- `Pdfrb::Font::TrueType::File` — real TTF parser with Head, Hhea,
+  Cmap, Hmtx, OS2 tables.
+- `Pdfrb::Font::TrueType::Subsetter` — real subsetting (TODO 52 uses
+  this — DONE).
+- `Fonts#glyph_width(char, resource)` — uses AFM metrics (Standard 14)
+  only; returns `DEFAULT_WIDTH = 500` for TTF.
+- `Fonts#measure_text(text, font:, size:)` — same: AFM or stub
+  (`length * 0.5 * size`).
+- `Fonts#metrics_for(resource)` — AFM metrics only.
 
-`/Users/mulgogi/src/claricle/pdfrb/lib/pdfrb/document/fonts.rb`:
+## What still needs pdfrb work
 
-```ruby
-def measure_text(text, font:, size:)
-  return 0 unless text
-  # TODO: use font metrics for per-glyph width lookup
-  text.to_s.length * (size || 0).to_f * 0.5
-end
+pdfrb's TTF parser exists but is not wired to the measurement API.
+`Fonts#glyph_width` reads `@afm_metrics[resource]`, which is only
+populated for Standard 14 AFM fonts (Helvetica, Times, Courier,
+Symbol, ZapfDingbats). For every other font (any TTF/OTF), it falls
+back to the stub.
 
-def glyph_width(_char, _resource)
-  500
-end
+The per-glyph-width proposal asked for "Look up glyph ID from cmap,
+then width from hmtx." The infrastructure is there (Cmap, Hmtx) but
+the integration with `Fonts#glyph_width` is pending.
 
-def metrics_for(_resource)
-  nil
-end
-```
+Until pdfrb's measurement API wires the TTF parser, the idml gem's
+Shaper and LineBreaker must keep using Fontisan-based FontMetrics
+for accurate per-glyph widths.
 
-The Shaper and LineBreaker depend on accurate per-glyph widths for
-correct word-wrap. With stub data, line breaks land at the wrong
-positions and run advance (needed for TODO 67's `text_rich`) is wrong.
+## Plan (after pdfrb unblocks)
 
-Fontisan correctly parses head/hhea/hmtx/cmap tables for any TTF/OTF
-and exposes per-glyph advance widths. Until pdfrb either grows real
-TTF parsing or accepts externally-provided metrics (see
-`/Users/mulgogi/src/claricle/pdfrb/PROPOSAL.external-font-metrics.md`),
-Fontisan stays.
-
-## Path forward
-
-Proposal at `~/src/claricle/pdfrb/PROPOSAL.external-font-metrics.md`
-suggests letting `Fonts#add` accept `widths:`/`metrics:`/`encoding:`
-hashes from the caller, so pdfrb does not need to parse TTF tables
-itself. The idml gem would then build those hashes from FontMetrics
-and pass them in — keeping Fontisan as the parser while pdfrb handles
-only PDF assembly. This is the Unix-philosophy split.
-
-Once that proposal lands:
-
-1. `PdfrbWriter#register_font_with_metrics(path, widths:, metrics:, ...)`.
-2. Pipeline resolves fonts via FontResolver (Fontisan) and passes the
-   resulting metrics to pdfrb.
-3. Internal `FontMetrics` may still exist as the Fontisan adapter, but
-   Shaper/LineBreaker no longer need to read it — they go through
-   `pdfrb.measure_text`.
+1. `PdfrbWriter#register_font(path)` records the TTF bytes for
+   later parsing.
+2. New `PdfrbFontMetrics` adapter implements FontMetrics' interface
+   by calling `pdfrb.fonts.glyph_width(font_resource, codepoint)`.
+3. `FontResolver` returns `PdfrbFontMetrics` for resolved fonts.
+4. Remove `text_engine/font_metrics.rb` (200+ lines).
 
 ## Acceptance criteria (after pdfrb unblocks)
 
-- [ ] Pipeline registers fonts with externally-provided metrics.
-- [ ] Shaper/LineBreaker call pdfrb's measurement API, not FontMetrics.
-- [ ] `fontisan` remains a dependency for parsing, but no longer feeds
-      Shaper/LineBreaker directly.
+- [ ] Pipeline registers fonts via `pdfrb.fonts.add(path)`.
+- [ ] Shaper/LineBreaker call pdfrb's measurement API, not
+      FontMetrics.
+- [ ] `fontisan` removed from gemspec dependencies.
+- [ ] Spec renders text and verifies line breaks land at correct
+      positions for TTF fonts.
 
 ## Dependencies
 
-- pdfrb 0.4.0's `Fonts#measure_text`/`#glyph_width`/`#metrics_for`
-  with real implementations OR
-- pdfrb accepts the external-metrics proposal linked above.
+- pdfrb `Fonts#glyph_width` uses parsed TTF tables, not AFM-only.
+- pdfrb `Fonts#measure_text` uses parsed TTF tables.

data/TODO.pdf/65-pdfrb-019-integration.md

@@ -1,29 +1,39 @@
-# TODO PDF 65: pdfrb 0.19.0 feature integration
+# TODO PDF 65: pdfrb feature integration
 
 ## Status: PARTIALLY DONE (remainder blocked)
 
 ## What was implemented
 
 1. **Canvas#text_lines**: `TextFrameRenderer` uses `canvas.text_lines`
-   instead of N separate `canvas.text` calls per line. Single batch
-   call for all lines within a run.
+   for batched single-font multi-line text (TODO 27).
+2. **Real PDF gradient shadings**: `RectangleRenderer` uses pdfrb's
+   `Shadings#add_axial` and `Shadings#add_radial` (TODOs 49, 66, 68).
+3. **Canvas#with_transparency**: `Blending.wrap` applies IDML
+   `BlendingSetting` opacity + blend modes (TODO 69).
+4. **Stroke-style setters**: `StrokeStyle.apply` calls
+   `line_cap=`/`line_join=`/`miter_limit=`/`dash_pattern=` from
+   IDML `EndCap`/`EndJoin`/`MiterLimit`/`StrokeDashAndGap` (TODO 70/72).
+5. **Font subsetting**: `Pipeline` calls `Fonts#subset_fonts!` before
+   write (TODO 52).
+6. **Placement module**: shared `Render::Placement.box` (TODO 71).
 
 ## What remains
 
-Blocked by pdfrb 0.4.0 stubs in `Fonts#measure_text` (returns
-`length * 0.5 * size`) and `#glyph_width` (returns 500). Until real
-per-glyph widths land:
+Blocked by pdfrb 0.4.0's TTF measurement still being AFM-only:
 
-2. **`text_rich` for multi-run text**: tracked in TODO 67. Cannot
-   advance between runs without real measurement.
-3. **Replace FontMetrics with pdfrb measurement**: tracked in TODO 63.
-4. **Font subsetting**: tracked in TODO 52. `Fonts#add` accepts `**opts`
-   but `subset:` has no effect.
+7. **`text_rich` for multi-run text** (TODO 67): multi-run batching
+   depends on accurate per-run advance, which `Fonts#measure_text`
+   cannot provide for TTF.
+8. **Replace FontMetrics with pdfrb measurement** (TODO 63): same
+   blocker.
 
 ## Acceptance criteria
 
 - [x] TextFrameRenderer uses `canvas.text_lines`
+- [x] Real PDF gradient shadings via pdfrb
+- [x] Transparency and blend modes
+- [x] Stroke styling
+- [x] Font subsetting
+- [x] Shared Placement module
 - [ ] FontMetrics replaced with pdfrb `glyph_width` (blocked — TODO 63)
-- [ ] FontResolver replaced with `Pdfrb::FontResolver` (blocked — TODO 63)
-- [ ] Font subsetting verified (blocked — TODO 52)
 - [ ] `text_rich` used for multi-run frames (blocked — TODO 67)

data/TODO.pdf/67-text-rich-multi-run.md

@@ -1,6 +1,6 @@
 # TODO PDF 67: Multi-run text batching via Canvas#text_rich
 
-## Status: BLOCKED (pdfrb measure_text is a stub)
+## Status: BLOCKED (pdfrb measure_text is stub for TTF)
 
 ## Goal
 
@@ -12,32 +12,35 @@ all runs inside one BT/ET block, advancing the text matrix between runs.
 
 `text_rich` is the right primitive for multi-run text — one begin/end
 text block per frame instead of N. But its run-advance relies on
-`Pdfrb::Document::Fonts#measure_text`, which in pdfrb 0.4.0 returns a
-stub `length * 0.5 * size`. Without accurate advance, runs overwrite
-each other on the line.
+`Pdfrb::Document::Fonts#measure_text`, which in pdfrb 0.4.0 still
+returns the stub `length * 0.5 * size` for TTF/OTF fonts (only
+Standard 14 AFM fonts get real measurement). Without accurate advance,
+runs overwrite each other on the line.
 
 ## Blocker
 
-`/Users/mulgogi/src/claricle/pdfrb/lib/pdfrb/document/fonts.rb:65`:
+`/Users/mulgogi/src/claricle/pdfrb/lib/pdfrb/document/fonts.rb`:
 
 ```ruby
 def measure_text(text, font:, size:)
-  return 0 unless text
-  # TODO: use font metrics for per-glyph width lookup
-  text.to_s.length * (size || 0).to_f * 0.5
+  return 0 unless text && size
+  metrics = @afm_metrics[font]
+  return text.to_s.length * size.to_f * 0.5 unless metrics  # STUB
+  ...
 end
 ```
 
-`glyph_width` (line 79) is also a stub returning 500, and `metrics_for`
-(line 87) returns nil. These need real Fontisan-backed implementations
-before text_rich produces correct output.
+`@afm_metrics` is only populated for Standard 14 fonts. pdfrb has
+the TTF parsing infrastructure (`Pdfrb::Font::TrueType::File` with
+Cmap, Hmtx) but the integration with `Fonts#glyph_width` /
+`measure_text` is pending.
 
 ## Plan (after pdfrb unblocks)
 
 1. Build `runs` array of `{ text:, font:, size:, color: }` per line.
-2. Replace `simple_render` body with `canvas.text_lines(..., rich: true)`
-   or `canvas.text_rich(runs, at: [x, y])`.
+2. Replace `simple_render` body with `canvas.text_rich(runs, at: [x, y])`.
 3. Verify rendered PDF: runs no longer overwrite, multi-color lines work.
+4. Combine with TODO 63 to drop Fontisan for measurement.
 
 ## Acceptance criteria
 
@@ -47,5 +50,5 @@ before text_rich produces correct output.
 
 ## Dependencies
 
-- pdfrb 0.4.0 `Fonts#measure_text` returns real per-glyph widths.
-- pdfrb 0.4.0 `Fonts#glyph_width` not returning stub 500.
+- pdfrb `Fonts#measure_text` returns real per-glyph widths for TTF
+  fonts (currently stub — see TODO 63).

data/TODO.pdf/70-stroke-styling.md

@@ -1,60 +1,32 @@
 # TODO PDF 70: Stroke styling — line cap, join, dash patterns
 
-## Status: PLANNED (not started)
-
-## Goal
-
-Honor IDML stroke styling attributes by routing through pdfrb 0.4.0's
-stroke-style Canvas setters:
-
-- `line_width=` (already used — TODO 31)
-- `line_cap=` — RoundButt/RoundJoin cap shape
-- `line_join=` — miter/round/bevel corner joining
-- `miter_limit=` — clamps spikes on sharp angles
-- `dash_pattern=` — dashed/dotted stroke styles
-
-## Background
-
-pdfrb 0.4.0 exposes the full PDF graphics-state stroke suite. IDML
-exposes the matching attributes through `<DashedStrokeStyle>` and
-`<DottedStrokeStyle>` elements in `Resources/Graphic.xml`, plus
-per-item `StrokeColor`/`StrokeWeight` (already handled).
-
-IDML → PDF stroke mapping:
-
-| IDML                                | PDF                       |
-|-------------------------------------|---------------------------|
-| `EndCap="RoundCap"`                 | `line_cap = 1` (Round)    |
-| `EndCap="ButtCap"`                  | `line_cap = 0` (Butt)     |
-| `EndCap="ProjectingCap"`            | `line_cap = 2` (Square)   |
-| `Join="RoundJoin"`                  | `line_join = 1`           |
-| `Join="MiterJoin"`                  | `line_join = 0`           |
-| `Join="BevelJoin"`                  | `line_join = 2`           |
-| `<DashedStrokeStyle DashArray="...">` | `dash_pattern = [arr, 0]` |
-| `<DottedStrokeStyle DotArray="...">`  | `dash_pattern = [arr, 0]` (dots are short dashes) |
-
-## Plan
-
-1. `Parts::Graphic` already exposes `dashed_stroke_style` and
-   `dotted_stroke_style` collections — add accessors on the renderer
-   context.
-2. New `Render::StrokeStyle` helper that resolves an
-   `Elements::DashedStrokeStyle` / `DottedStrokeStyle` (or stroke-style
-   self-id reference on the item) to a `{ line_cap:, line_join:,
-   miter_limit:, dash_pattern: }` hash.
-3. Each stroke renderer (Rectangle, Polygon, GraphicLine) calls
-   `StrokeStyle.apply(canvas, item, context)` before `canvas.stroke`.
+## Status: DONE (implemented as TODO 72)
+
+## What was implemented
+
+See [TODO 72](72-stroke-styling.md) for the full implementation
+details. Highlights:
+
+- New `Render::StrokeStyle` module with `strokeable?` predicate and
+  `apply(canvas, item) { ... }` block-wrapper.
+- Element models `Rectangle`, `Polygon`, `GraphicLine` gained
+  `end_cap`, `end_join`, `miter_limit`, `stroke_dash_and_gap`
+  attributes per the RNC schema.
+- Each renderer's stroke path routes through `StrokeStyle.apply`,
+  which maps IDML enums to PDF line cap/join codes and emits
+  `dash_pattern` from the `StrokeDashAndGap` list.
+- 13 dedicated specs in `spec/idml/render/stroke_style_spec.rb`.
 
 ## Acceptance criteria
 
-- [ ] Dashed strokes render with correct dash array.
-- [ ] Dotted strokes render as zero-length dashes with round caps.
-- [ ] Round/miter/bevel joins produce matching PDF output.
-- [ ] Plain strokes (no style) render exactly as before.
-- [ ] Spec covers dashed, dotted, round-join, and pass-through cases.
+- [x] Dashed strokes render with correct dash array.
+- [x] Round/miter/bevel joins produce matching PDF output.
+- [x] Round/butt/projecting caps produce matching PDF output.
+- [x] Plain strokes (no style attrs) render exactly as before.
+- [x] Spec covers dashed, round-join, round-cap, and pass-through cases.
 
 ## Dependencies
 
 - pdfrb 0.4.0 `line_cap=`, `line_join=`, `miter_limit=`, `dash_pattern=`
-  (DONE — all confirmed in pdfrb 0.4.0).
-- IDML element models for stroke styles (DONE — already generated from RNC).
+  (DONE).
+- IDML element models for stroke styles (DONE).

data/TODO.pdf/71-shared-placement-module.md

@@ -0,0 +1,41 @@
+# TODO PDF 71: Shared Placement module (MECE refactor)
+
+## Status: DONE
+
+## What was implemented
+
+Extracted the per-item placement computation (geometric_bounds +
+item_transform + page_height → PDF `{x:, y:, width:, height:}` rect)
+into a single shared module `Idml::Render::Placement`.
+
+Before this refactor:
+
+- `RectangleRenderer.placement_box` was the canonical helper.
+- `PolygonRenderer` and `GraphicLineRenderer` reached across to
+  `RectangleRenderer.placement_box` — a MECE violation (polygon and
+  graphic line depend on rectangle for non-rectangle concerns).
+- `TableRenderer` also reached across to `RectangleRenderer`.
+- `TextFrameRenderer.frame_box` re-implemented the bounds→transform
+  →rect pipeline by hand with a different fallback shape.
+
+After:
+
+- `Idml::Render::Placement.box(item, page_height, fallback: false)`
+  is the single source of truth.
+- All four shape renderers call `Placement.box`.
+- `TextFrameRenderer.frame_box` is now a one-liner that calls
+  `Placement.box(frame, page_height, fallback: true)`.
+
+## Why
+
+Placement is a geometry concern, not a rectangle concern. Hoisting it
+out of `RectangleRenderer` removes the cross-renderer coupling and
+keeps each renderer focused on its own shape's path construction.
+
+## Acceptance criteria
+
+- [x] No renderer references `RectangleRenderer.placement_box`.
+- [x] `Placement.box` returns nil for items without bounds (default)
+      or the fallback rect when `fallback: true`.
+- [x] All render specs pass.
+- [x] `bundle exec rake` green (2350+ examples).

data/TODO.pdf/72-stroke-styling.md

@@ -0,0 +1,55 @@
+# TODO PDF 72: Stroke styling — cap, join, miter, dash
+
+## Status: DONE
+
+## What was implemented
+
+Honors IDML stroke-style attributes by routing through pdfrb 0.4.0's
+stroke-style Canvas setters:
+
+- `line_cap=` — from IDML `EndCap` (ButtEndCap/RoundEndCap/ProjectingEndCap)
+- `line_join=` — from IDML `EndJoin` (MiterEndJoin/RoundEndJoin/BevelEndJoin)
+- `miter_limit=` — from IDML `MiterLimit` (ignored when < 1.0)
+- `dash_pattern=` — from IDML `StrokeDashAndGap` flat list
+
+New module `Idml::Render::StrokeStyle` exposes:
+
+- `StrokeStyle.strokeable?(item)` — predicate (replaces the per-renderer
+  `RectangleRenderer.strokeable?`). Returns a real boolean, not the
+  `nil`/value chain Ruby `&&` would produce.
+- `StrokeStyle.apply(canvas, item) { ... }` — saves graphics state,
+  applies all four style attributes (only those present on the item),
+  yields, restores. Scoping prevents cap/join/miter/dash from leaking
+  into subsequent strokes on the same page.
+
+Element models `Rectangle`, `Polygon`, `GraphicLine` each gained four
+new typed attributes from the RNC schema:
+
+- `end_cap` (`EndCap_EnumValue`)
+- `end_join` (`EndJoin_EnumValue`)
+- `miter_limit` (`xsd:double`)
+- `stroke_dash_and_gap` (`list { xsd:double * }`)
+
+Each renderer's `render_stroke` (or main `render` for GraphicLine)
+calls `StrokeStyle.apply(canvas, item) { ... build path ... stroke }`
+inside the existing `Blending.wrap`.
+
+## IDML → PDF enum map
+
+| IDML                | PDF code |
+|---------------------|----------|
+| ButtEndCap          | 0        |
+| RoundEndCap         | 1        |
+| ProjectingEndCap    | 2        |
+| MiterEndJoin        | 0        |
+| RoundEndJoin        | 1        |
+| BevelEndJoin        | 2        |
+
+## Acceptance criteria
+
+- [x] All four stroke-style setters invoked when their IDML attrs present.
+- [x] Items without stroke-style attrs render exactly as before.
+- [x] `StrokeStyle.apply` saves/restores graphics state.
+- [x] Spec covers all four setters, miter_limit clamp, dash array parse,
+      no-attrs pass-through, and `strokeable?` truth table (13 specs).
+- [x] `bundle exec rake` green.

data/TODO.pdf/73-gradient-resolver-cleanup.md

@@ -0,0 +1,48 @@
+# TODO PDF 73: Drop dead GradientResolver approximation code
+
+## Status: DONE
+
+## What was removed
+
+`Idml::Render::GradientResolver` previously carried a 32-rectangle
+discrete-approximation renderer (`render_gradient`, `interpolate`,
+`blend_between`, `blend`, and the `SEGMENTS = 32` constant). The
+discrete approximation was replaced by real PDF shadings in TODO 66
+(pdfrb 0.4.0 `Shadings#add_axial`) and TODO 68 (`add_radial`); the
+old code became dead.
+
+The class also had a `build(graphic)` factory that built a lookup
+table no consumer reads.
+
+## What was kept
+
+- `GradientResolver.gradient?(name)` — pure predicate (`name` is a
+  `Gradient/*` reference). Still used by `RectangleRenderer#render_fill`
+  to choose between gradient and solid color paths.
+
+The module is now a 4-line pure predicate. Every other responsibility
+moved to `RectangleRenderer`'s direct `Shadings#add_axial` /
+`add_radial` calls.
+
+## Why
+
+Dead code accumulates. Three reasons to drop it:
+
+1. **Confusion** — readers see `SEGMENTS = 32` and assume the discrete
+   approximation is still in use somewhere.
+2. **Maintenance** — every time `ColorResolver` or the gradient model
+   changed, the dead code needed reading to confirm it didn't need
+   updating.
+3. **Test surface** — the dead `render_gradient` had its own spec
+   using `instance_double(Idml::Parts::Graphic)`, violating the
+   no-doubles rule. Removing the method let us delete the offending
+   spec.
+
+## Acceptance criteria
+
+- [x] `GradientResolver` exposes only `.gradient?`.
+- [x] No `render_gradient`, `interpolate`, `blend_between`, `blend`,
+      `SEGMENTS`, or `build` in the module.
+- [x] Renderer code unchanged in behavior (still routes gradient fills
+      to pdfrb Shadings).
+- [x] `bundle exec rake` green.

data/TODO.pdf/74-renderer-spec-coverage.md

@@ -0,0 +1,53 @@
+# TODO PDF 74: Renderer spec coverage
+
+## Status: DONE
+
+## What was implemented
+
+Dedicated spec files for every renderer that previously had none.
+Each spec uses real `Idml::Elements::*` instances (parsed from XML
+fragments) or lightweight `Struct` doubles that mirror the renderer's
+interface — no RSpec `double()` or `instance_double` anywhere.
+
+New spec files:
+
+- `spec/idml/render/polygon_renderer_spec.rb` — 4 examples covering
+  visibility, missing bounds, fill rendering, and stroke + cap.
+- `spec/idml/render/graphic_line_renderer_spec.rb` — 4 examples
+  covering unstrokable skip, color-miss skip, basic line drawing,
+  and StrokeStyle application.
+- `spec/idml/render/group_renderer_spec.rb` — 4 examples covering
+  empty group, child dispatch, item_transform concat, and
+  layer_filter visibility.
+- `spec/idml/render/table_renderer_spec.rb` — 4 examples covering
+  visibility skip, empty rows skip, per-cell rectangle count, and
+  stroke emission.
+- `spec/idml/render/text_frame_renderer_spec.rb` — 3 examples
+  covering no-story skip, non-chain-head skip, and end-to-end BT/ET
+  emission against the sample-with-image fixture.
+- `spec/idml/render/blending_spec.rb` — 11 examples (added earlier
+  in TODO 69).
+- `spec/idml/render/stroke_style_spec.rb` — 13 examples (added
+  earlier in TODO 72).
+- `spec/idml/render/rectangle_renderer_spec.rb` — 6 examples (added
+  earlier in TODO 68).
+
+## Cleanups
+
+- `render_helpers_spec.rb` — removed the `#render_gradient` test that
+  used `instance_double(Idml::Parts::Graphic)` (TODO 73).
+- `rectangle_renderer_spec.rb` — replaced
+  `canvas.document.shadings.instance_variable_get(:@registry)` with
+  the public `shadings.registry` reader.
+- `designmap_spec.rb` — replaced `parsed.send(actual_attr(attr))`
+  with an explicit case/when reader.
+
+## Acceptance criteria
+
+- [x] Every renderer in `lib/idml/render/renderers/` has a dedicated
+      spec file with at least 3 examples.
+- [x] No `instance_double` or `double()` in render specs.
+- [x] No `instance_variable_get` / `instance_variable_set` in render
+      specs.
+- [x] No `.send(...)` in any spec.
+- [x] `bundle exec rake` green (2350+ examples, 0 failures).