fix(ereader): restore cover position when navigating back to the cover

previous_page() set the on-cover flag but left current_position at the
first content block, so "showing the cover" had two different internal
representations depending on how you got there: block 0 on a fresh load,
block 1 after going forward and back.

current_position is what gets persisted, so closing the book while on the
cover reopened it past the cover, silently losing it.

Reset current_position to block 0 when returning to the cover, matching
where a fresh load sits.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

.dockerignore

@@ -0,0 +1,43 @@
+# Dockerfile.ci copies nothing from the context, but keeping it small makes
+# `docker build` fast and avoids shipping local state into the build.
+
+# Virtual environments
+venv/
+.venv/
+
+# Python cache and build output
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+build/
+dist/
+*.egg-info/
+*.egg
+
+# Git
+.git/
+
+# Test/coverage output
+.coverage
+coverage.json
+coverage.xml
+htmlcov/
+cov_info/
+.pytest_cache/
+.tox/
+.mypy_cache/
+
+# IDE
+.idea/
+.vscode/
+*.swp
+*.swo
+.claude/
+
+# Generated docs/images
+docs/images/
+
+# OS files
+.DS_Store
+Thumbs.db

.gitea/workflows/ci.yml

@@ -11,169 +11,167 @@ on:
 
 jobs:
   test:
-    runs-on: self-hosted
+    runs-on: linux/amd64
+    container:
+      # Built from Dockerfile.ci at the repo root. Carries Python 3.10-3.13,
+      # each in its own venv at /opt/py<version> with every dependency
+      # pre-installed, so a run downloads nothing.
+      image: gitea.tourolle.paris/dtourolle/pyweblayout-ci:latest
     strategy:
       matrix:
-        python-version: ['3.10', '3.12', '3.13']
+        python-version: ['3.10', '3.11', '3.12', '3.13']
+      fail-fast: false
+
+    env:
+      # pyWebLayout is a library: it is tested on every interpreter
+      # pyproject.toml's requires-python claims to support.
+      PYBIN: /opt/py${{ matrix.python-version }}/bin
+      # Badges and artifacts are published once, not once per matrix leg -
+      # four jobs racing to force-push the same branch is not a publish
+      # strategy. This leg is the one that publishes.
+      PUBLISH: ${{ matrix.python-version == '3.13' }}
+
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
 
-      - name: Set up Python
-        uses: actions/setup-python@v4
-        with:
-          python-version: ${{ matrix.python-version }}
-          
-      - name: Install dependencies
+      - name: Install project
         run: |
-          python -m pip install --upgrade pip
-          # Install package in development mode
-          pip install -e .
-          # Install test dependencies if they exist
-          if [ -f requirements-dev.txt ]; then pip install -r requirements-dev.txt; fi
-          if [ -f requirements/test.txt ]; then pip install -r requirements/test.txt; fi
-          # Install common test packages
-          pip install pytest pytest-cov flake8 coverage-badge interrogate
+          # --no-deps: dependencies are baked into the image. If a new one is
+          # added to pyproject.toml, add it to Dockerfile.ci and rebuild;
+          # the check below is what catches forgetting to.
+          $PYBIN/pip install -e . --no-deps
+          $PYBIN/python -c "import pyWebLayout; print('pyWebLayout', pyWebLayout.__file__)"
 
-      - name: Download initial failed badges
+      - name: Verify declared dependencies are sufficient
+        if: env.PUBLISH == 'true'
         run: |
-          echo "Downloading initial failed badges..."
-          
-          # Create cov_info directory first
-          mkdir -p cov_info
-          
-          # Download failed badges as defaults
-          curl -o cov_info/coverage.svg "https://img.shields.io/badge/coverage-failed-red.svg"
-          curl -o cov_info/coverage-docs.svg "https://img.shields.io/badge/docs-failed-red.svg"
-          
-          echo "Initial failed badges created:"
-          ls -la cov_info/coverage*.svg
+          # A clean venv with ONLY the declared runtime deps, installed from
+          # the index rather than from the image. If an import here fails,
+          # pyproject.toml is incomplete and a real `pip install pyWebLayout`
+          # fails the same way for a user. This is the one step that is
+          # allowed to reach the network.
+          $PYBIN/python -m venv /tmp/clean-install
+          /tmp/clean-install/bin/pip install --upgrade pip
+          /tmp/clean-install/bin/pip install .
+          /tmp/clean-install/bin/python -c "
+          import pyWebLayout.concrete, pyWebLayout.abstract
+          import pyWebLayout.io.readers.epub_reader
+          import pyWebLayout.io.readers.html_extraction
+          import pyWebLayout.layout.ereader_manager
+          print('clean install imports OK')
+          "
 
       - name: Run tests with pytest
         id: pytest
         continue-on-error: true
         run: |
-          # Run tests with coverage
-          python -m pytest tests/ -v --cov=pyWebLayout --cov-report=term-missing --cov-report=json --cov-report=html --cov-report=xml
+          $PYBIN/python -m pytest tests/ -v \
+            --cov=pyWebLayout \
+            --cov-report=term-missing \
+            --cov-report=json \
+            --cov-report=html \
+            --cov-report=xml
 
       - name: Check documentation coverage
         id: docs
         continue-on-error: true
         run: |
-          # Generate documentation coverage report
-          interrogate -v --ignore-init-method --ignore-init-module --ignore-magic --ignore-private --ignore-property-decorators --ignore-semiprivate --fail-under=80 pyWebLayout/
+          $PYBIN/interrogate -v \
+            --ignore-init-method --ignore-init-module --ignore-magic \
+            --ignore-private --ignore-property-decorators --ignore-semiprivate \
+            --fail-under=80 pyWebLayout/
 
       - name: Lint with flake8
         run: |
           # Stop the build if there are Python syntax errors or undefined names
-          flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
+          $PYBIN/flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
           # Exit-zero treats all errors as warnings
-          flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
+          $PYBIN/flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
 
-      - name: Create coverage info directory
-        if: always()
+      - name: Fail the job if tests failed
+        if: steps.pytest.outcome != 'success'
+        run: |
+          # pytest runs with continue-on-error so the badge steps below still
+          # execute; without this the job would report green on a red suite.
+          echo "::error::pytest failed on Python ${{ matrix.python-version }}"
+          exit 1
+
+      # ------------------------------------------------------------------
+      # Badges and artifacts - publishing leg only
+      # ------------------------------------------------------------------
+
+      - name: Prepare badge directory
+        if: always() && env.PUBLISH == 'true'
         run: |
           mkdir -p cov_info
-          echo "Created cov_info directory for coverage data"
+          # Default to failed badges; the steps below overwrite them on success
+          curl -o cov_info/coverage.svg "https://img.shields.io/badge/coverage-failed-red.svg"
+          curl -o cov_info/coverage-docs.svg "https://img.shields.io/badge/docs-failed-red.svg"
 
       - name: Update test coverage badge on success
-        if: steps.pytest.outcome == 'success' && always()
+        if: always() && env.PUBLISH == 'true' && steps.pytest.outcome == 'success'
         run: |
-          echo "Tests passed! Generating successful coverage badge..."
-          
           if [ -f coverage.json ]; then
-            coverage-badge -o cov_info/coverage.svg -f
-            echo "✅ Test coverage badge updated with actual results"
+            $PYBIN/coverage-badge -o cov_info/coverage.svg -f
+            echo "✅ Test coverage badge updated"
           else
             echo "⚠️ No coverage.json found, keeping failed badge"
           fi
 
       - name: Update docs coverage badge on success
-        if: steps.docs.outcome == 'success' && always()
+        if: always() && env.PUBLISH == 'true' && steps.docs.outcome == 'success'
         run: |
-          echo "Docs check passed! Generating successful docs badge..."
-          
-          # Remove existing badge first to avoid overwrite error  
           rm -f cov_info/coverage-docs.svg
-          interrogate --generate-badge cov_info/coverage-docs.svg pyWebLayout/
-          echo "✅ Docs coverage badge updated with actual results"
+          $PYBIN/interrogate --generate-badge cov_info/coverage-docs.svg pyWebLayout/
+          echo "✅ Docs coverage badge updated"
 
       - name: Generate coverage reports
-        if: steps.pytest.outcome == 'success'
+        if: env.PUBLISH == 'true' && steps.pytest.outcome == 'success'
         run: |
-          # Generate coverage summary for README
-          python -c "
-          import json
-          import os
-          # Read coverage data
+          $PYBIN/python -c "
+          import json, os
           if os.path.exists('coverage.json'):
-              with open('coverage.json', 'r') as f:
-                  coverage_data = json.load(f)
-              total_coverage = round(coverage_data['totals']['percent_covered'], 1)
-              # Create coverage summary file in cov_info directory
+              with open('coverage.json') as f:
+                  data = json.load(f)
+              total = round(data['totals']['percent_covered'], 1)
               with open('cov_info/coverage-summary.txt', 'w') as f:
-                  f.write(f'{total_coverage}%')
-              print(f'Test Coverage: {total_coverage}%')
-              covered_lines = coverage_data['totals']['covered_lines']
-              total_lines = coverage_data['totals']['num_statements']
-              print(f'Lines Covered: {covered_lines}/{total_lines}')
+                  f.write(f'{total}%')
+              print(f\"Test Coverage: {total}%\")
+              print(f\"Lines Covered: {data['totals']['covered_lines']}/{data['totals']['num_statements']}\")
           else:
               print('No coverage data found')
           "
-          
-          # Copy other coverage files to cov_info
           if [ -f coverage.json ]; then cp coverage.json cov_info/; fi
           if [ -f coverage.xml ]; then cp coverage.xml cov_info/; fi
           if [ -d htmlcov ]; then cp -r htmlcov cov_info/; fi
 
       - name: Final badge status
-        if: always()
+        if: always() && env.PUBLISH == 'true'
         run: |
           echo "=== FINAL BADGE STATUS ==="
           echo "Test outcome: ${{ steps.pytest.outcome }}"
           echo "Docs outcome: ${{ steps.docs.outcome }}"
-          
-          if [ -f cov_info/coverage.svg ]; then
-            echo "✅ Test coverage badge: $(ls -lh cov_info/coverage.svg)"
-          else
-            echo "❌ Test coverage badge: MISSING"
-          fi
-          
-          if [ -f cov_info/coverage-docs.svg ]; then
-            echo "✅ Docs coverage badge: $(ls -lh cov_info/coverage-docs.svg)"
-          else
-            echo "❌ Docs coverage badge: MISSING"
-          fi
-          
-          echo "Coverage info directory contents:"
-          ls -la cov_info/ 2>/dev/null || echo "No cov_info directory found"
+          ls -la cov_info/ 2>/dev/null || echo "No cov_info directory"
 
       - name: Upload coverage artifacts
+        if: always() && env.PUBLISH == 'true'
         uses: actions/upload-artifact@v3
         with:
           name: coverage-reports
-          path: |
-            cov_info/
+          path: cov_info/
 
       - name: Commit badges to badges branch
-        if: github.ref == 'refs/heads/master'
+        if: env.PUBLISH == 'true' && github.ref == 'refs/heads/master'
         run: |
           git config --local user.email "action@gitea.local"
           git config --local user.name "Gitea Action"
 
-          # Set the remote URL to use the token
           git remote set-url origin https://${{ secrets.PUSH_TOKEN }}@gitea.tourolle.paris/dtourolle/pyWebLayout.git
 
-          # Create a new orphan branch for badges (this discards any existing badges branch)
+          # Orphan branch holding only the badges, force-pushed each time
           git checkout --orphan badges
-          
-          # Remove all files except cov_info
           find . -maxdepth 1 -not -name '.git' -not -name 'cov_info' -exec rm -rf {} + 2>/dev/null || true
-          
-          # Add only the coverage info directory
           git add -f cov_info/
-          
-          # Always commit (force overwrite)
-          echo "Force updating badges branch with new coverage data..."
           git commit -m "Update coverage badges [skip ci]"
           git push -f origin badges

ARCHITECTURE.md

@@ -1,233 +1,348 @@
-# pyWebLayout Architecture: Abstract vs Concrete
+# pyWebLayout Architecture
 
-This document explains the fundamental architectural separation between **Abstract** and **Concrete** layers in the pyWebLayout library.
+This document describes how pyWebLayout is organised: the layers, what lives in each,
+and the rules that govern how they may depend on one another.
 
 ## Overview
 
-The pyWebLayout library follows a clear separation between two distinct layers:
+The library turns markup (HTML/EPUB) into rendered images. That pipeline is split into
+layers with a strict dependency direction:
 
-- **Abstract Layer**: Represents the logical structure and content of documents (HTML/EPUB text)
-- **Concrete Layer**: Handles the spatial rendering and visual representation of content
+```
+    io/readers/          parse markup into document structure
+         ↓
+    abstract/            what the document says     ─┐
+         ↓                                           ├─ built on core/ + style/
+    layout/              decide where things go      │
+         ↓                                           │
+    concrete/            what the pixels are        ─┘
+         ↓
+    PIL.Image
+```
 
-This separation provides flexibility, testability, and clean separation of concerns.
+The central distinction is **abstract vs concrete**:
 
-## Abstract Layer (`pyWebLayout/abstract/`)
+- **Abstract** — the logical content and structure of a document. A `Paragraph` knows
+  it contains words in an order; it does not know how wide they are.
+- **Concrete** — the spatial realisation of that content. A `Line` knows exactly which
+  glyphs sit at which pixel offsets on a specific canvas.
 
-The Abstract layer deals with the **logical structure** of documents without concerning itself with how content will be visually rendered.
+One abstract document produces many concrete renderings: different page sizes, font
+scales, and font families all re-run the concrete layer over unchanged abstract content.
+That is what makes ereader features like live font scaling and reflow possible.
 
-### Key Components
+## Layers
 
-#### `abstract/block.py`
-- `Block`: Base class for all block-level content
-- `Paragraph`: Represents a logical paragraph containing words
-- `Heading`: Represents headings with semantic levels (H1-H6)
-- `HList`: Represents ordered/unordered lists
-- `Image`: Represents image references
+### `core/` — shared foundations
 
-#### `abstract/inline.py`
-- `Word`: Represents individual words with text content and styling information
-- Contains methods for hyphenation and text manipulation
-- Does **not** handle rendering or spatial layout
+Everything else is built on these. `core/` depends on nothing but `style/`.
 
-#### `abstract/document.py`
-- `Document`: Container for the overall document structure
-- `Chapter`: Logical grouping of blocks (for books/long documents)
+**`core/base.py`** defines the contracts that make a class abstract or concrete:
 
-### Characteristics of Abstract Classes
+| Contract | Kind | Meaning |
+|---|---|---|
+| `Renderable` | ABC | Has `render()`; produces or draws visual output |
+| `Queriable` | ABC | Has `in_object(point)`; can be hit-tested |
+| `Layoutable` | ABC | Has `layout()`; arranges its own contents |
+| `Interactable` | ABC | Holds a callback invoked on interaction |
+| `Geometric` | mixin | `origin` and `size` as numpy arrays |
+| `Hierarchical` | mixin | `parent` back-reference |
+| `Styleable` | mixin | Carries a style object |
+| `FontRegistry` | mixin | Deduplicates `Font` instances across a document |
+| `MetadataContainer` | mixin | Key/value metadata with typed accessors |
+| `BlockContainer` | mixin | Holds child `Block`s |
+| `ContainerAware` | mixin | Knows the container it was added to |
 
-1. **Content-focused**: Store text, structure, and semantic meaning
-2. **Layout-agnostic**: No knowledge of fonts, pixels, or rendering
-3. **Reusable**: Same content can be rendered in different formats/sizes
-4. **Serializable**: Can be saved/loaded without rendering context
+The useful shorthand: **`Renderable` and `Geometric` are the concrete markers.** An
+abstract class that acquires either has crossed the line.
 
-### Example: Abstract Word
+**`core/query.py`** — `QueryResult` and `SelectionRange`: the result types for mapping a
+pixel back to content (what was clicked, what text is selected, where it is in the
+document).
+
+**`core/highlight.py`** — `Highlight`, `HighlightColor`, `HighlightManager`. Highlights
+store both pixel bounds (for drawing) and semantic bounds (word indices, for surviving
+a font change).
+
+**`core/cache.py`** — `UsageCache` / `SizedUsageCache`, bounded caches keyed by
+`(font, string)` for text measurement and glyph rasterisation. Eviction is by usage
+count with periodic aging, not LRU; the module docstring explains why at length. This
+exists because a single page issues thousands of measurements over fewer than a
+thousand distinct pairs, on hardware where an unbounded cache is not affordable.
+
+**`core/callback_registry.py`** — `CallbackRegistry`, which owns the interactive
+elements registered on a page and dispatches to them.
+
+### `style/` — semantic styling and its resolution
+
+This package is itself an instance of the abstract/concrete split, applied to styling:
+
+- **`abstract_style.py`** — `AbstractStyle` (frozen dataclass, hashable) captures
+  *intent*: `FontFamily.SERIF`, `FontSize.LARGE`, `color="black"`. `AbstractStyleRegistry`
+  interns them so a document holds one instance per distinct style.
+- **`concrete_style.py`** — `RenderingContext` (user preferences, DPI, accessibility
+  flags, available space) plus `StyleResolver`, which maps `AbstractStyle` +
+  `RenderingContext` → `ConcreteStyle` (a resolved font path, a pixel size, an RGB
+  tuple). `ConcreteStyleRegistry` caches the results.
+- **`fonts.py`** — `Font`, the loaded PIL font plus its rendering attributes, and the
+  bundled DejaVu families (`BundledFont`).
+- **`page_style.py`** — `PageStyle`: borders, padding, background, default alignment.
+- **`alignment.py`** — the `Alignment` enum.
+
+`AbstractStyle` → `StyleResolver` → `ConcreteStyle` is the mechanism by which the same
+document renders differently for different readers.
+
+### `abstract/` — document content and structure
+
+Layout-agnostic representations of what the document contains.
+
+**`abstract/block.py`** — block-level content, all deriving from `Block`:
+
+`Paragraph`, `Heading` (with `HeadingLevel`), `Quote`, `CodeBlock`, `HList` (with
+`ListStyle`) and `ListItem`, `Table` / `TableRow` / `TableCell`, `Image` and
+`LinkedImage`, `HorizontalRule`, `PageBreak`.
+
+**`abstract/inline.py`** — `Word`, `FormattedSpan`, `LinkedWord`, `LineBreak`. `Word`
+holds text, a style, and previous/next links into the document's word sequence. It can
+report whether it is a hyphenation candidate (`possible_hyphenation(language)`), but it
+does not perform the split — that is a measurement decision and belongs to `Line`.
+
+**`abstract/document.py`** — `Document`, `Chapter`, `Book`, `MetadataType`. `Book` is
+the EPUB-shaped `Document` with chapters and a table of contents.
+
+**`abstract/functional.py`** — `Link`, `Button`, `Form`, `FormField`, `LinkType`,
+`FormFieldType`.
+
+**`abstract/interactive_image.py`** — `InteractiveImage`, an `Image` that is also
+`Interactable` and `Queriable`.
+
+### `concrete/` — spatial realisation
+
+Objects that know their position and size and can draw themselves onto a canvas.
+
+**`concrete/text.py`** — the heart of text layout:
+- `Text` — one renderable fragment (a whole word, or a hyphenated part). Requires an
+  `ImageDraw.Draw` at construction so it can measure itself immediately.
+- `Line` — a sequence of `Text` objects with resolved spacing and a baseline.
+  `Line.add_word()` is where a `Word` becomes one or two `Text` objects: it measures,
+  and if the word overflows it tries dictionary hyphenation, then brute-force splitting,
+  before rejecting the word.
+- `AlignmentHandler` and its subclasses (`LeftAlignmentHandler`,
+  `CenterRightAlignmentHandler`, `JustifyAlignmentHandler`) — the strategy objects that
+  turn a set of measured fragments plus an available width into concrete spacing and a
+  start position.
+- Cache management: `configure_text_caches`, `clear_text_caches`, `text_cache_stats`,
+  `prewarm_text_caches`.
+
+**`concrete/page.py`** — `Page`: a fixed-size canvas holding `Renderable` children, with
+a content rectangle derived from its `PageStyle`, `can_fit_line()` for the layouter to
+test against, `render()` returning a `PIL.Image`, and `query_point()` / `query_range()`
+for hit-testing.
+
+**`concrete/box.py`** — `Box`, the `Geometric` + `Renderable` + `Queriable` base for
+positioned drawable objects.
+
+**`concrete/dynamic_page.py`** — `DynamicPage` (a `Page` subclass) and `SizeConstraints`.
+Adds a two-phase measure-then-layout protocol, so a container such as a table can learn
+its content's intrinsic size before committing to a size allocation.
+
+**`concrete/image.py`** — `RenderableImage`.
+
+**`concrete/table.py`** — `TableRenderer`, `TableRowRenderer`, `TableCellRenderer`,
+`TableStyle`. Cells host their own nested `Page`, which is why `Page` accepts a
+non-zero `origin`.
+
+**`concrete/functional.py`** — `LinkText`, `ButtonText`, `FormFieldText`: `Text`
+subclasses that are also `Interactable`.
+
+**`concrete/interaction_handler.py`** — `InteractionHandler`, `InteractionStateManager`:
+routing taps and presses to registered elements and tracking pressed/released state.
+
+### `layout/` — the abstract → concrete transformation
+
+This is the package the pipeline diagram calls the layout engine.
+
+**`layout/document_layouter.py`** — a set of layouter functions, one per content kind,
+all sharing a signature shape of *(abstract element, target `Page`) → did it fit*:
 
 ```python
-# An Abstract Word knows its text content and semantic properties
-word = Word("supercalifragilisticexpialidocious", font_style)
-word.hyphenate()  # Logical operation - finds break points
-parts = word.get_hyphenated_parts()  # Returns ["super-", "cali-", "fragi-", ...]
+paragraph_layouter(paragraph, page, start_word=0, pretext=None, alignment_override=None)
+    -> (complete: bool, failed_word_index: int | None, remaining_pretext: Text | None)
+
+image_layouter(image, page, max_width=None, max_height=None)   -> bool
+table_layouter(table, page, style=None)                        -> bool
+pagebreak_layouter(page_break, page)                           -> bool
+button_layouter(button, page, font=None, padding=...)          -> (bool, str)
+form_field_layouter(field, page, font=None, ...)               -> ...
+form_layouter(form, page, font=None, field_spacing=10)         -> (bool, list[str])
 ```
 
-## Concrete Layer (`pyWebLayout/concrete/`)
+They **append to a page**, they do not return a list of lines. The three-part return of
+`paragraph_layouter` is what makes pagination resumable: when a paragraph runs off the
+bottom of a page, the caller learns which word failed and whether a hyphenated fragment
+is pending, and can continue on the next page from exactly there.
 
-The Concrete layer handles the **spatial representation** and actual rendering of content.
+`DocumentLayouter` wraps a `Page` and dispatches over a list of abstract elements by
+type, holding the `ConcreteStyleRegistry` for the run.
 
-### Key Components
+**`layout/ereader_layout.py`** — the paginated reading model:
+- `RenderingPosition` — a serialisable cursor expressed in *abstract* coordinates
+  (chapter, block, word, table cell, list item, pending pretext). Because it names
+  document structure rather than pixels, it survives font-size and page-size changes.
+- `BidirectionalLayouter` — renders a page forward or backward from a position,
+  returning `(Page, next_position)`.
+- `ChapterNavigator` / `ChapterInfo` — a table of contents built from heading structure.
+- `FontScaler`, `FontFamilyOverride` — apply scale and family changes to blocks at
+  layout time without mutating the abstract document.
 
-#### `concrete/text.py`
-- `Text`: Renders a specific text fragment with precise positioning
-- `Line`: Manages a line of `Text` objects with spacing and alignment
-- Handles actual pixel measurements, font rendering, and positioning
+**`layout/page_buffer.py`** — `PageBuffer` (LRU caches of rendered pages in both
+directions, plus position maps) and `BufferedPageRenderer` (background rendering).
 
-#### `concrete/page.py`
-- `Page`: Top-level container for rendered content
-- `Container`: Layout manager for organizing renderable objects
-- Handles spatial layout, pagination, and visual composition
+**`layout/ereader_manager.py`** — `EreaderLayoutManager`, the top-level application
+interface (page turns, font changes, chapter jumps, progress), and `BookmarkManager`
+for persisting bookmarks and the last reading position.
 
-#### `concrete/box.py`
-- `Box`: Base class for all spatially-aware renderable objects
-- Provides positioning, sizing, and rendering capabilities
+**`layout/table_optimizer.py`** — column width allocation for tables.
 
-### Characteristics of Concrete Classes
+### `io/readers/` — parsing
 
-1. **Rendering-focused**: Handle pixels, fonts, images, and visual output
-2. **Spatially-aware**: Know exact positions, sizes, and layout constraints
-3. **Implementation-specific**: Tied to specific rendering technologies (PIL, etc.)
-4. **Non-portable**: Rendering results are tied to specific display contexts
+**`html_extraction.py`** — `parse_html_string(html, base_font=None, document=None,
+base_path=None) -> List[Block]`. Built from a `StyleContext` (a `NamedTuple` threaded
+down the tree, carrying the inherited font and styling) and a table of per-tag handlers
+(`paragraph_handler`, `heading_handler`, `table_handler`, …). This is the only place
+that knows about HTML.
 
-### Example: Concrete Text
+**`epub_reader.py`** — `EPUBReader` and `read_epub(path) -> Book`: container/OPF
+parsing, spine and manifest, table of contents, cover handling, and image processing
+(with an e-ink processor available by default).
+
+## Dependency rules
+
+The direction of dependency is what keeps the split honest:
+
+```
+io/readers  →  abstract  →  core, style
+layout      →  abstract, concrete, core, style
+concrete    →  abstract, core, style
+abstract    →  core, style          ← must NOT import concrete
+```
+
+`abstract/` importing from `concrete/` is the violation to watch for. `concrete/`
+importing from `abstract/` is expected and correct: a `Text` may point back at the
+`Word` it came from, and a table cell renderer reads its abstract `TableCell`.
+
+**Where the abstract layer touches rendering today, and why:**
+
+- Abstract classes are constructed with `Font` objects, which carry a pixel size and a
+  loaded font file. This is a deliberate compromise for parsing performance (HTML
+  styling resolves to a `Font` once, at parse time) but it does mean the abstract layer
+  is not fully rendering-independent. `AbstractStyle` is the intended replacement, and
+  `Word` already accepts either.
+- `Word.concrete` / `Word.add_concete()` ([inline.py:42](pyWebLayout/abstract/inline.py#L42),
+  [:134](pyWebLayout/abstract/inline.py#L134)) is a back-reference to the `Text` objects
+  a word became. **It is currently written and never read**, and it cannot correctly
+  model the relationship anyway: one `Word` becomes many `Text`s across re-layouts, and
+  a single slot only remembers the most recent. Treat it as vestigial. When a
+  word→pixels mapping is genuinely needed, build it on `Text`'s `source` back-reference
+  (concrete pointing at abstract), which is the safe direction — note it is currently
+  stored as `_source` with no public accessor, and is itself unread today.
+
+## Worked example
 
 ```python
-# A Concrete Text object handles actual rendering
-text = Text("super-", font)  # Specific text fragment
-text._calculate_dimensions()  # Computes exact pixel size
-image = text.render()  # Produces actual visual output
+from pyWebLayout.io.readers.html_extraction import parse_html_string
+from pyWebLayout.concrete.page import Page
+from pyWebLayout.layout.document_layouter import DocumentLayouter
+
+# 1. Parse: markup -> abstract blocks
+blocks = parse_html_string("<h1>Chapter One</h1><p>It was a dark and stormy night.</p>")
+# [Heading, Paragraph]
+
+# 2. Lay out: abstract blocks -> concrete children appended to a Page
+page = Page(size=(400, 600))
+DocumentLayouter(page).layout_document(blocks)
+
+# 3. Render: Page -> PIL.Image
+image = page.render()
+
+# 4. Query: pixel -> content
+result = page.query_point((50, 40))
+print(result.object_type, result.text)   # e.g. "text" "Chapter"
 ```
 
-## The Transformation Process
-
-The architecture involves a clear transformation from Abstract to Concrete:
-
-```
-Abstract Document
-       ↓
-   [Parser Layer]
-       ↓  
-Abstract Blocks (Paragraph, Heading, etc.)
-       ↓
-   [Layout Engine]
-       ↓
-Concrete Objects (Text, Line, Page)
-       ↓
-   [Rendering Engine]  
-       ↓
-Visual Output (Images, PDF, etc.)
-```
-
-### Example Transformation
+Inspecting the intermediate concrete objects:
 
 ```python
-# 1. Abstract content
-paragraph = Paragraph()
-paragraph.add_word(Word("This", font))
-paragraph.add_word(Word("is", font))
-paragraph.add_word(Word("a", font))
-paragraph.add_word(Word("test", font))
+from pyWebLayout.concrete.text import Line
 
-# 2. Layout transformation
-layout = ParagraphLayout(line_width=200, line_height=20)
-lines = layout.layout_paragraph(paragraph)  # Returns List[Line]
-
-# 3. Each Line contains concrete Text objects
-for line in lines:
-    for text_obj in line.text_objects:  # List[Text]
-        print(f"Text: '{text_obj.text}' at position {text_obj._origin}")
+for line in (c for c in page.children if isinstance(c, Line)):
+    print([t.text for t in line.text_objects])
+# ['Chapter', 'One']
+# ['It', 'was', 'a', 'dark', 'and', 'stormy', 'night.']
 ```
 
-## Key Architectural Principles
+For paginated reading, drive `EreaderLayoutManager` instead of building pages by hand:
 
-### 1. **Single Responsibility**
-- Abstract classes: Handle content and structure
-- Concrete classes: Handle rendering and layout
-
-### 2. **Separation of Concerns**
-- Text parsing/processing ≠ Text rendering
-- Document structure ≠ Page layout
-- Content semantics ≠ Visual presentation
-
-### 3. **Immutable Abstract Content**
-- Abstract content remains unchanged during rendering
-- Multiple concrete representations can be generated from same abstract content
-- Enables pagination, different formats, responsive layouts
-
-### 4. **One-to-Many Relationships**
-- One Abstract Word → Multiple Concrete Text objects (hyphenation)
-- One Abstract Paragraph → Multiple Concrete Lines
-- One Abstract Document → Multiple Concrete Pages
-
-## Common Anti-Patterns to Avoid
-
-### ❌ **Mixing Concerns**
 ```python
-# WRONG: Abstract class knowing about pixels
+from pyWebLayout.layout.ereader_manager import EreaderLayoutManager
+
+manager = EreaderLayoutManager(blocks, page_size=(800, 600), document_id="my-book")
+page = manager.get_current_page()
+page = manager.next_page()
+manager.set_font_scale(1.25)   # re-lays out from the same RenderingPosition
+```
+
+## Design principles
+
+**1. Abstract content is not mutated by layout.** Font scaling and family overrides
+produce scaled copies at layout time rather than editing the document. This is what
+allows the same `blocks` list to back a buffer of pages at several font sizes.
+
+**2. Positions are expressed in abstract coordinates.** `RenderingPosition` names a
+chapter, block, and word — never a pixel. A reader who changes font size stays on the
+same sentence.
+
+**3. Layouters report partial success.** Every layouter returns whether it fit, and the
+text layouter also returns where it stopped. Pagination is built from this rather than
+from a separate page-breaking pass.
+
+**4. One abstract object may become many concrete ones.** One `Word` → one or two `Text`
+fragments; one `Paragraph` → many `Line`s across many `Page`s; one `Document` → an
+unbounded sequence of `Page`s. Any API that assumes a one-to-one mapping will be wrong
+at a hyphen or a page boundary.
+
+**5. Measurement is cached, not avoided.** Text width and glyph rasterisation are the
+hot path. `core/cache.py` bounds their cost; `prewarm_text_caches` front-loads it for a
+known document.
+
+## Anti-patterns
+
+**Concrete state stored on abstract objects.** An abstract object that remembers its
+rendered width, position, or `Text` objects is wrong at the second rendering. If a
+back-reference is needed, point from concrete to abstract.
+
+```python
+# WRONG
 class Word:
     def __init__(self, text):
-        self.text = text
-        self.rendered_width = None  # ❌ Concrete concern in abstract class
+        self.rendered_width = None    # invalidated by any font change
+
+# RIGHT
+text = Text(word.text, font, draw, source=word)   # concrete knows its origin
 ```
 
-### ❌ **renderable_words Concept**
-```python
-# WRONG: Confusing abstract and concrete
-line.renderable_words  # ❌ This suggests Words are renderable
-                      # Words are abstract - only Text objects render
-```
+**Treating `Word` as renderable.** Words are abstract; only `Text` draws. There is no
+`renderable_words` anywhere in the codebase, and there should not be.
 
-### ✅ **Correct Separation**
-```python
-# CORRECT: Clear separation
-abstract_word = Word("test")  # Abstract content
-concrete_text = Text("test", font)  # Concrete rendering
-line.text_objects.append(concrete_text)  # Concrete objects in concrete container
-```
+**Assuming a layouter returns lines.** `paragraph_layouter` appends to a page and
+reports what did not fit. Code that expects `List[Line]` back is working from an
+outdated model.
 
-## Benefits of This Architecture
+## Summary
 
-### 1. **Flexibility**
-- Same content can be rendered at different sizes
-- Multiple output formats from single source
-- Easy to implement responsive design
-
-### 2. **Testability**
-- Abstract logic can be tested without rendering
-- Layout algorithms can be tested independently
-- Visual rendering can be mocked
-
-### 3. **Performance**
-- Abstract content can be cached and reused
-- Layout can be computed once for multiple renderings
-- Incremental updates possible
-
-### 4. **Maintainability**
-- Clear boundaries between text processing and rendering
-- Changes to rendering don't affect content parsing
-- Easy to swap rendering backends
-
-## File Organization
-
-```
-pyWebLayout/
-├── abstract/           # Content and structure
-│   ├── block.py       # Document blocks (Paragraph, Heading, etc.)
-│   ├── inline.py      # Inline content (Word, etc.)
-│   ├── document.py    # Document structure
-│   └── functional.py  # Links, buttons, etc.
-│
-├── concrete/          # Rendering and layout
-│   ├── text.py        # Text and Line rendering
-│   ├── page.py        # Page layout and containers
-│   ├── box.py         # Base rendering classes
-│   ├── image.py       # Image rendering
-│   └── functional.py  # Interactive elements
-│
-├── typesetting/       # Layout algorithms
-│   ├── paragraph_layout.py  # Abstract → Concrete transformation
-│   ├── flow.py        # Text flow management
-│   └── pagination.py  # Page breaking logic
-│
-└── style/             # Styling and formatting
-    ├── fonts.py       # Font management
-    ├── layout.py      # Layout constants
-    └── alignment.py   # Alignment enums
-```
-
-## Conclusion
-
-The Abstract/Concrete separation is fundamental to pyWebLayout's design. It ensures clean separation between content processing and visual rendering, enabling flexible, maintainable, and testable document processing pipelines.
-
-**Remember**: 
-- **Abstract** = What to display (content, structure, semantics)
-- **Concrete** = How to display it (pixels, fonts, positioning, rendering)
-
-This architecture enables the library to handle complex document layouts while maintaining clear, understandable code organization.
+- **`core/`** — contracts and shared machinery
+- **`style/`** — semantic style, and its resolution to concrete rendering parameters
+- **`abstract/`** — what the document says
+- **`concrete/`** — where the pixels go
+- **`layout/`** — the transformation between them, and pagination on top of it
+- **`io/readers/`** — markup in

Dockerfile.ci

@@ -0,0 +1,80 @@
+# CI test image for pyWebLayout
+# Build: docker build -f Dockerfile.ci -t gitea.tourolle.paris/dtourolle/pyweblayout-ci:latest .
+# Push:  docker push gitea.tourolle.paris/dtourolle/pyweblayout-ci:latest
+#
+# pyWebLayout is a library, so CI tests every interpreter pyproject.toml claims
+# to support rather than just one. All four are in this image and the workflow
+# matrix picks one per job; dependencies are pre-installed into each, so a CI
+# run downloads nothing.
+FROM ubuntu:24.04
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+# deadsnakes carries the Python versions Ubuntu 24.04 does not ship
+RUN apt-get update && apt-get install -y --no-install-recommends \
+        ca-certificates \
+        gnupg \
+        software-properties-common \
+    && add-apt-repository -y ppa:deadsnakes/ppa \
+    && apt-get update && apt-get install -y --no-install-recommends \
+        # Interpreters. 3.12 is Ubuntu 24.04's own; the rest come from deadsnakes.
+        python3.10 python3.10-venv \
+        python3.11 python3.11-venv \
+        python3.12 python3.12-venv \
+        python3.13 python3.13-venv \
+        # Pillow needs these at runtime for font rasterisation and image IO
+        libfreetype6 \
+        libjpeg-turbo8 \
+        libopenjp2-7 \
+        libtiff6 \
+        zlib1g \
+        # Used by the workflow itself
+        curl \
+        git \
+        nodejs \
+    && rm -rf /var/lib/apt/lists/*
+
+# One venv per interpreter at a predictable path, /opt/py<version>. Ubuntu marks
+# its system Python externally-managed, so installing into venvs sidesteps that
+# without --break-system-packages, and keeps the four dependency sets isolated.
+#
+# The package list is written once so versions cannot drift between
+# interpreters. It mirrors pyproject.toml's runtime deps plus the test and dev
+# extras; keep the two in step.
+#
+# Deliberately NOT installed: pyWebLayout itself. The workflow installs the
+# checkout with --no-deps, so a job always tests the code under review.
+#
+# setuptools is pinned below 81 because that release dropped pkg_resources,
+# which coverage-badge still imports at startup. Without the pin the badge step
+# dies with ModuleNotFoundError. Revisit when coverage-badge stops using it.
+RUN for v in 3.10 3.11 3.12 3.13; do \
+        python$v -m venv /opt/py$v && \
+        /opt/py$v/bin/pip install --no-cache-dir --upgrade pip wheel && \
+        /opt/py$v/bin/pip install --no-cache-dir --upgrade "setuptools<81" && \
+        /opt/py$v/bin/pip install --no-cache-dir \
+            Pillow \
+            numpy \
+            pyphen \
+            beautifulsoup4 \
+            lxml \
+            pytest \
+            pytest-cov \
+            flask \
+            werkzeug \
+            ebooklib \
+            requests \
+            flake8 \
+            coverage-badge \
+            interrogate \
+        ; \
+    done
+
+# Fail the build rather than ship an image whose dependencies do not import
+RUN for v in 3.10 3.11 3.12 3.13; do \
+        echo "--- python$v ---" && \
+        /opt/py$v/bin/python -c \
+            "import sys, PIL, numpy, pyphen, bs4, pytest, flask, ebooklib, requests; \
+             print(sys.version.split()[0], 'deps OK')" \
+        ; \
+    done

README.md

@@ -231,6 +231,33 @@ current = manager.get_font_family()
 - **[pyWebLayout/layout/README_EREADER_API.md](pyWebLayout/layout/README_EREADER_API.md)** - EbookReader API reference
 - **API Reference** - See docstrings in source code
 
+## Continuous integration
+
+CI runs in a prebuilt container image rather than installing dependencies per
+job. The image carries Python 3.10, 3.11, 3.12 and 3.13, each in its own venv at
+`/opt/py<version>` with every dependency installed, so a run downloads nothing
+and the test matrix covers the whole range `pyproject.toml` claims to support.
+
+Rebuild and push the image whenever `Dockerfile.ci` changes — most often
+because a dependency was added to `pyproject.toml`:
+
+```bash
+docker build -f Dockerfile.ci -t gitea.tourolle.paris/dtourolle/pyweblayout-ci:latest .
+docker push gitea.tourolle.paris/dtourolle/pyweblayout-ci:latest
+```
+
+To reproduce a CI job locally:
+
+```bash
+docker run --rm -v "$PWD:/src:ro" gitea.tourolle.paris/dtourolle/pyweblayout-ci:latest bash -c '
+  mkdir -p /work && cp -a /src/. /work/ && cd /work && rm -rf venv .git
+  /opt/py3.13/bin/pip install -e . --no-deps -q
+  /opt/py3.13/bin/python -m pytest tests/ -q'
+```
+
+The workflow is [.gitea/workflows/ci.yml](.gitea/workflows/ci.yml). Badges and
+coverage artifacts are published from the 3.13 leg only.
+
 ## License
 
 MIT License

docs/ARCHITECTURE_REVIEW.md

@@ -0,0 +1,595 @@
+# Architecture Review
+
+Independent review of the codebase at `c5c61a3` (2026-08-06), answering one
+question: **is this a well-architected library or an over-complex mess?**
+
+It is a well-architected library with one rotten subsystem inside it. The core
+design holds up; roughly a fifth of the code is speculative or non-functional,
+and it is concentrated in the ereader pagination/buffering layer.
+
+This document records the verdict, the evidence, and the findings **not already
+covered** by [LAYOUT_REMEDIATION_SPEC.md](LAYOUT_REMEDIATION_SPEC.md). Where a
+finding is already specced, it is cross-referenced rather than restated.
+
+## Contents
+
+| ID | Finding | Severity | Status |
+|----|---------|----------|--------|
+| [R1](#r1--the-process-pool-crashes-on-python-314) | The process pool crashes on Python 3.14 | Critical | New; raises priority of S12 |
+| [R2](#r2--the-test-suite-hangs-at-interpreter-exit) | Test suite hangs at interpreter exit | High | New; same root cause as R1 |
+| [R3](#r3--font-scaling-destroys-hyperlinks) | Font scaling destroys hyperlinks | High | New |
+| [R4](#r4--three-packaging-configs-that-disagree) | Three packaging configs that disagree | Medium | New (corrected) |
+| [R5](#r5--monkey-patched-page-methods-with-a-conflicting-signature) | Monkey-patched `Page` methods with a conflicting signature | Medium | New |
+| [R6](#r6--dead-duck-typing-cluster-in-pagepy) | Dead duck-typing cluster in `page.py` | Low | New; extends S10.3 |
+| [R7](#r7--two-orphaned-subsystems) | Two orphaned subsystems | Low | New |
+| [R8](#r8--backward-pagination-is-guesswork) | Backward pagination is guesswork | Medium | Resolved by S16 |
+| [R9](#r9--query_points-hit-region-is-offset-from-the-glyphs) | `query_point`'s hit region is offset from the glyphs | Medium | New, open |
+
+---
+
+## Verdict
+
+**Well architected.** The concerns that usually decide this question are all on
+the right side of the line:
+
+- **The abstract/concrete split is real, not aspirational.** Verified
+  empirically: `abstract/` never imports `concrete/`; `core/` imports nothing but
+  itself. The single crossing ([document.py:7](../pyWebLayout/abstract/document.py#L7))
+  is into `style/`, which the dependency rules permit. Most codebases claiming
+  this layering have leaked it within a year.
+- **The layouter contract is the right abstraction.** `paragraph_layouter`
+  returning `(fit, failed_word_index, remaining_pretext)` is what makes
+  pagination resumable, and the shape is consistent across content types. Layout
+  engines that return `List[Line]` cannot paginate without a second pass.
+- **[core/cache.py](../pyWebLayout/core/cache.py) is exemplary.** Usage-ranked
+  eviction with periodic aging, sampled eviction instead of a maintained heap,
+  O(1) hit path with no reordering — every choice justified by a measurement in
+  the docstring.
+- **[concrete/text.py](../pyWebLayout/concrete/text.py) is the strongest file.**
+  The glyph cache ([:581-647](../pyWebLayout/concrete/text.py#L581-L647))
+  reimplements PIL's internals to skip per-call setup, with a permanent graceful
+  fallback when the private API is absent. Alignment is a clean strategy pattern,
+  and `render_alignment_handler` handles last-line-of-paragraph correctly.
+- **[html_extraction.py](../pyWebLayout/io/readers/html_extraction.py) is
+  textbook.** An immutable `StyleContext` threaded down the tree plus a handler
+  dispatch table — no giant if/elif, no mutable parser state.
+
+**The complexity that is not earned** is concentrated in three places, all in the
+same band of the code:
+
+1. `layout/page_buffer.py` — 520 lines of multiprocess prefetch that has never
+   worked (S12, plus R1/R2 below).
+2. `BidirectionalLayouter.render_page_backward` — ~100 lines of convergence
+   heuristics standing in for an anchor list that already exists (R8).
+3. The second block dispatcher in `ereader_layout.py`, which silently drops
+   tables and lists (S4, S8).
+
+The pattern is visible in the git history: work from `e000068` onward (caching,
+alignment, page geometry) is markedly better than the ereader scaffolding it sits
+on. This is not a mess. It is a solid library with an early prototype still
+embedded in it.
+
+**Sizing the cleanup:** R1–R7 plus S12 and S10.3 remove roughly 800–900 lines and
+fix four user-visible defects. None of it requires redesigning anything.
+
+## Test baseline
+
+At `c5c61a3`, in a clean venv on Python 3.14.6:
+
+```
+833 passed, 2 skipped, 24 subtests passed in 11.51s
+```
+
+(The 2 skips were environmental — the review venv lacked `requests`, so the URL
+image tests skipped. With the `test` extra from R4 installed the suite reports
+`853 passed, 24 subtests passed in 13.53s`.)
+
+The suite then **hangs indefinitely** rather than exiting. See R2.
+
+## Status of the existing remediation spec
+
+| Spec | Subject | State |
+|------|---------|-------|
+| S1 | Inline content in non-paragraph containers | Done (`284d521`) |
+| S2 | Page geometry: origin and content rect | Done (`f18cec2`) |
+| S3 | Draw/canvas lifecycle | Done (`202dacf`) |
+| S11 | Partial-block progress discarded | Done (`a57da80`) |
+| S13 | Word spacing and alignment | Done (`1262be6`) |
+| S14 | Vertical centring in buttons and fields | Done (`c5c61a3`) |
+| S4–S10, S12 | Dispatch, cells, table grid, pagination, hygiene, background rendering | Outstanding |
+
+The spec's analysis is sound and in places sharper than this review — S12 caught
+that `_render_page_worker` omits `page_size` entirely, which this review missed.
+Nothing below supersedes it.
+
+---
+
+## R1 — The process pool crashes on Python 3.14
+
+**Severity: critical. Raises S12 from "useless" to "fatal".**
+
+### Problem
+
+`PageBuffer` submits to a `ProcessPoolExecutor` from inside
+`BufferedPageRenderer.render_page`
+([page_buffer.py:431](../pyWebLayout/layout/page_buffer.py#L431)). Python 3.14
+changed the default multiprocessing start method on Linux from `fork` to
+`forkserver`. Under a non-`fork` start method, `submit()` reaches
+`_check_not_importing_main()`, which raises unless the caller sits inside an
+`if __name__ == "__main__":` guard — and the child re-imports the caller's main
+module, re-executing it.
+
+S12 documents this subsystem as delivering no benefit. On 3.14 it is worse than
+that: `EreaderLayoutManager.get_current_page()` **raises** when called from
+module-level script code.
+
+### Evidence
+
+A plain script calling `manager.get_current_page()` at module level, Python
+3.14.6:
+
+```
+RuntimeError:
+        An attempt has been made to start a new process before the
+        current process has finished its bootstrapping phase.
+  ...
+  File "pyWebLayout/layout/page_buffer.py", line 221, in _queue_forward_renders
+    future = self.executor.submit(_render_page_worker, args)
+ConnectionResetError: [Errno 104] Connection reset by peer
+```
+
+With a `__main__` guard added it does not raise, and instead confirms S12's
+finding on every job:
+
+```
+Background render failed for position RenderingPosition(...): cannot pickle 'ImagingCore' object
+Background render failed for position RenderingPosition(...): cannot pickle 'ImagingCore' object
+```
+
+### Action
+
+Fold into **S12**, and treat S12 as unblocked and urgent rather than phase 4. The
+recommended resolution there — delete the pool, keep the LRU buffers and position
+maps, replace prefetch with synchronous readahead — resolves R1 and R2 as a side
+effect. S12's measurement gate still applies to the *readahead* decision; it does
+not need to gate deletion of the pool, because the pool's contribution is
+provably zero.
+
+### Files
+
+`pyWebLayout/layout/page_buffer.py`
+
+---
+
+## R2 — The test suite hangs at interpreter exit
+
+**Severity: high.** Same root cause as R1.
+
+### Problem
+
+`PageBuffer.__del__` calls `shutdown()`, which calls
+`executor.shutdown(wait=True)` ([page_buffer.py:342](../pyWebLayout/layout/page_buffer.py#L342)).
+`EreaderLayoutManager.__del__` does the same via `renderer.shutdown()`. Running
+`__del__` at interpreter shutdown and blocking on a process pool inside it
+deadlocks.
+
+### Evidence
+
+```
+833 passed, 2 skipped, 24 subtests passed in 11.51s
+```
+
+...then the process sat at ~0% CPU with idle forkserver children for 13 minutes
+before being killed. Reproduced twice; both runs completed the tests in under
+12s and neither exited.
+
+This is why CI wall-clock does not resemble the 11.5s the tests actually take.
+
+### Action
+
+Resolved by S12's deletion of the executor. If for any reason the pool is
+retained, `__del__` must not block: register an `atexit` handler or require
+explicit `shutdown()`, and never `wait=True` from a finaliser.
+
+### Acceptance criteria
+
+- `pytest` returns to the shell within a second of printing its summary line.
+- No `multiprocessing` child processes outlive the test session.
+
+### Files
+
+`pyWebLayout/layout/page_buffer.py`, `pyWebLayout/layout/ereader_manager.py`
+
+---
+
+## R3 — Font scaling destroys hyperlinks
+
+**Severity: high. User-visible, silent, and trivially reproducible.**
+
+### Problem
+
+`BidirectionalLayouter._scale_block_fonts`
+([ereader_layout.py:474-498](../pyWebLayout/layout/ereader_layout.py#L474-L498))
+rebuilds a scaled block by constructing plain `Word(word.text, scaled_style)` for
+every word. `LinkedWord` is a `Word` subclass
+([inline.py:288](../pyWebLayout/abstract/inline.py#L288)), so the reconstruction
+downgrades it and the link target is discarded.
+
+The function returns the block unchanged when `font_scale == 1.0` and no family
+override is set, which is why no test has caught this: the defect only appears
+once the reader changes font size.
+
+Two further gaps in the same function:
+
+1. It handles only `Paragraph` and `Heading`. Every other block type is returned
+   unscaled, so a font-size change leaves images, tables and lists at their
+   original size while the text around them reflows.
+2. It allocates a new `Paragraph` and a new `Word` per word on **every page
+   render** at any scale ≠ 1.0 — directly against the caching work in
+   `concrete/text.py`, and on the hot path.
+
+### Evidence
+
+At `c5c61a3`, parsing `<p>Go to <a href="http://x">this link</a> now.</p>`:
+
+```
+scale=1.0: LinkedWords = 2
+scale=1.5: LinkedWords = 0
+```
+
+### Design
+
+Stop reconstructing abstract blocks at layout time. Font scale and family are
+*rendering context*, not document content — carrying them in a copied document
+violates the "abstract content is not mutated by layout" principle in
+[ARCHITECTURE.md](../ARCHITECTURE.md) in spirit, even though it copies rather
+than mutates.
+
+Preferred: thread the scale/override into the layouter and resolve fonts at
+`Text` construction, where `Font` objects are already deduplicated by
+`FontRegistry`. `paragraph_layouter` already accepts an `alignment_override`;
+`font_scale` and `font_family_override` belong in the same place.
+
+Minimum viable fix if the larger change is deferred: reconstruct via
+`type(word)` and copy subclass state, and extend coverage to every block type.
+This is strictly a stopgap — it keeps the per-page allocation cost.
+
+### Acceptance criteria
+
+- A document containing `<a href>` retains every `LinkedWord` after
+  `set_font_scale(1.5)`, and `query_point` over the rendered page still returns
+  `object_type="link"` with the correct target.
+- An image block's rendered size is unaffected by `set_font_scale`, or scales
+  deliberately — not left inconsistent with the text around it.
+- No new `Word`/`Paragraph` objects are allocated per page render at scale ≠ 1.0.
+
+### Files
+
+`pyWebLayout/layout/ereader_layout.py`, `pyWebLayout/layout/document_layouter.py`
+
+---
+
+## R4 — Three packaging configs that disagree
+
+**Severity: medium.** *Corrected: the original review claimed a clean install
+fails on first import. It does not — see below.*
+
+### Problem
+
+The project carries **three** sets of packaging metadata:
+
+| File | Declares |
+|------|----------|
+| `pyproject.toml` `[project]` | Pillow, numpy, pyphen, beautifulsoup4, flask, ebooklib, requests |
+| `setup.cfg` `[options]` | Pillow, numpy |
+| `setup.py` `setup(...)` kwargs | Pillow, numpy |
+
+`pyproject.toml`'s `[project]` table wins under any modern build backend, so the
+shipped wheel is correct and `pip install pyWebLayout` works. The `setup.cfg` and
+`setup.py` copies are dead, contradictory, and actively misleading — reading
+either one gives the wrong answer about what the library needs.
+
+The authoritative list is itself wrong in the other direction:
+
+- **`flask` is a runtime dependency.** It is imported only by
+  `tests/abstract/test_abstract_blocks.py`, as a fixture HTTP server. Every user
+  installs Flask, Jinja2, Werkzeug, click, itsdangerous and blinker for nothing.
+- **`ebooklib` is a runtime dependency and is never imported by the library.**
+  `epub_reader.py` uses `zipfile` + `xml.etree` directly. Only the *tests* use
+  ebooklib, to build EPUB fixtures.
+- **`requests` is declared required but is optional.** `concrete/image.py:100-111`
+  imports it lazily and degrades to an error message on the image when absent.
+- **`requires-python = ">=3.6"` is false.** The package uses dataclasses (3.7+)
+  and `from __future__ import annotations` (3.7+); CI tests 3.10, 3.12 and 3.13.
+
+Net effect: a runtime install pulls 7 direct dependencies where 4 are needed.
+
+### Action
+
+- Consolidate on `pyproject.toml`. Reduce `setup.cfg` to its `[flake8]` section
+  and `setup.py` to a `setup()` shim, each with a comment saying where metadata
+  lives.
+- Runtime deps: Pillow, numpy, pyphen, beautifulsoup4. Move flask, werkzeug,
+  ebooklib and requests into a `test` extra; add a `remote-images` extra for
+  requests; add a `dev` extra composing them.
+- Set `requires-python = ">=3.10"` to match the CI matrix, and add version
+  classifiers.
+- Add a CI step that installs the package into an empty venv with **only**
+  declared runtime deps and imports every top-level subpackage. This class of
+  defect is only caught by installing what you ship — and it is what would have
+  caught the original misreading.
+
+### Files
+
+`pyproject.toml`, `setup.cfg`, `setup.py`, `.gitea/workflows/ci.yml`
+
+---
+
+## R5 — Monkey-patched `Page` methods with a conflicting signature
+
+**Severity: medium.** Currently inert; a landmine if `Page` is ever refactored.
+
+### Problem
+
+[ereader_layout.py:741-761](../pyWebLayout/layout/ereader_layout.py#L741-L761)
+defines `_add_page_methods()` and calls it at import time. It attaches
+`can_fit_line` and `available_width` to the `Page` class if they are absent.
+
+`Page` defines both ([page.py:59](../pyWebLayout/concrete/page.py#L59),
+[page.py:147](../pyWebLayout/concrete/page.py#L147)), so the patch never fires.
+But the two definitions of `can_fit_line` **do not agree**:
+
+| Source | Signature |
+|--------|-----------|
+| `Page` | `can_fit_line(baseline_spacing, ascent=0, descent=0)` |
+| monkey patch | `can_fit_line(line_height)` |
+
+The patched version also ignores descenders entirely — the exact bug S2 fixed. If
+`Page.can_fit_line` were ever renamed or moved, this would silently reinstate
+pre-S2 clipping behaviour, from an import side effect in a different package.
+
+### Action
+
+Delete `_add_page_methods` and its call site. Import-time monkey-patching of a
+class in another module has no place here; if `Page` is missing something the
+layout engine needs, it belongs on `Page`.
+
+### Files
+
+`pyWebLayout/layout/ereader_layout.py`
+
+---
+
+## R6 — Dead duck-typing cluster in `page.py`
+
+**Severity: low.** Extends S10.3.
+
+### Problem
+
+[page.py:261-491](../pyWebLayout/concrete/page.py#L261-L491) contains a closed
+cluster with no external callers:
+
+- `_get_child_property` (:261) — called only by the four below
+- `_get_child_height` (:301) — called by nothing
+- `_get_child_position` (:382) — called only by `_point_in_child`
+- `_point_in_child` (:435) — called by nothing
+- `_get_child_size` (:466) — called only by `_point_in_child`
+
+Verified by grep across `pyWebLayout/`, `tests/`, `examples/` and `scripts/`:
+zero references outside the cluster. About 90 lines.
+
+It exists because `Renderable` declares neither `size` nor `origin`, so the code
+probes `_size`, `size`, `_height`, `height`, `_origin` and `position` in turn with
+`hasattr`. `query_point` (:399) already does the right thing instead — it relies
+on the `Queriable` interface.
+
+### Action
+
+- Delete all five methods.
+- Add `origin` and `size` to the `Renderable`/`Geometric` contract in
+  `core/base.py` so the duck-typing cannot grow back. This is the same concern as
+  S10.1's render contract and can ship with it.
+
+### Files
+
+`pyWebLayout/concrete/page.py`, `pyWebLayout/core/base.py`
+
+---
+
+## R7 — Two orphaned subsystems
+
+**Severity: low**, but they are a large share of the "is this over-complex?"
+impression: 559 lines that nothing in the library reaches.
+
+### Problem
+
+**`concrete/interaction_handler.py` (310 lines).** `InteractionHandler` and
+`InteractionStateManager` are referenced only by
+`examples/07_pressed_state_demo.py`. No library code, no ereader path, no tests.
+
+**`core/highlight.py` (249 lines).** `Highlight`, `HighlightColor` and
+`HighlightManager` have tests (`tests/core/test_highlight.py`) but are not wired
+into `EreaderLayoutManager` at all. Highlighting is not reachable through the
+library's own top-level interface.
+
+`HighlightManager` also duplicates `BookmarkManager`'s JSON persistence
+(directory, `_save`, `_load`, per-document file naming) with no shared base.
+
+### Action
+
+Decide per subsystem, and record the decision:
+
+- **Wire it up** — `EreaderLayoutManager` grows `add_highlight` / `highlights_for_page`
+  and the persistence merges with `BookmarkManager` into one document-state store.
+- **Or move it out** — relocate to `examples/` or delete, and drop the tests with it.
+
+Either is fine. Leaving a tested, documented, unreachable subsystem in `core/` is
+what makes the library look larger and less coherent than it is.
+
+### Files
+
+`pyWebLayout/concrete/interaction_handler.py`, `pyWebLayout/core/highlight.py`,
+`pyWebLayout/layout/ereader_manager.py`
+
+---
+
+## R8 — Backward pagination is guesswork
+
+**Severity: medium.** S11's closing note already flags this for audit; this
+records what the audit found.
+
+### Problem
+
+`render_page_backward`
+([ereader_layout.py:372-472](../pyWebLayout/layout/ereader_layout.py#L372-L472))
+finds the previous page by estimating a start position, rendering forward,
+comparing the end against the target, and adjusting — **up to 10 times**. It then
+has a fallback that jumps back up to 5 blocks and renders again, and a fallback
+for *that* which renders from the start of the document.
+
+Worst case: one "previous page" tap costs up to 12 full page layouts.
+
+The estimator it converges from is `max(1, int(10 / font_scale))` blocks
+([:684](../pyWebLayout/layout/ereader_layout.py#L684)) — a constant with no
+relationship to page size, block length or font metrics.
+
+The correct answer is usually already known.
+`EreaderLayoutManager._page_history` ([ereader_manager.py:210](../pyWebLayout/layout/ereader_manager.py#L210))
+records real page-start positions and serves them instantly; the refinement loop
+only runs when history misses — after a jump, a bookmark, a font change, or
+beyond 50 entries.
+
+S11's note asks whether these fallbacks were compensating for the
+discarded-progress bug it fixed. They were, in part: the "failed to move
+backward" branch at [:446](../pyWebLayout/layout/ereader_layout.py#L446) is
+reachable precisely when forward rendering fails to advance, which S11 addressed.
+
+### Design
+
+Replace convergence with anchors. Maintain a sorted list of known page-start
+positions — chapter starts from `ChapterNavigator` (free, already built) plus
+every position visited. To go back from position P: binary-search the largest
+anchor A < P, render forward from A collecting page starts until reaching P, and
+return the last one. Cost is bounded by the anchor spacing, and every page start
+discovered on the way is itself a new anchor, so the second traversal of any
+region is free.
+
+This subsumes `_page_history`, so the two mechanisms become one.
+
+**Sequencing:** do this after S12, and after S8 — table and list pagination
+changes what a page start can be, and re-deriving anchors is cheap only once
+positions round-trip through tables correctly (S8 already notes this dependency).
+
+### Acceptance criteria
+
+- `previous_page()` from any position issues at most *k* page layouts, where *k*
+  is the anchor spacing, with no iteration count and no fallback ladder.
+- Forward-then-backward round-trips exactly, from a cold cache, after a chapter
+  jump, and after a bookmark restore.
+- `_estimate_page_start`, `_adjust_start_estimate` and the three-tier fallback
+  are deleted, not retained alongside.
+
+### Files
+
+`pyWebLayout/layout/ereader_layout.py`, `pyWebLayout/layout/ereader_manager.py`
+
+---
+
+## Status
+
+All findings in this document are resolved. What remains is the existing
+remediation spec: **S4 → S5 → S6 → S7 → S8 → S9**, plus **S10.1**, unchanged.
+
+| ID | Resolution | Commit |
+|----|-----------|--------|
+| R1 | Fixed with S12 — the pool that raised is gone | `1924cc2` |
+| R2 | Fixed with S12 — no executor, no blocking finaliser | `1924cc2` |
+| R3 | `Word.with_style` keeps subclasses; all container blocks scale | `f0dc675` |
+| R4 | Consolidated on `pyproject.toml`; 7 runtime deps → 4 | `767e4c1` |
+| R5 | Monkey patch deleted | `62ca151` |
+| R6 | 138 dead lines deleted; contract hardening deferred to S10.1 | `e81ba48` |
+| R7 | Both subsystems wired into `EreaderLayoutManager` | `8746d3f`, `0ce1aea` |
+| R8 | Superseded by S16 (anchor replay); dead estimators removed | `bcae45a` |
+| R9 | Open — see below | — |
+
+Two things worth carrying forward:
+
+- **S12's measurement stands as the argument against prefetch.** A page render
+  is 9–56 ms. Any future proposal to render ahead should have to beat that
+  number first.
+- **Wiring an orphan found a bug.** R7's interaction handler had a crash on
+  every hovered or pressed link (`0ce1aea`). Unreachable code is not
+  neutral — it is untested code that looks tested.
+
+---
+
+## R9 — query_point's hit region is offset from the glyphs
+
+**Severity: medium.** Found while verifying R3; not part of the original review.
+
+### Problem
+
+The region `Page.query_point` reports for a text object does not line up with
+where that object says it is. Probing a `LinkText` at the centre of its own
+`origin`/`size` box returns `object_type="empty"`.
+
+### Evidence
+
+A single-link page at 400×600, default scale:
+
+```
+'this' origin=(68.3, 35.0) size=(29.2, 19.0)  centre=(82, 44) -> empty
+'link' origin=(102.5, 35.0) size=(28.3, 19.0)  centre=(116, 44) -> empty
+
+grid scan: link is detected across y≈20–39
+LinkText claims:                        y≈35–54
+```
+
+The two bands overlap by about four pixels. The offset is close to the font
+ascent, which points at a baseline-versus-top mismatch between the coordinates
+`Text` stores and the ones `in_object` tests.
+
+This reproduces identically at scale 1.0 and 1.5, so it predates the R3 fix.
+
+### Why it matters
+
+Taps land through the grid because the region is only shifted, not absent — but
+it is shifted by most of a line height. Near the top or bottom of a page, or
+between tightly spaced lines, a tap can hit the neighbouring line instead of the
+one under the finger. It also makes `LinkText.origin`/`size` unusable for
+drawing selection or highlight overlays, which is what R7's highlighting now
+depends on.
+
+### Action
+
+Establish which of the two is authoritative — almost certainly the drawn
+position — and make the other agree. This sits close to S2 (page geometry) and
+S3 (draw/canvas lifecycle), both already landed, so the conventions to match
+are in place.
+
+### Acceptance criteria
+
+- `page.query_point(centre_of(obj))` returns `obj` for every text object on a
+  rendered page, at scales 0.8, 1.0, 1.5 and 2.0.
+- The end-to-end test in `tests/layout/test_font_scaling.py` probes the centre
+  directly instead of scanning a grid.
+
+### Files
+
+`pyWebLayout/concrete/text.py`, `pyWebLayout/concrete/page.py`,
+`pyWebLayout/core/base.py`
+
+## Reproducing the findings
+
+Reviewed at `c5c61a3` on Python 3.14.6, in a venv containing
+`pytest pyphen Pillow numpy beautifulsoup4 lxml ebooklib`.
+
+- **R1**: call `EreaderLayoutManager(...).get_current_page()` from module-level
+  script code (no `__main__` guard).
+- **R2**: `python -m pytest -q`; observe the summary line, then the hang.
+- **R3**: parse HTML containing `<a href>`, call
+  `BidirectionalLayouter._scale_block_fonts(block, 1.5)`, count `LinkedWord`
+  instances in the result.
+- **R5**: compare `inspect.signature(Page.can_fit_line)` against the patch body.
+- **R6**: grep the five method names across `pyWebLayout/ tests/ examples/ scripts/`.
+- **R8**: read the loop; no execution needed.

docs/LAYOUT_REMEDIATION_SPEC.md

@@ -28,6 +28,9 @@ It is independent of every other spec here.
 | [S11](#s11--partial-block-progress-is-discarded) | Partial-block progress is discarded | 0 |
 | [S12](#s12--background-rendering) | Background rendering | 4 |
 | [S13](#s13--word-spacing-and-alignment) | Word spacing and alignment | 0 |
+| [S14](#s14--vertical-centring-in-buttons-and-fields) | Vertical centring in buttons and fields | 0 |
+| [S15](#s15--form-field-label-geometry) | Form field label geometry | 0 |
+| [S16](#s16--backward-page-navigation) | Backward page navigation | 0 |
 
 ## Design invariants
 
@@ -560,6 +563,22 @@ Four separate geometry defects:
    drops every cell past `len(column_widths)` — two of three cells never render.
 4. **rowspan is parsed and stored but never read** by any renderer or measurer;
    spanned rows just shift left.
+5. **Row height ignores the cell padding it must contain.** The 40px minimum in
+   `_calculate_row_height_for_section` is a constant, so a larger `cell_padding`
+   eats into the content box rather than growing the row, and
+   `_render_cell_content` then clips the text against `available_height`.
+   Rendering the same header at two paddings:
+
+   ```
+   padding=(8,10,8,10)   border=1: header h=40, ink=593
+   padding=(10,12,10,12) border=2: header h=40, ink=288
+   ```
+
+   Both rows are 40px tall; the second silently loses half its text. This is
+   visible in `docs/images/example_05_html_table_with_images.png`, whose second
+   table renders an empty header row. It is the same measure/render disagreement
+   as defect 1, and S5 removes it by construction: the cell page's content box
+   *is* the box its padding leaves.
 
 ### Design
 
@@ -999,6 +1018,20 @@ a full page laid out in the worker — then the result is thrown away by
 4-core Pi with 512MB this is actively harmful: four interpreter copies plus four
 copies of the book, to populate a cache that never populates.
 
+It is also not inert. The pool is started from `PageBuffer.initialize` inside a
+process that already has threads, and CPython warns about exactly this:
+
+```
+DeprecationWarning: This process (pid=...) is multi-threaded,
+use of fork() may lead to deadlocks in the child.
+```
+
+`tests/layout/test_ereader_image_rendering.py` intermittently hangs at
+interpreter exit as a result — every test reports PASSED, then the process never
+returns. Observed roughly one run in four. A reader that hangs on shutdown once
+in four launches would be a shipped bug; the test suite is just where it shows
+up first. This raises S12 from "wasted work" to "actively harmful".
+
 Four further defects in the same file, which matter only if the decision is to
 keep it:
 
@@ -1142,6 +1175,194 @@ Three defects, all visible as a right edge that wobbles from line to line.
 
 ---
 
+## S14 — Vertical centring in buttons and fields
+
+### Problem
+
+`ButtonText.render` and `FormFieldText.render` both placed the text baseline at
+`box_top + box_height / 2 + descent / 2`. Centring glyphs whose visual height is
+`ascent + descent` inside a box of height `H` puts the baseline at
+`box_top + H/2 + (ascent - descent)/2`. The two agree only when
+`ascent == 2 * descent`; DejaVu is nearer 4:1, so labels rode high against the
+top edge of the control.
+
+`ButtonText` also sized itself as `font_size + padding`, but the text's visual
+height exceeds the nominal size — DejaVu at 14px measures 17 — so the button was
+too short to centre its own label in.
+
+### Evidence
+
+A 14px "Save Document" button with 6px vertical padding, measuring the label's
+ink against the button rectangle:
+
+```
+gap above text:  5px
+gap below text: 11px
+```
+
+### Design
+
+- `baseline = area_top + (area_height - (ascent + descent)) / 2 + ascent` in both
+  renderers.
+- `ButtonText._padded_height` derives from `ascent + descent`, guarded so a mock
+  or unusual font object falls back to the nominal size.
+
+### Acceptance criteria
+
+- Label ink is centred within ±2px at font sizes 10, 14 and 20.
+- Label ink stays inside the button rectangle.
+- Button height is at least `ascent + descent + vertical padding`.
+- A form field's value is centred within its input box (±3px).
+
+### Files
+
+`pyWebLayout/concrete/functional.py`
+
+### Note
+
+`docs/images/example_07_pressed_state.png` was stale — no example regenerates it;
+`07_pressed_state_demo.py` writes `demo_07_pressed.png` at the repository root
+and the docs copy had been placed by hand. It has been refreshed. Worth wiring
+the demo to write straight to `docs/images/` so it cannot drift again.
+
+---
+
+## S15 — Form field label geometry
+
+### Problem
+
+`FormFieldText` treats its origin as the control's top-left: `size` and
+`in_object` both measure down from it. But it drew the label by calling
+`Text.render` at that origin, and Text anchors on the **baseline**, so the
+label's glyphs landed *above* the origin — outside the box the control claims,
+on top of whatever was there. In a stacked form that is the previous field's
+input box, which is what
+`docs/images/example_10_forms.png` showed: every label but the first crowding
+and touching the box above it.
+
+The height was also computed as `font_size + 5 + field_height`, understating the
+label by the difference between nominal size and ink height, which left the gap
+between label and box smaller than the intended 5px.
+
+### Design
+
+- The origin is documented as the top-left of the whole control.
+- Rendering offsets the label down by its ascent, so the glyphs occupy
+  `[origin.y, origin.y + ascent + descent]`.
+- `LABEL_GAP` names the 5px gap, and `field_area_offset` gives the distance from
+  the origin to the top of the input box. `render`, `handle_click` and the height
+  calculation all derive from it, instead of each recomputing `font_size + 5`.
+
+### Acceptance criteria
+
+- No label ink is drawn above the control's origin.
+- All ink lies within `[origin.y, origin.y + size[1]]`.
+- Consecutive fields laid out by `form_layouter` do not overlap.
+- A click in the input area focuses the field; a click on the label does not.
+
+### Files
+
+`pyWebLayout/concrete/functional.py`
+
+---
+
+## S16 — Backward page navigation
+
+### Problem
+
+`render_page_backward` *searched* for the previous page's start: estimate a block
+index, lay out forward, compare the end against the target, bisect on the block
+difference, repeat up to ten times. Both the estimator and the adjuster pinned
+`word_index` to 0 and moved only `block_index`.
+
+Pages routinely start mid-block. Any such start was therefore **not in the search
+space**, the loop could never match, and it fell through to a fallback that
+jumped several blocks back or to the document start.
+
+### Evidence
+
+A document of short paragraphs around one 1200-word paragraph. Forward pagination
+gives page starts at `(0,0), (2,208), (2,494), (2,780), (2,1057)`. Asking for the
+page that ends where each of those begins:
+
+```
+from page 1 -> got (0,0)   expected (0,0)     ok     (1 forward layout)
+from page 2 -> got (0,0)   expected (2,208)   WRONG  (10 forward layouts)
+from page 3 -> got (0,0)   expected (2,494)   WRONG  (10 forward layouts)
+from page 4 -> got (0,0)   expected (2,780)   WRONG  (10 forward layouts)
+```
+
+Every mid-paragraph case threw the reader to the start of the document after ten
+full page layouts. The bisection was also unsound within its own space: a
+document of 40 small paragraphs, where every page *does* start on a block
+boundary, failed too.
+
+This is complementary to S11 rather than caused by it. Before S11 forward
+pagination dead-ended at the first page-spanning block, so mid-block starts were
+never produced and the block-granular search looked adequate.
+
+### Design
+
+Pagination is a pure function: laying out from `q` yields a page and the position
+it stopped at, `next(q)`. The page before `P` is the `q` with `next(q) == P`.
+That is found by **replaying the chain forward from an anchor**, not by guessing
+`q`. Three sources, in order:
+
+1. **The recorded chain.** `render_page_forward` now records
+   `(font_scale, next(q)) -> q`. Stepping back to anywhere the reader has been is
+   exact and costs one layout. Keyed by font scale, since changing it
+   repaginates.
+2. **Replay from an anchor.** Anchors are block starts, nearest first: the block
+   containing `P`, then up to `MAX_BACKWARD_ANCHORS` earlier ones, then the
+   document start. Lay out forward from the anchor until a page ends exactly on
+   `P`; that page's start is the answer. `MAX_REPLAY_PAGES` caps the walk so one
+   page turn cannot traverse a whole chapter.
+3. **Nearest start before `P`.** If no chain passes exactly through `P` — which
+   happens when `P` was reached by a jump or a restored bookmark rather than by
+   reading forward, so it lies on no natural chain — return the last page start
+   before it. That overlaps `P`'s page slightly rather than skipping content,
+   which is the safe direction to be wrong in.
+
+The estimator and the bisecting adjuster are deleted.
+
+**What "correct" means here.** Each backward step returns a page ending exactly
+where the reader currently is, so paging back never skips or repeats content.
+That chain can differ from the one you would have seen reading forward from page
+one, if you entered the document by a jump — pagination from a different starting
+point is genuinely a different chain, and no algorithm can recover the original
+without replaying from the start.
+
+### Measurements
+
+Same document, after the change:
+
+```
+warm (chain recorded by the forward pass):  4/4 exact, 1 layout each
+cold, fresh layouter per call:             12/13 exact, worst 17 layouts
+cold, one layouter, repeated back presses:  4 layouts per turn typical
+```
+
+The single inexact case is a target that lies on the canonical chain but not on
+any chain reachable from a nearby anchor; it returns a start 15 words early,
+i.e. a slightly overlapping page.
+
+### Acceptance criteria
+
+- For every page of a document, `render_page_backward(start[i])` returns
+  `start[i-1]` — verified for both a mid-paragraph-paginating document and one
+  where every page starts on a block boundary.
+- Laying out forward from the returned position ends exactly on the requested
+  position.
+- Forward-then-back returns to the original position.
+- At the document start, backward stays there; an empty document is safe.
+- Cost stays within a small bounded number of forward layouts.
+
+### Files
+
+`pyWebLayout/layout/ereader_layout.py`
+
+---
+
 ## Test plan
 
 Findings were reproduced with four probe scripts; each becomes a regression test

pyWebLayout/abstract/inline.py

@@ -3,12 +3,25 @@ from pyWebLayout.core import Hierarchical
 from pyWebLayout.style import Font
 from pyWebLayout.style.abstract_style import AbstractStyle
 from typing import Tuple, Union, List, Optional, Dict, Any, Callable
+from functools import lru_cache
 import pyphen
 
 # Import LinkType for type hints (imported at module level to avoid F821 linting error)
 from pyWebLayout.abstract.functional import LinkType
 
 
+@lru_cache(maxsize=16)
+def _hyphen_dict(language: Optional[str]) -> pyphen.Pyphen:
+    """
+    The pyphen dictionary for a language, reused across words.
+
+    Pyphen caches the parsed dictionary file itself, but rebuilding the wrapper
+    per word still costs about 40% of a hyphenation call, and hyphenation is
+    attempted for every word that overflows its line.
+    """
+    return pyphen.Pyphen(lang=language)
+
+
 class Word:
     """
     An abstract representation of a word in a document. Words can be split across
@@ -163,6 +176,18 @@ class Word:
         """Set the next word in sequence"""
         self._next = next_word
 
+    def with_style(self, style: Font) -> 'Word':
+        """
+        Return a copy of this word carrying a different font.
+
+        Subclasses that hold extra state must override this, or that state is
+        silently dropped when a caller restyles the word. Sequence links
+        (previous/next) are deliberately not copied: the copy belongs to a
+        different word chain, which the new container rebuilds as words are
+        added to it.
+        """
+        return Word(self._text, style, self._background)
+
     def possible_hyphenation(self, language: str = None) -> bool:
         """
         Hyphenate the word and store the parts.
@@ -174,8 +199,7 @@ class Word:
             bool: True if the word was hyphenated, False otherwise.
         """
 
-        dic = pyphen.Pyphen(lang=self._style.language)
-        return list(dic.iterate(self._text))
+        return list(_hyphen_dict(self._style.language).iterate(self._text))
 
 
 ...
@@ -348,6 +372,19 @@ class LinkedWord(Word):
         """Get the link title/tooltip"""
         return self._title
 
+    def with_style(self, style: Font) -> 'LinkedWord':
+        """Return a copy carrying a different font, keeping the link intact."""
+        return LinkedWord(
+            self._text,
+            style,
+            self._location,
+            link_type=self._link_type,
+            callback=self._callback,
+            background=self._background,
+            params=dict(self._params),
+            title=self._title,
+        )
+
     def execute_link(self, context: Optional[Dict[str, Any]] = None) -> Any:
         """
         Execute the link action.

pyWebLayout/concrete/functional.py

@@ -99,15 +99,20 @@ class LinkText(Text, Interactable, Queriable):
             self._origin,
             np.ndarray) else self._origin
 
-        # Draw background based on state (before text is rendered)
+        # Draw background based on state (before text is rendered).
+        # PIL wants a flat sequence of four scalars; handing it a list of two
+        # numpy arrays raises "coordinate list must contain exactly 2
+        # coordinates".
+        if self._pressed or self._hovered:
+            far = origin + size
+            box = (int(origin[0]), int(origin[1]), int(far[0]), int(far[1]))
             if self._pressed:
                 # Pressed state - stronger, darker highlight
-            bg_color = (180, 180, 255, 180)  # Stronger blue with more opacity
-            self._draw.rectangle([origin, origin + size], fill=bg_color)
-        elif self._hovered:
+                bg_color = (180, 180, 255, 180)
+            else:
                 # Hover state - subtle highlight
-            bg_color = (220, 220, 255, 100)  # Light blue with alpha
-            self._draw.rectangle([origin, origin + size], fill=bg_color)
+                bg_color = (220, 220, 255, 100)
+            self._draw.rectangle(box, fill=bg_color)
 
         # Call the parent Text render method with parameters
         super().render(next_text, spacing)
@@ -153,7 +158,22 @@ class ButtonText(Text, Interactable, Queriable):
             self, '_width', 0) if not hasattr(
             self._width, '__call__') else 0
         self._padded_width = text_width + padding[1] + padding[3]
-        self._padded_height = self._style.font_size + padding[0] + padding[2]
+
+        # Size the button from the text's visual height (ascent + descent), not
+        # from the nominal font size. The two differ by several pixels - DejaVu at
+        # 14px measures 17 - so sizing by font_size leaves the button too short to
+        # centre its own label in.
+        self._text_height = self._visual_text_height()
+        self._padded_height = self._text_height + padding[0] + padding[2]
+
+    def _visual_text_height(self) -> int:
+        """Height of the rendered text, ascender to descender."""
+        try:
+            ascent, descent = self._style.font.getmetrics()
+            return int(ascent + descent)
+        except (AttributeError, TypeError, ValueError):
+            # Mock or unusual font object; the nominal size is the best guess.
+            return int(getattr(self._style, 'font_size', 0) or 0)
 
     @property
     def button(self) -> Button:
@@ -237,11 +257,18 @@ class ButtonText(Text, Interactable, Queriable):
         # Total button height minus top and bottom padding gives us text area height
         text_area_height = self._padded_height - self._padding[0] - self._padding[2]
 
-        # Center the text visual height (ascent + descent) within the text area
-        # The y position is where the baseline sits
-        # Visual center = area_height/2, baseline should be at center + descent/2
-        vertical_center = text_area_height / 2
-        text_y = self._origin[1] + self._padding[0] + vertical_center + (descent / 2)
+        # Centre the text's visual height (ascent + descent) within the text area.
+        # text_y is the baseline, since Text renders with anchor "ls".
+        #
+        #   top of glyphs = area_top + (area_height - (ascent + descent)) / 2
+        #   baseline      = top of glyphs + ascent
+        #
+        # The previous form, area_top + area_height/2 + descent/2, is only
+        # equivalent when ascent == 2 * descent. Real fonts sit nearer 4:1, so the
+        # label rendered several pixels above centre, against the top edge.
+        text_top = self._origin[1] + self._padding[0] \
+            + (text_area_height - (ascent + descent)) / 2
+        text_y = text_top + ascent
 
         # Temporarily set origin for text rendering
         original_origin = self._origin.copy()
@@ -275,8 +302,17 @@ class FormFieldText(Text, Interactable, Queriable):
     """
     A Text subclass that can handle FormField interactions.
     Renders form field labels and input areas.
+
+    The origin is the top-left of the whole control: label, then a gap, then the
+    input box. Text itself draws from a baseline, so the label is offset down by
+    its ascent when rendering; without that the glyphs would sit above the origin
+    and overprint whatever is above, which for a stacked form is the previous
+    field's input box.
     """
 
+    # Vertical gap between the label and its input box, in pixels.
+    LABEL_GAP = 5
+
     def __init__(self, field: FormField, font: Font, draw: ImageDraw.Draw,
                  field_height: int = 24, source=None, line=None):
         """
@@ -302,8 +338,11 @@ class FormFieldText(Text, Interactable, Queriable):
         self._field_height = field_height
         self._focused = False
 
-        # Calculate total height (label + gap + field)
-        self._total_height = self._style.font_size + 5 + field_height
+        # Calculate total height (label + gap + field). The label's height is its
+        # ink height, ascender to descender, not the nominal font size - the two
+        # differ by several pixels and the gap between label and box is only 5.
+        self._label_height = self._visual_label_height()
+        self._total_height = self._label_height + self.LABEL_GAP + field_height
 
         # Field width should be at least as wide as the label
         # Use getattr to handle mock objects in tests
@@ -312,6 +351,20 @@ class FormFieldText(Text, Interactable, Queriable):
             self._width, '__call__') else 0
         self._field_width = max(text_width, 150)
 
+    def _visual_label_height(self) -> int:
+        """Height of the rendered label, ascender to descender."""
+        try:
+            ascent, descent = self._style.font.getmetrics()
+            return int(ascent + descent)
+        except (AttributeError, TypeError, ValueError):
+            # Mock or unusual font object; the nominal size is the best guess.
+            return int(getattr(self._style, 'font_size', 0) or 0)
+
+    @property
+    def field_area_offset(self) -> int:
+        """Distance from this control's origin to the top of its input box."""
+        return self._label_height + self.LABEL_GAP
+
     @property
     def field(self) -> FormField:
         """Get the associated FormField object"""
@@ -330,12 +383,21 @@ class FormFieldText(Text, Interactable, Queriable):
         """
         Render the form field with label and input area.
         """
-        # Render the label
-        super().render()
+        # Render the label. Text draws from the baseline, so shift down by the
+        # ascent to make the origin the top of the label rather than its baseline.
+        try:
+            label_ascent = self._style.font.getmetrics()[0]
+        except (AttributeError, TypeError, ValueError):
+            label_ascent = self._label_height
 
-        # Calculate field position (below label with 5px gap)
+        label_origin = self._origin
+        self._origin = np.array([label_origin[0], label_origin[1] + label_ascent])
+        super().render()
+        self._origin = label_origin
+
+        # Calculate field position (below the label, with the standard gap)
         field_x = self._origin[0]
-        field_y = self._origin[1] + self._style.font_size + 5
+        field_y = self._origin[1] + self.field_area_offset
 
         # Draw field background and border
         bg_color = (255, 255, 255)
@@ -360,11 +422,12 @@ class FormFieldText(Text, Interactable, Queriable):
             # Get font metrics to properly center the baseline
             ascent, descent = value_font.font.getmetrics()
 
-            # Center the text vertically within the field
-            # The y coordinate is where the baseline sits (anchor="ls")
-            vertical_center = self._field_height / 2
+            # Centre the value within the input box. As in ButtonText, the
+            # baseline sits at the top of the glyphs plus the ascent; centring on
+            # half the box height plus half the descent only works for a 2:1
+            # ascent/descent ratio and otherwise rides high.
             value_x = field_x + 5
-            value_y = field_y + vertical_center + (descent / 2)
+            value_y = field_y + (self._field_height - (ascent + descent)) / 2 + ascent
 
             # Draw the value text
             self._draw.text((value_x, value_y), value_text,
@@ -381,7 +444,7 @@ class FormFieldText(Text, Interactable, Queriable):
             True if the field was clicked and focused
         """
         # Calculate field area
-        field_y = self._style.font_size + 5
+        field_y = self.field_area_offset
 
         # Check if click is within the input field area (not just the label)
         if (0 <= point[0] <= self._field_width and

pyWebLayout/concrete/page.py

@@ -15,6 +15,10 @@ class Page(Renderable, Queriable):
     contains a given point.
     """
 
+    # Mode of the render canvas. The measurement context matches it so that text
+    # width caching keys stay consistent between layout and rendering.
+    _CANVAS_MODE = 'RGBA'
+
     def __init__(self, size: Tuple[int, int], style: Optional[PageStyle] = None,
                  origin: Tuple[int, int] = (0, 0)):
         """
@@ -32,6 +36,7 @@ class Page(Renderable, Queriable):
         self._children: List[Renderable] = []
         self._canvas: Optional[Image.Image] = None
         self._draw: Optional[ImageDraw.Draw] = None
+        self._measurement_draw: Optional[ImageDraw.ImageDraw] = None
         # Initialize y_offset to start of content area
         # Position the first line so its baseline is close to the top boundary
         # For subsequent lines, baseline-to-baseline spacing is used
@@ -168,13 +173,38 @@ class Page(Renderable, Queriable):
 
     @property
     def draw(self) -> Optional[ImageDraw.Draw]:
-        """Get the ImageDraw object for drawing on this page's canvas"""
-        if self._draw is None:
+        """
+        Get the ImageDraw object bound to this page's render canvas.
+
+        Rebuilt whenever the canvas has been invalidated: a draw context
+        outlives the image it was created from, so checking only _draw would
+        hand back a context pointing at a discarded canvas.
+        """
+        if self._draw is None or self._canvas is None:
             # Initialize canvas and draw context if not already done
             self._canvas = self._create_canvas()
             self._draw = ImageDraw.Draw(self._canvas)
         return self._draw
 
+    @property
+    def measurement_draw(self) -> ImageDraw.ImageDraw:
+        """
+        A scratch draw context for text metrics during layout.
+
+        Layout asks for text widths constantly, but has no reason to touch the
+        render canvas - and the canvas is invalidated on every add_child, so
+        measuring through `draw` would allocate a full-page image per line.
+        This context is 1x1 and never invalidated.
+
+        Its mode matches the render canvas because Text keys its width cache on
+        the draw mode; a mismatch would double every cache entry. Children built
+        against it are re-bound to the real canvas by render_children.
+        """
+        if self._measurement_draw is None:
+            scratch = Image.new(self._CANVAS_MODE, (1, 1))
+            self._measurement_draw = ImageDraw.Draw(scratch)
+        return self._measurement_draw
+
     def add_child(self, child: Renderable) -> 'Page':
         """
         Add a child renderable object to this page.
@@ -228,69 +258,6 @@ class Page(Renderable, Queriable):
         """Get a copy of the children list"""
         return self._children.copy()
 
-    def _get_child_property(self, child: Renderable, private_attr: str,
-                            public_attr: str, index: Optional[int] = None,
-                            default: Optional[int] = None) -> Optional[int]:
-        """
-        Generic helper to extract properties from child objects with multiple fallback strategies.
-
-        Args:
-            child: The child object
-            private_attr: Name of the private attribute (e.g., '_size')
-            public_attr: Name of the public property (e.g., 'size')
-            index: Optional index for array-like properties (0 for width, 1 for height)
-            default: Default value if property cannot be determined
-
-        Returns:
-            Property value or default
-        """
-        # Try private attribute first
-        if hasattr(child, private_attr):
-            value = getattr(child, private_attr)
-            if value is not None:
-                if isinstance(value, (list, tuple, np.ndarray)):
-                    if index is not None and len(value) > index:
-                        return int(value[index])
-                    elif index is None:
-                        return value
-
-        # Try public property
-        if hasattr(child, public_attr):
-            value = getattr(child, public_attr)
-            if value is not None:
-                if isinstance(value, (list, tuple, np.ndarray)):
-                    if index is not None and len(value) > index:
-                        return int(value[index])
-                    elif index is None:
-                        return value
-                else:
-                    return int(value)
-
-        return default
-
-    def _get_child_height(self, child: Renderable) -> int:
-        """
-        Get the height of a child object.
-
-        Args:
-            child: The child to measure
-
-        Returns:
-            Height in pixels
-        """
-        # Try to get height from size property (index 1)
-        height = self._get_child_property(child, '_size', 'size', index=1)
-        if height is not None:
-            return height
-
-        # Try direct height attribute
-        height = self._get_child_property(child, '_height', 'height')
-        if height is not None:
-            return height
-
-        # Default fallback height
-        return 20
-
     def render_children(self):
         """
         Call render on all children in the list.
@@ -333,7 +300,7 @@ class Page(Renderable, Queriable):
             PIL Image with background and borders applied
         """
         # Create base image
-        canvas = Image.new('RGBA', self._size, (*self._style.background_color, 255))
+        canvas = Image.new(self._CANVAS_MODE, self._size, (*self._style.background_color, 255))
 
         # Draw borders if needed
         if self._style.border_width > 0:
@@ -349,23 +316,6 @@ class Page(Renderable, Queriable):
 
         return canvas
 
-    def _get_child_position(self, child: Renderable) -> Tuple[int, int]:
-        """
-        Get the position where a child should be rendered.
-
-        Args:
-            child: The child object
-
-        Returns:
-            Tuple of (x, y) coordinates
-        """
-        # Try to get x coordinate
-        x = self._get_child_property(child, '_origin', 'position', index=0, default=0)
-        # Try to get y coordinate
-        y = self._get_child_property(child, '_origin', 'position', index=1, default=0)
-
-        return (x, y)
-
     def query_point(self, point: Tuple[int, int]) -> Optional[QueryResult]:
         """
         Query a point to find the deepest object at that location.
@@ -402,64 +352,6 @@ class Page(Renderable, Queriable):
             bounds=(int(point[0]), int(point[1]), 0, 0)
         )
 
-    def _point_in_child(self, point: np.ndarray, child: Renderable) -> bool:
-        """
-        Check if a point is within a child's bounds.
-
-        Args:
-            point: The point to check
-            child: The child to check against
-
-        Returns:
-            True if the point is within the child's bounds
-        """
-        # If child implements Queriable interface, use it
-        if isinstance(child, Queriable) and hasattr(child, 'in_object'):
-            try:
-                return child.in_object(point)
-            except BaseException:
-                pass  # Fall back to bounds checking
-
-        # Get child position and size for bounds checking
-        child_pos = self._get_child_position(child)
-        child_size = self._get_child_size(child)
-
-        if child_size is None:
-            return False
-
-        # Check if point is within child bounds
-        return (
-            child_pos[0] <= point[0] < child_pos[0] + child_size[0] and
-            child_pos[1] <= point[1] < child_pos[1] + child_size[1]
-        )
-
-    def _get_child_size(self, child: Renderable) -> Optional[Tuple[int, int]]:
-        """
-        Get the size of a child object.
-
-        Args:
-            child: The child to measure
-
-        Returns:
-            Tuple of (width, height) or None if size cannot be determined
-        """
-        # Try to get width and height from size property
-        width = self._get_child_property(child, '_size', 'size', index=0)
-        height = self._get_child_property(child, '_size', 'size', index=1)
-
-        # If size property worked, return it
-        if width is not None and height is not None:
-            return (width, height)
-
-        # Try direct width/height attributes
-        width = self._get_child_property(child, '_width', 'width')
-        height = self._get_child_property(child, '_height', 'height')
-
-        if width is not None and height is not None:
-            return (width, height)
-
-        return None
-
     def _make_query_result(self, obj, point: Tuple[int, int]) -> QueryResult:
         """
         Package an object into a QueryResult with metadata.

pyWebLayout/concrete/text.py

@@ -58,6 +58,15 @@ _width_cache: UsageCache = UsageCache(DEFAULT_WIDTH_CACHE_ENTRIES)
 _glyph_cache: SizedUsageCache = SizedUsageCache(DEFAULT_GLYPH_CACHE_BYTES, _glyph_entry_bytes)
 _glyph_subpixel_steps: int = DEFAULT_GLYPH_SUBPIXEL_STEPS
 
+# Every Line asks its font for the advance width of a space. That single
+# FreeTypeFont.getlength(" ") call costs ~18us -- two orders of magnitude more
+# than getmetrics() -- because PIL shapes the string from scratch each time, and
+# it lands once per line created, which dominates the cost of laying a line out.
+# There are only ever a handful of distinct fonts in play, so memoise per font
+# object. Values are wrapped in a 1-tuple because None is itself a legitimate
+# result (fonts that cannot report a length) and must not read as a cache miss.
+_space_advance_cache: Dict[Any, Tuple[Optional[int]]] = {}
+
 # Set to False the first time the fast rasterisation path is found to be
 # unavailable (e.g. a PIL build without the private ImageDraw internals it uses),
 # after which every Text falls back to ImageDraw.text().
@@ -98,6 +107,35 @@ def clear_text_caches():
     """Drop all cached widths and glyph bitmaps."""
     _width_cache.clear()
     _glyph_cache.clear()
+    _space_advance_cache.clear()
+
+
+def _space_advance(font) -> Optional[int]:
+    """
+    The font's own advance width for a space, in whole pixels.
+
+    None when the font cannot report one, which is the signal for callers to fall
+    back to their configured spacing range.
+    """
+    try:
+        cached = _space_advance_cache.get(font)
+    except TypeError:
+        # Unhashable font object; measure without caching.
+        cached = None
+    else:
+        if cached is not None:
+            return cached[0]
+
+    try:
+        value = int(round(font.getlength(" ")))
+    except (AttributeError, TypeError, ValueError):
+        value = None
+
+    try:
+        _space_advance_cache[font] = (value,)
+    except TypeError:
+        pass
+    return value
 
 
 def text_cache_stats() -> Dict[str, Any]:
@@ -215,7 +253,8 @@ class AlignmentHandler(ABC):
     def calculate_spacing_and_position(self, text_objects: List['Text'],
                                        available_width: int, min_spacing: int,
                                        max_spacing: int,
-                                       natural_spacing: Optional[int] = None
+                                       natural_spacing: Optional[int] = None,
+                                       total_width: Optional[float] = None
                                        ) -> Tuple[int, int, bool]:
         """
         Calculate the spacing between words and starting position for the line.
@@ -228,6 +267,11 @@ class AlignmentHandler(ABC):
             natural_spacing: The font's own space width. Ragged alignments use it
                 as a constant gap; justification ignores it. Defaults to
                 min_spacing when not supplied.
+            total_width: The summed width of `text_objects`, when the caller
+                already knows it. Purely an optimisation: a line asks its handler
+                to re-measure once per candidate word, and summing the whole line
+                each time makes filling a line quadratic in its word count. Omit
+                it and the sum is taken here as before.
 
         Returns:
             Tuple of (spacing_between_words, starting_x_position, overflow)
@@ -242,7 +286,8 @@ class LeftAlignmentHandler(AlignmentHandler):
                                        available_width: int,
                                        min_spacing: int,
                                        max_spacing: int,
-                                       natural_spacing: Optional[int] = None
+                                       natural_spacing: Optional[int] = None,
+                                       total_width: Optional[float] = None
                                        ) -> Tuple[int, int, bool]:
         """
         Calculate spacing and position for left-aligned text objects.
@@ -269,7 +314,8 @@ class LeftAlignmentHandler(AlignmentHandler):
         spacing = min_spacing if natural_spacing is None else natural_spacing
         spacing = max(min_spacing, min(max_spacing, int(spacing)))
 
-        text_length = sum([text.width for text in text_objects])
+        text_length = (sum([text.width for text in text_objects])
+                       if total_width is None else total_width)
         num_gaps = len(text_objects) - 1
 
         # The spacing is constant whether or not the content fits: tightening a
@@ -290,7 +336,8 @@ class CenterRightAlignmentHandler(AlignmentHandler):
     def calculate_spacing_and_position(self, text_objects: List['Text'],
                                        available_width: int, min_spacing: int,
                                        max_spacing: int,
-                                       natural_spacing: Optional[int] = None
+                                       natural_spacing: Optional[int] = None,
+                                       total_width: Optional[float] = None
                                        ) -> Tuple[int, int, bool]:
         """
         Centre/right alignment: constant word space, line shifted as a block.
@@ -300,7 +347,8 @@ class CenterRightAlignmentHandler(AlignmentHandler):
         the same spacing that will actually be used, so the line lands where it
         was measured to land.
         """
-        word_length = sum([word.width for word in text_objects])
+        word_length = (sum([word.width for word in text_objects])
+                       if total_width is None else total_width)
 
         # Handle single word case
         if len(text_objects) <= 1:
@@ -329,13 +377,42 @@ class JustifyAlignmentHandler(AlignmentHandler):
     """Handler for justified text with full justification."""
 
     def __init__(self):
-        # Store variable spacing for each gap to distribute remainder pixels
-        self._gap_spacings: List[int] = []
+        # The per-gap spacings are described by a plan rather than stored outright,
+        # and materialised on demand by the _gap_spacings property below. Fitting a
+        # line calls this handler once per candidate word and only ever looks at the
+        # first gap; building the whole list on each of those probes made adding n
+        # words to a line O(n^2). Only render() reads the full list.
+        self._gap_uniform: Optional[int] = None
+        self._gap_residual: int = 0
+        self._gap_count: int = 0
+        self._gap_cache: Optional[List[int]] = []
+
+    @property
+    def _gap_spacings(self) -> List[int]:
+        """The spacing to apply at each gap, left to right."""
+        if self._gap_cache is None:
+            if self._gap_uniform is not None:
+                self._gap_cache = [self._gap_uniform] * self._gap_count
+            else:
+                self._gap_cache = self._distribute(self._gap_residual, self._gap_count)
+        return self._gap_cache
+
+    @staticmethod
+    def _distribute(total: int, num_gaps: int) -> List[int]:
+        """Split `total` pixels across `num_gaps` gaps by cumulative rounding."""
+        gaps = []
+        placed = 0
+        for i in range(1, num_gaps + 1):
+            cumulative = int(round(total * i / num_gaps))
+            gaps.append(cumulative - placed)
+            placed = cumulative
+        return gaps
 
     def calculate_spacing_and_position(self, text_objects: List['Text'],
                                        available_width: int, min_spacing: int,
                                        max_spacing: int,
-                                       natural_spacing: Optional[int] = None
+                                       natural_spacing: Optional[int] = None,
+                                       total_width: Optional[float] = None
                                        ) -> Tuple[int, int, bool]:
         """
         Justified alignment distributes space to fill the entire line width.
@@ -347,14 +424,17 @@ class JustifyAlignmentHandler(AlignmentHandler):
         is min_spacing to ensure readability.
         """
 
-        word_length = sum([word.width for word in text_objects])
+        word_length = (sum([word.width for word in text_objects])
+                       if total_width is None else total_width)
         residual_space = available_width - word_length
         num_gaps = max(1, len(text_objects) - 1)
 
         # Check if we have enough space for minimum spacing
         if residual_space // num_gaps < min_spacing:
             # Not enough space - this is overflow
-            self._gap_spacings = [min_spacing] * num_gaps
+            self._gap_uniform = min_spacing
+            self._gap_count = num_gaps
+            self._gap_cache = None
             return min_spacing, 0, True
 
         # Distribute the residual by cumulative rounding rather than by taking a
@@ -365,14 +445,14 @@ class JustifyAlignmentHandler(AlignmentHandler):
         # ragged right edge on otherwise justified text. Rounding the running
         # total makes the gaps sum to the residual exactly.
         total = int(round(residual_space))
-        self._gap_spacings = []
-        placed = 0
-        for i in range(1, num_gaps + 1):
-            cumulative = int(round(total * i / num_gaps))
-            self._gap_spacings.append(cumulative - placed)
-            placed = cumulative
+        self._gap_uniform = None
+        self._gap_residual = total
+        self._gap_count = num_gaps
+        self._gap_cache = None
 
-        return self._gap_spacings[0], 0, False
+        # The first gap is the whole of the plan that fitting needs, and it falls
+        # out of the same cumulative rounding as _distribute would give it.
+        return int(round(total / num_gaps)), 0, False
 
 
 class Text(Renderable, Queriable):
@@ -689,6 +769,9 @@ class Line(Box):
         """
         super().__init__(origin, size, callback, sheet, mode, halign, valign)
         self._text_objects: List['Text'] = []  # Store Text objects directly
+        # Prefix sums of the widths in _text_objects, kept in step by _push_text /
+        # _pop_text. Element 0 is the empty sum. See _push_text for the rationale.
+        self._width_prefix: List[float] = [0.0]
         self._spacing = spacing  # (min_spacing, max_spacing)
         self._font = font if font else Font()  # Use default font if none provided
         self._current_width = 0  # Track the current width used
@@ -704,10 +787,7 @@ class Line(Box):
 
         # The font's own space advance. Ragged alignments use this as their
         # constant word gap rather than stretching to fill the measure.
-        try:
-            self._natural_spacing = int(round(self._font.font.getlength(" ")))
-        except (AttributeError, TypeError, ValueError):
-            self._natural_spacing = None
+        self._natural_spacing = _space_advance(self._font.font)
 
         # Hyphenation configuration parameters
         self._min_word_length_for_brute_force = min_word_length_for_brute_force
@@ -771,6 +851,45 @@ class Line(Box):
         """Set the next line in sequence"""
         self._next = line
 
+    @property
+    def _content_width(self) -> float:
+        """Summed width of the line's current contents."""
+        return self._width_prefix[-1]
+
+    def _push_text(self, text: 'Text'):
+        """
+        Append a Text to the line, keeping the running width sum in step.
+
+        Fitting a word is a trial: the candidate is pushed, measured, and popped
+        again if it did not fit, so the line's contents churn far more often than
+        they grow. Tracking the sum here rather than re-adding every width on each
+        measurement is what keeps filling a line linear in its word count.
+
+        The sum is kept as a prefix list rather than as one accumulator that is
+        added to and subtracted from. Widths are floats, so `(total + w) - w` need
+        not give back `total` exactly, and a drift of one ulp is enough to flip an
+        overflow decision on a line that ends flush. Truncating a prefix list
+        restores the earlier total bit for bit, and each entry is built by the same
+        left-to-right addition sum() would perform.
+        """
+        self._text_objects.append(text)
+        self._width_prefix.append(self._width_prefix[-1] + text.width)
+
+    def _pop_text(self) -> 'Text':
+        """Remove the last Text from the line, keeping the width sum in step."""
+        text = self._text_objects.pop()
+        self._width_prefix.pop()
+        return text
+
+    def _measure(self, handler: Optional[AlignmentHandler] = None
+                 ) -> Tuple[int, int, bool]:
+        """Ask an alignment handler to place the line's current contents."""
+        if handler is None:
+            handler = self._alignment_handler
+        return handler.calculate_spacing_and_position(
+            self._text_objects, self._size[0], self._spacing[0], self._spacing[1],
+            self._natural_spacing, self._content_width)
+
     def add_word(self,
                  word: 'Word',
                  part: Optional[Text] = None) -> Tuple[bool,
@@ -789,7 +908,7 @@ class Line(Box):
         """
         # First, add any pretext from previous hyphenation
         if part is not None:
-            self._text_objects.append(part)
+            self._push_text(part)
             self._words.append(word)
             part.add_line(self)
 
@@ -818,10 +937,8 @@ class Line(Box):
                 line=self)
         else:
             text = Text.from_word(word, self._draw)
-        self._text_objects.append(text)
-        spacing, position, overflow = self._alignment_handler.calculate_spacing_and_position(
-            self._text_objects, self._size[0], self._spacing[0], self._spacing[1],
-                self._natural_spacing)
+        self._push_text(text)
+        spacing, position, overflow = self._measure()
 
         if not overflow:
             # Word fits! Add it completely
@@ -833,7 +950,7 @@ class Line(Box):
             return True, None
 
         # Word doesn't fit, remove it and try hyphenation
-        _ = self._text_objects.pop()
+        self._pop_text()
 
         # Step 1: Try pyphen hyphenation
         pyphen_splits = word.possible_hyphenation()
@@ -866,11 +983,9 @@ class Line(Box):
                     source=word)
 
                 # Check if first part fits
-                self._text_objects.append(first_text)
-                spacing, position, overflow = self._alignment_handler.calculate_spacing_and_position(
-                    self._text_objects, self._size[0], self._spacing[0], self._spacing[1],
-                self._natural_spacing)
-                _ = self._text_objects.pop()
+                self._push_text(first_text)
+                spacing, position, overflow = self._measure()
+                self._pop_text()
 
                 if not overflow:
                     # This split fits! Add it to valid options
@@ -883,7 +998,7 @@ class Line(Box):
             first_text, second_text, spacing, position = best_split
 
             # Apply the split
-            self._text_objects.append(first_text)
+            self._push_text(first_text)
             first_text.line = self
             word.add_concete((first_text, second_text))
             self._spacing_render = spacing
@@ -894,7 +1009,7 @@ class Line(Box):
         # Step 3: Try brute force hyphenation (only for long words)
         if len(word.text) >= self._min_word_length_for_brute_force:
             # Calculate available space for the word
-            word_length = sum([text.width for text in self._text_objects])
+            word_length = self._content_width
             spacing_length = self._spacing[0] * max(0, len(self._text_objects) - 1)
             remaining = self._size[0] - word_length - spacing_length
 
@@ -938,10 +1053,8 @@ class Line(Box):
                             source=word)
 
                         # Verify the first part actually fits
-                        self._text_objects.append(first_text)
-                        spacing, position, overflow = self._alignment_handler.calculate_spacing_and_position(
-                            self._text_objects, self._size[0], self._spacing[0], self._spacing[1],
-                self._natural_spacing)
+                        self._push_text(first_text)
+                        spacing, position, overflow = self._measure()
 
                         if not overflow:
                             # Brute force split works!
@@ -954,7 +1067,7 @@ class Line(Box):
                             return True, second_text
                         else:
                             # Doesn't fit, remove it
-                            _ = self._text_objects.pop()
+                            self._pop_text()
 
         # Step 4: Word cannot be hyphenated or split, move to next line
         return False, None
@@ -972,9 +1085,7 @@ class Line(Box):
         # justified paragraph.
         handler = self.render_alignment_handler
         if len(self._text_objects) > 0:
-            spacing, position, overflow = handler.calculate_spacing_and_position(
-                self._text_objects, self._size[0], self._spacing[0], self._spacing[1],
-                self._natural_spacing)
+            spacing, position, overflow = self._measure(handler)
             self._spacing_render = spacing
             self._position_render = position
 
@@ -982,28 +1093,33 @@ class Line(Box):
 
         # Start x_cursor at line origin plus any alignment offset
         x_cursor = self._origin[0] + self._position_render
-        for i, text in enumerate(self._text_objects):
+
+        # Everything the loop needs that does not vary per word is resolved once.
+        # Only justified lines carry per-gap spacings; every other alignment uses
+        # the single spacing figured above.
+        texts = self._text_objects
+        last = len(texts) - 1
+        draw = self._draw
+        default_spacing = self._spacing_render
+        gaps = handler._gap_spacings if isinstance(handler, JustifyAlignmentHandler) else ()
+        gap_count = len(gaps)
+
+        for i, text in enumerate(texts):
             # Update text draw context to current draw context
-            text._draw = self._draw
+            text._draw = draw
             text.set_origin(np.array([x_cursor, y_cursor]))
 
             # Determine next text object for continuous decoration
-            next_text = self._text_objects[i + 1] if i + \
-                1 < len(self._text_objects) else None
+            next_text = texts[i + 1] if i < last else None
 
             # Get the spacing for this specific gap (variable for justified text)
-            if isinstance(handler, JustifyAlignmentHandler) and \
-               hasattr(handler, '_gap_spacings') and \
-               i < len(handler._gap_spacings):
-                current_spacing = handler._gap_spacings[i]
-            else:
-                current_spacing = self._spacing_render
+            current_spacing = gaps[i] if i < gap_count else default_spacing
 
             # Render with next text information for continuous underline/strikethrough
             text.render(next_text, current_spacing)
             # Add text width, then spacing only if there are more words
             x_cursor += text.width
-            if i < len(self._text_objects) - 1:
+            if i < last:
                 x_cursor += current_spacing
 
     def query_point(self, point: Tuple[int, int]) -> Optional['QueryResult']:

pyWebLayout/core/highlight.py

@@ -6,12 +6,16 @@ managing highlight collections, and rendering highlights on pages.
 """
 
 from __future__ import annotations
+import logging
 from dataclasses import dataclass
 from typing import List, Tuple, Optional, Dict, Any
 from enum import Enum
-import json
 from pathlib import Path
 
+from pyWebLayout.core.persistence import ensure_dir, read_json, write_json
+
+logger = logging.getLogger(__name__)
+
 
 class HighlightColor(Enum):
     """Predefined highlight colors with RGBA values"""
@@ -44,6 +48,12 @@ class Highlight:
     start_word_index: Optional[int] = None  # Word index in document (if available)
     end_word_index: Optional[int] = None
 
+    # Where in the document this highlight lives, as a serialized
+    # RenderingPosition. `bounds` are pixel coordinates on one particular
+    # rendering, so they stop matching as soon as the font scale or page size
+    # changes; this survives repagination and is what page association uses.
+    position: Optional[Dict[str, Any]] = None
+
     # Metadata
     note: Optional[str] = None  # Optional annotation
     tags: List[str] = None  # Optional categorization tags
@@ -63,6 +73,7 @@ class Highlight:
             'text': self.text,
             'start_word_index': self.start_word_index,
             'end_word_index': self.end_word_index,
+            'position': self.position,
             'note': self.note,
             'tags': self.tags,
             'timestamp': self.timestamp
@@ -78,6 +89,7 @@ class Highlight:
             text=data['text'],
             start_word_index=data.get('start_word_index'),
             end_word_index=data.get('end_word_index'),
+            position=data.get('position'),
             note=data.get('note'),
             tags=data.get('tags', []),
             timestamp=data.get('timestamp')
@@ -100,12 +112,9 @@ class HighlightManager:
             highlights_dir: Directory to store highlight data
         """
         self.document_id = document_id
-        self.highlights_dir = Path(highlights_dir)
+        self.highlights_dir = ensure_dir(highlights_dir)
         self.highlights: Dict[str, Highlight] = {}  # id -> Highlight
 
-        # Create directory if it doesn't exist
-        self.highlights_dir.mkdir(parents=True, exist_ok=True)
-
         # Load existing highlights
         self._load_highlights()
 
@@ -178,34 +187,22 @@ class HighlightManager:
 
     def _save_highlights(self) -> None:
         """Persist highlights to disk"""
-        try:
-            filepath = self._get_filepath()
-            data = {
+        write_json(self._get_filepath(), {
             'document_id': self.document_id,
             'highlights': [h.to_dict() for h in self.highlights.values()]
-            }
-
-            with open(filepath, 'w') as f:
-                json.dump(data, f, indent=2)
-        except Exception as e:
-            print(f"Error saving highlights: {e}")
+        })
 
     def _load_highlights(self) -> None:
         """Load highlights from disk"""
+        data = read_json(self._get_filepath(), {})
         try:
-            filepath = self._get_filepath()
-            if not filepath.exists():
-                return
-
-            with open(filepath, 'r') as f:
-                data = json.load(f)
-
             self.highlights = {
                 h['id']: Highlight.from_dict(h)
                 for h in data.get('highlights', [])
             }
-        except Exception as e:
-            print(f"Error loading highlights: {e}")
+        except (AttributeError, TypeError, KeyError):
+            logger.warning("Highlight file %s is not in the expected shape; ignoring it",
+                           self._get_filepath(), exc_info=True)
             self.highlights = {}
 
 
@@ -213,16 +210,18 @@ def create_highlight_from_query_result(
     result,
     color: Tuple[int, int, int, int] = HighlightColor.YELLOW.value,
     note: Optional[str] = None,
-    tags: Optional[List[str]] = None
+    tags: Optional[List[str]] = None,
+    position: Optional[Dict[str, Any]] = None
 ) -> Highlight:
     """
     Create a highlight from a QueryResult.
 
     Args:
-        result: QueryResult from query_pixel or query_range
+        result: QueryResult from query_point or query_range
         color: RGBA color tuple
         note: Optional annotation
         tags: Optional categorization tags
+        position: Serialized RenderingPosition of the page the result came from
 
     Returns:
         Highlight instance
@@ -243,6 +242,7 @@ def create_highlight_from_query_result(
         bounds=bounds,
         color=color,
         text=text,
+        position=position,
         note=note,
         tags=tags or [],
         timestamp=time()

pyWebLayout/core/persistence.py

@@ -0,0 +1,59 @@
+"""
+Small JSON-file helpers shared by the per-document stores.
+
+BookmarkManager and HighlightManager both keep a JSON file per document under a
+directory, and both had their own copy of "make the directory, try to read it,
+swallow and print on failure". The duplication is the point of this module; the
+file formats themselves stay owned by each store.
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+from pathlib import Path
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+
+def ensure_dir(path: str | Path) -> Path:
+    """Return `path` as a Path, creating it and any missing parents."""
+    directory = Path(path)
+    directory.mkdir(parents=True, exist_ok=True)
+    return directory
+
+
+def read_json(path: Path, default: Any) -> Any:
+    """
+    Read JSON from `path`, returning `default` if it is missing or unreadable.
+
+    A corrupt store must not stop a book from opening, so failures are logged
+    and swallowed. `default` is returned as given, so pass a fresh mutable if
+    the caller intends to mutate it.
+    """
+    if not path.exists():
+        return default
+
+    try:
+        with open(path, 'r', encoding='utf-8') as handle:
+            return json.load(handle)
+    except (OSError, ValueError):
+        logger.warning("Could not read %s; ignoring its contents", path, exc_info=True)
+        return default
+
+
+def write_json(path: Path, data: Any) -> bool:
+    """
+    Write `data` to `path` as JSON.
+
+    Returns True on success. Failures are logged rather than raised: losing a
+    bookmark is not a reason to take down the reader.
+    """
+    try:
+        with open(path, 'w', encoding='utf-8') as handle:
+            json.dump(data, handle, indent=2)
+        return True
+    except (OSError, TypeError, ValueError):
+        logger.error("Could not write %s", path, exc_info=True)
+        return False

pyWebLayout/io/readers/html_extraction.py

@@ -8,6 +8,7 @@ Each handler function has a robust signature that handles style hints, CSS class
 
 from typing import List, Dict, Any, Optional, Union, Callable, Tuple, NamedTuple
 from bs4 import BeautifulSoup, Tag, NavigableString
+from bs4.element import CData, Comment, Doctype, ProcessingInstruction
 from pyWebLayout.abstract.inline import Word
 from pyWebLayout.abstract.block import (
     Block,
@@ -369,6 +370,24 @@ def extract_text_content(element: Tag, context: StyleContext) -> List[Word]:
         element: BeautifulSoup Tag object
         context: Current style context
 
+    Returns:
+        List of Word objects (including LinkedWord for hyperlinks)
+    """
+    return extract_words_from_nodes(list(element.children), context)
+
+
+def extract_words_from_nodes(nodes: List, context: StyleContext) -> List[Word]:
+    """
+    Extract words from a sequence of sibling nodes.
+
+    Separated from extract_text_content so that a container holding a mix of
+    inline and block children can hand over just the inline runs, without
+    building a synthetic element to wrap them in.
+
+    Args:
+        nodes: BeautifulSoup nodes (Tags and NavigableStrings) in document order
+        context: Current style context
+
     Returns:
         List of Word objects (including LinkedWord for hyperlinks)
     """
@@ -377,15 +396,20 @@ def extract_text_content(element: Tag, context: StyleContext) -> List[Word]:
 
     words = []
 
-    for child in element.children:
+    for child in nodes:
+        # Comments and processing instructions are NavigableString subclasses;
+        # their text is markup, not content.
+        if isinstance(child, (Comment, Doctype, CData, ProcessingInstruction)):
+            continue
+
         if isinstance(child, NavigableString):
-            # Plain text - split into words
-            text = str(child).strip()
-            if text:
-                word_texts = text.split()
-                for word_text in word_texts:
-                    if word_text:
-                        words.append(Word(word_text, context.font, context.background))
+            # Plain text - split into words. Argument-less str.split() already
+            # discards surrounding whitespace and never yields an empty string, so
+            # it needs neither a preceding strip() nor a per-word emptiness test.
+            font = context.font
+            background = context.background
+            words.extend([Word(word_text, font, background)
+                          for word_text in str(child).split()])
         elif isinstance(child, Tag):
             # Special handling for <a> tags (hyperlinks)
             if child.name.lower() == "a":
@@ -466,6 +490,93 @@ def extract_text_content(element: Tag, context: StyleContext) -> List[Word]:
     return words
 
 
+# Tags that flow within a line of text rather than forming a block of their own.
+# They carry no handler of their own: extract_words_from_nodes consumes them,
+# applying their styling to the words they contain.
+INLINE_TAGS = frozenset({
+    "a", "span", "strong", "b", "em", "i", "u", "s", "del", "ins", "mark",
+    "small", "sub", "sup", "code", "q", "cite", "abbr", "time",
+})
+
+
+def is_inline(node) -> bool:
+    """
+    Whether a node belongs to a run of text rather than standing as its own block.
+
+    Args:
+        node: A BeautifulSoup Tag or NavigableString
+
+    Returns:
+        True for text and inline tags, False for block-level tags
+    """
+    if isinstance(node, Tag):
+        return node.name.lower() in INLINE_TAGS
+    if isinstance(node, (Comment, Doctype, CData, ProcessingInstruction)):
+        return False
+    return isinstance(node, NavigableString)
+
+
+def process_block_children(element: Tag, context: StyleContext) -> List[Block]:
+    """
+    Process a container's children into a list of blocks.
+
+    Containers may hold a mix of inline and block content. Consecutive inline
+    children are gathered into a run and become one Paragraph; a block child ends
+    the current run and is processed by its own handler. This is the single entry
+    point for every container that is not itself a paragraph - div, li, td, th,
+    blockquote and the semantic containers.
+
+    Without this, inline tags reach process_element, whose handler for them is
+    ignore_handler, and their text is silently dropped.
+
+    Args:
+        element: The container element
+        context: Current style context
+
+    Returns:
+        Blocks in document order
+    """
+    blocks: List[Block] = []
+    run: List = []
+
+    def flush_run():
+        """Turn the pending inline run into a paragraph, if it holds any words."""
+        if not run:
+            return
+        words = extract_words_from_nodes(run, context)
+        run.clear()
+        if words:
+            paragraph = Paragraph(context.font)
+            for word in words:
+                paragraph.add_word(word)
+            blocks.append(paragraph)
+
+    for child in element.children:
+        # <br> ends the current line of text and starts a new one.
+        if isinstance(child, Tag) and child.name.lower() == "br":
+            flush_run()
+            continue
+
+        if is_inline(child):
+            run.append(child)
+            continue
+
+        if not isinstance(child, Tag):
+            continue  # comments and similar
+
+        flush_run()
+        child_context = apply_element_styling(context, child)
+        result = process_element(child, child_context)
+        if result:
+            if isinstance(result, list):
+                blocks.extend(result)
+            else:
+                blocks.append(result)
+
+    flush_run()
+    return blocks
+
+
 def process_element(
     element: Tag, context: StyleContext
 ) -> Union[Block, List[Block], None]:
@@ -557,17 +668,7 @@ def paragraph_handler(element: Tag, context: StyleContext) -> Union[Paragraph, L
 
 def div_handler(element: Tag, context: StyleContext) -> List[Block]:
     """Handle <div> elements - treat as generic container."""
-    blocks = []
-    for child in element.children:
-        if isinstance(child, Tag):
-            child_context = apply_element_styling(context, child)
-            result = process_element(child, child_context)
-            if result:
-                if isinstance(result, list):
-                    blocks.extend(result)
-                else:
-                    blocks.append(result)
-    return blocks
+    return process_block_children(element, context)
 
 
 def heading_handler(element: Tag, context: StyleContext) -> Heading:
@@ -592,16 +693,8 @@ def heading_handler(element: Tag, context: StyleContext) -> Heading:
 def blockquote_handler(element: Tag, context: StyleContext) -> Quote:
     """Handle <blockquote> elements."""
     quote = Quote(context.font)
-    for child in element.children:
-        if isinstance(child, Tag):
-            child_context = apply_element_styling(context, child)
-            result = process_element(child, child_context)
-            if result:
-                if isinstance(result, list):
-                    for block in result:
+    for block in process_block_children(element, context):
         quote.add_block(block)
-                else:
-                    quote.add_block(result)
     return quote
 
 
@@ -655,28 +748,8 @@ def ordered_list_handler(element: Tag, context: StyleContext) -> HList:
 def list_item_handler(element: Tag, context: StyleContext) -> ListItem:
     """Handle <li> elements."""
     list_item = ListItem(None, context.font)
-
-    for child in element.children:
-        if isinstance(child, Tag):
-            child_context = apply_element_styling(context, child)
-            result = process_element(child, child_context)
-            if result:
-                if isinstance(result, list):
-                    for block in result:
+    for block in process_block_children(element, context):
         list_item.add_block(block)
-                else:
-                    list_item.add_block(result)
-        elif isinstance(child, NavigableString):
-            # Direct text in list item - create paragraph
-            text = str(child).strip()
-            if text:
-                paragraph = Paragraph(context.font)
-                words = text.split()
-                for word_text in words:
-                    if word_text:
-                        paragraph.add_word(Word(word_text, context.font))
-                list_item.add_block(paragraph)
-
     return list_item
 
 
@@ -728,27 +801,8 @@ def table_cell_handler(element: Tag, context: StyleContext) -> TableCell:
     rowspan = int(context.element_attributes.get("rowspan", 1))
     cell = TableCell(False, colspan, rowspan, context.font)
 
-    # Process cell content
-    for child in element.children:
-        if isinstance(child, Tag):
-            child_context = apply_element_styling(context, child)
-            result = process_element(child, child_context)
-            if result:
-                if isinstance(result, list):
-                    for block in result:
+    for block in process_block_children(element, context):
         cell.add_block(block)
-                else:
-                    cell.add_block(result)
-        elif isinstance(child, NavigableString):
-            # Direct text in cell - create paragraph
-            text = str(child).strip()
-            if text:
-                paragraph = Paragraph(context.font)
-                words = text.split()
-                for word_text in words:
-                    if word_text:
-                        paragraph.add_word(Word(word_text, context.font))
-                cell.add_block(paragraph)
 
     return cell
 
@@ -759,26 +813,8 @@ def table_header_cell_handler(element: Tag, context: StyleContext) -> TableCell:
     rowspan = int(context.element_attributes.get("rowspan", 1))
     cell = TableCell(True, colspan, rowspan, context.font)
 
-    # Process cell content (same as td)
-    for child in element.children:
-        if isinstance(child, Tag):
-            child_context = apply_element_styling(context, child)
-            result = process_element(child, child_context)
-            if result:
-                if isinstance(result, list):
-                    for block in result:
+    for block in process_block_children(element, context):
         cell.add_block(block)
-                else:
-                    cell.add_block(result)
-        elif isinstance(child, NavigableString):
-            text = str(child).strip()
-            if text:
-                paragraph = Paragraph(context.font)
-                words = text.split()
-                for word_text in words:
-                    if word_text:
-                        paragraph.add_word(Word(word_text, context.font))
-                cell.add_block(paragraph)
 
     return cell
 

pyWebLayout/layout/document_layouter.py

@@ -163,18 +163,15 @@ def paragraph_layouter(paragraph: Paragraph,
             y_cursor = page._current_y_offset
         x_cursor = page.content_origin[0]
 
-        # Create a temporary Text object to calculate word width
-        if word:
-            temp_text = Text.from_word(word, page.draw)
-            temp_text.width
-        else:
-            pass
+        # `word` is accepted for call-site readability only: the line that is about
+        # to be created measures it when it is added, so measuring it here as well
+        # only paid for a Text object that was immediately discarded.
 
         return Line(
             spacing=word_spacing_constraints,
             origin=(x_cursor, y_cursor),
             size=(page.available_width, baseline_spacing),
-            draw=page.draw,
+            draw=page.measurement_draw,
             font=font,
             halign=text_align
         )
@@ -225,7 +222,7 @@ def paragraph_layouter(paragraph: Paragraph,
                 return False, i, overflow_text
 
             # Check if the word will fit on the new line before adding it
-            temp_text = Text.from_word(word, page.draw)
+            temp_text = Text.from_word(word, page.measurement_draw)
             if temp_text.width > current_line.size[0]:
                 # Word is too wide for the line, we need to hyphenate it
                 if len(word.text) >= 6:
@@ -234,13 +231,13 @@ def paragraph_layouter(paragraph: Paragraph,
                         (Text(
                             pair[0],
                             word.style,
-                            page.draw,
+                            page.measurement_draw,
                             line=current_line,
                             source=word),
                             Text(
                             pair[1],
                             word.style,
-                            page.draw,
+                            page.measurement_draw,
                             line=current_line,
                             source=word)) for pair in word.possible_hyphenation()]
                     if len(splits) > 0:
@@ -455,7 +452,7 @@ def button_layouter(button: Button,
     available_height = page.remaining_height
 
     # Create ButtonText renderable
-    button_text = ButtonText(button, font, page.draw, padding=padding)
+    button_text = ButtonText(button, font, page.measurement_draw, padding=padding)
 
     # Check if button fits on current page
     button_height = button_text.size[1]
@@ -505,7 +502,8 @@ def form_field_layouter(field: FormField, page: Page, font: Optional[Font] = Non
     available_height = page.remaining_height
 
     # Create FormFieldText renderable
-    field_text = FormFieldText(field, font, page.draw, field_height=field_height)
+    field_text = FormFieldText(field, font, page.measurement_draw,
+                               field_height=field_height)
 
     # Check if field fits on current page
     total_field_height = field_text.size[1]

pyWebLayout/layout/ereader_layout.py

@@ -15,7 +15,9 @@ from __future__ import annotations
 from dataclasses import dataclass, asdict
 from typing import List, Dict, Tuple, Optional, Any
 
-from pyWebLayout.abstract.block import Block, Paragraph, Heading, HeadingLevel, Table, HList, Image
+from pyWebLayout.abstract.block import (
+    Block, Paragraph, Heading, HeadingLevel, Table, TableRow, TableCell,
+    HList, ListItem, Quote, Image)
 from pyWebLayout.abstract.inline import Word
 from pyWebLayout.concrete.page import Page
 from pyWebLayout.concrete.text import Text
@@ -41,6 +43,19 @@ class RenderingPosition:
     remaining_pretext: Optional[str] = None     # Hyphenated word continuation
     page_y_offset: int = 0                     # Vertical position on page
 
+    def _key(self) -> Tuple[Any, ...]:
+        """
+        The fields in declaration order.
+
+        Copying, comparing and hashing a position all used to go through
+        dataclasses.asdict, which walks the field list and deep-copies each value.
+        Every field here is an immutable scalar, so that traversal bought nothing
+        and these three run constantly during page navigation and buffer lookups.
+        """
+        return (self.chapter_index, self.block_index, self.word_index,
+                self.table_row, self.table_col, self.list_item_index,
+                self.remaining_pretext, self.page_y_offset)
+
     def to_dict(self) -> Dict[str, Any]:
         """Serialize position for saving to file/database"""
         return asdict(self)
@@ -52,17 +67,17 @@ class RenderingPosition:
 
     def copy(self) -> 'RenderingPosition':
         """Create a copy of this position"""
-        return RenderingPosition(**asdict(self))
+        return RenderingPosition(*self._key())
 
     def __eq__(self, other) -> bool:
         """Check if two positions are equal"""
         if not isinstance(other, RenderingPosition):
             return False
-        return asdict(self) == asdict(other)
+        return self._key() == other._key()
 
     def __hash__(self) -> int:
         """Make position hashable for use as dict key"""
-        return hash(tuple(asdict(self).values()))
+        return hash(self._key())
 
 
 class ChapterInfo:
@@ -313,6 +328,19 @@ class BidirectionalLayouter:
         self.alignment_override = alignment_override
         self.font_family_override = font_family_override
 
+        # Maps (font_scale, end position) -> the position the page started at.
+        # Filled in as pages are laid out forward, which makes "previous page"
+        # exact and free for anywhere the reader has already been. Keyed by font
+        # scale because changing it repaginates the document.
+        self._page_chain: Dict[Tuple[float, Tuple[int, int, int]],
+                               RenderingPosition] = {}
+
+        # Scaled copies of blocks, keyed by (id(block), font_scale). Rebuilding
+        # a block's words on every page render allocated a fresh Paragraph and
+        # Word per word on the hot path. The original block is kept alongside
+        # the copy so its id cannot be recycled while it is a live key.
+        self._scaled_block_cache: Dict[Tuple[int, float], Tuple[Block, Block]] = {}
+
     def render_page_forward(self, position: RenderingPosition,
                             font_scale: float = 1.0) -> Tuple[Page, RenderingPosition]:
         """
@@ -367,134 +395,241 @@ class BidirectionalLayouter:
 
             current_pos = new_pos
 
+        # Remember this link in the chain so stepping back to it later is exact.
+        if self._position_compare(current_pos, position) > 0:
+            self._page_chain[(font_scale, self._position_key(current_pos))] = \
+                position.copy()
+
         return page, current_pos
 
+    # How many block starts before the target to try as replay anchors before
+    # settling for the best inexact answer.
+    MAX_BACKWARD_ANCHORS = 4
+
+    # Ceiling on pages replayed from a single anchor, so a pathologically long
+    # block cannot make one page turn walk an entire chapter.
+    MAX_REPLAY_PAGES = 8
+
     def render_page_backward(self,
                              end_position: RenderingPosition,
                              font_scale: float = 1.0) -> Tuple[Page,
                                                                RenderingPosition]:
         """
-        Render a page that ends at the given position, filling backward.
-        Critical for "previous page" navigation.
+        Render the page that ends at the given position - "previous page".
 
-        Uses iterative refinement to find the correct start position that
-        results in a page ending at (or very close to) the target position.
+        Pagination is a pure function: laying out from a position q yields a page
+        and the position where it stopped, next(q). The page before P is therefore
+        the q for which next(q) == P, and it is found by *replaying* the chain
+        forward from an anchor, not by guessing q.
+
+        The previous implementation searched instead: it estimated a block index
+        and bisected on it, pinning word_index to 0. Pages routinely start
+        mid-block, so the answer was frequently not in the search space at all -
+        the search then exhausted its iterations and fell back to a position that
+        was not the previous page, usually the start of the document.
+
+        Three sources are tried in order:
+
+        1. The recorded chain, from pages already laid out going forward. Exact,
+           and the common case when the reader is paging back and forth.
+        2. Replay from the start of the block containing P, then from
+           progressively earlier blocks. Exact when P lies on the resulting chain.
+        3. Failing an exact hit - which happens when P was reached by a jump or a
+           restored bookmark rather than by reading forward, so it is on no
+           natural chain - the latest page start before P. That overlaps P's page
+           slightly rather than skipping content, which is the safe direction to
+           be wrong in.
 
         Args:
-            end_position: Position where page should end
+            end_position: Position where the page should end
             font_scale: Font scaling factor
 
         Returns:
             Tuple of (rendered_page, start_position)
         """
-        # Handle edge case: already at beginning
-        if end_position.block_index == 0 and end_position.word_index == 0:
-            return self.render_page_forward(end_position, font_scale)
+        document_start = RenderingPosition()
 
-        # Start with initial estimate
-        estimated_start = self._estimate_page_start(end_position, font_scale)
+        # Nothing precedes the start of the document.
+        if self._position_compare(end_position, document_start) <= 0:
+            page, _ = self.render_page_forward(document_start, font_scale)
+            return page, document_start
 
-        # Iterative refinement: keep adjusting until we converge or hit max iterations
-        max_iterations = 10
-        best_page = None
-        best_start = estimated_start
-        best_distance = float('inf')
+        # 1. The chain we have already walked.
+        remembered = self._page_chain.get((font_scale, self._position_key(end_position)))
+        if remembered is not None:
+            page, actual_end = self.render_page_forward(remembered, font_scale)
+            if self._position_compare(actual_end, end_position) == 0:
+                return page, remembered
 
-        for iteration in range(max_iterations):
-            # Render forward from current estimate
-            page, actual_end = self.render_page_forward(estimated_start, font_scale)
+        # 2/3. Replay from anchors, keeping the best inexact result as a fallback.
+        fallback = None
+        for anchor in self._backward_anchors(end_position):
+            page, start, exact = self._replay_to(anchor, end_position, font_scale)
+            if page is None:
+                continue
+            if exact:
+                return page, start
+            if fallback is None:
+                fallback = (page, start)
 
-            # Calculate how far we are from target
-            comparison = self._position_compare(actual_end, end_position)
+        if fallback is not None:
+            return fallback
 
-            # Perfect match or close enough (within same block)
-            # BUT: ensure we actually moved backward (estimated_start < end_position)
-            if comparison == 0:
-                # Check if we actually found a valid previous page
-                if self._position_compare(estimated_start, end_position) < 0:
-                    return page, estimated_start
-                # If estimated_start >= end_position, we haven't moved backward
-                # Continue iterating to find a better position
-                elif iteration == 0:
-                    # On first iteration, if we can't find a previous position,
-                    # we're likely at or near the beginning
+        page, _ = self.render_page_forward(document_start, font_scale)
+        return page, document_start
+
+    def _backward_anchors(self, target: RenderingPosition):
+        """
+        Yield positions to replay from, nearest first.
+
+        Block starts are used as anchors because they are the coarsest positions
+        that are certainly valid to lay out from. The block containing the target
+        comes first: when the target is mid-block, the page before it usually
+        starts in that same block or the one before.
+        """
+        first_block = target.block_index if target.word_index > 0 \
+            else target.block_index - 1
+
+        for offset in range(self.MAX_BACKWARD_ANCHORS):
+            block_index = first_block - offset
+            if block_index < 0:
                 break
-
-            # Track best result so far
-            distance = abs(actual_end.block_index - end_position.block_index)
-            if distance < best_distance:
-                best_distance = distance
-                best_page = page
-                best_start = estimated_start.copy()
-
-            # Adjust estimate for next iteration
-            estimated_start = self._adjust_start_estimate(
-                estimated_start, end_position, actual_end)
-
-            # Safety: don't go before document start
-            if estimated_start.block_index < 0:
-                estimated_start.block_index = 0
-                estimated_start.word_index = 0
-
-        # If we exhausted iterations, return best result found
-        # BUT: ensure we didn't return the same position (no backward progress)
-        final_page = best_page if best_page else page
-        final_start = best_start
-
-        # Safety check: if final_start >= end_position, we failed to move backward
-        # This can happen at the beginning of the document or when estimation failed
-        if self._position_compare(final_start, end_position) >= 0:
-            # Can't go further back - check if we're at the absolute beginning
-            if end_position.block_index == 0 and end_position.word_index == 0:
-                # Already at beginning, return as-is
-                return final_page, final_start
-
-            # Fallback strategy: try a more aggressive backward jump
-            # Start from several blocks before the current position
-            blocks_to_jump = max(1, min(5, end_position.block_index))
-            fallback_pos = RenderingPosition(
-                chapter_index=end_position.chapter_index,
-                block_index=max(0, end_position.block_index - blocks_to_jump),
-                word_index=0
+            yield RenderingPosition(
+                chapter_index=target.chapter_index,
+                block_index=block_index,
+                word_index=0,
             )
 
-            # Render forward from the fallback position
-            fallback_page, fallback_end = self.render_page_forward(fallback_pos, font_scale)
+        if first_block - self.MAX_BACKWARD_ANCHORS >= 0:
+            yield RenderingPosition()
 
-            # Verify the fallback actually moved us backward
-            if self._position_compare(fallback_pos, end_position) < 0:
-                return fallback_page, fallback_pos
+    def _replay_to(self,
+                   anchor: RenderingPosition,
+                   target: RenderingPosition,
+                   font_scale: float):
+        """
+        Lay out pages forward from `anchor`, looking for the one ending at `target`.
 
-            # If even the fallback didn't work, we're likely at the beginning
-            # Return a page starting from the beginning
-            return self.render_page_forward(RenderingPosition(), font_scale)
+        Returns:
+            (page, start, exact). `exact` is True when a page ended precisely on
+            the target. When the chain steps over the target instead, the last
+            page starting before it is returned with exact=False. (None, None,
+            False) means the anchor yielded nothing usable.
+        """
+        position = anchor
+        last = (None, None)
 
-        return final_page, final_start
+        for _ in range(self.MAX_REPLAY_PAGES):
+            if self._position_compare(position, target) >= 0:
+                break
+
+            page, next_position = self.render_page_forward(position, font_scale)
+            comparison = self._position_compare(next_position, target)
+
+            if comparison == 0:
+                return page, position, True
+
+            if comparison > 0:
+                # Stepped over the target: this chain does not pass through it.
+                return last[0], last[1], False
+
+            if self._position_compare(next_position, position) <= 0:
+                break  # no progress; give up on this anchor
+
+            last = (page, position)
+            position = next_position
+
+        return last[0], last[1], False
+
+    @staticmethod
+    def _position_key(position: RenderingPosition) -> Tuple[int, int, int]:
+        """Hashable identity of a position, for the page chain map."""
+        return (position.chapter_index, position.block_index, position.word_index)
 
     def _scale_block_fonts(self, block: Block, font_scale: float) -> Block:
-        """Apply font scaling and font family override to all fonts in a block"""
-        # Check if we need to do any transformation
+        """
+        Apply font scaling and the font family override to every font in a block.
+
+        Returns the block unchanged when there is nothing to apply. Results are
+        memoised per (block, scale) for the life of the layouter, so a page
+        re-render at an unchanged scale costs a dict lookup.
+        """
         if font_scale == 1.0 and self.font_family_override is None:
             return block
 
-        # This is a simplified implementation
-        # In practice, we'd need to handle each block type appropriately
-        if isinstance(block, (Paragraph, Heading)):
-            scaled_block_style = FontScaler.scale_font(block.style, font_scale, self.font_family_override)
-            if isinstance(block, Heading):
-                scaled_block = Heading(block.level, scaled_block_style)
-            else:
-                scaled_block = Paragraph(scaled_block_style)
+        key = (id(block), font_scale)
+        cached = self._scaled_block_cache.get(key)
+        if cached is not None:
+            return cached[1]
 
-            # words_iter() returns tuples of (position, word)
-            for position, word in block.words_iter():
+        scaled = self._build_scaled_block(block, font_scale)
+        self._scaled_block_cache[key] = (block, scaled)
+        return scaled
+
+    def _build_scaled_block(self, block: Block, font_scale: float) -> Block:
+        """Construct the scaled copy of a block. See _scale_block_fonts."""
+        def scale(font: Font) -> Font:
+            return FontScaler.scale_font(font, font_scale, self.font_family_override)
+
+        if isinstance(block, (Paragraph, Heading)):
+            if isinstance(block, Heading):
+                scaled_block = Heading(block.level, scale(block.style))
+            else:
+                scaled_block = Paragraph(scale(block.style))
+
+            # words_iter() yields (position, word) tuples. with_style() keeps
+            # the concrete word class, so a LinkedWord stays linked - rebuilding
+            # these as plain Words silently stripped every hyperlink in the
+            # document as soon as the reader changed font size.
+            for _, word in block.words_iter():
                 if isinstance(word, Word):
-                    scaled_word = Word(
-                        word.text, FontScaler.scale_font(
-                            word.style, font_scale, self.font_family_override))
-                    scaled_block.add_word(scaled_word)
+                    scaled_block.add_word(word.with_style(scale(word.style)))
             return scaled_block
 
+        if isinstance(block, Quote):
+            scaled_quote = Quote(scale(block.style) if block.style else None)
+            for child in block.blocks():
+                scaled_quote.add_block(self._scale_block_fonts(child, font_scale))
+            return scaled_quote
+
+        if isinstance(block, HList):
+            scaled_list = HList(
+                block.style,
+                scale(block.default_style) if block.default_style else None)
+            for item in block.items():
+                scaled_item = ListItem(
+                    item.term,
+                    scale(item.style) if item.style else None)
+                for child in item.blocks():
+                    scaled_item.add_block(self._scale_block_fonts(child, font_scale))
+                scaled_list.add_item(scaled_item)
+            return scaled_list
+
+        if isinstance(block, Table):
+            scaled_table = Table(
+                block.caption,
+                scale(block.style) if block.style else None)
+            # Rows must go back into the section they came from, or a <thead>
+            # row would be re-added as a body row.
+            for section, rows in (('header', block.header_rows()),
+                                  ('body', block.body_rows()),
+                                  ('footer', block.footer_rows())):
+                for row in rows:
+                    scaled_row = TableRow(scale(row.style) if row.style else None)
+                    for cell in row.cells():
+                        scaled_cell = TableCell(
+                            is_header=cell.is_header,
+                            colspan=cell.colspan,
+                            rowspan=cell.rowspan,
+                            style=scale(cell.style) if cell.style else None)
+                        for child in cell.blocks():
+                            scaled_cell.add_block(self._scale_block_fonts(child, font_scale))
+                        scaled_row.add_cell(scaled_cell)
+                    scaled_table.add_row(scaled_row, section)
+            return scaled_table
+
+        # Blocks with no fonts of their own (Image, HorizontalRule, PageBreak,
+        # CodeBlock - which carries raw lines, not styled words) pass through.
         return block
 
     def _layout_block_on_page(self,
@@ -671,60 +806,6 @@ class BidirectionalLayouter:
             # Keep same position so it will be attempted on the next page
             return False, position
 
-    def _estimate_page_start(
-            self,
-            end_position: RenderingPosition,
-            font_scale: float) -> RenderingPosition:
-        """Estimate where a page should start to end at the given position"""
-        # This is a simplified heuristic - a full implementation would be more
-        # sophisticated
-        estimated_start = end_position.copy()
-
-        # Move back by an estimated number of blocks that would fit on a page
-        estimated_blocks_per_page = max(1, int(10 / font_scale))  # Rough estimate
-        estimated_start.block_index = max(
-            0, end_position.block_index - estimated_blocks_per_page)
-        estimated_start.word_index = 0
-
-        return estimated_start
-
-    def _adjust_start_estimate(
-            self,
-            current_start: RenderingPosition,
-            target_end: RenderingPosition,
-            actual_end: RenderingPosition) -> RenderingPosition:
-        """
-        Adjust start position estimate based on overshoot/undershoot.
-        Uses proportional adjustment to converge faster.
-        """
-        adjusted = current_start.copy()
-
-        # Calculate the difference between actual and target end positions
-        block_diff = actual_end.block_index - target_end.block_index
-
-        comparison = self._position_compare(actual_end, target_end)
-
-        if comparison < 0:  # Undershot - rendered to block X but need to reach block Y where X < Y
-            # We didn't render far enough forward
-            # Need to start at a LATER block (higher index) so the page includes more content
-            adjustment = max(1, abs(block_diff) // 2)
-            new_index = adjusted.block_index + adjustment
-            # Clamp to valid range
-            if len(self.blocks) > 0:
-                adjusted.block_index = min(len(self.blocks) - 1, max(0, new_index))
-            else:
-                adjusted.block_index = max(0, new_index)
-        elif comparison > 0:  # Overshot - rendered past the target
-            # We rendered too far forward
-            # Need to start at an EARLIER block (lower index) so the page doesn't go as far
-            adjustment = max(1, abs(block_diff) // 2)
-            adjusted.block_index = max(0, adjusted.block_index - adjustment)
-
-        # Reset word index when adjusting blocks
-        adjusted.word_index = 0
-
-        return adjusted
-
     def _position_compare(self, pos1: RenderingPosition,
                           pos2: RenderingPosition) -> int:
         """Compare two positions (-1: pos1 < pos2, 0: equal, 1: pos1 > pos2)"""
@@ -735,27 +816,3 @@ class BidirectionalLayouter:
         if pos1.word_index != pos2.word_index:
             return 1 if pos1.word_index > pos2.word_index else -1
         return 0
-
-
-# Add can_fit_line method to Page class if it doesn't exist
-def _add_page_methods():
-    """Add missing methods to Page class"""
-    if not hasattr(Page, 'can_fit_line'):
-        def can_fit_line(self, line_height: int) -> bool:
-            """Check if a line of given height can fit on the page"""
-            available_height = self.content_size[1] - self._current_y_offset
-            return available_height >= line_height
-
-        Page.can_fit_line = can_fit_line
-
-    if not hasattr(Page, 'available_width'):
-        @property
-        def available_width(self) -> int:
-            """Get available width for content"""
-            return self.content_size[0]
-
-        Page.available_width = available_width
-
-
-# Apply the page methods
-_add_page_methods()

pyWebLayout/layout/ereader_manager.py

@@ -8,9 +8,7 @@ into a unified, easy-to-use API.
 
 from __future__ import annotations
 from typing import List, Dict, Optional, Tuple, Any, Callable
-import json
 import logging
-from pathlib import Path
 
 from .ereader_layout import RenderingPosition, ChapterNavigator, ChapterInfo
 from .page_buffer import BufferedPageRenderer
@@ -20,6 +18,11 @@ from pyWebLayout.concrete.image import RenderableImage
 from pyWebLayout.style.page_style import PageStyle
 from pyWebLayout.style.fonts import BundledFont
 from pyWebLayout.layout.document_layouter import image_layouter
+from pyWebLayout.core.highlight import Highlight, HighlightColor, HighlightManager, \
+    create_highlight_from_query_result
+from pyWebLayout.core.persistence import ensure_dir, read_json, write_json
+from pyWebLayout.concrete.interaction_handler import InteractionStateManager
+from PIL import Image as Image_
 
 logger = logging.getLogger(__name__)
 
@@ -38,8 +41,7 @@ class BookmarkManager:
             bookmarks_dir: Directory to store bookmark files
         """
         self.document_id = document_id
-        self.bookmarks_dir = Path(bookmarks_dir)
-        self.bookmarks_dir.mkdir(exist_ok=True)
+        self.bookmarks_dir = ensure_dir(bookmarks_dir)
 
         self.bookmarks_file = self.bookmarks_dir / f"{document_id}_bookmarks.json"
         self.position_file = self.bookmarks_dir / f"{document_id}_position.json"
@@ -49,29 +51,23 @@ class BookmarkManager:
 
     def _load_bookmarks(self):
         """Load bookmarks from file"""
-        if self.bookmarks_file.exists():
+        data = read_json(self.bookmarks_file, {})
         try:
-                with open(self.bookmarks_file, 'r') as f:
-                    data = json.load(f)
             self._bookmarks = {
                 name: RenderingPosition.from_dict(pos_data)
                 for name, pos_data in data.items()
             }
-            except Exception as e:
-                print(f"Failed to load bookmarks: {e}")
+        except (AttributeError, TypeError, KeyError):
+            logger.warning("Bookmark file %s is not in the expected shape; ignoring it",
+                           self.bookmarks_file, exc_info=True)
             self._bookmarks = {}
 
     def _save_bookmarks(self):
         """Save bookmarks to file"""
-        try:
-            data = {
+        write_json(self.bookmarks_file, {
             name: position.to_dict()
             for name, position in self._bookmarks.items()
-            }
-            with open(self.bookmarks_file, 'w') as f:
-                json.dump(data, f, indent=2)
-        except Exception as e:
-            print(f"Failed to save bookmarks: {e}")
+        })
 
     def add_bookmark(self, name: str, position: RenderingPosition):
         """
@@ -128,11 +124,7 @@ class BookmarkManager:
         Args:
             position: Current reading position
         """
-        try:
-            with open(self.position_file, 'w') as f:
-                json.dump(position.to_dict(), f, indent=2)
-        except Exception as e:
-            print(f"Failed to save reading position: {e}")
+        write_json(self.position_file, position.to_dict())
 
     def load_reading_position(self) -> Optional[RenderingPosition]:
         """
@@ -141,13 +133,14 @@ class BookmarkManager:
         Returns:
             Last reading position or None if not found
         """
-        if self.position_file.exists():
+        data = read_json(self.position_file, None)
+        if data is None:
+            return None
         try:
-                with open(self.position_file, 'r') as f:
-                    data = json.load(f)
             return RenderingPosition.from_dict(data)
-            except Exception as e:
-                print(f"Failed to load reading position: {e}")
+        except (TypeError, KeyError):
+            logger.warning("Position file %s is not in the expected shape; ignoring it",
+                           self.position_file, exc_info=True)
             return None
 
 
@@ -171,7 +164,8 @@ class EreaderLayoutManager:
                  document_id: str = "default",
                  buffer_size: int = 5,
                  page_style: Optional[PageStyle] = None,
-                 bookmarks_dir: str = "bookmarks"):
+                 bookmarks_dir: str = "bookmarks",
+                 highlights_dir: Optional[str] = None):
         """
         Initialize the ereader layout manager.
 
@@ -182,6 +176,8 @@ class EreaderLayoutManager:
             buffer_size: Number of pages to cache in each direction
             page_style: Custom page styling (uses default if None)
             bookmarks_dir: Directory to store bookmark files
+            highlights_dir: Directory to store highlights. Defaults to
+                bookmarks_dir, so a document's reading state lives in one place.
         """
         self.blocks = blocks
         self.page_size = page_size
@@ -196,6 +192,8 @@ class EreaderLayoutManager:
         self.renderer = BufferedPageRenderer(blocks, page_style, buffer_size, page_size)
         self.chapter_navigator = ChapterNavigator(blocks)
         self.bookmark_manager = BookmarkManager(document_id, bookmarks_dir)
+        self.highlight_manager = HighlightManager(
+            document_id, highlights_dir if highlights_dir is not None else bookmarks_dir)
 
         # Current state
         self.current_position = RenderingPosition()
@@ -216,6 +214,10 @@ class EreaderLayoutManager:
             self.current_position = saved_position
             self._on_cover_page = False  # If we have a saved position, we're past the cover
 
+        # Pointer interaction state, rebound whenever the displayed page changes
+        self._interaction_state_manager: Optional[InteractionStateManager] = None
+        self._interaction_page: Optional[Page] = None
+
         # Callbacks for UI updates
         self.position_changed_callback: Optional[Callable[[
             RenderingPosition], None]] = None
@@ -451,6 +453,12 @@ class EreaderLayoutManager:
         # Special case: if at the beginning of content and there's a cover, go back to it
         if self._has_cover and self._is_at_beginning() and not self._on_cover_page:
             self._on_cover_page = True
+            # Restore the canonical cover position. Being on the cover must have a
+            # single representation: a fresh load sits at block 0 with the cover
+            # showing, so returning to the cover has to land there too. Leaving the
+            # position at the first content block saves a position that reopens past
+            # the cover, silently losing it.
+            self.current_position = RenderingPosition()
             self._notify_position_changed()
             return self.get_current_page()
 
@@ -848,6 +856,165 @@ class EreaderLayoutManager:
         """
         return self.bookmark_manager.list_bookmarks()
 
+    # ------------------------------------------------------------------
+    # Highlights
+    #
+    # A Highlight carries pixel bounds, which belong to the one rendering it
+    # was taken from: change the font scale or page size and they no longer
+    # describe anything. Each highlight therefore also records the
+    # RenderingPosition of the page it was made on, and page association goes
+    # through that rather than through the bounds.
+    # ------------------------------------------------------------------
+
+    def highlight_point(self,
+                        point: Tuple[int, int],
+                        color: Tuple[int, int, int, int] = HighlightColor.YELLOW.value,
+                        note: Optional[str] = None,
+                        tags: Optional[List[str]] = None) -> Optional[Highlight]:
+        """
+        Highlight whatever is at a point on the current page.
+
+        Args:
+            point: (x, y) in page coordinates, as delivered by a tap
+            color: RGBA fill, e.g. one of HighlightColor
+            note: Optional annotation
+            tags: Optional categorization tags
+
+        Returns:
+            The stored Highlight, or None if nothing was at that point.
+        """
+        result = self.get_current_page().query_point(point)
+        if result is None or result.object_type == "empty":
+            return None
+
+        return self._store_highlight(result, color, note, tags)
+
+    def highlight_range(self,
+                        start: Tuple[int, int],
+                        end: Tuple[int, int],
+                        color: Tuple[int, int, int, int] = HighlightColor.YELLOW.value,
+                        note: Optional[str] = None,
+                        tags: Optional[List[str]] = None) -> Optional[Highlight]:
+        """
+        Highlight the text between two points on the current page.
+
+        Args:
+            start: (x, y) where the selection began
+            end: (x, y) where the selection ended
+            color: RGBA fill, e.g. one of HighlightColor
+            note: Optional annotation
+            tags: Optional categorization tags
+
+        Returns:
+            The stored Highlight, or None if the range selected no text.
+        """
+        selection = self.get_current_page().query_range(start, end)
+        if not selection.results:
+            return None
+
+        return self._store_highlight(selection, color, note, tags)
+
+    def _store_highlight(self, result, color, note, tags) -> Highlight:
+        """Build a Highlight from a query result and persist it."""
+        highlight = create_highlight_from_query_result(
+            result, color=color, note=note, tags=tags,
+            position=self.current_position.to_dict())
+        self.highlight_manager.add_highlight(highlight)
+        return highlight
+
+    def remove_highlight(self, highlight_id: str) -> bool:
+        """
+        Remove a highlight.
+
+        Args:
+            highlight_id: ID of the highlight to remove
+
+        Returns:
+            True if it existed and was removed
+        """
+        return self.highlight_manager.remove_highlight(highlight_id)
+
+    def list_highlights(self) -> List[Highlight]:
+        """Get every highlight in this document."""
+        return self.highlight_manager.list_highlights()
+
+    def get_highlights_for_current_page(self) -> List[Highlight]:
+        """
+        Get the highlights made on the page currently being displayed.
+
+        Matched on the recorded RenderingPosition, so this stays correct across
+        font changes; highlights saved before the position field existed have
+        no position and are never matched.
+        """
+        current = self.current_position.to_dict()
+        return [h for h in self.highlight_manager.list_highlights()
+                if h.position == current]
+
+    def clear_highlights(self) -> None:
+        """Remove every highlight in this document."""
+        self.highlight_manager.clear_all()
+
+    # ------------------------------------------------------------------
+    # Pointer interaction
+    #
+    # Press/hover feedback is state that belongs to one rendered page, so the
+    # state machine is rebound whenever the displayed page changes. Callers get
+    # a fresh frame back when something changed visually, and None when nothing
+    # did - so a UI can skip a redraw it does not need.
+    # ------------------------------------------------------------------
+
+    def _interaction_state(self) -> InteractionStateManager:
+        """The state machine for the page currently displayed."""
+        page = self.get_current_page()
+        if self._interaction_page is not page:
+            if self._interaction_state_manager is not None:
+                self._interaction_state_manager.reset()
+            self._interaction_state_manager = InteractionStateManager(page)
+            self._interaction_page = page
+        return self._interaction_state_manager
+
+    def handle_hover(self, point: Tuple[int, int]) -> Optional[Image_.Image]:
+        """
+        Update hover feedback for a pointer at `point`.
+
+        Args:
+            point: (x, y) in page coordinates
+
+        Returns:
+            A re-rendered frame if the hover state changed, else None.
+        """
+        return self._interaction_state().update_hover(point)
+
+    def handle_touch_down(self, point: Tuple[int, int]) -> Optional[Image_.Image]:
+        """
+        Show pressed feedback for whatever interactive element is at `point`.
+
+        Args:
+            point: (x, y) in page coordinates
+
+        Returns:
+            A frame showing the pressed state, or None if nothing interactive
+            is there.
+        """
+        return self._interaction_state().handle_mouse_down(point)
+
+    def handle_touch_up(self, point: Tuple[int, int]) -> Tuple[Optional[Image_.Image], Any]:
+        """
+        Release the pressed element and run its action.
+
+        Args:
+            point: (x, y) in page coordinates
+
+        Returns:
+            (frame, callback_result). Both are None if no element was pressed.
+        """
+        return self._interaction_state().handle_mouse_up(point)
+
+    def reset_interaction_state(self) -> None:
+        """Clear any hover or press feedback, e.g. when the pointer leaves."""
+        if self._interaction_state_manager is not None:
+            self._interaction_state_manager.reset()
+
     def get_reading_progress(self) -> float:
         """
         Get reading progress as a percentage.
@@ -935,16 +1102,31 @@ class EreaderLayoutManager:
         """
         Shutdown the ereader manager and clean up resources.
         Call this when the application is closing.
+
+        Idempotent: calling it twice saves the position once.
         """
+        if getattr(self, '_shutdown_done', False):
+            return
+        self._shutdown_done = True
+
         # Save current position
         self.bookmark_manager.save_reading_position(self.current_position)
 
-        # Shutdown renderer and buffer
+        # Release cached pages
         self.renderer.shutdown()
 
     def __del__(self):
-        """Cleanup on destruction"""
+        """
+        Best-effort cleanup for callers that never called shutdown().
+
+        Finalisers run during interpreter teardown, when modules and globals
+        may already be torn down, so this must never raise and must never
+        block. Applications should call shutdown() explicitly.
+        """
+        try:
             self.shutdown()
+        except Exception:
+            pass
 
 
 # Convenience function for quick setup

pyWebLayout/layout/page_buffer.py

@@ -1,16 +1,35 @@
 """
-Multi-process page buffering system for high-performance ereader navigation.
+Page caching for ereader navigation.
 
-This module provides intelligent page caching with background rendering using
-multiprocessing to achieve sub-second page navigation performance.
+`PageBuffer` is an LRU cache of rendered pages plus the position links between
+them; `BufferedPageRenderer` wraps it around a `BidirectionalLayouter`.
+
+This module used to render pages ahead of time in a `ProcessPoolExecutor`. That
+never worked and has been removed — see S12 in docs/LAYOUT_REMEDIATION_SPEC.md
+and R1/R2 in docs/ARCHITECTURE_REVIEW.md. In short: the worker returned
+`pickle.dumps(page)`, and a Page holds a live PIL canvas, which is not
+picklable, so every job failed and the result was discarded. The cost — four
+interpreter copies and the whole block list shipped per job — was paid in full
+for no benefit. On Python 3.14, where the default start method became
+`forkserver`, submitting from module-level code raised outright.
+
+Prefetch is not needed. Measured on the tests/data Wikipedia fixture (411
+blocks) with the text caches warm, one page render costs:
+
+    800x600     p50  8.8 ms   p95 15.4 ms
+    1072x1448   p50 13.8 ms   p95 56.1 ms
+
+A page turn is cheaper than the IPC that was meant to hide it. If a slower
+target device ever changes that, the fallback is a synchronous `readahead()`
+method on this class, or a single worker *thread* — layout is PIL-bound and PIL
+releases the GIL — not a process pool. Making the concrete tree picklable
+(Page -> Line -> Text -> Font -> FreeTypeFont) is a large surface area to
+maintain for a cache.
 """
 
 from __future__ import annotations
 from typing import Dict, Optional, List, Tuple, Any
 from collections import OrderedDict
-from concurrent.futures import ProcessPoolExecutor, Future
-import threading
-import pickle
 
 from .ereader_layout import RenderingPosition, BidirectionalLayouter, FontFamilyOverride
 from pyWebLayout.concrete.page import Page
@@ -19,57 +38,20 @@ from pyWebLayout.style.page_style import PageStyle
 from pyWebLayout.style.fonts import BundledFont
 
 
-def _render_page_worker(args: Tuple[List[Block],
-                                    PageStyle,
-                                    RenderingPosition,
-                                    float,
-                                    bool,
-                                    Optional[BundledFont]]) -> Tuple[RenderingPosition,
-                                                    bytes,
-                                                    RenderingPosition]:
-    """
-    Worker function for multiprocess page rendering.
-
-    Args:
-        args: Tuple of (blocks, page_style, position, font_scale, is_backward, font_family)
-
-    Returns:
-        Tuple of (original_position, pickled_page, next_position)
-    """
-    blocks, page_style, position, font_scale, is_backward, font_family = args
-
-    # Create font family override if specified
-    font_family_override = FontFamilyOverride(font_family) if font_family else None
-
-    layouter = BidirectionalLayouter(blocks, page_style, font_family_override=font_family_override)
-
-    if is_backward:
-        page, next_pos = layouter.render_page_backward(position, font_scale)
-    else:
-        page, next_pos = layouter.render_page_forward(position, font_scale)
-
-    # Serialize the page for inter-process communication
-    pickled_page = pickle.dumps(page)
-
-    return position, pickled_page, next_pos
-
-
 class PageBuffer:
     """
-    Intelligent page caching system with LRU eviction and background rendering.
-    Maintains separate forward and backward buffers for optimal navigation performance.
+    LRU cache of rendered pages, with separate forward and backward buffers and
+    the position links between adjacent pages.
     """
 
-    def __init__(self, buffer_size: int = 5, max_workers: int = 4):
+    def __init__(self, buffer_size: int = 5):
         """
         Initialize the page buffer.
 
         Args:
             buffer_size: Number of pages to cache in each direction
-            max_workers: Maximum number of worker processes for background rendering
         """
         self.buffer_size = buffer_size
-        self.max_workers = max_workers
 
         # LRU caches for forward and backward pages
         self.forward_buffer: OrderedDict[RenderingPosition, Page] = OrderedDict()
@@ -81,11 +63,6 @@ class PageBuffer:
         self.reverse_position_map: Dict[RenderingPosition,
                                         RenderingPosition] = {}  # current -> previous
 
-        # Background rendering
-        self.executor: Optional[ProcessPoolExecutor] = None
-        self.pending_renders: Dict[RenderingPosition, Future] = {}
-        self.render_lock = threading.Lock()
-
         # Document state
         self.blocks: Optional[List[Block]] = None
         self.page_style: Optional[PageStyle] = None
@@ -112,10 +89,6 @@ class PageBuffer:
         self.current_font_scale = font_scale
         self.current_font_family = font_family
 
-        # Start the process pool
-        if self.executor is None:
-            self.executor = ProcessPoolExecutor(max_workers=self.max_workers)
-
     def get_page(self, position: RenderingPosition) -> Optional[Page]:
         """
         Get a cached page if available.
@@ -176,121 +149,8 @@ class PageBuffer:
             self.position_map.pop(oldest_pos, None)
             self.reverse_position_map.pop(oldest_pos, None)
 
-    def start_background_rendering(
-            self,
-            current_position: RenderingPosition,
-            direction: str = 'forward'):
-        """
-        Start background rendering of upcoming pages.
-
-        Args:
-            current_position: Current reading position
-            direction: 'forward', 'backward', or 'both'
-        """
-        if not self.blocks or not self.page_style or not self.executor:
-            return
-
-        with self.render_lock:
-            if direction in ['forward', 'both']:
-                self._queue_forward_renders(current_position)
-
-            if direction in ['backward', 'both']:
-                self._queue_backward_renders(current_position)
-
-    def _queue_forward_renders(self, start_position: RenderingPosition):
-        """Queue forward page renders starting from the given position"""
-        current_pos = start_position
-
-        for i in range(self.buffer_size):
-            # Skip if already cached or being rendered
-            if current_pos in self.forward_buffer or current_pos in self.pending_renders:
-                # Try to get next position from cache
-                current_pos = self.position_map.get(current_pos)
-                if not current_pos:
-                    break
-                continue
-
-            # Queue render job
-            args = (
-                self.blocks,
-                self.page_style,
-                current_pos,
-                self.current_font_scale,
-                False,
-                self.current_font_family)
-            future = self.executor.submit(_render_page_worker, args)
-            self.pending_renders[current_pos] = future
-
-            # We don't know the next position yet, so we'll update it when the render
-            # completes
-            break
-
-    def _queue_backward_renders(self, start_position: RenderingPosition):
-        """Queue backward page renders ending at the given position"""
-        current_pos = start_position
-
-        for i in range(self.buffer_size):
-            # Skip if already cached or being rendered
-            if current_pos in self.backward_buffer or current_pos in self.pending_renders:
-                # Try to get previous position from cache
-                current_pos = self.reverse_position_map.get(current_pos)
-                if not current_pos:
-                    break
-                continue
-
-            # Queue render job
-            args = (
-                self.blocks,
-                self.page_style,
-                current_pos,
-                self.current_font_scale,
-                True,
-                self.current_font_family)
-            future = self.executor.submit(_render_page_worker, args)
-            self.pending_renders[current_pos] = future
-
-            # We don't know the previous position yet, so we'll update it when the
-            # render completes
-            break
-
-    def check_completed_renders(self):
-        """Check for completed background renders and cache the results"""
-        if not self.pending_renders:
-            return
-
-        completed = []
-
-        with self.render_lock:
-            for position, future in self.pending_renders.items():
-                if future.done():
-                    try:
-                        original_pos, pickled_page, next_pos = future.result()
-
-                        # Deserialize the page
-                        page = pickle.loads(pickled_page)
-
-                        # Cache the page
-                        self.cache_page(original_pos, page, next_pos, is_backward=False)
-
-                        completed.append(position)
-
-                    except Exception as e:
-                        print(f"Background render failed for position {position}: {e}")
-                        completed.append(position)
-
-            # Remove completed renders
-            for pos in completed:
-                self.pending_renders.pop(pos, None)
-
     def invalidate_all(self):
-        """Clear all cached pages and cancel pending renders"""
-        with self.render_lock:
-            # Cancel pending renders
-            for future in self.pending_renders.values():
-                future.cancel()
-            self.pending_renders.clear()
-
-            # Clear caches
+        """Clear all cached pages"""
         self.forward_buffer.clear()
         self.backward_buffer.clear()
         self.position_map.clear()
@@ -323,7 +183,6 @@ class PageBuffer:
         return {
             'forward_buffer_size': len(self.forward_buffer),
             'backward_buffer_size': len(self.backward_buffer),
-            'pending_renders': len(self.pending_renders),
             'position_mappings': len(self.position_map),
             'reverse_position_mappings': len(self.reverse_position_map),
             'current_font_scale': self.current_font_scale,
@@ -331,28 +190,20 @@ class PageBuffer:
         }
 
     def shutdown(self):
-        """Shutdown the page buffer and clean up resources"""
-        if self.executor:
-            # Cancel pending renders
-            with self.render_lock:
-                for future in self.pending_renders.values():
-                    future.cancel()
+        """
+        Release cached pages.
 
-            # Shutdown executor
-            self.executor.shutdown(wait=True)
-            self.executor = None
-
-        # Clear all caches
+        Cheap and idempotent. There is deliberately no __del__ calling this:
+        blocking work in a finaliser is what deadlocked the interpreter at exit
+        while the process pool existed.
+        """
         self.invalidate_all()
 
-    def __del__(self):
-        """Cleanup on destruction"""
-        self.shutdown()
-
 
 class BufferedPageRenderer:
     """
-    High-level interface for buffered page rendering with automatic background caching.
+    High-level interface for page rendering with an LRU cache in front of the
+    layouter.
     """
 
     def __init__(self,
@@ -390,7 +241,7 @@ class BufferedPageRenderer:
     def render_page(self, position: RenderingPosition,
                     font_scale: float = 1.0) -> Tuple[Page, RenderingPosition]:
         """
-        Render a page with intelligent caching.
+        Render a page, serving it from cache when possible.
 
         Args:
             position: Position to render from
@@ -407,32 +258,18 @@ class BufferedPageRenderer:
         # Check cache first
         cached_page = self.buffer.get_page(position)
         if cached_page:
-            # Get next position from position map
+            # Only use the cache if we also know where the next page starts;
+            # otherwise fall through and compute it.
             next_pos = self.buffer.position_map.get(position)
-
-            # Only use cache if we have the forward position mapping
-            # Otherwise, we need to compute it
             if next_pos is not None:
-                # Start background rendering for upcoming pages
-                self.buffer.start_background_rendering(position, 'forward')
-
                 return cached_page, next_pos
 
-            # Cache hit for the page, but we don't have the forward position
-            # Fall through to compute it below
-
         # Render the page directly
         page, next_pos = self.layouter.render_page_forward(position, font_scale)
 
         # Cache the result
         self.buffer.cache_page(position, page, next_pos)
 
-        # Start background rendering
-        self.buffer.start_background_rendering(position, 'both')
-
-        # Check for completed background renders
-        self.buffer.check_completed_renders()
-
         return page, next_pos
 
     def render_page_backward(self,
@@ -440,7 +277,8 @@ class BufferedPageRenderer:
                              font_scale: float = 1.0) -> Tuple[Page,
                                                                RenderingPosition]:
         """
-        Render a page ending at the given position with intelligent caching.
+        Render a page ending at the given position, serving it from cache when
+        possible.
 
         Args:
             end_position: Position where page should end
@@ -457,32 +295,18 @@ class BufferedPageRenderer:
         # Check cache first
         cached_page = self.buffer.get_page(end_position)
         if cached_page:
-            # Get previous position from reverse position map
+            # Only use the cache if we also know where the previous page
+            # starts; otherwise fall through and compute it.
             prev_pos = self.buffer.reverse_position_map.get(end_position)
-
-            # Only use cache if we have the reverse position mapping
-            # Otherwise, we need to compute it
             if prev_pos is not None:
-                # Start background rendering for previous pages
-                self.buffer.start_background_rendering(end_position, 'backward')
-
                 return cached_page, prev_pos
 
-            # Cache hit for the page, but we don't have the reverse position
-            # Fall through to compute it below
-
         # Render the page directly
         page, start_pos = self.layouter.render_page_backward(end_position, font_scale)
 
         # Cache the result
         self.buffer.cache_page(start_pos, page, end_position, is_backward=True)
 
-        # Start background rendering
-        self.buffer.start_background_rendering(end_position, 'both')
-
-        # Check for completed background renders
-        self.buffer.check_completed_renders()
-
         return page, start_pos
 
     def set_font_family(self, font_family: Optional[BundledFont]):
@@ -516,5 +340,5 @@ class BufferedPageRenderer:
         return self.buffer.get_cache_stats()
 
     def shutdown(self):
-        """Shutdown the renderer and clean up resources"""
+        """Release cached pages"""
         self.buffer.shutdown()

pyWebLayout/style/abstract_style.py

@@ -112,7 +112,17 @@ class AbstractStyle:
         Since this is a frozen dataclass, it should be hashable by default,
         but we provide a custom implementation to ensure all fields are
         properly considered and to handle the Union types correctly.
+
+        The result is memoised on first use. Styles are used as dictionary keys
+        throughout parsing and style resolution, and five of the fields are enum
+        members whose own __hash__ is a Python-level call, so rebuilding the
+        15-tuple on every lookup was a measurable share of document parsing. The
+        class is frozen, so the value cannot go stale.
         """
+        cached = self.__dict__.get('_hash_cache')
+        if cached is not None:
+            return cached
+
         # Convert all values to hashable forms
         hashable_values = (
             self.font_family,
@@ -132,7 +142,9 @@ class AbstractStyle:
             self.parent_style_id
         )
 
-        return hash(hashable_values)
+        result = hash(hashable_values)
+        object.__setattr__(self, '_hash_cache', result)
+        return result
 
     def merge_with(self, other: 'AbstractStyle') -> 'AbstractStyle':
         """

pyproject.toml

@@ -4,24 +4,56 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pyWebLayout"
+version = "0.1.1"
 description = "A Python library for HTML-like layout and rendering"
 readme = "README.md"
-requires-python = ">=3.6"
+requires-python = ">=3.10"
 license = {file = "LICENSE"}
 authors = [
     {name = "Duncan Tourolle", email = "duncan@tourolle.paris"}
 ]
-dynamic = ["version"]
+classifiers = [
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "License :: OSI Approved :: MIT License",
+    "Operating System :: OS Independent",
+]
 dependencies = [
     "Pillow",
     "numpy",
     "pyphen",
     "beautifulsoup4",
-    "flask",
-    "ebooklib",
-    "requests"
 ]
 
+[project.urls]
+Homepage = "https://gitea.tourolle.paris/pyWebLayout"
+
+[project.optional-dependencies]
+# Loading images from http(s) URLs. concrete.image imports requests lazily and
+# degrades to an error message on the image when it is absent, so it is not a
+# hard requirement.
+remote-images = ["requests"]
+test = [
+    "pytest>=6.0",
+    "pytest-cov",
+    "flask",      # fixture HTTP server in tests/abstract/test_abstract_blocks.py
+    "werkzeug",   # make_server, same fixture
+    "ebooklib",   # builds EPUB fixtures; the reader itself uses zipfile + ElementTree
+    "requests",   # exercises the remote-images path
+]
+dev = [
+    "pyWebLayout[test,remote-images]",
+    "flake8",
+    "coverage-badge",
+    "interrogate",
+]
+
+[tool.setuptools.packages.find]
+include = ["pyWebLayout*"]
+
 [tool.coverage.run]
 source = ["pyWebLayout"]
 branch = true

setup.cfg

@@ -1,26 +1,6 @@
-[metadata]
-name = pyWebLayout
-version = 0.1.1
-author = Duncan Tourolle
-author_email = duncan@tourolle.paris
-description = A Python library for HTML-like layout and rendering
-long_description = file: README.md
-long_description_content_type = text/markdown
-url = https://gitea.tourolle.paris/pyWebLayout
-classifiers =
-    Programming Language :: Python :: 3
-    License :: OSI Approved :: MIT License
-    Operating System :: OS Independent
-
-[options]
-packages = find:
-python_requires = >=3.6
-install_requires =
-    Pillow
-    numpy
-
-[options.packages.find]
-include = pyWebLayout*
+# Packaging metadata lives in pyproject.toml ([project]), which takes
+# precedence over anything declared here. This file keeps only tool config
+# that has nowhere better to live.
 
 [flake8]
 exclude =

setup.py

@@ -1,32 +1,10 @@
-from setuptools import setup, find_packages
+"""Shim for legacy `python setup.py` invocations.
 
-setup(
-    name="pyWebLayout",
-    version="0.1.1",
-    packages=find_packages(),
-    install_requires=[
-        "Pillow",
-        "numpy",
-    ],
-    extras_require={
-        "test": [
-            "coverage>=5.0",
-        ],
-        "dev": [
-            "coverage>=5.0",
-            "pytest>=6.0",
-        ],
-    },
-    author="Duncan Tourolle",
-    author_email="duncan@tourolle.paris",
-    description="A Python library for HTML-like layout and rendering",
-    long_description=open("README.md").read(),
-    long_description_content_type="text/markdown",
-    url="https://gitea.tourolle.paris/pyWebLayout",
-    classifiers=[
-        "Programming Language :: Python :: 3",
-        "License :: OSI Approved :: MIT License",
-        "Operating System :: OS Independent",
-    ],
-    python_requires=">=3.6",
-)
+All packaging metadata lives in setup.cfg. Keeping a second copy here was an
+active hazard: keyword arguments passed to setup() override setup.cfg, so the
+two could disagree silently and the setup.py copy would win.
+"""
+
+from setuptools import setup
+
+setup()

tests/abstract/test_abstract_blocks.py

@@ -307,6 +307,8 @@ class TestImagePIL(unittest.TestCase):
 
         if FLASK_AVAILABLE and hasattr(cls, 'flask_thread'):
             cls.flask_server_running = False
+            cls.flask_server.shutdown()
+            cls.flask_server.server_close()
             cls.flask_thread.join(timeout=2)
 
     @classmethod
@@ -350,10 +352,9 @@ class TestImagePIL(unittest.TestCase):
         """Start a Flask server for URL testing."""
         import urllib.request
         import urllib.error
+        from werkzeug.serving import make_server
 
         cls.flask_app = Flask(__name__)
-        cls.flask_port = 5555  # Use a specific port for testing
-        cls.flask_server_running = True
 
         @cls.flask_app.route('/test.jpg')
         def serve_test_image():
@@ -363,15 +364,20 @@ class TestImagePIL(unittest.TestCase):
         def health_check():
             return 'OK', 200
 
-        def run_flask():
-            cls.flask_app.run(host='127.0.0.1', port=cls.flask_port, debug=False,
-                              use_reloader=False, threaded=True)
+        # Bind to an ephemeral port so concurrent/leftover test runs can't clash
+        cls.flask_server = make_server('127.0.0.1', 0, cls.flask_app, threaded=True)
+        cls.flask_port = cls.flask_server.server_port
+        cls.flask_server_running = True
 
-        cls.flask_thread = threading.Thread(target=run_flask, daemon=True)
+        cls.flask_thread = threading.Thread(target=cls.flask_server.serve_forever, daemon=True)
         cls.flask_thread.start()
 
-        # Wait for server to be ready with health check
-        max_wait = 5  # Maximum 5 seconds
+        # Wait for server to be ready with health check.
+        # Generous, because this now raises rather than falling through
+        # silently: on a loaded CI runner the accept loop can take several
+        # seconds to get scheduled, and a spurious failure here is worse than
+        # a slow one. The loop exits as soon as the server answers.
+        max_wait = 30
         wait_interval = 0.1  # Check every 100ms
         elapsed = 0
 
@@ -379,12 +385,15 @@ class TestImagePIL(unittest.TestCase):
             try:
                 with urllib.request.urlopen(f'http://127.0.0.1:{cls.flask_port}/health', timeout=1) as response:
                     if response.status == 200:
-                        break
+                        return
             except (urllib.error.URLError, ConnectionRefusedError, OSError):
                 pass
             time.sleep(wait_interval)
             elapsed += wait_interval
 
+        raise RuntimeError(
+            f"Test Flask server did not become ready on port {cls.flask_port} within {max_wait}s")
+
     def test_image_url_detection(self):
         """Test URL detection functionality."""
         img = Image()

tests/concrete/test_canvas_lifecycle.py

@@ -0,0 +1,129 @@
+"""
+Regression tests for the page draw/canvas lifecycle (spec S3).
+
+add_child invalidates the canvas but left _draw pointing at it, and the draw
+property only rebuilt when _draw was None. Callers therefore received a context
+bound to a discarded image while page._canvas stayed None - which is how images
+inside table cells ended up as grey placeholders: table_layouter passed
+canvas=None through to the cell renderer.
+
+Fixing that alone would make layout allocate a full-page canvas per line, since
+layout measures text through the page. Measurement now goes through a dedicated
+scratch context.
+"""
+
+import pytest
+from PIL import Image
+
+from pyWebLayout.abstract.block import Image as AbstractImage, Paragraph
+from pyWebLayout.abstract.inline import Word
+from pyWebLayout.concrete.page import Page
+from pyWebLayout.layout.document_layouter import DocumentLayouter
+from pyWebLayout.style import Font
+from pyWebLayout.style.page_style import PageStyle
+
+
+@pytest.fixture
+def font():
+    return Font(font_size=12)
+
+
+@pytest.fixture
+def page():
+    return Page(size=(400, 600), style=PageStyle())
+
+
+def paragraph_of(font, count=40):
+    paragraph = Paragraph(font)
+    for i in range(count):
+        paragraph.add_word(Word(f"word{i}", font))
+    return paragraph
+
+
+class TestDrawIsNeverStale:
+
+    def test_draw_matches_canvas_after_add_child(self, page, font):
+        page.draw  # force canvas creation
+        DocumentLayouter(page).layout_paragraph(paragraph_of(font))
+
+        assert page.draw.im is page._canvas.im, \
+            "draw must be bound to the page's current canvas"
+
+    def test_canvas_is_present_after_layout(self, page, font):
+        DocumentLayouter(page).layout_paragraph(paragraph_of(font))
+        page.draw
+
+        assert page._canvas is not None
+
+    def test_repeated_draw_access_is_stable(self, page):
+        first = page.draw
+        assert page.draw is first, "draw must not be rebuilt while the canvas stands"
+
+
+class TestMeasurementDoesNotAllocateCanvases:
+
+    def test_layout_allocates_no_page_canvas(self, page, font, monkeypatch):
+        calls = []
+        original = Page._create_canvas
+
+        def counting(self):
+            calls.append(1)
+            return original(self)
+
+        monkeypatch.setattr(Page, "_create_canvas", counting)
+        DocumentLayouter(page).layout_paragraph(paragraph_of(font, 400))
+
+        assert calls == [], \
+            f"layout allocated {len(calls)} full-page canvases; it should allocate none"
+
+    def test_measurement_context_is_tiny_and_matches_canvas_mode(self, page):
+        scratch = page.measurement_draw
+        assert scratch.im.size == (1, 1)
+        assert scratch.mode == Page._CANVAS_MODE
+
+    def test_measurement_context_is_stable(self, page, font):
+        first = page.measurement_draw
+        DocumentLayouter(page).layout_paragraph(paragraph_of(font))
+        assert page.measurement_draw is first, \
+            "the scratch context must survive canvas invalidation"
+
+
+class TestRenderIsRepeatable:
+
+    def test_two_renders_are_identical(self, page, font):
+        DocumentLayouter(page).layout_paragraph(paragraph_of(font))
+
+        first = page.render().copy()
+        second = page.render().copy()
+
+        assert first.tobytes() == second.tobytes()
+
+
+class TestImageInCellGetsARealCanvas:
+    """The concrete symptom: table images degraded to placeholders."""
+
+    @pytest.fixture
+    def image_path(self, tmp_path):
+        path = tmp_path / "swatch.png"
+        Image.new("RGB", (40, 30), (10, 200, 10)).save(path)
+        return str(path)
+
+    def test_table_after_paragraph_receives_a_canvas(self, page, font, image_path):
+        from pyWebLayout.abstract.block import Table, TableCell, TableRow
+        from pyWebLayout.layout.document_layouter import table_layouter
+
+        layouter = DocumentLayouter(page)
+        layouter.layout_paragraph(paragraph_of(font, 10))
+
+        table = Table()
+        row = TableRow()
+        cell = TableCell()
+        cell.add_block(AbstractImage(image_path))
+        row.add_cell(cell)
+        table.add_row(row)
+
+        # The canvas is invalidated by the preceding add_child; the table must
+        # still be handed a real one.
+        assert table_layouter(table, page) or True  # placement may fail on space
+        assert page._canvas is not None, \
+            "table layout must not run against a None canvas"

tests/concrete/test_concrete_functional.py

@@ -334,8 +334,11 @@ class TestFormFieldText(unittest.TestCase):
         """Test size property includes field area"""
         renderable = FormFieldText(self.text_field, self.font, self.mock_draw)
 
-        # Size should include label height + gap + field height
-        expected_height = renderable._style.font_size + 5 + renderable._field_height
+        # Size should include label height + gap + field height. The label's
+        # height is its ink height (ascent + descent), not the nominal font size.
+        ascent, descent = renderable._style.font.getmetrics()
+        expected_height = (ascent + descent) + FormFieldText.LABEL_GAP \
+            + renderable._field_height
         expected_width = renderable._field_width  # Use the calculated field width
 
         np.testing.assert_array_equal(

tests/concrete/test_form_field_geometry.py

@@ -0,0 +1,140 @@
+"""
+Regression tests for form field label geometry (spec S15).
+
+Text renders with a baseline anchor, so drawing the label at the field's origin
+put its glyphs above that origin - outside the box the field claims through size
+and in_object. Stacked fields therefore had each label overprinting the input box
+of the field before it.
+"""
+
+import numpy as np
+import pytest
+from PIL import Image, ImageDraw
+
+from pyWebLayout.abstract.functional import Form, FormField, FormFieldType
+from pyWebLayout.concrete.functional import FormFieldText
+from pyWebLayout.concrete.page import Page
+from pyWebLayout.layout.document_layouter import form_layouter
+from pyWebLayout.style import Font
+from pyWebLayout.style.page_style import PageStyle
+
+
+ORIGIN = (10, 40)
+FIELD_HEIGHT = 24
+
+
+@pytest.fixture
+def font():
+    return Font(font_size=12, colour=(0, 0, 0))
+
+
+@pytest.fixture
+def canvas():
+    image = Image.new("RGB", (300, 200), (255, 255, 255))
+    return image, ImageDraw.Draw(image)
+
+
+def make_field(font, draw, label="Email Address"):
+    field = FormField(name="email", field_type=FormFieldType.TEXT, label=label)
+    renderable = FormFieldText(field, font, draw, field_height=FIELD_HEIGHT)
+    renderable.set_origin(np.array(list(ORIGIN)))
+    return renderable
+
+
+def ink_rows(image, x_range, y_range):
+    pixels = image.convert("RGB").load()
+    return [y for y in y_range
+            if any(sum(pixels[x, y]) < 400 for x in x_range)]
+
+
+class TestLabelStaysInsideTheFieldBox:
+
+    def test_label_ink_is_below_the_origin(self, font, canvas):
+        image, draw = canvas
+        renderable = make_field(font, draw)
+        renderable.render()
+
+        rows = ink_rows(image, range(ORIGIN[0], ORIGIN[0] + 140),
+                        range(0, ORIGIN[1]))
+        assert not rows, \
+            f"label drew above its own origin, at rows {rows}"
+
+    def test_label_and_box_do_not_overlap(self, font, canvas):
+        image, draw = canvas
+        renderable = make_field(font, draw)
+        renderable.render()
+
+        ascent, descent = font.font.getmetrics()
+        label_bottom = ORIGIN[1] + ascent + descent
+        box_top = renderable.field_area_offset + ORIGIN[1]
+
+        assert box_top >= label_bottom, \
+            "the input box must start below the label's descenders"
+
+    def test_reported_height_covers_everything_drawn(self, font, canvas):
+        image, draw = canvas
+        renderable = make_field(font, draw)
+        renderable.render()
+
+        top, bottom = ORIGIN[1], ORIGIN[1] + int(renderable.size[1])
+        rows = ink_rows(image, range(ORIGIN[0], ORIGIN[0] + 200), range(0, 200))
+        assert min(rows) >= top, "ink above the field's declared box"
+        assert max(rows) < bottom, "ink below the field's declared box"
+
+
+class TestStackedFieldsDoNotCollide:
+
+    def test_form_layout_leaves_labels_clear(self, font):
+        page = Page(size=(300, 400), style=PageStyle())
+        form = Form("signup")
+        for name in ["Username", "Email Address", "Password"]:
+            form.add_field(FormField(name=name.lower().replace(" ", "_"),
+                                     field_type=FormFieldType.TEXT, label=name))
+
+        ok, ids = form_layouter(form, page, font)
+        assert ok and len(ids) == 3
+
+        fields = [c for c in page.children if isinstance(c, FormFieldText)]
+        assert len(fields) == 3
+
+        for earlier, later in zip(fields, fields[1:]):
+            earlier_bottom = earlier.origin[1] + earlier.size[1]
+            assert later.origin[1] >= earlier_bottom, \
+                "fields overlap: a label would print over the preceding input box"
+
+    def test_rendered_form_has_no_ink_collisions(self, font):
+        """Every field's ink stays within its own declared bounds."""
+        page = Page(size=(300, 400), style=PageStyle())
+        form = Form("signup")
+        for name in ["Username", "Email Address"]:
+            form.add_field(FormField(name=name.lower(), field_type=FormFieldType.TEXT,
+                                     label=name))
+        form_layouter(form, page, font)
+        image = page.render()
+
+        fields = [c for c in page.children if isinstance(c, FormFieldText)]
+        for field in fields:
+            top = int(field.origin[1])
+            bottom = top + int(field.size[1])
+            rows = ink_rows(image, range(int(field.origin[0]),
+                                         int(field.origin[0] + field.size[0])),
+                            range(max(0, top - 6), top))
+            assert not rows, f"ink found just above a field at y={top}"
+
+
+class TestClickTargetsFollowTheLayout:
+
+    def test_click_in_the_input_area_focuses(self, font, canvas):
+        _, draw = canvas
+        renderable = make_field(font, draw)
+
+        inside = (5, renderable.field_area_offset + FIELD_HEIGHT // 2)
+        assert renderable.handle_click(inside) is True
+        assert renderable._focused is True
+
+    def test_click_on_the_label_does_not_focus(self, font, canvas):
+        _, draw = canvas
+        renderable = make_field(font, draw)
+
+        on_label = (5, 2)
+        assert renderable.handle_click(on_label) is False

tests/concrete/test_functional_text_centering.py

@@ -0,0 +1,124 @@
+"""
+Regression tests for vertical centring of text in buttons and form fields.
+
+Both placed the baseline at `top + height/2 + descent/2`. Centring text whose
+visual height is ascent+descent inside a box of height H puts the baseline at
+`top + H/2 + (ascent-descent)/2`; the two agree only when ascent == 2*descent.
+Real fonts have a much larger ratio - DejaVu is nearer 4:1 - so the text sat
+several pixels high, hugging the top edge of the button.
+
+The button was also sized from the nominal font size rather than the text's
+actual visual height, leaving it too short to centre anything in.
+"""
+
+import numpy as np
+import pytest
+from PIL import Image, ImageDraw
+
+from pyWebLayout.abstract.functional import Button, FormField, FormFieldType
+from pyWebLayout.concrete.functional import ButtonText, FormFieldText
+from pyWebLayout.style import Font
+
+
+CANVAS = (300, 120)
+PADDING = (6, 10, 6, 10)  # top, right, bottom, left
+
+
+@pytest.fixture
+def draw_ctx():
+    image = Image.new("RGB", CANVAS, (255, 255, 255))
+    return image, ImageDraw.Draw(image)
+
+
+def ink_rows(image, box):
+    """
+    Rows within box that carry text ink.
+
+    Only the central columns are sampled: the button has rounded corners, so the
+    page background shows through at the extremes of every row and would read as
+    white text on all of them.
+    """
+    x0, y0, x1, y1 = box
+    inset = (x1 - x0) // 4
+    pixels = image.convert("RGB").load()
+    rows = []
+    for y in range(y0, y1):
+        for x in range(x0 + inset, x1 - inset):
+            r, g, b = pixels[x, y]
+            # Button text is white on a blue fill; look for near-white ink.
+            if r > 240 and g > 240 and b > 240:
+                rows.append(y)
+                break
+    return rows
+
+
+class TestButtonTextCentring:
+
+    @pytest.mark.parametrize("font_size", [10, 14, 20])
+    def test_text_is_vertically_centred(self, draw_ctx, font_size):
+        image, draw = draw_ctx
+        font = Font(font_size=font_size, colour=(255, 255, 255))
+        button = ButtonText(Button(label="Save Document", callback=lambda p: None),
+                            font, draw, padding=PADDING)
+        button.set_origin(np.array([20, 20]))
+        button.render()
+
+        x0, y0 = 20, 20
+        x1 = x0 + int(button.size[0])
+        y1 = y0 + int(button.size[1])
+        rows = ink_rows(image, (x0, y0, x1, y1))
+        assert rows, "the button should have visible text"
+
+        gap_above = min(rows) - y0
+        gap_below = y1 - max(rows) - 1
+
+        assert abs(gap_above - gap_below) <= 2, (
+            f"text not centred at size {font_size}: "
+            f"{gap_above}px above, {gap_below}px below")
+
+    def test_button_is_tall_enough_for_its_text(self):
+        font = Font(font_size=14, colour=(255, 255, 255))
+        image = Image.new("RGB", CANVAS, (255, 255, 255))
+        draw = ImageDraw.Draw(image)
+        button = ButtonText(Button(label="Cancel", callback=lambda p: None),
+                            font, draw, padding=PADDING)
+
+        ascent, descent = font.font.getmetrics()
+        assert int(button.size[1]) >= ascent + descent + PADDING[0] + PADDING[2], \
+            "button height must accommodate the text's visual height, not the nominal size"
+
+    def test_text_stays_inside_the_button(self, draw_ctx):
+        image, draw = draw_ctx
+        font = Font(font_size=14, colour=(255, 255, 255))
+        button = ButtonText(Button(label="Save Document", callback=lambda p: None),
+                            font, draw, padding=PADDING)
+        button.set_origin(np.array([20, 20]))
+        button.render()
+
+        y0, y1 = 20, 20 + int(button.size[1])
+        rows = ink_rows(image, (20, y0, 20 + int(button.size[0]), y1))
+        assert min(rows) >= y0, "text escaped above the button"
+        assert max(rows) < y1, "text escaped below the button"
+
+
+class TestFormFieldValueCentring:
+
+    def test_value_is_centred_in_the_input_box(self):
+        image = Image.new("RGB", (300, 120), (0, 0, 0))
+        draw = ImageDraw.Draw(image)
+        font = Font(font_size=12, colour=(0, 0, 0))
+        field = FormField(name="who", field_type=FormFieldType.TEXT, value="Hello")
+        renderable = FormFieldText(field, font, draw, field_height=28)
+        renderable.set_origin(np.array([10, 10]))
+        renderable.render()
+
+        field_y = 10 + font.font_size + 5
+        pixels = image.convert("RGB").load()
+        rows = [y for y in range(field_y, field_y + 28)
+                if any(pixels[x, y] == (0, 0, 0) for x in range(12, 200))]
+        assert rows, "the field value should be visible"
+
+        gap_above = min(rows) - field_y
+        gap_below = (field_y + 28) - max(rows) - 1
+        assert abs(gap_above - gap_below) <= 3, (
+            f"field value not centred: {gap_above}px above, {gap_below}px below")

tests/io_tests/test_inline_content_in_containers.py

@@ -0,0 +1,169 @@
+"""
+Regression tests for inline content inside block containers (spec S1).
+
+Inline tags are registered to ignore_handler because they are meant to be
+consumed by extract_text_content. Only <p> and <h1>-<h6> ever called it, so
+every other container - div, li, td, th, blockquote - iterated its children as
+blocks, and inline tags returned None. Their text was silently discarded, and
+bare text nodes each became a separate paragraph.
+"""
+
+import pytest
+
+from pyWebLayout.abstract.block import (
+    HList,
+    Paragraph,
+    Quote,
+    Table,
+)
+from pyWebLayout.abstract.inline import LinkedWord, Word
+from pyWebLayout.io.readers.html_extraction import parse_html_string
+
+
+def words_of(block):
+    return [w.text for w in getattr(block, 'words', [])]
+
+
+def all_words(blocks):
+    out = []
+    for block in blocks:
+        out.extend(words_of(block))
+    return out
+
+
+def cell_blocks(table):
+    for _, row in table.all_rows():
+        for cell in row.cells():
+            yield list(cell.blocks())
+
+
+EXPECTED = ["hello", "world", "again"]
+
+
+class TestInlineContentIsKept:
+    """The same markup must survive in every container."""
+
+    def test_paragraph_control(self):
+        """<p> already worked - this is the reference behaviour."""
+        blocks = parse_html_string("<p>hello <b>world</b> again</p>")
+        assert all_words(blocks) == EXPECTED
+
+    def test_div(self):
+        blocks = parse_html_string("<div>hello <b>world</b> again</div>")
+        assert all_words(blocks) == EXPECTED
+
+    def test_list_item(self):
+        blocks = parse_html_string("<ul><li>hello <b>world</b> again</li></ul>")
+        hlist = next(b for b in blocks if isinstance(b, HList))
+        item = list(hlist.items())[0]
+        assert all_words(item.blocks()) == EXPECTED
+
+    def test_table_cell(self):
+        blocks = parse_html_string(
+            "<table><tr><td>hello <b>world</b> again</td></tr></table>")
+        table = next(b for b in blocks if isinstance(b, Table))
+        assert all_words(next(cell_blocks(table))) == EXPECTED
+
+    def test_table_header_cell(self):
+        blocks = parse_html_string(
+            "<table><tr><th>hello <b>world</b> again</th></tr></table>")
+        table = next(b for b in blocks if isinstance(b, Table))
+        assert all_words(next(cell_blocks(table))) == EXPECTED
+
+    def test_blockquote(self):
+        blocks = parse_html_string("<blockquote>hello <b>world</b> again</blockquote>")
+        quote = next(b for b in blocks if isinstance(b, Quote))
+        assert all_words(quote.blocks()) == EXPECTED
+
+
+class TestInlineRunsCoalesce:
+    """A run of inline content is one paragraph, not one per text node."""
+
+    def test_div_yields_a_single_paragraph(self):
+        blocks = parse_html_string("<div>a <b>b</b> c</div>")
+        paragraphs = [b for b in blocks if isinstance(b, Paragraph)]
+        assert len(paragraphs) == 1, f"expected one paragraph, got {len(blocks)} blocks"
+        assert words_of(paragraphs[0]) == ["a", "b", "c"]
+
+    def test_cell_yields_a_single_paragraph(self):
+        blocks = parse_html_string("<table><tr><td>a <b>b</b> c</td></tr></table>")
+        table = next(b for b in blocks if isinstance(b, Table))
+        cell = next(cell_blocks(table))
+        assert len(cell) == 1
+        assert words_of(cell[0]) == ["a", "b", "c"]
+
+    def test_block_child_splits_the_run(self):
+        """Inline runs either side of a block child stay separate, in order."""
+        blocks = parse_html_string(
+            "<table><tr><td>before<p>middle</p>after</td></tr></table>")
+        table = next(b for b in blocks if isinstance(b, Table))
+        cell = next(cell_blocks(table))
+        assert [words_of(b) for b in cell] == [["before"], ["middle"], ["after"]]
+
+    def test_line_break_splits_the_run(self):
+        blocks = parse_html_string("<div>first<br>second</div>")
+        paragraphs = [b for b in blocks if isinstance(b, Paragraph)]
+        assert [words_of(p) for p in paragraphs] == [["first"], ["second"]]
+
+    def test_whitespace_between_blocks_makes_no_paragraph(self):
+        blocks = parse_html_string("<div>\n  <p>one</p>\n  <p>two</p>\n</div>")
+        assert [words_of(b) for b in blocks] == [["one"], ["two"]]
+
+
+class TestLinksSurvive:
+    """<a href> must produce LinkedWord wherever it appears."""
+
+    def test_link_in_cell(self):
+        blocks = parse_html_string(
+            '<table><tr><td><a href="http://x">link</a> text</td></tr></table>')
+        table = next(b for b in blocks if isinstance(b, Table))
+        cell = next(cell_blocks(table))
+        found = [w for b in cell for w in getattr(b, 'words', [])]
+
+        assert [w.text for w in found] == ["link", "text"]
+        linked = [w for w in found if isinstance(w, LinkedWord)]
+        assert len(linked) == 1
+        assert linked[0].location == "http://x"
+
+    def test_link_in_div(self):
+        blocks = parse_html_string('<div>see <a href="#s2">Section 2</a> now</div>')
+        found = [w for b in blocks for w in getattr(b, 'words', [])]
+        assert [w.text for w in found] == ["see", "Section", "2", "now"]
+        assert all(isinstance(w, LinkedWord) for w in found[1:3])
+
+    def test_link_in_list_item(self):
+        blocks = parse_html_string('<ul><li><a href="u">click</a> here</li></ul>')
+        hlist = next(b for b in blocks if isinstance(b, HList))
+        item = hlist._items[0]
+        found = [w for b in item.blocks() for w in getattr(b, 'words', [])]
+        assert [w.text for w in found] == ["click", "here"]
+        assert isinstance(found[0], LinkedWord)
+
+
+class TestNestedContainers:
+
+    def test_div_in_div(self):
+        blocks = parse_html_string("<div>outer <div>inner</div> tail</div>")
+        assert [words_of(b) for b in blocks] == [["outer"], ["inner"], ["tail"]]
+
+    def test_block_children_still_pass_through(self):
+        blocks = parse_html_string("<div><h1>Title</h1><p>Body</p></div>")
+        assert len(blocks) == 2
+        assert words_of(blocks[0]) == ["Title"]
+        assert words_of(blocks[1]) == ["Body"]
+
+    def test_cell_containing_a_list(self):
+        blocks = parse_html_string(
+            "<table><tr><td>intro<ul><li>item</li></ul></td></tr></table>")
+        table = next(b for b in blocks if isinstance(b, Table))
+        cell = next(cell_blocks(table))
+        assert isinstance(cell[0], Paragraph)
+        assert words_of(cell[0]) == ["intro"]
+        assert isinstance(cell[1], HList)
+
+
+class TestComments:
+
+    def test_comment_text_is_not_content(self):
+        blocks = parse_html_string("<div>real<!-- hidden note -->text</div>")
+        assert all_words(blocks) == ["real", "text"]

tests/layout/test_backward_navigation.py

@@ -0,0 +1,171 @@
+"""
+Regression tests for backward page navigation (spec S16).
+
+The previous page of P is the position q for which laying out forward from q ends
+exactly at P. The old implementation searched for q by guessing a block index and
+bisecting, with word_index pinned to 0 - so a page starting mid-paragraph was not
+in the search space at all. It exhausted its ten iterations and fell back to a
+position that was not the previous page, typically the start of the document.
+"""
+
+import pytest
+
+from pyWebLayout.abstract.block import Paragraph
+from pyWebLayout.abstract.inline import Word
+from pyWebLayout.layout.ereader_layout import BidirectionalLayouter, RenderingPosition
+from pyWebLayout.style import Font
+from pyWebLayout.style.page_style import PageStyle
+
+
+PAGE_SIZE = (800, 600)
+
+
+@pytest.fixture
+def font():
+    return Font(font_size=16)
+
+
+def paragraph(font, count, tag):
+    block = Paragraph(font)
+    for i in range(count):
+        block.add_word(Word(f"{tag}{i}", font))
+    return block
+
+
+@pytest.fixture
+def long_document(font):
+    """Short paragraphs around one that spans several pages."""
+    return [
+        paragraph(font, 60, "a"),
+        paragraph(font, 80, "b"),
+        paragraph(font, 1200, "long"),
+        paragraph(font, 70, "c"),
+        paragraph(font, 90, "d"),
+    ]
+
+
+@pytest.fixture
+def block_document(font):
+    """Many small blocks, so every page starts on a block boundary."""
+    return [paragraph(font, 40, f"p{i}") for i in range(40)]
+
+
+def forward_chain(layouter, limit=30):
+    """The page start positions a reader would visit going forward."""
+    starts = []
+    pos = RenderingPosition()
+    for _ in range(limit):
+        starts.append(pos)
+        _, nxt = layouter.render_page_forward(pos, 1.0)
+        if nxt.block_index >= len(layouter.blocks):
+            break
+        if (nxt.block_index, nxt.word_index) == (pos.block_index, pos.word_index):
+            pytest.fail("forward pagination made no progress")
+        pos = nxt
+    return starts
+
+
+def key(position):
+    return (position.chapter_index, position.block_index, position.word_index)
+
+
+class TestBackwardMatchesForward:
+    """The defining invariant: forward from the answer lands exactly on P."""
+
+    @pytest.mark.parametrize("document", ["long_document", "block_document"])
+    def test_previous_page_is_the_forward_predecessor(self, document, request):
+        blocks = request.getfixturevalue(document)
+        layouter = BidirectionalLayouter(blocks, PageStyle(), PAGE_SIZE)
+        starts = forward_chain(layouter)
+        assert len(starts) > 2, "need a few pages to test against"
+
+        for i in range(1, len(starts)):
+            _, got = layouter.render_page_backward(starts[i], 1.0)
+            assert key(got) == key(starts[i - 1]), (
+                f"page {i}: expected to land on page {i - 1} "
+                f"{key(starts[i - 1])}, got {key(got)}")
+
+    def test_result_lays_out_to_the_target(self, long_document):
+        """Independent of the recorded chain: replaying the answer must reach P."""
+        layouter = BidirectionalLayouter(long_document, PageStyle(), PAGE_SIZE)
+        starts = forward_chain(layouter)
+
+        for target in starts[1:]:
+            _, start = layouter.render_page_backward(target, 1.0)
+            _, end = layouter.render_page_forward(start, 1.0)
+            assert key(end) == key(target), (
+                f"a page starting at {key(start)} ends at {key(end)}, "
+                f"not at the requested {key(target)}")
+
+    def test_mid_paragraph_targets_are_reachable(self, long_document):
+        """The specific regression: starts inside a block, not on its boundary."""
+        layouter = BidirectionalLayouter(long_document, PageStyle(), PAGE_SIZE)
+        starts = forward_chain(layouter)
+        mid = [s for s in starts if s.word_index > 0]
+        assert mid, "this document should paginate mid-paragraph"
+
+        for target in mid:
+            _, got = layouter.render_page_backward(target, 1.0)
+            assert key(got) != (0, 0, 0) or key(target) == key(starts[1]), \
+                "backward navigation fell back to the document start"
+
+
+class TestRoundTrip:
+
+    def test_forward_then_back_returns_to_the_same_place(self, long_document):
+        layouter = BidirectionalLayouter(long_document, PageStyle(), PAGE_SIZE)
+        pos = RenderingPosition()
+
+        for _ in range(4):
+            _, nxt = layouter.render_page_forward(pos, 1.0)
+            _, back = layouter.render_page_backward(nxt, 1.0)
+            assert key(back) == key(pos), \
+                f"round trip drifted: {key(pos)} -> {key(nxt)} -> {key(back)}"
+            pos = nxt
+
+
+class TestEdges:
+
+    def test_at_document_start_stays_there(self, long_document):
+        layouter = BidirectionalLayouter(long_document, PageStyle(), PAGE_SIZE)
+        _, got = layouter.render_page_backward(RenderingPosition(), 1.0)
+        assert key(got) == (0, 0, 0)
+
+    def test_second_page_goes_back_to_the_first(self, long_document):
+        layouter = BidirectionalLayouter(long_document, PageStyle(), PAGE_SIZE)
+        _, second = layouter.render_page_forward(RenderingPosition(), 1.0)
+        _, got = layouter.render_page_backward(second, 1.0)
+        assert key(got) == (0, 0, 0)
+
+    def test_empty_document_is_safe(self):
+        layouter = BidirectionalLayouter([], PageStyle(), PAGE_SIZE)
+        page, got = layouter.render_page_backward(RenderingPosition(), 1.0)
+        assert page is not None
+        assert key(got) == (0, 0, 0)
+
+
+class TestCost:
+
+    def test_backward_is_not_wildly_more_expensive_than_forward(self, long_document):
+        """The old path burned ten full layouts per call and still got it wrong."""
+        layouter = BidirectionalLayouter(long_document, PageStyle(), PAGE_SIZE)
+        starts = forward_chain(layouter)
+
+        calls = {"n": 0}
+        original = BidirectionalLayouter.render_page_forward
+
+        def counting(self, position, font_scale=1.0):
+            calls["n"] += 1
+            return original(self, position, font_scale)
+
+        BidirectionalLayouter.render_page_forward = counting
+        try:
+            worst = 0
+            for target in starts[1:]:
+                calls["n"] = 0
+                layouter.render_page_backward(target, 1.0)
+                worst = max(worst, calls["n"])
+        finally:
+            BidirectionalLayouter.render_page_forward = original
+
+        assert worst <= 10, f"backward navigation cost {worst} forward layouts"

tests/layout/test_ereader_highlights.py

@@ -0,0 +1,168 @@
+"""
+Tests for the highlight API on EreaderLayoutManager (R7).
+
+core/highlight.py was fully implemented and tested but unreachable: the manager
+had no highlight API, so highlighting could not be used through the library's
+own interface. These tests cover the wiring, not the dataclass - that is
+tests/core/test_highlight.py.
+"""
+
+import pytest
+
+from pyWebLayout.core.highlight import Highlight, HighlightColor
+from pyWebLayout.io.readers.html_extraction import parse_html_string
+from pyWebLayout.layout.ereader_manager import EreaderLayoutManager
+
+
+@pytest.fixture
+def manager(tmp_path):
+    blocks = parse_html_string(
+        "<p>" + " ".join(f"word{i}" for i in range(300)) + "</p>")
+    manager = EreaderLayoutManager(blocks, page_size=(400, 600),
+                                   document_id="highlights",
+                                   bookmarks_dir=str(tmp_path))
+    yield manager
+    manager.shutdown()
+
+
+def text_points(page, limit=None):
+    """Points on the rendered page that land on a text object."""
+    found = []
+    for y in range(0, 120, 2):
+        for x in range(0, 400, 2):
+            result = page.query_point((x, y))
+            if result is not None and result.object_type == "text" and result.text:
+                found.append((x, y))
+                if limit and len(found) >= limit:
+                    return found
+    return found
+
+
+@pytest.fixture
+def point_on_text(manager):
+    page = manager.get_current_page()
+    page.render()
+    return text_points(page, limit=1)[0]
+
+
+class TestHighlightPoint:
+    def test_highlighting_a_word_returns_a_stored_highlight(self, manager, point_on_text):
+        highlight = manager.highlight_point(point_on_text)
+
+        assert isinstance(highlight, Highlight)
+        assert highlight.text
+        assert manager.list_highlights() == [highlight]
+
+    def test_colour_note_and_tags_are_kept(self, manager, point_on_text):
+        highlight = manager.highlight_point(
+            point_on_text, color=HighlightColor.GREEN.value,
+            note="a note", tags=["review"])
+
+        assert highlight.color == HighlightColor.GREEN.value
+        assert highlight.note == "a note"
+        assert highlight.tags == ["review"]
+
+    def test_highlighting_empty_space_returns_none(self, manager):
+        manager.get_current_page().render()
+
+        assert manager.highlight_point((399, 599)) is None
+        assert manager.list_highlights() == []
+
+    def test_the_originating_position_is_recorded(self, manager, point_on_text):
+        highlight = manager.highlight_point(point_on_text)
+
+        assert highlight.position == manager.current_position.to_dict()
+
+
+class TestHighlightRange:
+    def test_a_selection_spans_multiple_words(self, manager):
+        page = manager.get_current_page()
+        page.render()
+        points = text_points(page)
+
+        highlight = manager.highlight_range(points[0], points[-1])
+
+        assert highlight is not None
+        assert len(highlight.text.split()) > 1
+        assert len(highlight.bounds) > 1
+
+    def test_a_selection_hitting_no_text_returns_none(self, manager):
+        manager.get_current_page().render()
+
+        assert manager.highlight_range((398, 596), (399, 599)) is None
+
+
+class TestHighlightsAreScopedToTheirPage:
+    def test_current_page_highlights_do_not_leak_across_pages(self, manager, point_on_text):
+        manager.highlight_point(point_on_text)
+        assert len(manager.get_highlights_for_current_page()) == 1
+
+        manager.next_page()
+
+        assert manager.get_highlights_for_current_page() == []
+        assert len(manager.list_highlights()) == 1, "still in the document, just not here"
+
+    def test_returning_to_the_page_finds_it_again(self, manager, point_on_text):
+        highlight = manager.highlight_point(point_on_text)
+        manager.next_page()
+        manager.previous_page()
+
+        assert manager.get_highlights_for_current_page() == [highlight]
+
+
+class TestPersistence:
+    def test_highlights_survive_a_restart(self, manager, point_on_text, tmp_path):
+        highlight = manager.highlight_point(point_on_text, note="kept")
+        manager.shutdown()
+
+        reopened = EreaderLayoutManager(
+            manager.blocks, page_size=(400, 600), document_id="highlights",
+            bookmarks_dir=str(tmp_path))
+        try:
+            restored = reopened.list_highlights()
+            assert len(restored) == 1
+            assert restored[0].id == highlight.id
+            assert restored[0].note == "kept"
+            assert restored[0].position == highlight.position
+        finally:
+            reopened.shutdown()
+
+    def test_highlights_share_the_bookmarks_directory_by_default(self, manager,
+                                                                 point_on_text, tmp_path):
+        manager.highlight_point(point_on_text)
+
+        assert (tmp_path / "highlights_highlights.json").exists()
+
+    def test_removing_a_highlight_persists(self, manager, point_on_text, tmp_path):
+        highlight = manager.highlight_point(point_on_text)
+
+        assert manager.remove_highlight(highlight.id) is True
+        assert manager.remove_highlight(highlight.id) is False
+
+        reopened = EreaderLayoutManager(
+            manager.blocks, page_size=(400, 600), document_id="highlights",
+            bookmarks_dir=str(tmp_path))
+        try:
+            assert reopened.list_highlights() == []
+        finally:
+            reopened.shutdown()
+
+    def test_clear_removes_everything(self, manager, point_on_text):
+        manager.highlight_point(point_on_text)
+
+        manager.clear_highlights()
+
+        assert manager.list_highlights() == []
+
+    def test_a_corrupt_store_does_not_stop_the_book_opening(self, tmp_path):
+        (tmp_path / "broken_highlights.json").write_text("{not json")
+        blocks = parse_html_string("<p>hello world</p>")
+
+        manager = EreaderLayoutManager(blocks, page_size=(400, 600),
+                                       document_id="broken",
+                                       bookmarks_dir=str(tmp_path))
+        try:
+            assert manager.list_highlights() == []
+            assert manager.get_current_page() is not None
+        finally:
+            manager.shutdown()

tests/layout/test_ereader_interaction.py

@@ -0,0 +1,129 @@
+"""
+Tests for pointer interaction on EreaderLayoutManager (R7).
+
+concrete/interaction_handler.py was 310 lines reachable only from
+examples/07_pressed_state_demo.py - no library code, no tests. These cover the
+wiring; the press/hover state on the elements themselves lives in
+tests/concrete/.
+"""
+
+import pytest
+from PIL import Image
+
+from pyWebLayout.io.readers.html_extraction import parse_html_string
+from pyWebLayout.layout.ereader_manager import EreaderLayoutManager
+
+
+@pytest.fixture
+def manager(tmp_path):
+    blocks = parse_html_string(
+        '<p>Tap <a href="action:go">this link</a> please.</p>'
+        '<p>' + " ".join(f"w{i}" for i in range(400)) + '</p>')
+    manager = EreaderLayoutManager(blocks, page_size=(400, 600),
+                                   document_id="interaction",
+                                   bookmarks_dir=str(tmp_path))
+    yield manager
+    manager.shutdown()
+
+
+@pytest.fixture
+def link_point(manager):
+    """A page coordinate that lands on the interactive link."""
+    page = manager.get_current_page()
+    page.render()
+    for y in range(0, 120, 2):
+        for x in range(0, 400, 2):
+            result = page.query_point((x, y))
+            if result is not None and result.is_interactive:
+                return (x, y)
+    pytest.fail("fixture document rendered no interactive element")
+
+
+EMPTY_POINT = (399, 599)
+
+
+class TestHover:
+    def test_hovering_an_element_produces_a_frame(self, manager, link_point):
+        assert isinstance(manager.handle_hover(link_point), Image.Image)
+
+    def test_hovering_the_same_element_again_reports_no_change(self, manager, link_point):
+        manager.handle_hover(link_point)
+
+        assert manager.handle_hover(link_point) is None, \
+            "an unchanged hover should not force the caller to redraw"
+
+    def test_moving_off_the_element_clears_the_hover(self, manager, link_point):
+        manager.handle_hover(link_point)
+
+        assert isinstance(manager.handle_hover(EMPTY_POINT), Image.Image)
+
+
+class TestPress:
+    def test_pressing_an_element_produces_a_frame(self, manager, link_point):
+        assert isinstance(manager.handle_touch_down(link_point), Image.Image)
+
+    def test_pressing_empty_space_does_nothing(self, manager):
+        manager.get_current_page().render()
+
+        assert manager.handle_touch_down(EMPTY_POINT) is None
+
+    def test_release_runs_the_link_action(self, manager, link_point):
+        manager.handle_touch_down(link_point)
+
+        frame, result = manager.handle_touch_up(link_point)
+
+        assert isinstance(frame, Image.Image)
+        assert result == "action:go"
+
+    def test_release_without_a_press_is_a_no_op(self, manager):
+        manager.get_current_page().render()
+
+        assert manager.handle_touch_up(EMPTY_POINT) == (None, None)
+
+    def test_a_full_press_release_cycle_leaves_no_state(self, manager, link_point):
+        manager.handle_touch_down(link_point)
+        manager.handle_touch_up(link_point)
+
+        assert manager.handle_touch_up(link_point) == (None, None)
+
+
+class TestStateFollowsTheDisplayedPage:
+    def test_navigating_rebinds_the_state_machine(self, manager, link_point):
+        before = manager._interaction_state()
+
+        manager.next_page()
+
+        assert manager._interaction_state() is not before, \
+            "press state belongs to one rendered page"
+
+    def test_state_survives_repeated_access_on_one_page(self, manager, link_point):
+        assert manager._interaction_state() is manager._interaction_state()
+
+    def test_reset_is_safe_before_any_interaction(self, manager):
+        manager.reset_interaction_state()  # must not raise
+
+    def test_reset_clears_a_pending_press(self, manager, link_point):
+        manager.handle_touch_down(link_point)
+
+        manager.reset_interaction_state()
+
+        assert manager.handle_touch_up(link_point) == (None, None)
+
+
+class TestPressedRenderingRegression:
+    """
+    LinkText.render passed [origin, origin + size] - two numpy arrays - to
+    PIL's draw.rectangle, which needs a flat four-scalar box. Rendering any
+    hovered or pressed link raised TypeError. Nothing caught it because the
+    only caller was an example.
+    """
+
+    def test_rendering_a_hovered_link_does_not_raise(self, manager, link_point):
+        manager.handle_hover(link_point)
+
+        assert isinstance(manager.get_current_page().render(), Image.Image)
+
+    def test_rendering_a_pressed_link_does_not_raise(self, manager, link_point):
+        manager.handle_touch_down(link_point)
+
+        assert isinstance(manager.get_current_page().render(), Image.Image)

tests/layout/test_ereader_layout.py

@@ -570,30 +570,6 @@ class TestBidirectionalLayouter:
         # Should return same block
         assert scaled == paragraph
 
-    def test_estimate_page_start(self):
-        """Test estimation of page start position."""
-        layouter = BidirectionalLayouter([], PageStyle())
-
-        end_pos = RenderingPosition(chapter_index=0, block_index=20, word_index=0)
-
-        estimated = layouter._estimate_page_start(end_pos, 1.0)
-
-        # Should estimate some blocks before the end position
-        assert estimated.block_index < end_pos.block_index
-        assert estimated.block_index >= 0
-
-    def test_estimate_page_start_with_font_scale(self):
-        """Test that font scale affects page start estimation."""
-        layouter = BidirectionalLayouter([], PageStyle())
-
-        end_pos = RenderingPosition(chapter_index=0, block_index=20, word_index=0)
-
-        est_normal = layouter._estimate_page_start(end_pos, 1.0)
-        est_large = layouter._estimate_page_start(end_pos, 2.0)
-
-        # Larger font should estimate fewer blocks
-        assert est_large.block_index >= est_normal.block_index
-
     def test_scale_block_fonts_paragraph(self, sample_font):
         """Test scaling fonts in a paragraph block."""
         layouter = BidirectionalLayouter([], PageStyle())
@@ -784,50 +760,6 @@ class TestBidirectionalLayouter:
         # Start position should be before or at end position
         assert start_pos.block_index <= end_position.block_index
 
-    def test_adjust_start_estimate_overshot(self):
-        """Test adjustment when forward render overshoots target."""
-        layouter = BidirectionalLayouter([], PageStyle())
-
-        current_start = RenderingPosition(block_index=5)
-        target_end = RenderingPosition(block_index=10)
-        actual_end = RenderingPosition(block_index=12)  # Overshot (went too far)
-
-        adjusted = layouter._adjust_start_estimate(
-            current_start, target_end, actual_end)
-
-        # Overshot means we rendered too far forward
-        # So we need to start EARLIER (decrease block_index) to not go as far
-        assert adjusted.block_index < current_start.block_index
-
-    def test_adjust_start_estimate_undershot(self):
-        """Test adjustment when forward render undershoots target."""
-        layouter = BidirectionalLayouter([], PageStyle())
-
-        current_start = RenderingPosition(block_index=5)
-        target_end = RenderingPosition(block_index=10)
-        actual_end = RenderingPosition(block_index=8)  # Undershot (didn't go far enough)
-
-        adjusted = layouter._adjust_start_estimate(
-            current_start, target_end, actual_end)
-
-        # Undershot means we didn't render far enough forward
-        # So we need to start LATER (increase block_index) to include more content
-        assert adjusted.block_index > current_start.block_index
-
-    def test_adjust_start_estimate_exact(self):
-        """Test adjustment when forward render hits target exactly."""
-        layouter = BidirectionalLayouter([], PageStyle())
-
-        current_start = RenderingPosition(block_index=5)
-        target_end = RenderingPosition(block_index=10)
-        actual_end = RenderingPosition(block_index=10)  # Exact
-
-        adjusted = layouter._adjust_start_estimate(
-            current_start, target_end, actual_end)
-
-        # Should return same or similar position
-        assert adjusted.block_index >= 0
-
     def test_layout_paragraph_on_page_with_pretext(
             self, sample_font, sample_page_style):
         """Test paragraph layout with pretext (hyphenated word continuation)."""
@@ -899,5 +831,43 @@ class TestBidirectionalLayouter:
         assert next_pos == position  # No progress possible
 
 
+class TestNoPageMonkeyPatching:
+    """
+    R5: importing this module used to run _add_page_methods(), which attached
+    can_fit_line/available_width to Page if they were absent. They are not
+    absent, so it never fired - but its can_fit_line took (line_height) and
+    ignored descenders, while Page's takes (baseline_spacing, ascent, descent).
+    Had Page's ever been renamed, the import would have silently reinstated the
+    pre-S2 clipping bug from another package.
+    """
+
+    def test_module_does_not_patch_page(self):
+        import pyWebLayout.layout.ereader_layout as ereader_layout
+
+        assert not hasattr(ereader_layout, '_add_page_methods')
+
+    def test_page_owns_its_geometry_methods(self):
+        from pyWebLayout.concrete.page import Page
+
+        assert 'can_fit_line' in vars(Page)
+        assert 'available_width' in vars(Page)
+
+    def test_can_fit_line_still_accounts_for_descenders(self, sample_page_style):
+        """
+        The patched version took a single line_height and had no way to express
+        descent, so a descender hanging past the content box counted as fitting.
+        """
+        from pyWebLayout.concrete.page import Page
+
+        page = Page(size=(200, 100), style=sample_page_style)
+        content_y, content_h = page.content_rect[1], page.content_rect[3]
+        available = content_y + content_h - page._current_y_offset
+
+        assert page.can_fit_line(0, ascent=available, descent=0)
+        assert not page.can_fit_line(0, ascent=available, descent=1), \
+            "a descender past the content box must not be reported as fitting"
+        assert page.can_fit_line(0, ascent=available - 1, descent=1)
+
+
 if __name__ == "__main__":
     pytest.main([__file__, "-v"])

tests/layout/test_font_scaling.py

@@ -0,0 +1,262 @@
+"""
+Tests for font scaling in the ereader layout path (R3).
+
+_scale_block_fonts rebuilds a block with scaled fonts. It used to construct a
+plain Word for every word, which downgraded LinkedWord and silently discarded
+every hyperlink in the document as soon as the reader changed font size. It
+also handled only Paragraph and Heading, so quotes, lists and tables kept their
+original size while the text around them reflowed.
+"""
+
+import tempfile
+
+import pytest
+
+from pyWebLayout.abstract.block import Paragraph, Heading, Quote, HList, Table
+from pyWebLayout.abstract.inline import LinkedWord, Word
+from pyWebLayout.concrete.functional import LinkText
+from pyWebLayout.io.readers.html_extraction import parse_html_string
+from pyWebLayout.layout.ereader_layout import BidirectionalLayouter
+from pyWebLayout.layout.ereader_manager import EreaderLayoutManager
+from pyWebLayout.style import Font
+from pyWebLayout.style.page_style import PageStyle
+
+
+HTML = """
+<p>Go to <a href="http://example.com" title="Tooltip">this link</a> now.</p>
+<blockquote><p>Quoted <a href="http://q.example">qlink</a> text.</p></blockquote>
+<ul><li>item <a href="http://l.example">llink</a> one</li></ul>
+<table>
+  <thead><tr><th>head <a href="http://h.example">hlink</a></th></tr></thead>
+  <tbody><tr><td>cell <a href="http://c.example">clink</a></td></tr></tbody>
+</table>
+"""
+
+
+def collect_links(block, out=None):
+    """Every LinkedWord reachable in a block, at any nesting depth."""
+    out = [] if out is None else out
+    if isinstance(block, Paragraph):          # covers Heading
+        for _, word in block.words_iter():
+            if isinstance(word, LinkedWord):
+                out.append(word)
+    elif isinstance(block, Quote):
+        for child in block.blocks():
+            collect_links(child, out)
+    elif isinstance(block, HList):
+        for item in block.items():
+            for child in item.blocks():
+                collect_links(child, out)
+    elif isinstance(block, Table):
+        for rows in (block.header_rows(), block.body_rows(), block.footer_rows()):
+            for row in rows:
+                for cell in row.cells():
+                    for child in cell.blocks():
+                        collect_links(child, out)
+    return out
+
+
+def collect_sizes(block, out=None):
+    """Every font size reachable in a block, at any nesting depth."""
+    out = [] if out is None else out
+    if isinstance(block, Paragraph):
+        for _, word in block.words_iter():
+            out.append(word.style.font_size)
+    elif isinstance(block, Quote):
+        for child in block.blocks():
+            collect_sizes(child, out)
+    elif isinstance(block, HList):
+        for item in block.items():
+            for child in item.blocks():
+                collect_sizes(child, out)
+    elif isinstance(block, Table):
+        for rows in (block.header_rows(), block.body_rows(), block.footer_rows()):
+            for row in rows:
+                for cell in row.cells():
+                    for child in cell.blocks():
+                        collect_sizes(child, out)
+    return out
+
+
+@pytest.fixture
+def blocks():
+    return parse_html_string(HTML)
+
+
+@pytest.fixture
+def layouter(blocks):
+    return BidirectionalLayouter(blocks, PageStyle(), (400, 600))
+
+
+# ============================================================================
+# Word.with_style
+# ============================================================================
+
+class TestWithStyle:
+    def test_word_keeps_its_text_and_takes_the_new_font(self):
+        word = Word("hello", Font(font_size=16))
+
+        copy = word.with_style(Font(font_size=24))
+
+        assert copy.text == "hello"
+        assert copy.style.font_size == 24
+        assert word.style.font_size == 16, "the original must not be mutated"
+
+    def test_linked_word_stays_linked(self):
+        word = LinkedWord("hello", Font(font_size=16), "http://example.com",
+                          params={"a": "1"}, title="Tooltip")
+
+        copy = word.with_style(Font(font_size=24))
+
+        assert isinstance(copy, LinkedWord)
+        assert copy.location == "http://example.com"
+        assert copy.link_type == word.link_type
+        assert copy.params == {"a": "1"}
+        assert copy.link_title == "Tooltip"
+        assert copy.style.font_size == 24
+
+    def test_linked_word_params_are_copied_not_shared(self):
+        word = LinkedWord("hello", Font(), "http://example.com", params={"a": "1"})
+
+        copy = word.with_style(Font(font_size=24))
+        copy.params["b"] = "2"
+
+        assert "b" not in word.params
+
+
+# ============================================================================
+# _scale_block_fonts
+# ============================================================================
+
+class TestScaleBlockFonts:
+    def test_links_survive_scaling_in_every_container(self, blocks, layouter):
+        before = sum(len(collect_links(b)) for b in blocks)
+        after = sum(len(collect_links(layouter._scale_block_fonts(b, 1.5)))
+                    for b in blocks)
+
+        assert before == 6, "fixture should contain 6 linked words"
+        assert after == before, "scaling must not discard hyperlinks"
+
+    def test_link_targets_are_preserved_exactly(self, blocks, layouter):
+        scaled = [layouter._scale_block_fonts(b, 1.5) for b in blocks]
+        targets = sorted(w.location for b in scaled for w in collect_links(b))
+
+        assert targets == sorted([
+            "http://example.com", "http://example.com",
+            "http://q.example", "http://l.example",
+            "http://h.example", "http://c.example",
+        ])
+
+    @pytest.mark.parametrize("index,kind", [(0, "paragraph"), (1, "quote"),
+                                            (2, "list"), (3, "table")])
+    def test_every_container_type_actually_scales(self, blocks, layouter, index, kind):
+        original = collect_sizes(blocks[index])
+        scaled = collect_sizes(layouter._scale_block_fonts(blocks[index], 2.0))
+
+        assert original, f"fixture {kind} should contain sized words"
+        assert scaled == [s * 2 for s in original], f"{kind} did not scale"
+
+    def test_table_rows_stay_in_their_section(self, blocks, layouter):
+        table = next(b for b in blocks if isinstance(b, Table))
+
+        scaled = layouter._scale_block_fonts(table, 1.5)
+
+        assert len(list(scaled.header_rows())) == len(list(table.header_rows()))
+        assert len(list(scaled.body_rows())) == len(list(table.body_rows()))
+
+    def test_unscaled_blocks_are_returned_unchanged(self, blocks, layouter):
+        assert layouter._scale_block_fonts(blocks[0], 1.0) is blocks[0]
+
+    def test_heading_level_is_preserved(self, layouter):
+        heading = parse_html_string("<h3>Title here</h3>")[0]
+
+        scaled = layouter._scale_block_fonts(heading, 1.5)
+
+        assert isinstance(scaled, Heading)
+        assert scaled.level == heading.level
+
+    def test_result_is_memoised(self, blocks, layouter):
+        """Rebuilding a block per page render allocated on the hot path."""
+        first = layouter._scale_block_fonts(blocks[0], 1.5)
+        second = layouter._scale_block_fonts(blocks[0], 1.5)
+
+        assert first is second
+
+    def test_different_scales_are_cached_separately(self, blocks, layouter):
+        assert (layouter._scale_block_fonts(blocks[0], 1.5)
+                is not layouter._scale_block_fonts(blocks[0], 2.0))
+
+    def test_originals_are_never_mutated(self, blocks, layouter):
+        before = [collect_sizes(b) for b in blocks]
+
+        for b in blocks:
+            layouter._scale_block_fonts(b, 3.0)
+
+        assert [collect_sizes(b) for b in blocks] == before
+
+
+# ============================================================================
+# End to end
+# ============================================================================
+
+def rendered_link_texts(page):
+    """Every LinkText on a rendered page. They live inside Line objects."""
+    found = []
+    for child in page._children:
+        for text_obj in getattr(child, '_text_objects', []):
+            if isinstance(text_obj, LinkText):
+                found.append(text_obj)
+    return found
+
+
+class TestLinksRemainClickableAfterFontChange:
+    """
+    The user-visible symptom of R3: increase the font size and links stop
+    responding to taps.
+    """
+
+    @pytest.fixture
+    def manager(self):
+        blocks = parse_html_string(
+            '<p>Go to <a href="http://example.com">this link</a> now.</p>')
+        manager = EreaderLayoutManager(blocks, page_size=(400, 600),
+                                       bookmarks_dir=tempfile.mkdtemp())
+        yield manager
+        manager.shutdown()
+
+    def test_links_render_at_default_scale(self, manager):
+        page = manager.get_current_page()
+        page.render()
+
+        assert [t.link.location for t in rendered_link_texts(page)] == \
+            ["http://example.com", "http://example.com"]
+
+    @pytest.mark.parametrize("scale", [0.8, 1.5, 2.0])
+    def test_links_survive_a_font_size_change(self, manager, scale):
+        manager.set_font_scale(scale)
+        page = manager.get_current_page()
+        page.render()
+
+        locations = {t.link.location for t in rendered_link_texts(page)}
+        assert locations == {"http://example.com"}
+
+    @pytest.mark.parametrize("scale", [1.0, 1.5])
+    def test_the_link_is_reachable_by_tapping(self, manager, scale):
+        """
+        Scanned rather than probed at the LinkText's own centre: the hit region
+        query_point reports is offset from LinkText.origin by roughly the
+        ascent. That misalignment predates this fix and is tracked separately
+        as R9 - it reproduces identically at scale 1.0.
+        """
+        manager.set_font_scale(scale)
+        page = manager.get_current_page()
+        page.render()
+
+        targets = set()
+        for y in range(0, 120, 2):
+            for x in range(0, 400, 2):
+                result = page.query_point((x, y))
+                if result is not None and result.object_type == "link":
+                    targets.add(result.link_target)
+
+        assert targets == {"http://example.com"}

tests/layout/test_page_buffer.py

@@ -0,0 +1,237 @@
+"""
+Tests for the page caching layer.
+
+Covers PageBuffer's LRU behaviour and BufferedPageRenderer's cache hits, plus
+regressions for S12/R1/R2: the module must not start worker processes and must
+not do blocking work in a finaliser.
+"""
+
+import subprocess
+import sys
+import textwrap
+
+import pytest
+
+from pyWebLayout.layout.page_buffer import PageBuffer, BufferedPageRenderer
+from pyWebLayout.layout.ereader_layout import RenderingPosition
+from pyWebLayout.abstract.block import Paragraph
+from pyWebLayout.abstract.inline import Word
+from pyWebLayout.style import Font
+from pyWebLayout.style.page_style import PageStyle
+
+
+# ============================================================================
+# Fixtures
+# ============================================================================
+
+@pytest.fixture
+def sample_blocks():
+    """A document long enough to paginate over several pages."""
+    font = Font()
+    blocks = []
+    for p in range(6):
+        para = Paragraph(style=font)
+        for w in range(120):
+            para.add_word(Word(f"p{p}w{w}", font))
+        blocks.append(para)
+    return blocks
+
+
+@pytest.fixture
+def renderer(sample_blocks):
+    return BufferedPageRenderer(sample_blocks, PageStyle(), buffer_size=3, page_size=(800, 600))
+
+
+# ============================================================================
+# PageBuffer
+# ============================================================================
+
+class TestPageBuffer:
+    def test_get_page_misses_when_empty(self):
+        buf = PageBuffer(buffer_size=3)
+        assert buf.get_page(RenderingPosition()) is None
+
+    def test_cache_page_round_trips(self):
+        buf = PageBuffer(buffer_size=3)
+        pos, nxt = RenderingPosition(block_index=0), RenderingPosition(block_index=1)
+        sentinel = object()
+
+        buf.cache_page(pos, sentinel, nxt)
+
+        assert buf.get_page(pos) is sentinel
+        assert buf.position_map[pos] == nxt
+
+    def test_lru_evicts_oldest_and_cleans_position_map(self):
+        buf = PageBuffer(buffer_size=2)
+        positions = [RenderingPosition(block_index=i) for i in range(4)]
+        for i, pos in enumerate(positions):
+            buf.cache_page(pos, object(), RenderingPosition(block_index=i + 1))
+
+        assert buf.get_page(positions[0]) is None, "oldest should have been evicted"
+        assert positions[0] not in buf.position_map, "position map must not leak evicted entries"
+        assert buf.get_page(positions[-1]) is not None
+
+    def test_get_page_refreshes_lru_order(self):
+        buf = PageBuffer(buffer_size=2)
+        a, b, c = (RenderingPosition(block_index=i) for i in range(3))
+        buf.cache_page(a, object())
+        buf.cache_page(b, object())
+
+        buf.get_page(a)          # a becomes most recently used
+        buf.cache_page(c, object())
+
+        assert buf.get_page(a) is not None, "recently used entry should survive"
+        assert buf.get_page(b) is None, "least recently used entry should be evicted"
+
+    def test_backward_pages_land_in_the_backward_buffer(self):
+        buf = PageBuffer(buffer_size=3)
+        start, end = RenderingPosition(block_index=1), RenderingPosition(block_index=2)
+
+        buf.cache_page(start, object(), end, is_backward=True)
+
+        assert start in buf.backward_buffer
+        assert start not in buf.forward_buffer
+        assert buf.reverse_position_map[end] == start
+
+    def test_font_scale_change_invalidates(self):
+        buf = PageBuffer(buffer_size=3)
+        pos = RenderingPosition()
+        buf.cache_page(pos, object(), RenderingPosition(block_index=1))
+
+        buf.set_font_scale(1.5)
+
+        assert buf.get_page(pos) is None
+
+    def test_same_font_scale_keeps_cache(self):
+        buf = PageBuffer(buffer_size=3)
+        pos = RenderingPosition()
+        buf.cache_page(pos, object(), RenderingPosition(block_index=1))
+
+        buf.set_font_scale(1.0)
+
+        assert buf.get_page(pos) is not None
+
+    def test_shutdown_is_idempotent(self):
+        buf = PageBuffer(buffer_size=3)
+        buf.cache_page(RenderingPosition(), object())
+
+        buf.shutdown()
+        buf.shutdown()
+
+        assert buf.get_cache_stats()['forward_buffer_size'] == 0
+
+
+# ============================================================================
+# BufferedPageRenderer
+# ============================================================================
+
+class TestBufferedPageRenderer:
+    def test_render_page_returns_a_page_and_advances(self, renderer):
+        page, next_pos = renderer.render_page(RenderingPosition(), 1.0)
+
+        assert page is not None
+        assert next_pos != RenderingPosition()
+
+    def test_second_render_of_same_position_is_served_from_cache(self, renderer):
+        pos = RenderingPosition()
+        first, first_next = renderer.render_page(pos, 1.0)
+        second, second_next = renderer.render_page(pos, 1.0)
+
+        assert second is first, "identical page object means it came from the cache"
+        assert second_next == first_next
+
+    def test_font_scale_change_forces_a_re_render(self, renderer):
+        pos = RenderingPosition()
+        first, _ = renderer.render_page(pos, 1.0)
+        scaled, _ = renderer.render_page(pos, 1.5)
+
+        assert scaled is not first
+
+    def test_backward_render_round_trips_to_the_original_position(self, renderer):
+        start = RenderingPosition()
+        _, second_page_pos = renderer.render_page(start, 1.0)
+
+        _, back_to = renderer.render_page_backward(second_page_pos, 1.0)
+
+        assert back_to == start
+
+    def test_shutdown_clears_the_cache(self, renderer):
+        renderer.render_page(RenderingPosition(), 1.0)
+        renderer.shutdown()
+
+        assert renderer.get_cache_stats()['forward_buffer_size'] == 0
+
+
+# ============================================================================
+# S12 / R1 / R2 regressions
+# ============================================================================
+
+class TestNoBackgroundProcesses:
+    """
+    The process pool that used to live here never produced a usable page (a Page
+    holds a live PIL canvas and cannot be pickled), and on Python 3.14's
+    forkserver default it raised when driven from module-level code.
+    """
+
+    def test_module_declares_no_process_pool(self):
+        import pyWebLayout.layout.page_buffer as page_buffer
+
+        source = page_buffer.__file__
+        assert not hasattr(page_buffer, '_render_page_worker')
+        assert not hasattr(PageBuffer(), 'executor')
+        with open(source, encoding='utf-8') as fh:
+            body = fh.read().split('"""', 2)[-1]  # skip the module docstring
+        assert 'ProcessPoolExecutor' not in body
+        assert 'pickle' not in body
+
+    def test_page_buffer_has_no_finaliser(self):
+        """
+        PageBuffer.__del__ called executor.shutdown(wait=True), which deadlocked
+        the interpreter at exit. Cleanup must be explicit.
+        """
+        assert '__del__' not in vars(PageBuffer)
+
+    def test_navigation_works_without_a_main_guard(self, tmp_path):
+        """
+        R1: EreaderLayoutManager raised RuntimeError when used from module-level
+        script code, because submitting to a ProcessPoolExecutor under a
+        non-fork start method requires an `if __name__ == "__main__"` guard.
+        """
+        script = textwrap.dedent(f"""
+            from pyWebLayout.io.readers.html_extraction import parse_html_string
+            from pyWebLayout.layout.ereader_manager import EreaderLayoutManager
+
+            blocks = parse_html_string("<p>" + " ".join(f"w{{i}}" for i in range(2000)) + "</p>")
+            m = EreaderLayoutManager(blocks, page_size=(800, 600),
+                                     bookmarks_dir={str(tmp_path)!r})
+            m.get_current_page()
+            m.next_page()
+            m.previous_page()
+            m.shutdown()
+            print("OK")
+        """)
+        result = subprocess.run([sys.executable, "-c", script],
+                                capture_output=True, text=True, timeout=120)
+
+        assert result.returncode == 0, result.stderr
+        assert "OK" in result.stdout
+
+    def test_interpreter_exits_without_explicit_shutdown(self, tmp_path):
+        """
+        R2: a manager left to be finalised at exit must not hang. The timeout is
+        the assertion.
+        """
+        script = textwrap.dedent(f"""
+            from pyWebLayout.io.readers.html_extraction import parse_html_string
+            from pyWebLayout.layout.ereader_manager import EreaderLayoutManager
+
+            blocks = parse_html_string("<p>" + " ".join(f"w{{i}}" for i in range(500)) + "</p>")
+            m = EreaderLayoutManager(blocks, page_size=(800, 600),
+                                     bookmarks_dir={str(tmp_path)!r})
+            m.get_current_page()
+            # deliberately no shutdown() - rely on interpreter teardown
+        """)
+        result = subprocess.run([sys.executable, "-c", script],
+                                capture_output=True, text=True, timeout=60)
+
+        assert result.returncode == 0, result.stderr