Docs site (MkDocs), web-installer split, per-module tests

[ewowi]

What this does

Stands up a rendered documentation site and de-overloads docs/, closes the test-coverage gaps the doc links exposed, and wires the pieces together.

Docs site (Phase 0 of the docs overhaul)

• Material for MkDocs at the Pages root — top-down nav (user guide → developer reference), instant search, Material icons, use_directory_urls: false so relative image/source links resolve on both the rendered site and GitHub's raw view.
• New docs/index.md landing page (replaces the single-file docs/landing/index.html); Flash button routes to /install/.
• release.yml builds the site into the Pages root via a new scripts/docs/build_docs.py (uv wrapper, PEP-723 inline deps); history/ + backlog/ excluded (internal, transient).
• Full design study: docs/backlog/docs-system-overhaul.md (phases 1–4 — nav split, generated-test folding, fact de-duplication, Doxide source drill-down — each need separate approval).

docs/install/ → top-level web-installer/

• The web installer is an app, not documentation — moved out of docs/ (git mv, history preserved). Deployed URL unchanged at /install/ (release.yml maps web-installer/ → pages/install/), so QR codes, deployed-device OTA URLs, and existing links keep working.
• ~104 references swept across scripts, both workflows, checks, tests, CLAUDE.md gate triggers, and docs.

Per-module tests (24 new)

• The broken [Tests] links on effects/modifiers were a missing-coverage signal, not a doc bug. Added behaviour-specific unit tests for 19 effects + 5 modifiers (~90 cases): audio effects fed via AudioModule simulate + frozen clock with RAII mic-seat guards; degenerate-grid no-crash cases enforcing the every-grid-size rule. Regenerated docs/tests/*.md — all 25 broken test-anchor links now resolve.

Firmware ? links → rendered site

• Module ? help links open moonmodules.org/projectMM/moonmodules/<path>.html (was GitHub blob URLs); all 85 registered docPaths verified to resolve to real pages + anchors. These 404 until the site's first Pages deploy (on merge to main); the installer stays live throughout.

MoonDeck + dev-server hygiene

• New "Build Docs Site" card (PC → docs group).
• build_docs.py --serve self-cleans a stray mkdocs serve; the three local dev servers use adjacent ports — MoonDeck 8420, installer preview 8421, docs preview 8422 — so all three coexist.

Verification (all green pre-commit)

Desktop build (0 warnings) · ctest 577 cases / 12,219 assertions · scenarios · ESP32 S3 build · platform boundary · KPI · device-catalog · firmware-list · JS host tests 31/31 · Python host tests 22/22 · spec-check · docs build (0 broken anchors).

Follow-ups (not in this PR)

• scripts/ → moondeck/ rename — spec'd in docs/backlog/rename-scripts-to-moondeck.md, its own next commit.
• Reflash boards after merge to pick up the new ? links.

🤖 Generated with Claude Code

Summary by CodeRabbit

• New Features

  • Added a new documentation site with a clearer homepage, navigation, search, and improved pages for getting started, architecture, testing, and module references.
  • Expanded the browser-based installer preview and docs-building workflow for easier local review.

• Bug Fixes

  • Corrected many documentation links and references throughout the site.
  • Updated release and CI automation to use the new installer/docs locations.

• Tests

  • Added broad new unit-test coverage for many light effects and modifiers, including edge cases and empty-grid handling.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: 1f024d2c-dbbd-40cd-aff7-b2d69dfed405

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• ✅ Review completed - (🔄 Check again to review again)

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch next-iteration

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 14

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

src/ui/app.js (1)

510-527: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value

Fragile .md→.html conversion.

docPath.replace(".md", ".html") replaces only the first literal match. It works today because docPath always contains exactly one .md (at the end, before an optional #anchor), but it's easy to regress silently if a future docPath segment ever contains .md elsewhere (e.g. a directory name). A suffix-anchored regex would make the intent explicit and be immune to that.

♻️ Proposed fix

-        help.href = "https://moonmodules.org/projectMM/moonmodules/" + docPath.replace(".md", ".html");
+        help.href = "https://moonmodules.org/projectMM/moonmodules/" + docPath.replace(/\.md(?=$|#)/, ".html");

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/ui/app.js` around lines 510 - 527, The help-link URL building in app.js
uses a fragile docPath.replace(".md", ".html") conversion that only replaces the
first literal match. Update the URL construction in the docPath/help anchor
logic to use a suffix-anchored replacement so only the trailing .md before any
`#anchor` is converted, keeping the intent explicit and avoiding accidental
changes if docPath ever contains .md elsewhere.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In @.claude/workflows/write-behaviour-tests.js:
- Line 57: The generated prompt currently hardcodes a developer-specific
absolute repo path, which will break for other machines and CI. Update the
prompt निर्माण in write-behaviour-tests.js so it derives the repository location
dynamically from the current workspace/root instead of embedding a fixed
/Users/... value. Make the repo reference generic and reusable in the prompt
template so the same workflow works across contributors and runners.
- Around line 93-98: The module-level return in write-behaviour-tests.js is
invalid JavaScript/ESM and prevents the workflow from loading. Remove the
top-level return and either wrap this result-building block in a named function
or export the object directly from the workflow module, using the existing
wrote, failed, and results logic as the source for the returned data.

In @.github/workflows/release.yml:
- Around line 29-36: The release workflow path filter is missing the web
installer directory, so installer-only changes will not trigger the Pages
deploy. Update the path list in the release workflow to include web-installer/**
alongside the existing docs/** and mkdocs.yml entries, using the same trigger
block that already covers src/ui/install-picker*.js.

In `@docs/architecture.md`:
- Line 250: The rendered link text is stale and still says
install/deviceModels.json even though the target now points to
web-installer/deviceModels.json. Update the visible label in
docs/architecture.md to match the current destination, and keep the surrounding
reference to the catalog schema consistent with the existing
deviceModel/deviceModels.json wording.

In `@docs/backlog/docs-system-overhaul.md`:
- Around line 13-67: Add the missing blank lines around the tables and headings
in the docs/backlog/docs-system-overhaul.md content to satisfy markdownlint
MD058 and MD022. Update the affected Markdown sections in place so each table is
separated by a blank line before and after, and each heading has a blank line
below it; check the top fact table plus the later section headings like the ones
under “The core insight,” “The tools,” and “Phased transition” for consistency.

In `@docs/history/plans/Plan-20260702` - Docs site Phase 0 (MkDocs Material).md:
- Around line 13-14: Update the historical note in the plan so it matches what
was actually shipped: the docs build in CI does not use build_docs.py with
--strict, and link/anchor validation remains warn-level in mkdocs.yml. Adjust
the CI wiring description around deploy-pages and build_docs.py to state that CI
fails on nav/build issues, while broken links/anchors are only documented as a
local strict-mode check in scripts/MoonDeck.md.

In `@docs/index.md`:
- Line 5: The link in the docs index uses an absolute /install/ path, which will
break under the /projectMM/ subpath deployment. Update the link target in the
docs index to use the project-prefixed path consistently with the mkdocs
navigation fix, and ensure any other references in this page that point to
/install/ are updated to the same resolved subpath. Keep the change aligned with
the root-cause fix so the same path behavior is used throughout the site.

In `@docs/tests/unit-tests.md`:
- Line 835: The regression note in the unit test docs is written in past-tense,
change-history language instead of present tense. Update the text in the
affected documentation comment to describe the behavior in present tense only,
and remove phrases like “Before the fix” while keeping the same meaning; use the
surrounding regression note wording in docs/tests/unit-tests.md to locate it.

In `@mkdocs.yml`:
- Around line 101-107: The installer links are using a root-relative /install/
path, which skips the deployed /projectMM/ prefix. Update the Web installer
entry in mkdocs.yml and the homepage links in docs/index.md so they point to
/projectMM/install/ instead, keeping the existing installer CTA text and
placement unchanged.

In `@scripts/MoonDeck.md`:
- Around line 157-173: The new build_docs subsection is interrupting the
existing update_module_docs description, causing the unreferenced-screenshots
sentence to read like part of build_docs. Move the “Reports unreferenced
screenshots…” sentence back so it stays with the update_module_docs content and
the MoonDeck.md section boundaries remain correct; use the update_module_docs
and build_docs headings plus the “Reports unreferenced screenshots” sentence to
locate the affected text.

In `@test/unit/light/unit_BlockModifier.cpp`:
- Around line 67-77: The 1x1x1 expectation in BlockModifier survives degenerate
grids is inconsistent with the BlockModifier algorithm used elsewhere in
unit_BlockModifier.cpp. Update the assertions in the TEST_CASE to match the
actual fold/blockSize results for the {1,1,1} case, using the existing fold and
blockSize helpers as the reference points. Keep the degenerate 0x0x0 coverage,
but ensure the expected size values reflect the real logical size growth
produced by mm::BlockModifier.

In `@test/unit/light/unit_FreqMatrixEffect.cpp`:
- Around line 23-176: The new effect tests repeat the same
Layouts/GridLayout/Layer setup sequence in every TEST_CASE, so extract that
boilerplate into a shared helper such as makeGridLayer or buildEffectLayer.
Centralize the repeated steps in the helper using the existing Layouts,
GridLayout, Layer, and FreqMatrixEffect setup pattern, then update each test to
call it with the desired dimensions/channels and add the effect afterward if
needed.

In `@test/unit/light/unit_GEQ3DEffect.cpp`:
- Around line 44-91: Move the active-mic cleanup out of the manual tail call in
GEQ3DEffect test setup by using the same RAII pattern as AudioGuard from the
other unit tests. The issue is that mic.teardown() is only reached after
REQUIRE, so a failed assertion leaves AudioModule::active_ set; fix this by
introducing an automatic guard around mm::AudioModule in the test so teardown
runs in the destructor even on early abort, and remove the explicit teardown
call at the end.

In `@test/unit/light/unit_StarFieldEffect.cpp`:
- Around line 10-50: The test cases in unit_StarFieldEffect.cpp manually reset
the virtual clock with a trailing setTestNowMs(0), which can be skipped if a
REQUIRE fails and leave later tests polluted. Update each TEST_CASE that calls
setTestNowMs(...) to use the same RAII ClockGuard pattern used in
unit_BouncingBallsEffect.cpp, so the clock is restored automatically by scope
exit. Remove the manual end-of-test reset and keep the guard local to each test
body, especially around the StarFieldEffect checks and the other affected
TEST_CASEs in this file.

---

Outside diff comments:
In `@src/ui/app.js`:
- Around line 510-527: The help-link URL building in app.js uses a fragile
docPath.replace(".md", ".html") conversion that only replaces the first literal
match. Update the URL construction in the docPath/help anchor logic to use a
suffix-anchored replacement so only the trailing .md before any `#anchor` is
converted, keeping the intent explicit and avoiding accidental changes if
docPath ever contains .md elsewhere.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: b2199448-d1d8-42d5-a99a-b9cb858dc4ee

📥 Commits

Reviewing files that changed from the base of the PR and between 766c7ed and 28e0854.

⛔ Files ignored due to path filters (5)

• scripts/build/build_esp32.py is excluded by !**/build/**
• scripts/build/generate_firmwares.py is excluded by !**/build/**
• scripts/build/improv_provision.py is excluded by !**/build/**
• scripts/build/improv_smoke_test.py is excluded by !**/build/**
• web-installer/favicon.png is excluded by !**/*.png

📒 Files selected for processing (85)

• .claude/workflows/write-behaviour-tests.js
• .github/workflows/release.yml
• .github/workflows/test.yml
• .gitignore
• CLAUDE.md
• docs/architecture.md
• docs/backlog/docs-system-overhaul.md
• docs/backlog/rename-scripts-to-moondeck.md
• docs/building.md
• docs/history/plans/Plan-20260702 - Docs site Phase 0 (MkDocs Material).md
• docs/index.md
• docs/landing/index.html
• docs/moonmodules/core/DevicesModule.md
• docs/moonmodules/core/I2cScanModule.md
• docs/moonmodules/core/NetworkModule.md
• docs/moonmodules/core/SystemModule.md
• docs/moonmodules/light/BlendMap.md
• docs/moonmodules/light/Drivers.md
• docs/moonmodules/light/Layers.md
• docs/moonmodules/light/Layouts.md
• docs/moonmodules/light/drivers/LcdLedDriver.md
• docs/moonmodules/light/drivers/NetworkSendDriver.md
• docs/moonmodules/light/drivers/ParlioLedDriver.md
• docs/moonmodules/light/drivers/RmtLedDriver.md
• docs/moonmodules/light/drivers/drivers.md
• docs/performance.md
• docs/reference/esp32-s31-coreboard.md
• docs/testing.md
• docs/tests/scenario-tests.md
• docs/tests/unit-tests.md
• mkdocs.yml
• scripts/MoonDeck.md
• scripts/check/check_devices.py
• scripts/check/check_firmwares.py
• scripts/docs/build_docs.py
• scripts/moondeck.py
• scripts/moondeck_config.json
• scripts/moondeck_ui/app.js
• scripts/run/preview_installer.py
• src/core/I2cScanModule.h
• src/ui/app.js
• src/ui/install-picker-boards.js
• src/ui/install-picker.js
• test/js/config-ops.test.mjs
• test/js/improv-frame.test.mjs
• test/js/installer-eth-only.test.mjs
• test/js/installer-s31-webflash.test.mjs
• test/python/test_improv_frame.py
• test/python/test_installer_manifests.py
• test/scenarios/light/scenario_GridLayout_resize.json
• test/scenarios/light/scenario_MultiplyModifier_pipeline.json
• test/unit/light/unit_BlockModifier.cpp
• test/unit/light/unit_BlurzEffect.cpp
• test/unit/light/unit_BouncingBallsEffect.cpp
• test/unit/light/unit_CircleModifier.cpp
• test/unit/light/unit_FixedRectangleEffect.cpp
• test/unit/light/unit_FreqMatrixEffect.cpp
• test/unit/light/unit_FreqSawsEffect.cpp
• test/unit/light/unit_GEQ3DEffect.cpp
• test/unit/light/unit_GEQEffect.cpp
• test/unit/light/unit_LissajousEffect.cpp
• test/unit/light/unit_MirrorModifier.cpp
• test/unit/light/unit_Noise2DEffect.cpp
• test/unit/light/unit_NoiseMeterEffect.cpp
• test/unit/light/unit_PaintBrushEffect.cpp
• test/unit/light/unit_PraxisEffect.cpp
• test/unit/light/unit_RandomEffect.cpp
• test/unit/light/unit_RippleXZModifier.cpp
• test/unit/light/unit_RubiksCubeEffect.cpp
• test/unit/light/unit_SolidEffect.cpp
• test/unit/light/unit_SphereMoveEffect.cpp
• test/unit/light/unit_StarFieldEffect.cpp
• test/unit/light/unit_StarSkyEffect.cpp
• test/unit/light/unit_TetrixEffect.cpp
• test/unit/light/unit_TransposeModifier.cpp
• web-installer/README.md
• web-installer/config-ops.js
• web-installer/deviceModels.json
• web-installer/devices.js
• web-installer/firmwares.json
• web-installer/improv-frame.js
• web-installer/index.html
• web-installer/install-orchestrator.js
• web-installer/install.css
• web-installer/install.js

💤 Files with no reviewable changes (1)

• docs/landing/index.html

[coderabbitai]

🎯 Functional Correctness | 🟠 Major | ⚡ Quick win

Hardcoded developer-specific absolute path in generated prompt.

Repo: /Users/ewoud/Developer/GitHub/MoonModules/projectMM bakes one developer's local machine path into the agent prompt. This workflow will produce broken/misleading prompts for any other contributor or CI runner that isn't at that exact path.

🔧 Proposed fix

-Repo: /Users/ewoud/Developer/GitHub/MoonModules/projectMM
+Repo: ${process.cwd()}

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	Repo: /Users/ewoud/Developer/GitHub/MoonModules/projectMM
	Repo: ${process.cwd()}

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In @.claude/workflows/write-behaviour-tests.js at line 57, The generated prompt
currently hardcodes a developer-specific absolute repo path, which will break
for other machines and CI. Update the prompt निर्माण in write-behaviour-tests.js
so it derives the repository location dynamically from the current
workspace/root instead of embedding a fixed /Users/... value. Make the repo
reference generic and reusable in the prompt template so the same workflow works
across contributors and runners.

[coderabbitai]

🩺 Stability & Availability | 🟠 Major | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
# Check for other workflow files using the same top-level phase()/agent()/parallel()/return pattern,
# and find the runner that loads/executes .claude/workflows/*.js to see if it wraps the module body.
fd . .claude/workflows --type f
rg -n 'return \{' .claude/workflows -A2
rg -nP '\bnew Function\b|vm\.(Script|createContext|runInContext)|\bimport\(' .claude -g '*.js' -C3

Repository: MoonModules/projectMM

Length of output: 427

---

🏁 Script executed:

#!/bin/bash
set -euo pipefail

echo '--- outline .claude/workflows/write-behaviour-tests.js ---'
wc -l .claude/workflows/write-behaviour-tests.js
sed -n '1,170p' .claude/workflows/write-behaviour-tests.js

echo
echo '--- search for workflow runner / execution wrapper ---'
rg -n --hidden --glob '!**/node_modules/**' --glob '!**/dist/**' --glob '!**/build/**' \
  'phase\(|agent\(|parallel\(|write-behaviour-tests|\.claude/workflows|new Function|vm\.(Script|createContext|runInContext)|Function\(' \
  . -C 2

Repository: MoonModules/projectMM

Length of output: 13910

---

Remove the top-level return here. This file is parsed as JavaScript/ESM, so return { ... } at module scope is a syntax error and will stop the workflow from loading. Wrap the workflow body in a function or export the result instead.

🧰 Tools

🪛 Biome (2.5.1)

[error] 93-98: Illegal return statement outside of a function

(parse)

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In @.claude/workflows/write-behaviour-tests.js around lines 93 - 98, The
module-level return in write-behaviour-tests.js is invalid JavaScript/ESM and
prevents the workflow from loading. Remove the top-level return and either wrap
this result-building block in a named function or export the object directly
from the workflow module, using the existing wrote, failed, and results logic as
the source for the returned data.

Source: Linters/SAST tools

[coderabbitai]

🎯 Functional Correctness | 🟠 Major | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
rg -n -B3 -A3 "paths:" .github/workflows/release.yml | rg -n -A15 "on:" 
rg -n "web-installer" .github/workflows/release.yml

Repository: MoonModules/projectMM

Length of output: 1369

---

🏁 Script executed:

#!/bin/bash
cat -n .github/workflows/release.yml | sed -n '1,70p'

Repository: MoonModules/projectMM

Length of output: 3530

---

🏁 Script executed:

#!/bin/bash
git ls-files | rg '^web-installer(/|$)|/web-installer(/|$)'

Repository: MoonModules/projectMM

Length of output: 467

---

Add web-installer/** to the release workflow paths .github/workflows/release.yml still triggers on docs/** and mkdocs.yml, but not on web-installer/**; installer-only changes will skip the Pages deploy.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In @.github/workflows/release.yml around lines 29 - 36, The release workflow
path filter is missing the web installer directory, so installer-only changes
will not trigger the Pages deploy. Update the path list in the release workflow
to include web-installer/** alongside the existing docs/** and mkdocs.yml
entries, using the same trigger block that already covers
src/ui/install-picker*.js.

[coderabbitai]

📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win

Markdownlint: missing blank lines around tables/headings.

Static analysis flags MD058 (tables not surrounded by blank lines) at lines 13 and 19, and MD022 (headings need a blank line below) at lines 44, 50, 55, 60, and 67.

🧹 Example fix pattern (apply to each flagged table/heading)

 - **Duplication hotspots** (each fact lives in N places, changing one forces the others):
+
   | Fact | Lives in | Sync today |
   |---|---|---|
@@
   | Architectural fact (buffer persists; AudioModule respects `enabled`) | `.h` comment + module `.md` + `architecture.md` | none |
+
 - **No structured doc-comments in source.**
@@
 ### Phase 0 — Stand up the site (no content changes) — *now, ~half a day*
+
 - Add `mkdocs.yml` with Material, `pymdownx.snippets`, instant search.

🧰 Tools

🪛 markdownlint-cli2 (0.22.1)

[warning] 13-13: Tables should be surrounded by blank lines

(MD058, blanks-around-tables)

---

[warning] 19-19: Tables should be surrounded by blank lines

(MD058, blanks-around-tables)

---

[warning] 44-44: Headings should be surrounded by blank lines
Expected: 1; Actual: 0; Below

(MD022, blanks-around-headings)

---

[warning] 50-50: Headings should be surrounded by blank lines
Expected: 1; Actual: 0; Below

(MD022, blanks-around-headings)

---

[warning] 55-55: Headings should be surrounded by blank lines
Expected: 1; Actual: 0; Below

(MD022, blanks-around-headings)

---

[warning] 60-60: Headings should be surrounded by blank lines
Expected: 1; Actual: 0; Below

(MD022, blanks-around-headings)

---

[warning] 67-67: Headings should be surrounded by blank lines
Expected: 1; Actual: 0; Below

(MD022, blanks-around-headings)

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/backlog/docs-system-overhaul.md` around lines 13 - 67, Add the missing
blank lines around the tables and headings in the
docs/backlog/docs-system-overhaul.md content to satisfy markdownlint MD058 and
MD022. Update the affected Markdown sections in place so each table is separated
by a blank line before and after, and each heading has a blank line below it;
check the top fact table plus the later section headings like the ones under
“The core insight,” “The tools,” and “Phased transition” for consistency.

Source: Linters/SAST tools

[coderabbitai]

🎯 Functional Correctness | 🔴 Critical | ⚡ Quick win

Wrong expected value for the 1×1×1 degenerate case.

Applying BlockModifier's actual formula (validated by the other three TEST_CASEs in this file) to box={1,1,1}: centerX = (1+1)/2-1 = 0, dx = |1-0| = 1, dy = 1, distance = 1, so the logical size becomes {0,1,0} then each axis grows by one → {1,2,1}, not {1,1,1}. This assertion will fail against the real implementation.

🐛 Proposed fix

-    // 1x1x1: single light is the centre -> distance 0; logical box {1,1,1}.
+    // 1x1x1: centre offset (1-0)=1 on each axis -> distance 1; logical box {1,2,1}.
     CHECK(fold(b, 0, 0, 0, {1, 1, 1}) == mm::Coord3D{0, 0, 0});
-    CHECK(blockSize(b, {1, 1, 1}) == mm::Coord3D{1, 1, 1});
+    CHECK(blockSize(b, {1, 1, 1}) == mm::Coord3D{1, 2, 1});

As per path instructions, "Verify tests cover edge cases and match the specifications in docs/moonmodules/" — this edge case doesn't match the modifier's own algorithm.

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	TEST_CASE("BlockModifier survives degenerate grids") {
	mm::BlockModifier b;
	// 1x1x1: single light is the centre -> distance 0; logical box {1,1,1}.
	CHECK(fold(b, 0, 0, 0, {1, 1, 1}) == mm::Coord3D{0, 0, 0});
	CHECK(blockSize(b, {1, 1, 1}) == mm::Coord3D{1, 1, 1});
	// 0x0x0: no crash; the fold and size are still finite (each axis grows by one).
	CHECK_NOTHROW(fold(b, 0, 0, 0, {0, 0, 0}));
	CHECK(blockSize(b, {0, 0, 0}).x >= 1);
	CHECK(blockSize(b, {0, 0, 0}).y >= 1);
	CHECK(blockSize(b, {0, 0, 0}).z >= 1);
	}
	TEST_CASE("BlockModifier survives degenerate grids") {
	mm::BlockModifier b;
	// 1x1x1: centre offset (1-0)=1 on each axis -> distance 1; logical box {1,2,1}.
	CHECK(fold(b, 0, 0, 0, {1, 1, 1}) == mm::Coord3D{0, 0, 0});
	CHECK(blockSize(b, {1, 1, 1}) == mm::Coord3D{1, 2, 1});
	// 0x0x0: no crash; the fold and size are still finite (each axis grows by one).
	CHECK_NOTHROW(fold(b, 0, 0, 0, {0, 0, 0}));
	CHECK(blockSize(b, {0, 0, 0}).x >= 1);
	CHECK(blockSize(b, {0, 0, 0}).y >= 1);
	CHECK(blockSize(b, {0, 0, 0}).z >= 1);
	}

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@test/unit/light/unit_BlockModifier.cpp` around lines 67 - 77, The 1x1x1
expectation in BlockModifier survives degenerate grids is inconsistent with the
BlockModifier algorithm used elsewhere in unit_BlockModifier.cpp. Update the
assertions in the TEST_CASE to match the actual fold/blockSize results for the
{1,1,1} case, using the existing fold and blockSize helpers as the reference
points. Keep the degenerate 0x0x0 coverage, but ensure the expected size values
reflect the real logical size growth produced by mm::BlockModifier.

Source: Path instructions

[MoonModules]

@coderabbitai, I asked for review but it is not scheduled in, so again: please review 49c16cb and prosent your findings in a copy pastable prompt for agents

[coderabbitai]

Oops, something went wrong! Please try again later. 🐰 💔

[MoonModules]

@coderabbitai, I asked for review but it is not scheduled in, so again: please review 49c16cb and prosent your findings in a copy pastable prompt for agents

[coderabbitai]

Oops, something went wrong! Please try again later. 🐰 💔

[MoonModules]

@coderabbitai, I asked for review but it is not scheduled in, so again: please review 49c16cb and prosent your findings in a copy pastable prompt for agents

[coderabbitai]

Oops, something went wrong! Please try again later. 🐰 💔