[jtravs]

Refs #40. This shortens the build a lot, but it does not explain the stalls, so it leaves #40 open until builds show the result.

What changed

The 763 source record pages are now :orphan:, reached through the alphabetical index instead of a hidden toctree. The conf.py setting that narrowed their sidebar is removed, because they no longer have one.

Readers: URLs and search entries are unchanged. Record pages lose the sidebar and previous and next links, and keep the top navigation and their link to the catalogue. the-data/sources.html drops from 125 KB to 33 KB, because it no longer lists 829 sidebar entries.

Why

Sphinx copies the whole navigation tree for every page it writes. A profile of a full build, on an Apple Silicon Mac at c63a648, without re-running the gallery:

	Time
Whole build, under the profiler	271 s
Writing pages	222 s
The theme's generate_toctree_html	177 s

That is about 35 million node copies, nearly all for records the sidebar then hides.

Unprofiled builds, same machine, also without re-running the gallery:

Build	Time
Before	164 s
collapse_navigation across the whole manual	62 s
This change	39 s
-j 2, without this change	141 s

collapse_navigation was rejected because it also changes the sidebar on every other page. The 62 s and 39 s runs ran side by side.

Verification

• uv run poe check: passed, 21539 passed, 1 skipped, 100% coverage, including a clean docs build with warnings as errors and the gallery executing.
• tests/test_source_catalogue.py now requires that records are orphans and that the index holds no toctree.
• The public-wording scan shows no new hits.
• The Read the Docs build of this PR is the check that matters. Results will follow here and on Read the Docs builds have been killed after about 24 minutes since 30 October #40.

Not checked

• Whether this stops the kills. They began on 2026-10-29 with no change in optmat, and solprop's builds changed on the same date, so this reduces exposure to the problem rather than explaining it.
• Record pages were checked in the generated HTML, not looked at in a browser.

🤖 Generated with Claude Code

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-04T20:29:40.492236Z	0b747d9	PR opened
🔒 Security Review	✅ Completed 2026-11-04T20:30:32.261902Z	0b747d9	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[read-the-docs-community]

Documentation build overview

📚 optmat | 🛠️ Build #35653969 | 📁 Comparing 24d4484 against latest (24672b1)

  🔍 Preview build  

865 files changed · + 2 added · ± 863 modified

+ Added

• wanted.html
• developers/releasing.html

± Modified

• accuracy.html
• changelog.html
• contributing.html
• data.html
• nonlinear_accuracy.html
• propagation.html
• roadmap.html
• sg_execution_times.html
• api/index.html
• developers/documentation.html
• and 853 more...

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[jtravs]

Read the Docs: 6 of 6 builds of this tree passed, at 230 to 341 s, on the unchanged config that had passed 0 of 10 since 30 October. Details are on #40. A later commit changes only the changelog fragment, from "about a third" to "less than half", because the whole build fell from about 600 s to a mean of 257 s.