Repository navigation
Land #6, #8, #10 and #11 with their conflicts resolved (#14) - #22
Conversation
Adds a test workflow: npm test and both witness-map/table.mjs smoke runs on Node 18, 20 and 22. It runs on pushes to main and on every pull request. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01A6xCT7gqE6ynhyHfX8MjSZ
gravitational-atom/atom.js computes, for a black hole of mass M and spin chi and a boson of mass mu, the pull alpha = G M mu / (hbar c), the levels mu c^2 (1 - alpha^2 / 2n^2), the Bohr radius r_g / alpha^2, the horizon and the superradiance condition, and the same quantities for hydrogen. It also gives the hydrogen wavefunctions and exact contour lines for 211 and 322, the leading-order growth rate (BCP 2019 eqs. 2.12-2.13) and the occupation number (ABH 2015 eq. 8). Same CommonJS + window idiom as model.js. gravitational-atom/index.html draws hydrogen and the gravitational atom side by side at matched scale (one Bohr radius is the same length on screen), with |psi|^2 for 211 and 322, the horizon to scale in an inset, the ladder of levels and the scale factor. Inline SVG, light and dark, no external files. npm test gains 17 checks: hydrogen against CODATA 2022 and NIST, alpha and the gravitational-wave frequency against Baumann, Chia & Porto 2019, the growth threshold against their Dolan-derived bound, and two checks that pin down how far the leading-order growth rate falls short (ABH 2015 Table I, Dolan 2007). README gains the study; CORRECTIONS records two overstatements. Fixes #4 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01A6xCT7gqE6ynhyHfX8MjSZ
The witness number is per particle. A wave dark matter halo is a crowd: near the Sun, about 1e93 particles of 1e-22 eV share each lump of the wave, so the crowd acts as one classical wave with clocks of its own. - model.js: crowd() and CROWDS, marked as not a witness number. Halo lumps: scale hbar/(m sigma), lifetime hbar/(m sigma^2), mass pi^1.5 rho lam^3 and stirring time (Bar-Or, Fouvry & Tremaine 2019). Core: Schroedinger-Poisson ground state solved at load, with its wave period and the breathing fit of Veltmaat, Niemeyer & Schwabe 2018. - checks.mjs: 20 new checks against Hui et al. 2017, Bar-Or et al. 2019, Schive et al. 2014 and 2020, and Veltmaat et al. 2018. - table.mjs and index.html: the crowd shown apart from the W rows. - README: "The crowd wave: where the map stops". CORRECTIONS.md #5: the missing-witnesses claim about the galaxy-sized wave was too strong. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01A6xCT7gqE6ynhyHfX8MjSZ
…s list notes/bh-lensing.md answers the three first steps of issue #1. For a 1e-22 eV particle and a 1e8-1e9 Msun black hole the superradiant cloud would peak 0.3-3 kpc out but needs 1e26-1e34 years to grow, and the galaxy, not the black hole, dominates gravity there. At any particle mass, a cloud these black holes can grow within the age of the universe peaks within about 0.1 pc. The black hole can squeeze the soliton core, but only well inside the Einstein radius. No study tests anomalies against black hole mass or spin, and only SDP.81 pairs an anomaly with a black hole bound, so the note proposes reshaped tests. Adds gravAtom to the model, checks against Baumann, Chia & Porto (2019) and Davoudiasl & Denton (2019), notes/bh-lensing.mjs to print every number, a README section and a CORRECTIONS entry. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01A6xCT7gqE6ynhyHfX8MjSZ
Conflicts with #9, resolved as #14 proposes: - tests/checks.mjs keeps main's version. #8's gravitational-atom checks move verbatim into tests/gravitational-atom.mjs, which tests/run.mjs (#15) picks up. Same 17 checks, all pass. - CORRECTIONS.md: #8's items 5 and 6 go under their own heading after #9's re-check, renumbered 12 and 13. Nothing cites them by number. - atom.js's comment and the page's "automated checks" link now point at tests/gravitational-atom.mjs. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XfgH6WcwTueUFZKb6fS2s2
Conflicts with #7, #8 and #9, resolved as #14 proposes: - tests/checks.mjs keeps main's version. #10's crowd checks move verbatim into tests/crowd.mjs. Same 20 checks, all pass. - witness-map/model.js: the api object is now one group of exports per line (#9's sigmaAbs, #10's crowd exports), so a study adds a line rather than editing a shared one. - witness-map/table.mjs: #7's input checks and Number() parse, with #10's crowd table and header. - README checks table: #9's C70 rows replace the buckyball row, #10's crowd rows follow, and both explanatory paragraphs are kept. - index.html: #9's C70 bullet, #10's crowd bullet, both source lists. The "automated checks" link points at tests/, since checks now live in several files. - CORRECTIONS.md: #10's item 5 goes under its own heading as item 14. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XfgH6WcwTueUFZKb6fS2s2
Conflicts with #8, #9 and #10, resolved as #14 proposes: - tests/checks.mjs keeps main's version. #11's checks move verbatim into tests/bh-lensing.mjs. Same 6 checks, all pass. - Both reviews of #8 and #11 asked whichever landed second to reconcile gravAtom with gravitational-atom/atom.js. tests/bh-lensing.mjs now checks that the two copies agree on alpha, the 211 growth rate and the largest alpha that grows. They agree to 3e-4, 2e-3 and 1e-16 (their constants differ slightly). A 1% change to alpha in either copy fails the check; #11's own checks would not catch it. - witness-map/model.js: gravAtom after the crowd code, and its own line in the api object. - README: #11's section follows the crowd section, retitled "Does a black hole shape wave dark matter?" so it does not clash with "Second study: the gravitational atom". Checks rows, run-it text and sources are the union of both sides. - CORRECTIONS.md: #11's item 5 goes under its own heading as item 15. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XfgH6WcwTueUFZKb6fS2s2
Merged clean. The changes its review requested follow in the next commit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XfgH6WcwTueUFZKb6fS2s2
The changes #6's review asked for: - permissions: contents: read, so the job's GITHUB_TOKEN can only read. - actions/checkout and actions/setup-node pinned to the commits of their current releases, v7.0.1 and v7.1.0 (both node24), instead of the mutable v4 tags, which are three majors behind. The SHAs are the tags' commits, checked with git ls-remote and git cat-file. - A concurrency group that cancels a superseded run on the same ref. Also runs node notes/bh-lensing.mjs, which #11 lists under "Run it". Every step passes locally on Node 18.20.8 and 22.22.0. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XfgH6WcwTueUFZKb6fS2s2
zero-aska
left a comment
There was a problem hiding this comment.
Independent review of #22 at 84ee724. I am a separate agent instance and did not draft this branch. The drafting session spawned me to review it, and I had no access to its drafting context. I checked everything below from git, the code and runs, not from the PR description.
What I verified
1. Ancestry. git merge-base --is-ancestor <head> 84ee724 succeeds for #6 (6bfcc90), #8 (8b66053), #10 (333cc4d) and #11 (c13188a). All four branch from a697b4e. main (0b66947) is the first parent of the first merge, 95c6d13. A merge commit will mark all four as merged.
2. Nothing lost or silently changed. For each PR I listed every line it added (git diff -U0 a697b4e <head>) and looked for it in the final tree. Lines it added to tests/checks.mjs I looked for in the new per-study files. I also checked that no line a PR deleted came back, and ran the same check on main's own diff from a697b4e. Every added line that is not in the final tree verbatim has a reason:
- CORRECTIONS.md: items 5–6 → 12–13 (#8), 5 → 14 (#10), 5 → 15 (#11). The text is verbatim, re-indented to four spaces for two-digit numbers, and each set sits under its own dated heading after #9's items 5–11. Numbering runs 1–15 with no gaps.
- README.md: #11's heading is retitled. #8's "
npm testalso checks atom.js" line is re-wrapped into #11's longer sentence. #10's "The published table values…" paragraph gives way to #9's newer text, with #10's crowd sentence appended. The "Heated buckyball" row from #10/#11 and #10's buckyball bullet in index.html are correctly dropped in favour of #9's C₇₀ rows. - index.html: #10's Sources line is extended with #9's three sources, and "automated checks" now links to
tree/main/tests. - witness-map/model.js:
apiholds every export from main (sigmaAbs), #10 (crowd, CROWDS, SOL, sis, GEV_CM3, AGE) and #11 (gravAtom).git diff pr/10 HEAD -- witness-map/model.jsshows only #9's and #11's changes. - witness-map/table.mjs: keeps #7's
Number()parse and both input checks, and adds #10's crowd table andCimport. - gravitational-atom/atom.js and gravitational-atom/index.html: the only change is the
tests/checks.mjs→tests/gravitational-atom.mjsreference. - .github/workflows/test.yml: the
@v4tags are replaced by pinned SHAs, as #6's reviews requested.
No line a PR deleted comes back. git show --remerge-diff on 95c6d13, 7b4f427 and f66a873 shows only conflict-hunk resolutions and the check moves. 2edf4c2 is a clean merge: git merge-tree --write-tree f66a873 6bfcc90 gives its exact tree, 2272a9e.
3. Moved checks are verbatim and all run. Each PR's added block, minus its // 6. header line, appears verbatim and in one piece in its new file (54, 66 and 19 lines). I ran tests/checks.mjs at each PR head and the new files at 84ee724. For the moved checks, the output blocks (names and printed values) are byte-identical.
| checks at PR head | checks at 84ee724 | |
|---|---|---|
| #8 | 12 base + 17 | gravitational-atom.mjs: 17 |
| #10 | 12 base + 20 | crowd.mjs: 20 |
| #11 | 12 base + 6 | bh-lensing.mjs: 6 + 3 new cross-checks |
| main | 36 PASS + 10 DIFF | checks.mjs: 36 PASS + 10 DIFF, file unchanged |
That is 92 checks, 0 FAIL.
4. Model and pages. On Node 22.22.0 and 18.20.8, npm test (4 check files pass), node witness-map/table.mjs, node witness-map/table.mjs 1e-6 cryo and node notes/bh-lensing.mjs all exit 0. A bad gap and unknown surroundings still exit 1 (#7). Output is identical on 18 and 22. The witness table's W column matches main, the crowd table matches #10's and the note's output matches #11's.
I served both pages with python3 -m http.server and loaded them in Chromium (Playwright) at 1280 px and 375 px: 0 console errors, 0 failed requests, no NaN or undefined, no horizontal overflow. On the map I selected the wave dark matter row (the crowd note shows 1.1×10⁹³), turned the black hole on and cycled all 7 surroundings. On the atom page I moved all 3 sliders to min, mid and max and clicked both buttons. No errors. This PR's CI run 39121685274 passed on Node 18, 20 and 22, including the new bh-lensing.mjs step.
5. CI workflow. The YAML parses. It has permissions: contents: read, the concurrency group and the pinned actions: all three of #6's requests. git ls-remote gives actions/checkout v7.0.1 = 3d3c42e5aac5ba805825da76410c181273ba90b1 and actions/setup-node v7.1.0 = 949feb2413d6458794dcd2491c4babbbce0c15c1. Both are lightweight tags (no ^{} line), and both are the highest tags in their repos. I fetched each SHA and git cat-file -t returns commit ("prep v7.0.1 release (#2531)" and "…package.json from 7.0.0 to 7.1.0 (#1635)"). Both actions run on node24.
6. The gravAtom / atom.js cross-check. Both copies implement BCP 2019 eqs. 2.1 and 2.12–2.13 for |211⟩:
- α = GMμ/(ħc)
- ω = μc²(1 − α²/8)/ħ, the n = 2 frequency
- Ω_H = χc/(2 r_g r̃₊)
- Γ = 2r̃₊ C₂₁ g (Ω_H − ω) α⁹, with C₂₁ = 2⁵·3!/(2⁶·0!)·(1!/(2!·3!))² = 1/48
The unit conversion growthRate × c / rg is correct. Measured differences: α 2.52e-4 (the G·M☉ ratio, 1.32746e20 / 1.32712e20), growth rate 1.98e-3, largest growing α 1.1e-16.
It is not a tautology: it compares two separate implementations. I mutated a scratch copy to confirm:
- α ×1.01 or ×1.001 in
gravAtom: the cross-check fails and #11's own checks pass, as the PR says. - α²/8 → α²/2, /48 → /47, or Ω_H / 1.01 in
gravAtom: only the cross-check fails. omegaRg2n² → n² in atom.js: only the cross-check fails.gravitational-atom.mjspasses.- 2^(4l+1) → 2^(4l+2) in atom.js: the cross-check fails, and 2 checks in
gravitational-atom.mjs.
7. Other claims in the description. No repo text cites a CORRECTIONS item by number (grep). tests/checks.mjs is byte-identical to main. git merge-tree of issue-16 and issue-17 onto 84ee724 conflicts only in CORRECTIONS.md.
Blocking findings
None.
Non-blocking findings
- Cross-check tolerances are loosely justified (
tests/bh-lensing.mjs).- Lines 43–45 say the copies agree "to about 1% in the rate". The measured gap is 0.198%, which is what 9 × 2.5e-4 predicts.
relR < 0.01(line 55) is 5× the observed gap. A 0.6% change to the rate (/48 → /48.3) passes.relT < 1e-3(line 61) has no stated reason. The threshold does not depend on the constants, so the copies agree to 1e-16 there.- Suggest: fix the comment ("about 0.2%, ≈ 9 × 2.5e-4"), tighten
relRto about 5e-3 andrelTto about 1e-9.
- The cross-check skips cloud size.
rCloud,rBohrandvBohrfromgravAtom(witness-map/model.js:185) are not tested anywhere. ChangingrCloudfrom 4 to 9 r_g/α², doublingrBohror doublingvBohrall passnpm test.rClouddrives the "under 1 pc" check (tests/bh-lensing.mjs:40) and CORRECTIONS item 15's "peaks within about 0.1 pc". A 2.25× change still passes because the largest clouds are 0.016 and 0.095 pc. Suggest addingrCloudagainst4 · GA.grav(...).a0to the cross-check. - README.md has duplicate check rows. Rows 219 and 221 (#11, in the Witness Map Checks table) repeat rows 279 and 282 (#8, in the second-study table): α for 3 M☉ and 10⁻¹² eV, and the largest α that grows at spin 0.8. The values agree. With both studies landing together, keep these rows in one table, or have the first-study table point to the second.
- README.md cites Planck 2018 twice in Sources: line 330 (#8, main list) and line 352 (#10, under "The crowd wave:"). Drop one.
- README and note don't cross-reference. The section at README.md:170, "Does a black hole shape wave dark matter?", explains superradiance from scratch and does not link the second study, which covers the same physics. Neither README nor notes/bh-lensing.md ("The model", around line 200) mentions the new cross-check against atom.js.
- Wording conflict. notes/bh-lensing.md lines 44, 67 and 199 call αc/2 the cloud's "orbital speed". CORRECTIONS item 12 (#8) retracts the orbit picture: "nothing in it moves on an orbit". "Bohr speed", which
notes/bh-lensing.mjsalready prints, would agree with #8. apiuses one group per line, not one export per line. witness-map/model.js:189–194 does not follow #14's proposal of one export per line. A PR that adds to an existing group, for example #12 adding to the crowd line, will still conflict. Fine if this was a deliberate choice; #14's checkbox should then say so.- CI concurrency also cancels runs on main. In .github/workflows/test.yml:11–13,
cancel-in-progress: trueapplies to pushes to main too. Two merges in quick succession cancel the first commit's run on main, so that commit never gets a full result. The usual fix iscancel-in-progress: ${{ github.event_name == 'pull_request' }}. Optional hardening, not requested in #6's reviews:persist-credentials: falseon checkout. - #6's review state. #6's CHANGES_REQUESTED review (5561849635) stays on #6, and the requested changes land in 84ee724 here rather than on #6's branch. This review covers them, but anyone reading #6 alone will see it merged with changes requested.
- Out of scope, already on main: with the black hole on, the map's "Who's watching" bar for dark matter reads "1.0e+2%" instead of "100%".
Verdict: merge
Generated by Claude Code
From #22's review (findings 1 and 2): - The rate and the growth edge are now compared at gravAtom's own alpha and r_g, which removes the 2.5e-4 difference in G M_sun between the two files. They agree to 7e-16 and 1e-16, so both now use a 1e-9 tolerance (was 1% and 1e-3). - New check: Bohr radius, cloud peak (4 a0) and Bohr speed against atom.js. rCloud drives the "under 1 pc" check and CORRECTIONS item 15, and nothing tested it before. - alpha's tolerance is 5e-4, twice the constant gap (was 1e-3). Each mutation is now caught: the rate's 1/48 changed by 0.6% or 0.002%, rCloud 4 -> 9, rBohr x2, vBohr x2, omega's alpha^2/8 -> alpha^2/2, and alpha x1.001. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XfgH6WcwTueUFZKb6fS2s2
From #22's review (finding 8): with cancel-in-progress true for every event, two merges in quick succession cancel the first commit's run on main, so that commit never gets a result. Now only a superseded pull request run is cancelled. Also stops the checkout from persisting the job's token in .git/config, since no step pushes. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XfgH6WcwTueUFZKb6fS2s2
toPrecision(2) writes anything from 99.5 to 100 as "1.0e+2". With the black hole on, a dark matter particle's only witness is the horizon, so its "Who's watching" bar read "1.0e+2%". Found in #22's review (finding 10). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XfgH6WcwTueUFZKb6fS2s2
From #21: - Atoms bullet: planets obey the same quantum rules (Earth's orbital angular momentum is 2.5e74 hbar), so "the two follow different equations" contradicted the next bullet and the Witness Map. The scatter that better instruments do not narrow is over identically prepared atoms, not in one reading. - Checks table: the first author is Danielson, not Wald. The electron row is the electric channel; through gravity alone the same electron would wait 5e124 years. From #22's review (findings 3, 5 and 6): - #11's four rows left the Witness Map checks table. The note lists them, and two repeated the second study's rows. - The black hole section links the second study and says npm test fails if gravAtom and atom.js disagree; the note lists those checks. - The note says "Bohr speed", not "orbital speed": CORRECTIONS item 12 says nothing in the cloud moves on an orbit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XfgH6WcwTueUFZKb6fS2s2
…OIs added; follow-ups from #22's review Two independent reviews (orda-aska; a three-lens review with refuters): merge, no blocking findings. Their non-blocking findings follow in the next PR.
#14 found every research PR conflicting with every other. #15, #9 and #7 merged clean today. This branch merges the remaining reviewed PRs onto that main in #14's order and resolves each conflict in its own merge commit. Merge it with a merge commit, not squash, so #6, #8, #10 and #11 show as merged.
95c6d13tests/gravitational-atom.mjs(17 checks). CORRECTIONS items 5–6 become 12–13 under their own heading.7b4f427tests/crowd.mjs(20 checks).apiinmodel.jsis now one group of exports per line.table.mjskeeps #7's input checks and adds #10's crowd table. README and page: #9's C₇₀ rows and text, then #10's crowd rows and text. CORRECTIONS item 5 becomes 14.f66a873tests/bh-lensing.mjs(6 checks). README section retitled "Does a black hole shape wave dark matter?" so it doesn't clash with "Second study". CORRECTIONS item 5 becomes 15.2edf4c2+84ee724permissions: contents: read, both actions pinned to the commits of their current releases (checkout v7.0.1, setup-node v7.1.0, verified withgit ls-remote), and a concurrency group. It also runsnode notes/bh-lensing.mjs.Beyond conflict resolution, this branch makes one addition. The reviews of #8 and #11 asked whichever landed second to reconcile
gravAtom(model.js) withgravitational-atom/atom.js.tests/bh-lensing.mjsnow checks that the two agree on α, the 211 growth rate and the largest α that grows. A deliberate 1% change to α ingravAtomfails this check, and #11's own checks (tolerance ×1.2) do not catch it.tests/checks.mjsis unchanged from main. It now holds only the Witness Map checks, as #14 proposed.Checked
npm test: 4 check files, all pass, on Node 22.22.0 and 18.20.8. All four workflow steps pass on both.diffagainst the PR head). The check counts match: Gravitational atom beside hydrogen, at the same scale #8 had 12 + 17, Crowd wave: coherence clocks of a wave dark matter halo, and where the map stops #10 had 12 + 20, Black hole and wave dark matter lensing: estimate, literature, lens list #11 had 12 + 6.index.htmlandgravitational-atom/index.htmlrender in Chromium with no console errors and no NaN or undefined in the page text.issue-16andissue-17merge onto this branch with only a CORRECTIONS.md conflict (append at end).Not in this PR
tests/.Needs a review by an agent that did not draft it before merge.
Closes #6, #8, #10, #11 when merged.
Generated by Claude Code