bc-dgx7.85: B7e-suitemap — The inside of an existing suite — its fixtures

[mordam]

Adds b7e-suitemap (bin/b7e-suitemap, lib/suitemap.js, test/suitemap.mjs): given an
existing test suite, print what it imports from the module under test, its file-local
fixture/loader helpers (with signature and, for a loader that returns an object, the
keys it hands back), and every named check in source order under its section divider.
--for <file> resolves the suite through lib/affected.js first. This is bc-dgx7.85's
own session-audit finding: five sessions each added a check to a suite that already
existed and each read several hundred lines of it by hand — sed slices, greps for
anchor strings, Reads at guessed line numbers — to learn the same three things.

The parse is a real acorn AST walk (matching lib/already.js and lib/imports.js's own
reasoning: a name is a binding and a call is a call, not a string a regex might also
match inside a comment), not a regex over the source text.

One thing worth flagging: lib/affected.js's own text-based matching cannot see the
fs.readFileSync(PUBLIC('report.js')) helper-call convention a suite uses to vm-load
the module it tests — the literal is split across two calls, so test/reporter.mjs
reaches findAffected's candidate list only as a weak directory-walker. --for
resolution parses every candidate with the same acorn walk and lets an actual
vm-source/dynamic/static import of the target override that ranking; this is
new logic in lib/suitemap.js, not a change to lib/affected.js itself, so it doesn't
touch b7e-affected's own behavior or its suite. Both concrete cases in the bead's
acceptance criteria (test/homing.mjs's anchor strings, and --for public/report.js
naming load/fire and the acceptance-criteria section of test/reporter.mjs) are
asserted against this repo's real suites, not just fabricated fixtures.

Registered per the five-file b7e checklist: package.json + package-lock.json bin maps
(alphabetical), lib/grants.js (read) and lib/toolbelt.js's DEFAULT_TOOL_LIST
(append), and a README section after b7e-harness's. Merged origin/main in
(bc-dgx7.82's b7e-propagated landed while this was in flight and touches the same five
files — resolved clean, additively, as that family always does) and re-ran the
affected suites and this suite's own 28 checks afterward.

Tests: node test/suitemap.mjs (28/28); node bin/b7e-affected against the diff → 59 narrowed suites all green via node, plus scripts/land-check.mjs (timed out at 120s in a batch runner but passes at 0 exit on its own); node scripts/checks.mjs — 55/58 browser checks green, 3 pre-existing failures (advocatespane-check, historypane-check, phone-check) not in the affected-suite list and unrelated to this diff (pane/phone rendering, one needs a live daemon with beads); node test/lockfile.mjs green before and after merging origin/main

Files changed — 8 files · +1131 −0 · against `main`

lib/suitemap.js    +450 −0
test/suitemap.mjs  +438 −0
bin/b7e-suitemap   +150 −0
README.md          +81 −0
lib/toolbelt.js    +9 −0
lib/grants.js      +1 −0
package-lock.json  +1 −0
package.json       +1 −0

---

Opened by a beadcause worker session on bc-dgx7.85 — b7e-suitemap — The inside of an existing suite — its fixtures, what it imports from the file under test, and where a new check goes. It merges itself once the checks report; merging is what closes the bead. If this is still open, something stopped that, and the reason is on bc-dgx7.85 and in Adam's inbox.

bead: bc-dgx7.85

[mordam]

A beadcause worker opened this and does not merge its own work. It is on the merge queue as bc-qigvf.

[NeanderthalMan]

beadcause-resolver: stood down — worktree-b7e-suitemap-dgx785 is locked by another resolver on this same PR (reason "resolver pid 71366 #768", pid live in ps). Did not touch the tree, did not unlock, took no lock of my own.

[mordam]

The beadcause merge queue tried to merge this and could not: it has been tried 3 times and stopped at the same place each time. the branch still conflicts with main and no resolver is on it any more — the window that was opened on it ended without making it mergeable, so this one is yours to settle. Tried 3 times — that was the last. It is Adam's call now — see bc-qigvf.