[hivecommons-hive]

Test Improvement

Adds one regression test to tests/e2e-coverage-report.test.mjs pinning that a
single-line zero-count region is not folded away by a covered region that
encloses it, in the union collectE2ECoverage builds across artifacts.

Three separate proposals would fold exactly that shape:

• [quality] e2e region union still counts three demonstrably-executed single-line arms uncovered; their covered twin differs in START column, which #1066's proposed key would miss #1079 step 1 — "fold a zero-count region into a covered one when the covered
  region's span contains it"
• [quality] e2e region union does not fold single-line '||' fallback arms across the real and variant builds, so correct variant tests lower the gate #1066 — keying a region on startLine:startColumn plus branch ordinal
• line-level containment, the weaker form of the same idea

The test builds two coverage artifacts of one script whose source map gives
line 1 a mid-line mapping. The union they produce is:

artifact 0   1:0:3:12 = 1
artifact 1   1:0:1:12 = 1   1:6:1:12 = 0   1:6:3:12 = 1

The zero region 1:6:1:12 is enclosed three different ways — by 1:0:3:12
from the other artifact (whole-span containment, the #1079 rule), by
1:0:1:12 (same line), and by 1:6:3:12 (same start column, differing end —
the #1066 shape). Every line of the file is covered, so the line view sees
nothing and the region has to survive on its own. The test asserts it does.

Why this boundary matters

Measured on the published e2e-coverage artifact for run
37412914368
(artifact 11390121619), re-rendered locally at main 7ab301e and
reproducing CI exactly at src files 100.00% lines / 91.29% regions
(419/459): every uncovered region in the report is single-line and enclosed
by some covered region from another artifact of the same script. That includes
the arms open work is writing real tests for — GroupLinkStatus's checkedAt
guard (#1094, PR #1113), DirectoryFreshness's plural arms (PR #1104) and
RadarReports' empty-corpus arms (#1097). Enclosure is not evidence of drift,
and a fold keyed on it would erase those gaps rather than fix a report defect.

Related Issue

Refs #1079 (needs-human: the issue asks for a containment fold that the
measured data refutes — the same enclosure holds for genuinely uncovered arms,
and its headline case MemberProfile.js 10:40:10:48 is not span-contained by
its covered twin at all, so no containment rule folds it. This PR lands the
guard the issue's own step 2 asks for; deciding what replaces step 1, and
re-deriving --check-source-regions in .github/workflows/ci.yml, both need a
human. Full measurements are in a comment on #1079.)

Coordination

Touches tests/e2e-coverage-report.test.mjs only. No open PR touches that file
(#1090 scripts/lib/svg-active-content.mjs, #1112 package.json, #1104 and
#1113 e2e variant fixtures and specs, #1110 tests/e2e-data-fixtures.test.mjs
and tests/tools/e2e-data-fixtures.cjs). It does not change
tests/tools/e2e-coverage-report.mjs, so no reported number moves.

---

Filed by quality agent (hold-gated mode). Human review required.

— hive: agent=quality backend=copilot model=claude-opus-5 copilot=1.0.88

[hivecommons-hive]

Important

Held for human review by the hive's ACMM level gate.

This PR was opened by the "quality" agent while Hive policy required a human checkpoint for that agent. Non-outreach agents are held at ACMM L3–L5; the outreach agent is always held because it publishes project-facing communication.

Hive will keep the hold label until a human removes it. Operators can make a deliberate one-off release during an ACMM level change with release_level_holds=true, but level changes never release this hold automatically.