`/debugAuth` env feature flag — verify whether actually needed (likely close as not-applicable)

[TaprootFreak]

Context

/debugAuth route is kDebugMode-gated at two places:

• lib/setup/routing/router_config.dart:90-95 (route registration)
• lib/screens/welcome/welcome_page.dart:120-125 (welcome card that opens the route)

This means the route is present in debug builds and absent in release builds. Maestro CI runs debug builds by default, so for Maestro testing alone, the kDebugMode gate is sufficient — no additional env flag needed.

The README's "Coverage infrastructure roadmap" however explicitly lists a build-time feature-flag mechanism ("Build-time feature-flag mechanism (analogous to EXPO_PUBLIC_ENABLE_* in dfx-wallet) so non-MVP features can be gated out of the activated surface") as an unchecked item — useful for things beyond /debugAuth, e.g. excluding non-MVP feature surfaces from the activated-surface coverage gate.

This issue scopes whether the flag mechanism is needed and, if so, implements the minimum that unblocks the existing referencing issues.

Status of dependent issues

• Tier 3: extend Maestro handbook to cover every user-facing feature #543 Tier 3 Maestro: previously listed 57-debug-auth.yaml as deferred-due-to-flag. Re-audit shows this is wrong: Maestro runs debug builds, so the existing kDebugMode gate is sufficient. 57-debug-auth.yaml is unblocked at this layer. Drop the "needs env flag" pre-condition from Tier 3: extend Maestro handbook to cover every user-facing feature #543 entirely.
• README roadmap: still lists the flag as future work. The unblocker for the activated-surface scope move (post-testWidgets: spec every *_page.dart (12 untested + 5 special-handling) #544/testWidgets: spec every shared widget under lib/widgets/ (8 testable + 1 dead code) #545/Visual regression: scale from 5-screen pilot to every *_page.dart #547) is the .test-coverage-allowlist mechanism from CI: file-pair existence guard (every page/widget needs a test or an allow-list entry) #551 — not env flags.

Recommendation

Close as not-needed for the current test-coverage push. The kDebugMode gate covers /debugAuth for Maestro; the .test-coverage-allowlist from #551 covers scope carve-outs.

Open a fresh issue only if a non-Maestro use case for build-time flags surfaces (e.g. shipping a --enable-experimental-feature-X build of the app).

Acceptance criteria

• Either close as not-applicable (recommended), or scope an env-flag mechanism with concrete consumers
• If closing: remove the "env feature flag" pre-condition from Tier 3: extend Maestro handbook to cover every user-facing feature #543's body
• If keeping: list at least 2 concrete consumers beyond /debugAuth

Estimated effort

Path	Days
Close as not-needed + update #543 body	0.1
Implement --dart-define-based flag system	1-2

Recommended: close. Total: 0.1 day.

Related

• Tier 3: extend Maestro handbook to cover every user-facing feature #543 — Tier 3 Maestro (currently lists this as a pre-condition; can be removed)
• README "Coverage infrastructure roadmap"
• lib/setup/routing/router_config.dart:90-95
• lib/screens/welcome/welcome_page.dart:120-125

[TaprootFreak]

Closing as not-applicable, verified by deep audit against the codebase.

Three independent reasons

1. The stated coverage-gating use case is already solved by lcov-filter + inline // coverage:ignore-* annotations + Port-pattern refactors. PRs test: cover the four Drift repositories + tighten coverage scope #538 (merged) and test: coverage uplift batch (storage + dfx services + config + sell_bitbox + kyc cubits) #539 (in flight) ratcheted the floor from 51 → 94 without ever needing an env-flag. Toolkit used: lcov --remove '*.g.dart', ~30+ inline coverage:ignore-line/start/end annotations on Drift schema getters / StateError programmer-error paths / platform-channel adapter boundaries, plus Port-pattern refactors (BiometricPort, DocumentsDirectoryPort, SumsubIdentPort).

2. The dfx-wallet precedent (EXPO_PUBLIC_ENABLE_*) is also an engineering tool, not a product/region/beta gate. Its 17 flags are MVP-triage + test-coverage gates that flip on by default once every gated function is full in the test matrix. No product or region use case in the upstream precedent either. Plus Dart's String.fromEnvironment has no tree-shake equivalent — the JS / Babel / Expo pattern doesn't port cleanly.

3. No product surface in realunit-app benefits. Zero existing uses of String.fromEnvironment / --dart-define. Five kDebugMode/kReleaseMode hits across 4 files, all well-scoped (/debugAuth route + welcome card; verify_seed_cubit pre-fill; main.dart release-mode error reporting). No screen named experiment/beta/staging/preview/wip. All 12 defer-tagged features intentionally ship in user builds today — "defer" is a coverage-deadline label, not a ship gate. /debugAuth is already correctly kDebugMode-gated at both router and welcome-card layers.

Follow-up actions in the closing PR

• Strike README.md:28 — the unchecked roadmap line Build-time feature-flag mechanism (analogous to EXPO_PUBLIC_ENABLE_* in dfx-wallet). README currently advertises a roadmap item that has been declared not-needed.
• Drop the "env feature flag" pre-condition from Tier 3: extend Maestro handbook to cover every user-facing feature #543's body if still present there (already done in V3).

Mild caveat — reopen if needed

If a future "ship-but-hide" feature appears (regulator-gated jurisdiction, TestFlight-only experimental flow, staging-vs-prod compile-time endpoint switch), the flag becomes useful. Reopen this issue at that point. Today: not applicable.