refactor(ui): flatten nested ternaries in apps/mobile (Sonar typescript:S3358)

[thomasluizon]

What

Flattens every nested conditional expression flagged by SonarCloud rule typescript:S3358 (Extract this nested ternary operation into an independent statement) that is scoped to apps/mobile — 44 files / 59 instances.

apps/web + packages/shared S3358 instances are handled by a sibling slice, so cross-platform parity is preserved at the batch level.

How (behavior- and render-preserving)

Each nested ternary was removed with the smallest structural change that keeps rendered output identical:

• Value chains → extract the inner ternary into a preceding named const (or an if/else helper for 3+ levels), leaving only single-level ternaries.
• Style-callback nested ternaries → convert the arrow to a block body and hoist the inner ternary into a local const.
• cond ? (inner) : null in JSX/style arrays → cond && (inner) (falsy is ignored identically), used only where the discarded branch is null/falsy and the guard is a boolean.
• Multi-state JSX chains (loading/error/empty/content) → hoist into a let x: ReactNode computed with if/else if before return, rendered as {x}.

No any, no narration comments, no i18n or logic changes.

Verification

• npx tsc --noEmit — clean
• npx vitest run — 1056 passed (200 files)
• npx expo lint — 0 errors (1 pre-existing warning in untouched calendar-sync.tsx)
• react-doctor --blocking error --scope changed — 0 blocking errors
• jscpd at Sonar's 10-line threshold — 0 new cross-file clones among the refactored files

Refs #243

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
orbit-ui-mobile-web	Ignored		Jul 13, 2026 1:38pm

[claude]

Code Review: PR #484 — refactor(ui): flatten nested ternaries in apps/mobile (Sonar typescript:S3358)

Scope: PR #484 in thomasluizon/orbit-ui-mobile (44 files, 59 nested-ternary instances flattened)
Recommendation: APPROVE

Summary

This is a pure, mechanical, behavior-preserving refactor that eliminates every SonarCloud typescript:S3358 (nested ternary) instance scoped to apps/mobile. Each transform is one of four small, well-understood patterns (hoist inner ternary to a named const, convert to an if/else block before return, cond ? X : null → cond && X where the guard is boolean/nullable-object, or multi-branch JSX chains hoisted into a let x: ReactNode variable). I walked all 44 changed files diff-hunk by diff-hunk and traced every extracted variable/branch back to the original ternary to confirm identical runtime behavior. No logic, styling, token, or i18n-key changes were found.

Findings

Critical

None.

High

None.

Medium

None.

Low / Info

• Info: A few of the cond && (...) conversions (apps/mobile/app/(tabs)/calendar.tsx:9-30, apps/mobile/app/accountability-pair.tsx:378, apps/mobile/components/goal-card.tsx:1134) rely on the guard being a strict boolean or a nullable object rather than a number/string, which is the classic React "renders a stray 0" footgun for this pattern. Verified each guard's type (activeError/rangeError are query-error objects, isPending/isSelectMode are booleans, trackingDot is {...} | null) — none can be a falsy-but-truthy-rendering primitive, so this is not a live bug, just worth naming as the one thing this transform class needs to keep checking each time it's applied.

Subagents

Agent	Verdict
parity-checker	PAIRED — see note below (checked manually, not delegated)
i18n-syncer	N/A — no i18n keys added, removed, or renamed; every t(...) call in the diff is a pre-existing key relocated verbatim into the new if/const structure
contract-aligner	N/A — no packages/shared/src/types/* or endpoints.ts touched
security-reviewer	N/A — no orbit-api code in this diff
design-reviewer	PASS — confirmed every color/style computation still resolves through the same semantic tokens as before (tokens.fg1/fg2/fg4, tintFromPrimary, tokens.primary, etc.); no new AI-slop tell, no raw --slate-*/hardcoded rgba, no em dash introduced

Parity note: this PR only touches apps/mobile. The PR body states the apps/web + packages/shared S3358 instances were fixed in a sibling PR. I confirmed this directly: PR #483 ("chore(ui): flatten nested ternaries (Sonar typescript:S3358)"), already merged to main ahead of this PR, fixes the corresponding web-side files for the same components touched here (calendar-range-view, calendar-week-view, social/*, upgrade/page, goal-card, goal-detail-drawer, edit-goal-modal, empty-state, pricing-section, habit-row-check-circle, move-parent-overlay, retrospective-dashboard, login-sections). Parity is preserved at the batch level, as claimed — not a finding.

Validation

Check	Result
Lint	N/A in CI review — Build/Unit Tests/SonarCloud run as separate required checks on this PR per repo CI config
Type check	N/A in CI review — same as above
Tests	N/A in CI review — PR body reports vitest run 1056/1056 passing (200 files) and tsc --noEmit clean; not independently re-run here
Build (api)	N/A — orbit-api not touched, not checked out in this job

Deferred — N/A dimensions & files not verdicted

• FEATURES.md parity (#14): N/A — pure refactor, no user-facing behavior change, no screen/route/tool/gating change.
• Backend hard rules (#13): N/A — orbit-api not touched.
• Contract drift (#11): N/A — no shared-type or DTO surface touched.
• i18n (#10): N/A — no string/key changes (see Subagents).
• Validation (Lint/Type/Test): not run in this session — per this repo's CI wrapper instructions, Build/Unit Tests/SonarCloud already run as separate required checks on this PR, so re-running here would be redundant.
• Cross-repo dimensions requiring orbit-api: not verifiable in this job (sibling repo not checked out).
• Every one of the 44 changed files got an explicit verdict (traced against its pre-refactor version) — nothing in the diff was skipped.

What's good

• Every single-variable extraction and if/else chain is traceable 1:1 back to the original ternary's branch order and condition — zero behavior drift found across 59 instances.
• No any, no console.log, no new comments (narration or otherwise) introduced anywhere in the diff — clean on rules 3, 4, and 5.
• The PR body is unusually thorough: it names the four transform patterns used, discloses the sibling-PR split for parity, and reports its own verification commands (tsc, vitest, expo lint, react-doctor, jscpd) with actual output — exactly the kind of self-auditing a pure mechanical refactor should ship with.
• Correctly resisted the temptation to "clean up while here" — no drive-by logic changes riding along with the Sonar fix.

Recommendation

Approve as-is. This PR does exactly what it says, nothing more, and every transformation checks out against the original code.

[thomasluizon]

CI status — ready to merge (SonarCloud coverage-only override)

All required checks pass. The only red is SonarCloud → new_coverage 0.0% (< 80%), which needs an admin override at merge.

SonarCloud PR quality-gate conditions:

Condition	Status	Actual	Threshold
new_reliability_rating	✅ OK	1 (A)	1
new_security_rating	✅ OK	1 (A)	1
new_maintainability_rating	✅ OK	1 (A)	1
new_duplicated_lines_density	✅ OK	0.0%	3%
new_security_hotspots_reviewed	✅ OK	100%	100%
new_coverage	❌ ERROR	0.0%	80%

• typescript:S3358 on this PR: 0 (all 59 apps/mobile instances across 44 files resolved).
• No new smells, bugs, security issues, or duplication — the flattening is behavior- and render-preserving (extracted nested ternaries into named consts / if/else assignments / hoisted ReactNode variables).
• The 0% new-code coverage is expected: the touched files are presentational React Native components with no unit tests, and a behavior-preserving refactor adds no new covered lines. Full mobile suite stays green (1056 tests passing).

The transient Lint / Type Check / Contract Drift / Dependency Audit reds were a GitHub Actions action-resolution outage ("Failed to resolve action download info") and all pass on re-run.

[sonarqubecloud]

Quality Gate failed

Failed conditions
0.0% Coverage on New Code (required ≥ 80%)

See analysis details on SonarQube Cloud

[claude]

/pr-review — PR #484

Scope: refactor(ui): flatten nested ternaries in apps/mobile (Sonar typescript:S3358) — 44 files, 59 instances.

No diff since last review. This bot already posted an APPROVE on this exact head commit (04615549629166bc74b8f5873c811a93e8c1747f) at 2026-07-13T13:32:19Z. gh pr view 484 confirms headRefOid is byte-identical and the PR is now MERGED — zero new commits, zero new hunks since that review. Re-walking the full rubric would just reproduce the prior pass verbatim.

Recommendation: APPROVE (carried forward, unchanged)

Severity	Count
Critical	0
High	0
Medium	0
Low / Info	1 (carried forward — cond && (...) guard-type pattern noted, assessed as not a live bug)

Findings, still valid at the current head:

• Pure, mechanical, behavior-preserving refactor; all 44 files traced 1:1 against pre-refactor ternary logic.
• No any, no console.log, no new comments introduced; no i18n or logic changes.
• Parity: corresponding apps/web / packages/shared S3358 instances were handled in the already-merged sibling PR #483 — confirmed directly.
• No packages/shared or orbit-api surface touched → contract-drift and backward-compat-guard dimensions are not applicable; cross-repo dimensions requiring the orbit-api checkout are not verifiable in this CI job but are also not triggered by this diff.
• Validation (lint/type-check/tests) intentionally not re-run — owned by CI's Build / Unit Tests / SonarCloud required checks per this workflow's scope.