CPD follow-up: preserve goal carry choice across next-year target confirmation #3161

Activity

1. BigSimmo commented on Oct 15, 2026

   

   BigSimmo

   on Oct 15, 2026 – with MiniMax Code App

   OwnerAuthor

   More actions

   PR draft ready (2026-10-15) — local branch mavis/3161-cme-goal-carry-choice, not pushed, no PR opened, no label added.

   Shape (4 files, 267 +/3 -):

   • src/lib/cme/return-target.ts (new, 66 lines) — zod-validated { fromYear, goalId? } round-trip via base64url JSON. decodeCmeCarryReturnTarget returns null on any failure rather than throwing — worst case is the pre-fix behaviour, never an unsafe deep-link.
   • src/components/cme/cme-setup-route.tsx — reads ?returnTo= via useSearchParams, decodes, after confirmation navigates to /cme/plan?year=<fromYear>[&goal=<goalId>]#cme-carry-goals when valid, else falls back to the original /cme/setup?year=<year> (no regression).
   • src/components/cme/cme-plan-page.tsx:354 — the "Confirm <year+1> targets first" link now appends &returnTo=<encoded> so the carry context survives the round trip.
   • tests/cme-setup-route-return-target.dom.test.tsx (new, 178 lines) — 11 cases: round-trip with/without goal, malformed payload, tampered payload (out-of-range year, overlong goalId), missing target, programme-page no-confirm, codec edge cases.

   Verification (all run on Node 24.22, npm 11.19):

   • ✅ tsc -p tsconfig.changed.json --noEmit (focused on the 4 changed files): 0 errors.
   • ✅ tests/cme-setup-route-return-target.dom.test.tsx: 11/11 pass.
   • ✅ Related CPD tests (no regressions): cme-setup.dom.test.tsx 10/10, cme-plan-page.dom.test.tsx 14/14, cme-plan-carry.test.ts 10/10, cme-one-tap-routine-log.dom.test.tsx 3/3, cme-catch-up-plan.test.ts 11/11. 48/48 pass.

   Open for owner decision:

   • The goalId field is in the schema but only fromYear is set in cme-plan-page.tsx. A future flow that carries a specific goal could pass goalId too. Currently the link is invoked from the general "Confirm <year+1> targets first" prompt and the doctor re-selects the goal on return. Acceptable, or do you want the specific goal to flow through now?
   • The cmeCarryReturnTargetHref always appends #cme-carry-goals (the carry section anchor on the plan page). If the doctor came from elsewhere (e.g., the dashboard), this still lands them on the plan page at the carry section. Confirm the landing is right for those flows before they adopt the link.

   Diff: /workspace/issue-3161-pr-diff.patch (327 lines including the test). Branch is local at mavis/3161-cme-goal-carry-choice in /workspace/PsychSift, staged, not committed, not pushed. Push + PR + merge is yours.
2. BigSimmo commented on Oct 15, 2026

   

   BigSimmo

   on Oct 15, 2026 – with MiniMax Code App

   OwnerAuthor

   More actions

   Rebase + re-verification (2026-10-15, 14:25 local):

   Main moved forward 5 commits (a07dec0 → e0c62df) since the PR draft was first staged. Rebased branch mavis/3161-cme-goal-carry-choice onto the new main. The 4 files in the diff are unchanged in shape; no conflicts with the 5 new commits.

   Re-verification on current main (e0c62df):

   • ✅ tsc -p tsconfig.changed.json --noEmit (4 changed files): 0 errors.
   • ✅ tests/cme-setup-route-return-target.dom.test.tsx: 11/11 pass.
   • ✅ Related CPD tests: 48/48 pass (cme-setup 10, cme-plan-page 14, cme-plan-carry 10, cme-one-tap-routine-log 3, cme-catch-up-plan 11).

   No open PR addresses this issue (verified by is:pr is:open search). The 4 currently open PRs on PsychSift (#3555 migrations rule, #3554 search cutoff, #3553 #3538 walk-back, #3546 admin ingestion) are unrelated.

   Diff at /workspace/issue-3161-pr-diff.patch — 327 lines, 267+/3-, 4 files. Branch is local, not pushed, no PR opened, no label added.
3. BigSimmo

   added a commit that references this issue on Oct 21, 2026

   fix(cpd): return to the carry choice after confirming next year's tar…

   ...

   9f9d2ea
4. BigSimmo

   closed this as completedin #3753on Oct 22, 2026