Migration 009:
- Simplified uuid_generate_v7() to avoid integer overflow on bit shifts
- Uses double precision for timestamp then converts to hex
Seeder:
- Changed roles from 'admin','company' to 'companyAdmin'
- Matches users table CHECK constraint: superadmin, companyAdmin, recruiter, jobSeeker
- Rename 021_create_uuid_v7_function.sql to 009_create_uuid_v7_function.sql
- Add CREATE EXTENSION IF NOT EXISTS pgcrypto for gen_random_bytes()
- Remove obsolete 022_migrate_to_uuid_v7.sql (already handled by table def)
This fixes the error where migration 017 tried to use uuid_generate_v7()
before it was created.
- IDs are now auto-generated by DB SERIAL columns
- No need to generate UUID in application code
- Let repository handle ID assignment from DB response
- Removed unused import
Migrations:
- 016, 017, 019: Replace gen_random_uuid() with uuid_generate_v7()
- All UUID tables now use custom uuid_generate_v7() function
Backend:
- Create internal/utils/uuid/uuid.go with V7() function (RFC 9562)
- Update storage_handler.go to use internal uuid.V7()
- Remove dependency on google/uuid for file naming
All new UUIDs in the system are now UUID v7 (time-ordered)
Migrations:
- Fix 010_seed_super_admin.sql: only use columns from migration 001
- Add 021_create_uuid_v7_function.sql: PostgreSQL uuid_generate_v7() function
- Add 022_migrate_to_uuid_v7.sql: update notifications, tickets, job_payments to use v7
Seeder:
- Create seeder-api/src/utils/uuid.js with uuidv7() function
- Update notifications.js to use uuidv7() instead of randomUUID()
Docs:
- Update DATABASE.md with UUID v7 section and benefits
UUID v7 benefits:
- Time-ordered (sortable by creation time)
- Better index performance than v4
- RFC 9562 compliant
BREAKING CHANGE: Removed core_companies, core_users, core_user_roles tables
Migrations:
- Create 020_unify_schema.sql: adds tenant_id, email, name to users table
- Create user_roles table (replaces core_user_roles)
- Disable 009_create_core_tables.sql (renamed to .disabled)
- Update 010_seed_super_admin.sql to use unified tables
Backend Repositories:
- company_repository.go: use companies table with INT id
- user_repository.go: use users/user_roles with INT id conversion
Seeders:
- All seeders now use companies/users/user_roles tables
- Removed all core_* table insertions
- Query companies by slug to get SERIAL id
This eliminates the redundancy between core_* and legacy tables.
- job_id changed from UUID to INT to match jobs.id SERIAL
- user_id changed from UUID to INT to match users.id SERIAL
- Added user_id FK to users table
- Add JWT auth guard with Bearer token and cookie support
- Update .env.example files with PASSWORD_PEPPER documentation
- Update seeder to use PASSWORD_PEPPER for password hashing
- Update seeder README with hash verification examples
- Fix frontend auth and page components
- Update backend JWT service and seed migration
Backend fixes:
- Fix FK violation in candidate registration by creating company first
- Add CompanyRepository to RegisterCandidateUseCase
- Add handler integration tests for validation
Frontend improvements:
- Add registerCompany function in auth.ts
- Connect company registration form to backend API
- Replace mockJobs with API call in job detail page
- Add loading/error states to job detail page
- Add Jest tests for auth module
- impl(frontend): server-side pagination for jobs listing
- impl(frontend): standardized api error handling and sonner integration
- test(frontend): added unit tests for JobCard
- impl(backend): added SanitizeMiddleware for XSS protection
- test(backend): added table-driven tests for JobService
- docs: updated READMES, created ROADMAP.md and DATABASE.md
- fix(routing): redirected landing page buttons to /jobs
