7.8.0
📦 prismaView on GitHub →
✨ 1 features🐛 9 fixes🔧 4 symbols
Summary
The 7.8.0 release introduces fine-grained control over the query plan cache via `queryPlanCacheMaxSize` in `PrismaClient` and resolves numerous bugs across PostgreSQL, SQL Server, and migration tooling.
✨ New Features
- Added a `queryPlanCacheMaxSize` option to the `PrismaClient` constructor for fine-grained control over the query plan cache. Pass `0` to disable the cache entirely, or omit it to use the default cache size.
🐛 Bug Fixes
- Fixed an equality filter panic and incorrect "::jsonb" cast when filtering on PostgreSQL JSON list columns using `where: { jsonListField: { equals: [...] } }`.
- Fixed case-insensitive JSON field filtering (`mode: insensitive`), allowing `where: { jsonField: { equals: "...", mode: "insensitive" } }` to work correctly.
- Fixed incorrect parameterization of enum values that have a custom database name set via `@map`.
- Fixed a database parameter limit check (`P2029`), which could incorrectly reject or miss over-limit queries.
- Fixed a regression that caused missing SQL Server `VARCHAR` casts for parameterized values.
- Fixed a misleading error message in `prisma migrate diff` that referenced the `--shadow-database-url` CLI flag, which was removed in Prisma 7.
- Fixed `prisma migrate dev` (and shadow database migration replay in general) failing with `CREATE INDEX CONCURRENTLY cannot run inside a transaction block` when a migration contained concurrent index creation statements on PostgreSQL.
- Fixed PostgreSQL introspection silently dropping sequence defaults when the database returns the schema-qualified form `pg_catalog.nextval('sequence_name'::regclass)` instead of the bare `nextval(...)`. Columns backed by sequences now correctly appear as `@default(autoincrement())` in the Prisma schema in all cases.
- Savepoint operations (`createSavepoint`, `rollbackToSavepoint`, `releaseSavepoint`) in `@prisma/adapter-d1` now silently no-op with debug logging instead of executing SQL statements.