Bug fixes for 2.4.0 in sqllin-dsl and sqllin-processor

CHANGELOG.md

@@ -11,10 +11,33 @@
 * Migrate the Android instrumented tests to `Robolectric`, they now run on the JVM as host unit tests against `API 26` and `API 37`, and no longer need an emulator
 * Move the `sqllin-driver` tests back into the `sqllin-driver` module's `commonTest`, and remove the `sqllin-driver-test` module
 
+### sqllin-dsl
+
+* **Breaking change**: The parameter of annotation `@PrimaryKey` renamed from `isAutoincrement` to `autoIncrement`, aligning it with the name already used in the documentation and with the naming of the other annotations. Call sites using the named argument `@PrimaryKey(isAutoincrement = true)` must be updated to `@PrimaryKey(autoIncrement = true)`; positional usage such as `@PrimaryKey(true)` is unaffected
+* **Breaking change**: The nullability of a `@PrimaryKey` property now decides who supplies its value. A `Long?` key is assigned by the database, as before. A non-null `Long` key is now allowed: it remains an `INTEGER PRIMARY KEY`, a rowid alias, but is supplied by the caller and written by every `INSERT`. A key of any other type must be non-null and is declared `NOT NULL`. Previously every `@PrimaryKey` was forced to be nullable, against the annotation's own documentation, and because SQLite does not let `PRIMARY KEY` imply `NOT NULL` on such a column, a `String` key could hold `NULL` in any number of rows. `autoIncrement = true` now requires a `Long?` key, and a `ULong?` key, which used to be stored as `NULL` because it was left out of `INSERT` without being a rowid alias, is now rejected. To migrate, drop the `?` from any non-`Long` `@PrimaryKey`; a single-column `@CompositePrimaryKey` that only existed to hold a caller-supplied `Long` key can become `@PrimaryKey val id: Long`
+* **Breaking change**: `@CompositePrimaryKey` now requires at least two properties. A single-column primary key is declared with `@PrimaryKey`, which for a `Long` key maps to `INTEGER`, a rowid alias, where a single-column `@CompositePrimaryKey` mapped it to `BIGINT`. To migrate, replace a lone `@CompositePrimaryKey` with `@PrimaryKey`
+* **Breaking change**: The DSL APIs `ALERT_ADD_COLUMN` and `ALERT_RENAME_TABLE_TO` renamed to `ALTER_ADD_COLUMN` and `ALTER_RENAME_TABLE_TO`, correcting a misspelling of the SQL keyword `ALTER`. The internal `Alert` operation object is renamed to `Alter` accordingly
+* Fix: the ALTER operations emitted the invalid keyword `ALERT TABLE` instead of `ALTER TABLE`, so `ALTER_ADD_COLUMN`, `ALTER_RENAME_TABLE_TO`, `RENAME_COLUMN` and `DROP_COLUMN` all failed at runtime and had never worked. The tests that covered them swallowed the failure, which is why it went unnoticed
+* Fix documentation: the KDoc of `DatabaseScope` now states that statement execution is deferred until the scope exits, and its example no longer reads a `SelectStatement`'s results while the scope is still open, which throws
+* Fix documentation: the `CREATE_INDEX` and `CREATE_UNIQUE_INDEX` examples referenced a `KClass.table` extension that does not exist in the library, and referred to columns by property reference instead of through the generated table object
+* Fix: the SQL string functions added in 2.2.0, `substr`, `trim`, `ltrim`, `rtrim`, `replace`, `instr` and `printf`, now carry the same DSL marker as the other SQL functions, so IntelliJ IDEA highlights their calls the same way
+* Fix documentation: the installation guide now declares the task dependencies the generated code needs, as SQLlin's own builds always did. Without them Gradle fails the build, in particular when another KSP processor runs in the same module. It also states that each generated object is named after its class with a `Table` suffix, not after the table
+
 ### sqllin-driver
 
 * Update `sqlite-jdbc`'s version to `3.53.4.0`
 
+### sqllin-processor
+
+* Fix: the visibility of the class annotated with `@DBRow` is now propagated to the generated table object. An `internal` `@DBRow` class used to produce a `public` object, which failed to compile with `EXPOSED_SUPER_CLASS`, `EXPOSED_FUNCTION_RETURN_TYPE` and `EXPOSED_RECEIVER_TYPE`. A `@DBRow` class that is neither `public` nor `internal` is now reported as an error
+* Fix: every `SetClause` property generated for a column declared after a nullable `Long` `@PrimaryKey` was typed nullable regardless of the column's own declaration, so `UPDATE ... SET { column = null }` compiled against `NOT NULL` columns and failed only at runtime. Each property now takes the nullability its own column declares
+* Fix: the columns of a `@CompositePrimaryKey` are now declared `NOT NULL`. SQLite, unlike standard SQL, does not let a table-level `PRIMARY KEY` imply it on a rowid table, so such a key used to accept `NULL`, and any number of rows sharing the same key once a `NULL` was part of it. SQLlin itself could not write those `NULL`s, but anything else writing to the database could, and SQLlin then read them back as `0` or an empty string. This only changes the schema of tables created from now on; an existing table keeps its schema, as SQLite cannot add `NOT NULL` to an existing column
+* **Breaking change**: an `ON_DELETE_SET_DEFAULT` or `ON_UPDATE_SET_DEFAULT` foreign key now requires the column to declare `@Default`, as the documentation always said. On `@References` the check was inverted: it accepted a non-null column without a default, whose parent row then could not be deleted (`NOT NULL constraint failed`), and rejected a nullable one. On `@ForeignKey` groups there was no check at all, so a nullable column without a default compiled and was set to `NULL`; that is now rejected too. To migrate, add `@Default` to the column, or use `ON_DELETE_SET_NULL` / `ON_UPDATE_SET_NULL` where setting it to `NULL` is the intent. Both checks hold whichever order `@Default` and the foreign key annotation are written in
+* Fix: a computed property of a `@DBRow` class, one without a backing field such as `val title: String get() = ...`, no longer becomes a column. kotlinx.serialization doesn't serialize such a property, but it was given a `NOT NULL` column that `INSERT` never wrote, so every insert failed with `NOT NULL constraint failed`, and an accessor that looked its column up past the end of the serializer's descriptor
+* Fix: a `@DBRow` property of a type no column can hold, such as a `List`, is now a compile-time error naming the property. It used to be skipped silently: left out of `CREATE TABLE` while its serializer still wrote and read it, so `INSERT` and `SELECT` failed at runtime with "no column named", and as the last property it left a trailing comma that made `CREATE TABLE` itself fail. Annotate such a property with `kotlinx.serialization.Transient` to keep it out of the table
+* Fix: the generated table objects no longer produce a `DSL_MARKER_APPLIED_TO_WRONG_TARGET` warning for every column, which Kotlin 2.3.20 and later report in the module that compiles them. Their `@ColumnNameDslMaker` is there for IntelliJ IDEA's DSL highlighting rather than for the compiler's DSL scope control, so the warning is now suppressed on each generated object
+* Fix: the generated `SetClause` setter of a non-null enum column no longer uses a safe call, `value?.ordinal`, which the module compiling the generated code reported as unnecessary. Only a nullable enum column keeps it
+
 ## 2.3.0 / 2026-08-20
 
 ### All

sqllin-dsl-test/src/androidHostTest/kotlin/com/ctrip/sqllin/dsl/test/AndroidTest.kt

@@ -87,6 +87,9 @@ class AndroidTest {
     @Test
     fun testSchemaModification() = commonTest.testSchemaModification()
 
+    @Test
+    fun testPrimaryKeyNullability() = commonTest.testPrimaryKeyNullability()
+
     @Test
     fun testStringOperators() = commonTest.testStringOperators()