Major refactor to introduce safer patterns

README.md

@@ -110,39 +110,166 @@ have been introduced during the migration.
 
 ## Connection Pooling
 
-By default, one connection will be used for both reading and writing, and only one thread can acquire that connection 
-at a time. If you have WAL enabled, you could (and should) set the amount of pooled reader connections that will be used:
+SQLite supports several concurrency models that can significantly impact your application's performance. This driver
+provides flexible connection pooling through the `AndroidxSqliteConcurrencyModel` interface.
+
+### Available Concurrency Models
+
+#### 1. SingleReaderWriter
+
+The simplest model with one connection handling all operations:
 
 ```kotlin
-AndroidxSqliteDriver(
-  ...,
-  readerConnections = 4,
-  ...,
+AndroidxSqliteConfiguration(
+  concurrencyModel = AndroidxSqliteConcurrencyModel.SingleReaderWriter
 )
 ```
 
-On Android you can defer to the system to determine how many reader connections there should be<sup>[1]</sup>:
+**Best for:**
+
+- Simple applications with minimal database usage
+- Testing and development
+- When memory usage is a primary concern
+- Single-threaded applications
+
+#### 2. MultipleReaders
+
+Dedicated reader connections for read-only access:
+
+```kotlin
+AndroidxSqliteConfiguration(
+  concurrencyModel = AndroidxSqliteConcurrencyModel.MultipleReaders(
+    readerCount = 3  // Number of concurrent reader connections
+  )
+)
+```
+
+**Best for:**
+
+- Read-only applications (analytics dashboards, reporting tools)
+- Data visualization and content browsing applications
+- Scenarios where all writes happen externally (data imports, ETL processes)
+- Applications that only query pre-populated databases
+
+**Important:** This model is designed for **read-only access**. No write operations (INSERT, UPDATE, DELETE) should be
+performed. If you need write capabilities, use `MultipleReadersSingleWriter` in WAL mode instead.
+
+#### 3. MultipleReadersSingleWriter (Recommended)
+
+The most flexible model that adapts based on journal mode:
+
+```kotlin
+AndroidxSqliteConfiguration(
+  concurrencyModel = AndroidxSqliteConcurrencyModel.MultipleReadersSingleWriter(
+    isWal = true,        // Enable WAL mode for true concurrency
+    walCount = 4,        // Reader connections when WAL is enabled
+    nonWalCount = 0      // Reader connections when WAL is disabled
+  )
+)
+```
+
+**Best for:**
+
+- Most production applications
+- Mixed read/write workloads
+- When you want to leverage WAL mode benefits
+- Applications requiring optimal performance
+
+### WAL Mode Benefits
+
+- **True Concurrency**: Readers and writers don't block each other
+- **Better Performance**: Concurrent operations improve throughput
+- **Consistency**: ACID properties are maintained (when `PRAGMA synchronous = FULL` is used)
+- **Scalability**: Handles higher concurrent load
+
+### Choosing Reader Connection Count
+
+The optimal number of reader connections depends on your use case:
+
+```kotlin
+// Conservative (default)
+AndroidxSqliteConcurrencyModel.MultipleReadersSingleWriter(
+  isWal = true,
+  walCount = 4,
+  nonWalCount = 0,
+)
+
+// High-concurrency applications
+AndroidxSqliteConcurrencyModel.MultipleReadersSingleWriter(
+  isWal = true, 
+  walCount = 8
+)
+
+// Memory-conscious applications
+AndroidxSqliteConcurrencyModel.MultipleReadersSingleWriter(
+  isWal = true,
+  walCount = 2
+)
+```
+
+### Platform-Specific Configuration
+
+On Android, you can use system-determined connection pool sizes:
 
 ```kotlin
 // Based on SQLiteGlobal.getWALConnectionPoolSize()
-fun getWALConnectionPoolSize() {
+fun getWALConnectionPoolSize(): Int {
   val resources = Resources.getSystem()
-  val resId =
-    resources.getIdentifier("db_connection_pool_size", "integer", "android")
+  val resId = resources.getIdentifier("db_connection_pool_size", "integer", "android")
   return if (resId != 0) {
     resources.getInteger(resId)
   } else {
-    2
+    2  // Fallback default
   }
 }
+
+AndroidxSqliteConfiguration(
+  concurrencyModel = AndroidxSqliteConcurrencyModel.MultipleReadersSingleWriter(
+    isWal = true,
+    walCount = getWALConnectionPoolSize(),
+    nonWalCount = 0,
+  )
+)
 ```
 
-See [WAL & Dispatchers] for more information about how to configure dispatchers to use for reads and writes.
+### Performance Considerations
+
+| Model                                 | Memory Usage | Read Concurrency | Write Capability | Best Use Case      |
+|---------------------------------------|--------------|------------------|------------------|--------------------|
+| SingleReaderWriter                    | Lowest       | None             | Full             | Simple apps        |
+| MultipleReaders                       | Medium       | Excellent        | None (read-only) | Read-only apps     |
+| MultipleReadersSingleWriter (WAL)     | Higher       | Excellent        | Full             | Production         |
+| MultipleReadersSingleWriter (non-WAL) | Medium       | Limited          | Full             | Legacy/constrained |
+
+### Special Database Types
 
 > [!NOTE]  
-> In-Memory and temporary databases will always use 0 reader connections i.e. there will be a single connection 
+> In-Memory and temporary databases automatically use `SingleReaderWriter` model regardless of configuration, as
+> connection pooling provides no benefit for these database types.
+
+### Journal Mode
+
+If `PRAGMA journal_mode = ...` is used, the connection pool will:
+
+1. Acquire the writer connection
+2. Acquire all reader connections
+3. Close all reader connections
+4. Run the `PRAGMA` statement
+5. Recreate the reader connections
+
+This ensures all connections use the same journal mode and prevents inconsistencies.
+
+### Best Practices
+
+1. **Start with defaults**: Uses `MultipleReadersSingleWriter` in WAL mode
+2. **Monitor performance**: Profile your specific workload to determine optimal reader count
+3. **Consider memory**: Each connection has overhead - balance performance vs memory usage
+4. **Test thoroughly**: Verify your concurrency model works under expected load
+5. **Platform differences**: Android may have different optimal settings than JVM/Native
+
+See [WAL & Dispatchers] for more information about how to configure dispatchers to use for reads and writes.
 
-[1]: https://blog.p-y.wtf/parallelism-with-android-sqlite#heading-secondary-connections
 [AndroidX Kotlin Multiplatform SQLite]: https://developer.android.com/kotlin/multiplatform/sqlite
 [SQLDelight]: https://github.com/sqldelight/sqldelight
 [WAL & Dispatchers]: https://blog.p-y.wtf/parallelism-with-android-sqlite#heading-wal-amp-dispatchers
+[Write-Ahead Logging]: https://sqlite.org/wal.html

integration/src/commonTest/kotlin/com/eygraber/sqldelight/androidx/driver/integration/AndroidxSqliteConcurrencyIntegrationTest.kt

@@ -1,6 +1,7 @@
 package com.eygraber.sqldelight.androidx.driver.integration
 
 import app.cash.sqldelight.coroutines.asFlow
+import com.eygraber.sqldelight.androidx.driver.AndroidxSqliteConcurrencyModel.MultipleReadersSingleWriter
 import com.eygraber.sqldelight.androidx.driver.AndroidxSqliteConfiguration
 import com.eygraber.sqldelight.androidx.driver.AndroidxSqliteDatabaseType
 import kotlinx.coroutines.delay
@@ -23,7 +24,10 @@ class AndroidxSqliteConcurrencyIntegrationTest : AndroidxSqliteIntegrationTest()
     // having 2 readers instead of the default 4 makes it more
     // likely to have concurrent readers using the same cached statement
     configuration = AndroidxSqliteConfiguration(
-      readerConnectionsCount = 2,
+      concurrencyModel = MultipleReadersSingleWriter(
+        isWal = true,
+        walCount = 2,
+      ),
     )
 
     launch {

integration/src/commonTest/kotlin/com/eygraber/sqldelight/androidx/driver/integration/AndroidxSqliteIntegrationTest.kt

@@ -22,8 +22,8 @@ abstract class AndroidxSqliteIntegrationTest {
 
   @OptIn(ExperimentalCoroutinesApi::class, DelicateCoroutinesApi::class)
   private fun readDispatcher(): CoroutineDispatcher? = when {
-    configuration.readerConnectionsCount >= 1 -> newFixedThreadPoolContext(
-      nThreads = configuration.readerConnectionsCount,
+    configuration.concurrencyModel.readerCount >= 1 -> newFixedThreadPoolContext(
+      nThreads = configuration.concurrencyModel.readerCount,
       name = "db-reads",
     )
     else -> null

library/src/commonMain/kotlin/com/eygraber/sqldelight/androidx/driver/AndroidxSqliteConcurrencyModel.kt

@@ -0,0 +1,110 @@
+package com.eygraber.sqldelight.androidx.driver
+
+/**
+ * Defines the concurrency model for SQLite database connections, controlling how many
+ * reader and writer connections are maintained in the connection pool.
+ *
+ * SQLite supports different concurrency models depending on the journal mode and application needs:
+ * - Single connection for simple use cases
+ * - Multiple readers with WAL (Write-Ahead Logging) for better read concurrency
+ * - Configurable reader counts for fine-tuned performance
+ *
+ * @property readerCount The number of reader connections to maintain in the pool
+ */
+public sealed interface AndroidxSqliteConcurrencyModel {
+  public val readerCount: Int
+
+  /**
+   * Single connection model - one connection handles both reads and writes.
+   *
+   * This is the simplest and most conservative approach, suitable for:
+   * - Applications with low concurrency requirements
+   * - Simple database operations
+   * - Testing scenarios
+   * - When database contention is not a concern
+   *
+   * **Performance characteristics:**
+   * - Lowest memory overhead
+   * - No connection pooling complexity
+   * - Sequential read/write operations only
+   * - Suitable for single-threaded or low-concurrency scenarios
+   */
+  public data object SingleReaderWriter : AndroidxSqliteConcurrencyModel {
+    override val readerCount: Int = 0
+  }
+
+  /**
+   * Multiple readers model - allows concurrent read operations only.
+   *
+   * This model creates a pool of dedicated reader connections for read-only access.
+   * **No write operations should be performed** when using this model.
+   *
+   * **Use cases:**
+   * - Read-only applications (analytics dashboards, reporting tools)
+   * - Data visualization and content browsing applications
+   * - Scenarios where all writes happen externally (e.g., data imports)
+   * - Applications that only query pre-populated databases
+   *
+   * **Performance characteristics:**
+   * - Excellent read concurrency
+   * - Higher memory overhead due to connection pooling
+   * - No write capability - reads only
+   * - Optimal for read-heavy workloads with no database modifications
+   *
+   * **Important:** This model is designed for read-only access. If your application
+   * needs to perform any write operations (INSERT, UPDATE, DELETE, schema changes),
+   * use `MultipleReadersSingleWriter` in WAL mode instead.
+   *
+   * @param readerCount Number of reader connections to maintain (typically 2-8)
+   */
+  public data class MultipleReaders(
+    override val readerCount: Int,
+  ) : AndroidxSqliteConcurrencyModel
+
+  /**
+   * Multiple readers with single writer model - optimized for different journal modes.
+   *
+   * This is the most flexible model that adapts its behavior based on whether
+   * Write-Ahead Logging (WAL) mode is enabled:
+   *
+   * **WAL Mode (isWal = true):**
+   * - Enables true concurrent reads and writes
+   * - Readers don't block writers and vice versa
+   * - Best performance for mixed read/write workloads
+   * - Uses `walCount` reader connections
+   *
+   * **Non-WAL Mode (isWal = false):**
+   * - Falls back to traditional SQLite locking
+   * - Reads and writes are still serialized
+   * - Uses `nonWalCount` reader connections (typically 0)
+   *
+   * **Recommended configuration:**
+   * ```kotlin
+   * // For WAL mode
+   * MultipleReadersSingleWriter(
+   *   isWal = true,
+   *   walCount = 4  // Good default for most applications
+   * )
+   *
+   * // For non-WAL mode
+   * MultipleReadersSingleWriter(
+   *   isWal = false,
+   *   nonWalCount = 0  // Single connection is often sufficient
+   * )
+   * ```
+   *
+   * @param isWal Whether WAL (Write-Ahead Logging) journal mode is enabled
+   * @param nonWalCount Number of reader connections when WAL is disabled (default: 0)
+   * @param walCount Number of reader connections when WAL is enabled (default: 4)
+   */
+  public data class MultipleReadersSingleWriter(
+    public val isWal: Boolean,
+    public val nonWalCount: Int = 0,
+    public val walCount: Int = 4,
+  ) : AndroidxSqliteConcurrencyModel {
+    override val readerCount: Int = when {
+      isWal -> walCount
+      else -> nonWalCount
+    }
+  }
+}

library/src/commonMain/kotlin/com/eygraber/sqldelight/androidx/driver/AndroidxSqliteConfigurableDriver.kt

@@ -2,21 +2,23 @@ package com.eygraber.sqldelight.androidx.driver
 
 import app.cash.sqldelight.db.QueryResult
 import app.cash.sqldelight.db.SqlCursor
+import app.cash.sqldelight.db.SqlDriver
 import app.cash.sqldelight.db.SqlPreparedStatement
 
 public class AndroidxSqliteConfigurableDriver(
-  private val driver: AndroidxSqliteDriver,
+  private val driver: SqlDriver,
 ) {
   public fun setForeignKeyConstraintsEnabled(isForeignKeyConstraintsEnabled: Boolean) {
-    driver.setForeignKeyConstraintsEnabled(isForeignKeyConstraintsEnabled)
+    val foreignKey = if(isForeignKeyConstraintsEnabled) "ON" else "OFF"
+    executePragma("foreign_keys = $foreignKey")
   }
 
   public fun setJournalMode(journalMode: SqliteJournalMode) {
-    driver.setJournalMode(journalMode)
+    executePragma("journal_mode = ${journalMode.value}")
   }
 
   public fun setSync(sync: SqliteSync) {
-    driver.setSync(sync)
+    executePragma("synchronous = ${sync.value}")
   }
 
   public fun executePragma(
@@ -27,10 +29,10 @@ public class AndroidxSqliteConfigurableDriver(
     driver.execute(null, "PRAGMA $pragma;", parameters, binders)
   }
 
-  public fun <T> executePragmaQuery(
+  public fun <R> executePragmaQuery(
     pragma: String,
-    mapper: (SqlCursor) -> QueryResult<T>,
+    mapper: (SqlCursor) -> QueryResult<R>,
     parameters: Int = 0,
     binders: (SqlPreparedStatement.() -> Unit)? = null,
-  ): QueryResult.Value<T> = driver.executeQuery(null, "PRAGMA $pragma;", mapper, parameters, binders)
+  ): QueryResult<R> = driver.executeQuery(null, "PRAGMA $pragma;", mapper, parameters, binders)
 }