@okikio/observables 1.0.2

package/script/error.js

@@ -0,0 +1,414 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.ObservableError = void 0;
+exports.assertObservableError = assertObservableError;
+exports.isObservableError = isObservableError;
+// @filename: error.ts
+/**
+ * Error primitives and guards for Observable pipelines.
+ *
+ * This entrypoint explains the error values that travel through this library
+ * when an operator uses pass-through error handling. It exports the
+ * `ObservableError` class plus helper functions for narrowing and asserting
+ * those wrapped failures without losing the original error object, stack, or
+ * operator context.
+ *
+ * Reach for this module when you want to inspect failures as data, recover from
+ * an upstream step without throwing away buffered values, or surface richer
+ * debugging information than a plain `Error` can carry on its own.
+ *
+ * @module
+ */
+require("./_dnt.polyfills.js");
+/**
+ * Represents an error that occurred during Observable operations,
+ * with the ability to aggregate multiple underlying errors.
+ *
+ * This class extends AggregateError to provide additional context about
+ * where and how errors occurred in an Observable pipeline. It can collect
+ * multiple errors that occur during a chain of operations while preserving
+ * the contextual information about each error.
+ *
+ * In addition, this class solves a crucial problem with error handling in ReadableStreams. When we call
+ * `controller.error()` on a ReadableStream, it immediately puts the stream in an errored state,
+ * which can cause values emitted before the error to be lost. By wrapping errors as special
+ * values that flow through the normal value channel, we ensure all values emitted before an
+ * error are properly processed.
+ *
+ * Key features:
+ * - Tracks which operator caused the error
+ * - Captures the value being processed when the error occurred
+ * - Aggregates multiple errors from a pipeline
+ * - Preserves the original error objects
+ * - Builds an error chain showing the full path of error propagation
+ */
+class ObservableError extends AggregateError {
+    /**
+     * Creates a new ObservableError.
+     *
+     * @param errors - The error(s) that caused this error
+     * @param message - The error message
+     * @param options - Additional error context
+     */
+    constructor(errors, message, options) {
+        // Normalize errors to an array of Error objects
+        const errorArray = Array.isArray(errors) ? errors : [errors];
+        const normalizedErrors = errorArray.map((err) => err instanceof Error ? err : new Error(String(err)));
+        super(normalizedErrors, message, { cause: options?.cause });
+        /** The operator where the error occurred */
+        Object.defineProperty(this, "operator", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
+        /** The value being processed when the error occurred */
+        Object.defineProperty(this, "value", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
+        /** Helpful potential fixes for errors */
+        Object.defineProperty(this, "tip", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
+        this.name = "ObservableError";
+        this.operator = options?.operator;
+        this.value = options?.value;
+        this.tip = options?.tip;
+    }
+    /**
+     * Returns a string representation of the error including the operator
+     * and value context if available.
+     */
+    toString() {
+        let result = `${this.name}: ${this.message}`;
+        if (this.operator) {
+            result += `\n  in operator: ${this.operator}`;
+        }
+        if (this.value !== undefined) {
+            const valueStr = typeof this.value === "object"
+                ? JSON.stringify(this.value).slice(0, 100) // Truncate long objects
+                : String(this.value);
+            result += `\n  processing value: ${valueStr}`;
+        }
+        if (this.errors.length > 0) {
+            result += "\n  with errors:";
+            this.errors.forEach((err, i) => {
+                result += `\n    ${i + 1}) ${err}`;
+            });
+        }
+        if (this.tip) {
+            result += `\n  tip: ${this.tip}`;
+        }
+        return result;
+    }
+    /**
+     * Creates an ObservableError from any error that occurs during
+     * operator execution.
+     *
+     * @param error - The original error
+     * @param operator - The operator name
+     * @param value - The value being processed
+     * @returns An ObservableError
+     */
+    static from(error, operator, value, tip) {
+        if (error instanceof ObservableError) {
+            // If it's already an ObservableError, add context if not present
+            if (!error.operator && operator) {
+                return new ObservableError(error.errors, error.message, {
+                    operator,
+                    value: error.value || value,
+                    cause: error.cause,
+                    tip: error.tip,
+                });
+            }
+            return error;
+        }
+        // Create a new ObservableError
+        return new ObservableError(error, error instanceof Error ? error.message : String(error), { operator, value, cause: error, tip });
+    }
+}
+exports.ObservableError = ObservableError;
+/**
+ * Asserts that a value is not an ObservableError and narrows the TypeScript type.
+ *
+ * This function acts as a TypeScript assertion function that:
+ * 1. **Type Narrowing**: If the function returns normally, TypeScript knows the value is definitely T (not T | ObservableError)
+ * 2. **Error Handling**: If the value is an ObservableError, either delegates to observer.error or throws
+ * 3. **Never Returns on Error**: When value is ObservableError, this function never returns normally
+ *
+ * **Key Behavior Changes**:
+ * - Now uses TypeScript's `asserts value is T` for proper type narrowing
+ * - When observer handles error, the function still doesn't return normally (assertion still fails)
+ * - Only returns normally when value is definitely not an ObservableError
+ *
+ * **Intent**: Provide type-safe error checking with automatic TypeScript type narrowing.
+ *
+ * **Usage Patterns**:
+ * ```typescript
+ * // Type narrowing in operator results
+ * const result: string | ObservableError = someOperation();
+ * assertObservableError(result); // Throws if error
+ * // TypeScript now knows result is string, not string | ObservableError
+ * console.log(result.toUpperCase()); // ✅ No type error
+ *
+ * // With observer error handling
+ * const observer = { error: err => console.error('Error:', err) };
+ * assertObservableError(result, observer);
+ * // Still throws/doesn't return normally on error, but observer is notified first
+ *
+ * // In subscribe callbacks
+ * observable.subscribe({
+ *   next(value) { // value is T | ObservableError
+ *     assertObservableError(value);
+ *     // value is now narrowed to T
+ *     processCleanValue(value);
+ *   }
+ * });
+ * ```
+ *
+ * @template T - The expected type of the value when it's not an error
+ * @param value - The value to check, which may be either T or ObservableError
+ * @param obs - Optional observer that may contain an error handler function
+ *
+ * @throws {ObservableError} When value is an ObservableError (after notifying observer if provided)
+ *
+ * @example Basic type narrowing
+ * ```typescript
+ * const mixed: string | ObservableError = getValue();
+ * assertObservableError(mixed); // Throws if ObservableError
+ * console.log(mixed.length); // ✅ TypeScript knows mixed is string
+ * ```
+ *
+ * @example With error observer
+ * ```typescript
+ * const observer = {
+ *   error: (err) => analytics.track('error', { message: err.message })
+ * };
+ *
+ * const mixed: User | ObservableError = fetchUser();
+ * assertObservableError(mixed, observer); // Observer notified, then throws
+ * console.log(mixed.name); // ✅ TypeScript knows mixed is User
+ * ```
+ *
+ * @example In operator pipeline
+ * ```typescript
+ * import { pipe, map, tap } from './helpers/mod.ts';
+ *
+ * pipe(
+ *   source,
+ *   map(x => processX(x)), // Returns T | ObservableError
+ *   tap(result => {
+ *     assertObservableError(result); // Type narrows from T | ObservableError to T
+ *     sendAnalytics(result); // ✅ result is definitely T
+ *   })
+ * )
+ * ```
+ */
+function assertObservableError(value, obs) {
+    if (value instanceof ObservableError) {
+        // Notify observer first if available
+        if (typeof obs?.error === "function") {
+            try {
+                obs.error(value);
+            }
+            catch (observerError) {
+                // If observer throws, still throw the original error
+                // but log the observer error to avoid losing it
+                console.error("Observer error handler threw:", observerError);
+            }
+        }
+        // Always throw when value is ObservableError
+        // This ensures the function never returns normally for errors
+        throw value;
+    }
+    // If we reach here, value is definitely T (not T | ObservableError)
+    // TypeScript assertion function automatically narrows the type
+}
+/**
+ * Checks if a value is an ObservableError without throwing exceptions.
+ *
+ * When working with Observable pipelines, you often receive values that could be either successful
+ * results or errors. This creates a dilemma: how do you safely check what you got without risking
+ * crashes or poor performance? That's exactly what this function solves.
+ *
+ * **Why This Function Exists**:
+ *
+ * Imagine you're processing a stream of data where some items might be errors. Without a proper
+ * way to distinguish between good data and errors, you'd have to either:
+ * - Risk calling methods on errors (which crashes your app)
+ * - Write repetitive `instanceof` checks everywhere (which clutters your code)
+ * - Use try/catch blocks for control flow (which hurts performance)
+ *
+ * This function eliminates all those problems by giving you a clean, fast way to ask:
+ * "Is this thing an error?" If yes, you can handle it appropriately. If no, you can
+ * safely process it as valid data.
+ *
+ * **How It Fits With Other Functions**:
+ *
+ * Think of this as the gentle cousin of `assertObservableError()`. While `assertObservableError()`
+ * says "this better not be an error or I'm throwing an exception," this function politely asks
+ * "excuse me, are you an error?" and gives you a yes/no answer.
+ *
+ * This makes it perfect for situations where you want to handle both success and error cases
+ * gracefully, rather than just crashing when something goes wrong.
+ *
+ * **Performance Story**:
+ *
+ * Under the hood, this function uses a simple `instanceof` check. Now, you might wonder:
+ * "Why not just use `instanceof` directly?" The answer lies in convenience and consistency.
+ *
+ * Here's what actually happens performance-wise:
+ * - This function IS an `instanceof` check (no performance difference there)
+ * - Modern JavaScript engines are extremely good at optimizing `instanceof`
+ * - The real performance win comes from avoiding try/catch blocks for control flow
+ * - When the engine inlines this function, the overhead becomes virtually zero
+ *
+ * **Performance Reality Check**:
+ * - `isObservableError(value)` and `value instanceof ObservableError` perform identically
+ * - Both are fast enough that you'll never notice the difference in real applications
+ * - The real performance benefit is architectural: avoiding exceptions for normal control flow
+ * - Exception throwing/catching can be 10-100x slower, but that's comparing apples to oranges
+ *
+ * In practice, use this function because it's clearer and more consistent with the library's
+ * design patterns, not because of micro-optimizations.
+ *
+ * **Common Ways to Use This Function**:
+ *
+ * Let's walk through the most common scenarios where this function shines:
+ *
+ * ```typescript
+ * // Scenario 1: Branching logic in data processing
+ * function processItem<T>(item: T | ObservableError) {
+ *   if (isObservableError(item)) {
+ *     // TypeScript now knows item is an ObservableError
+ *     console.error('Something went wrong:', item.message);
+ *     return null; // or however you want to handle errors
+ *   }
+ *
+ *   // TypeScript now knows item is T - no more type errors!
+ *   return transformData(item);
+ * }
+ *
+ * // Scenario 2: Filtering out errors from a collection
+ * // Note: You need proper type predicates for TypeScript to understand
+ * const cleanData = mixedResults.filter((item): item is T => !isObservableError(item));
+ * // cleanData is now properly typed as T[]
+ *
+ * // Scenario 3: Separating errors from successes
+ * const errors = results.filter(isObservableError);
+ * const successes = results.filter((item): item is T => !isObservableError(item));
+ * // Now you can handle each group with proper typing
+ *
+ * // Scenario 4: Early exit optimization
+ * function expensiveCalculation<T, U>(input: T | ObservableError): U | ObservableError {
+ *   if (isObservableError(input)) {
+ *     return input; // Don't waste time processing errors
+ *   }
+ *
+ *   // Only do expensive work on valid data
+ *   return performComplexOperation(input);
+ * }
+ * ```
+ *
+ * **What Makes This Function Safe**:
+ *
+ * Unlike functions that might throw exceptions, this one is designed to never fail.
+ * It gracefully handles all the weird edge cases you might encounter:
+ * - Null or undefined values? Returns false (they're not errors)
+ * - Numbers, strings, or other primitives? Returns false (can't be ObservableErrors)
+ * - Objects that aren't ObservableErrors? Returns false (not what we're looking for)
+ * - Subclasses of ObservableError? Returns true (proper inheritance support)
+ *
+ * This means you can safely call it on anything without worrying about crashes.
+ *
+ * **When to Use This vs Other Options**:
+ *
+ * Choose `isObservableError()` when:
+ * - You want to handle both error and success cases in your code
+ * - You're filtering or sorting mixed arrays of results and errors
+ * - You're building conditional logic that branches based on error state
+ * - You want to avoid exceptions and prefer explicit error handling
+ *
+ * Choose `assertObservableError()` instead when:
+ * - You expect the value to NOT be an error, and want to crash if it is
+ * - You're in a context where errors should stop processing immediately
+ * - You want TypeScript to automatically narrow types via assertion
+ *
+ * @template T - The expected type for successful (non-error) values
+ * @param value - Any value that might or might not be an ObservableError
+ * @returns true if the value is an ObservableError, false otherwise
+ *
+ * @example Simple error checking
+ * ```typescript
+ * const result: string | ObservableError = fetchData();
+ *
+ * if (isObservableError(result)) {
+ *   console.error('Oops, something went wrong:', result.message);
+ *   // result is typed as ObservableError here
+ *   return;
+ * }
+ *
+ * // result is typed as string here
+ * console.log('Success! Got:', result.toUpperCase());
+ * ```
+ *
+ * @example Filtering errors from a list
+ * ```typescript
+ * const mixedResults: (User | ObservableError)[] = await fetchAllUsers();
+ *
+ * // Get only the successful results (with proper type predicate)
+ * const validUsers = mixedResults.filter((item): item is User => !isObservableError(item));
+ * // validUsers is now correctly typed as User[]
+ *
+ * // Get only the errors
+ * const errors = mixedResults.filter(isObservableError);
+ * // errors is now typed as ObservableError[]
+ *
+ * // Process each group separately
+ * if (errors.length > 0) {
+ *   console.error(`Found ${errors.length} errors:`, errors);
+ * }
+ * console.log(`Processing ${validUsers.length} valid users`);
+ * ```
+ *
+ * @example Building error-resilient operators
+ * ```typescript
+ * import { pipe } from './helpers/mod.ts';
+ *
+ * function safeMap<T, U>(transform: (value: T) => U) {
+ *   return (source: Observable<T | ObservableError>) => {
+ *     return pipe(
+ *       source,
+ *       map(value => {
+ *         // Skip processing errors - just pass them through
+ *         if (isObservableError(value)) return value;
+ *
+ *         // Only transform valid values
+ *         try {
+ *           return transform(value);
+ *         } catch (error) {
+ *           return ObservableError.from(error, 'safeMap', value);
+ *         }
+ *       })
+ *     );
+ *   };
+ * }
+ *
+ * // Usage
+ * const result = pipe(
+ *   mixedDataStream,
+ *   safeMap(user => user.name.toUpperCase())
+ * );
+ * ```
+ */
+function isObservableError(value) {
+    // This is just a straightforward instanceof check
+    // Modern JavaScript engines optimize this extremely well
+    return value instanceof ObservableError;
+}