2.3 KiB
2.3 KiB
Changelog
All notable changes to TypedFetch will be documented in this file.
[0.2.0] - 2025-01-20
Added
- SSE Streaming Support: Native Server-Sent Events with
streamSSE()method - POST Body Support for Streaming: All streaming methods now accept request bodies for AI/ML workloads
- Race Method:
race()for getting the first successful response from multiple endpoints - Batch Processing:
batch()method with concurrency control for efficient bulk operations - Parallel Requests:
parallel()method using Web Workers for true parallelism - Resumable Uploads:
uploadResumable()with adaptive chunking and progress tracking - Streaming Uploads:
uploadStream()for large file/model uploads - Bandwidth Throttling:
throttled()method using token bucket algorithm - Auto-Reconnecting Streams:
streamWithReconnect()for resilient streaming
Changed
- Bundle size increased to 11KB gzipped (from 8KB) due to new features
- Improved TypeScript types for all new methods
- Enhanced error handling for streaming operations
Removed
- Removed duplicate
withExponentialBackoffimplementation (already in retry logic) - Removed AI-specific
trackTokenUsagemethod (too specialized) - Removed
parallelInferencemethod (replaced with genericparallel())
Fixed
- Fixed streaming methods to properly resolve relative URLs with baseURL
- Fixed TypeScript strict mode compatibility issues
- Ensured
throttled()enforces bandwidth limits for streamed responses instead of being a no-op
[0.1.3] - 2025-01-19
Fixed
- Fixed streaming methods (
stream()andstreamJSON()) to support baseURL for relative URLs - Fixed "Failed to parse URL" errors when using streaming with configured instances
[0.1.2] - 2025-01-18
Added
- Initial streaming support with
stream()andstreamJSON()methods - Circuit breaker pattern for fault tolerance
- Request deduplication
- W-TinyLFU caching algorithm
[0.1.1] - 2025-01-17
Added
- Basic retry logic with exponential backoff
- Request/response interceptors
- Cache management
[0.1.0] - 2025-01-16
Added
- Initial release
- Type-safe HTTP client with Proxy-based API
- Zero dependencies
- Support for all HTTP methods
- Automatic JSON parsing
- Error handling with detailed context