Implement non-blocking I/O with AsyncStream and MessageChannel in Pulp.
npx skills add https://github.com/danielraffel/pulp --skill streams-danielraffel
Please help me install this Agent Skill. Skill: streams Source: https://github.com/danielraffel/pulp/tree/main/.agents/skills/streams Command: npx skills add https://github.com/danielraffel/pulp --skill streams-danielraffel
This Skill provides a structured approach to using the Pulp streaming framework to read, write, and transport data across different I/O backends without deadlocks or race conditions, reducing complexity in async code.
Instantiate an AsyncStream with your event loop and select a transport to begin reading and writing data.
High-intent search queries and answers about installing and using this skill.
Backpressure in async I/O streams is handled by returning a WouldBlock status during read or write operations, which pauses execution and prevents unbounded buffer growth while waiting for the transport to drain.
Preventing race conditions in async callbacks is achieved by using an executor closure to dispatch callbacks onto your own event loop, avoiding library cycles and ensuring safe cross-thread dispatch.
Adding custom transport protocols to an async streaming framework requires implementing the required channel interfaces, enabling new transports like WebSocket or OSC to integrate seamlessly with the unified AsyncStream model.
Yes, you can use a unified AsyncStream and MessageChannel model for local file I/O, inter-process pipes, and TCP or HTTP transports, providing a consistent interface for reading and writing data across different backends.
Safely canceling queued writes in a non-blocking I/O system involves cancelling drains and ensuring clean completion of pending operations, which maintains data integrity and prevents deadlocks during shutdown.
You need a MessageChannel for structured communication when transmitting data over any transport layer, as it provides a unified interface that abstracts the underlying I/O backend and ensures reliable cross-thread dispatch.
Generous-CorpManage asynchronous I/O streams and network transport protocols in the Pulp framework.
dodygBuild network I/O pipelines with System.IO.Pipelines for protocol parsing and backpressure.
tassiovaleStore and manage large N-dimensional arrays in chunked, compressed format.
archibateImplement asyncio patterns for concurrent I/O and task orchestration.
