You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I start a child process using fork() and { serialization: 'advanced' }. Then transfer messages between the child and the parent process.
Sometimes the transfer hangs at the receiving side.
How often does it reproduce? Is there a required condition?
Rarely, after several hours of high rate messaging
What is the expected behavior? Why is that the expected behavior?
Transmitting messages without hangups
What do you see instead?
The queue (kMessageBuffer) on the receiving side is constantly growing without processing.
The length of the first buffer in the kMessageBuffer is less than 4 bytes, so new messages are not processed and are only added to the end of queue.
Additional information
I suppose that this PR led to this behavior.
Instead of concatenating the queue and then checking the length, only the first buffer is checked.
This does not take into account that the first buffer may be smaller than 4 bytes after processing the previous message (code)
If the first buffer is less than 4 bytes, it must be concat with with the next one, if there is one.
The text was updated successfully, but these errors were encountered:
Version
20.12.2
Platform
Subsystem
child_process
What steps will reproduce the bug?
I start a child process using fork() and { serialization: 'advanced' }. Then transfer messages between the child and the parent process.
Sometimes the transfer hangs at the receiving side.
How often does it reproduce? Is there a required condition?
Rarely, after several hours of high rate messaging
What is the expected behavior? Why is that the expected behavior?
Transmitting messages without hangups
What do you see instead?
The queue (kMessageBuffer) on the receiving side is constantly growing without processing.
The length of the first buffer in the kMessageBuffer is less than 4 bytes, so new messages are not processed and are only added to the end of queue.
Additional information
I suppose that this PR led to this behavior.
Instead of concatenating the queue and then checking the length, only the first buffer is checked.
This does not take into account that the first buffer may be smaller than 4 bytes after processing the previous message (code)
If the first buffer is less than 4 bytes, it must be concat with with the next one, if there is one.
The text was updated successfully, but these errors were encountered: