Marek
Assuming a single read request waiting on tap is voiding the "io_uring gives you batching" argument. It's 100% possible to have more than one packet delivered onto tap in one...
ACPI is indeed not needed (apart from the described halt problem), BLK_DEV_INITRD is needed though. Without it: ``` ~/src/virtme$ sudo ./virtme-run --rw --pwd --kimg ../linux/arch/x86/boot/bzImage --memory 1G --force-initramfs List of...
The "wait" after connect should block until all IP's on all protocols (v4 v6) are tried out and failed. When it fails it should throw an exception, socket.error is fine....
Is that a big deal?
There is some skeleton there, maybe adding that would not be hard: https://github.com/majek/puka/blob/master/puka/machine.py#L150
apologies, I havent looked at the code for years. Perhaps someone on rabbitmq-users could help https://groups.google.com/g/rabbitmq-users
"type" is a keyword in python.
Puka treats your message delivery serious. When you publish with pubacks=False, what do you actually mean by saying "synchronous"? Usually people mean: being sent off the networks. For puka this...
hmm If pubacks=True puka uses publisher acks to synchronize. If pubacks=False puka uses some other obscure mechanism to synchronize, since publishes are asycn.
Ha! Nag lcamtuf, as he neglects to review this patch :)