multicast example panic (Mac) about message-io HOT 5 CLOSED

Hi @dabat, thanks for the comment, I'm glad you like it 😃

It's really weird since on my computer this is working and this functionality is the base of termchat and anybody complains about it...

I checked the code and the reuse address is set to true correctly in multicast to avoid these issues.

Could you tell me the OS version are you using? Thanks!

from message-io.

Hi, what's the state of this issue? Did it work?

from message-io.

hi @lemunozm my apologies for not getting back to you, the emails were buried in my inbox.
i am still seeing this error in the multicast example.

i'm running on macOS 11.4 and i pulled latest on master branch.
as you can see in the error message,
thread 'main' panicked at 'called Result::unwrap()on anErr value: Os { code: 48, kind: AddrInUse, message: "Address already in use" }', examples/multicast/main.rs:23:70
this is the line that is panicing

from message-io.

Hi @dabat , thanks so much for your answer.

I made a integration test to run this use-case in CI: #94

As you noticed, multicast is not working in newer versions of Mac. Mac requires to add also the SO_REUSEPORT option in addition to SO_REUSEADDR.

Two interesting resources about it:

• https://stackoverflow.com/questions/32661091/behavior-of-so-reuseaddr-and-so-reuseport-changed
• https://stackoverflow.com/questions/14388706/how-do-so-reuseaddr-and-so-reuseport-differ
• https://stackoverflow.com/questions/67541705/ios-socket-with-so-reuseport-does-not-receive-all-multicast-packets (possible future problem?)

Solution overview:

net2 (used currently to enable SO_REUSEADDR) do not support SO_REUSE_PORT. Use socket2 instead that support it

from message-io.

Fixed in version 0.14.2

Since the test is already working in Mac I close the issue. If you continue having problems, feel free to reopen it.

Thanks for notifying the error!

from message-io.