jeffparsons commented on issue #7148:
Is this meant to support broadcast? I couldn't find any mention of it in the spec (maybe I just wasn't looking properly?) so I'm wondering whether this is just a question of whether the host is configured to allow
the guest to send to a broadcast address.
rvolosatovs commented on issue #7148:
Is this meant to support broadcast? I couldn't find any mention of it in the spec (maybe I just wasn't looking properly?) so I'm wondering whether this is just a question of whether the host is configured to allow the guest to send to a broadcast address.
I feel like this is something that belongs in https://github.com/WebAssembly/wasi-sockets/blob/main/wit/udp.wit, e.g.
set-broadcast
method onudp-socket
alexcrichton commented on issue #7148:
I might also recommend opening an issue on https://github.com/WebAssembly/wasi-sockets/issues for support if you're interested
rvolosatovs commented on issue #7148:
@alexcrichton I believe I've addressed everything apart from the https://github.com/bytecodealliance/wasmtime/pull/7148#discussion_r1348810978
rvolosatovs commented on issue #7148:
Any idea what went wrong with the merge? Did I miss something?
rvolosatovs edited a comment on issue #7148:
Any idea what went wrong with the merge? Did I miss something?
Update: I see it now https://github.com/bytecodealliance/wasmtime/actions/runs/6435495256/job/17476888343#step:13:1
I'm away from my laptop already, but I will fix it as soon as I'm back
rvolosatovs commented on issue #7148:
@alexcrichton this should be good to merge now
rvolosatovs commented on issue #7148:
Ah, looks like there's also a wit syntax change to address after a rebase
rvolosatovs commented on issue #7148:
Rebased/updated yet again
badeend commented on issue #7148:
Apologies for the late comment, I just noticed this PR.
- Why does
start_connect
perform an explicit bind to0.0.0.0:0
? This shouldn't be necessary. The first call toconnect
will implicitly bind the socket for you, right?- On UDP sockets, the native
connect
call doesn't perform any IO and always succeeds or fails synchronously. It never returns EINPROGRESS. I suspect you can drop a lot of complexity instart_connect
&finish_connect
.I have the changes ready to go if you want: https://github.com/rvolosatovs/wasmtime/pull/1
rvolosatovs commented on issue #7148:
Apologies for the late comment, I just noticed this PR.
- Why does
start_connect
perform an explicit bind to0.0.0.0:0
? This shouldn't be necessary. The first call toconnect
will implicitly bind the socket for you, right?- On UDP sockets, the native
connect
call doesn't perform any IO and always succeeds or fails synchronously. It never returns EINPROGRESS. I suspect you can drop a lot of complexity instart_connect
&finish_connect
.I have the changes ready to go if you want: https://github.com/rvolosatovs/wasmtime/pull/1
Thank you, integrated into this PR!
badeend commented on issue #7148:
#7120 was just merged and is causing some conflicts with this one. I've resolved them for you in https://github.com/rvolosatovs/wasmtime/pull/2. Also I've revised the test case; that may help with the CI failure.
rvolosatovs commented on issue #7148:
@badeend thank you for working on this. Yes, my initial thought was to wait for #7120 actually, but back then it was not ready for merge, and I hoped to get this one in quick (which did not really work out as planned)
I've integrated your changes
rvolosatovs edited a comment on issue #7148:
@badeend thank you for working on this. Yes, my initial thought was to wait for #7120 actually, but back then it was not ready for merge, and I hoped to get this one in quick (which did not really work out as planned)
I've integrated and tested your changes
Last updated: Dec 23 2024 at 12:05 UTC