Conversation
|
Is this still relevant? If so, what is blocking it? Is there anything you can do to help move it forward? |
|
@christianbundy also this |
|
@christianbundy I think maybe this just needs to be merged. check if the tests of sbot and multiserver still work when this is linked in. |
|
if it does then hit merge |
|
This branch isn't passing tests for me locally, and Travis is reporting the same problem. It seems to be hanging here: |
|
It works if you link in the upgraded pull-ws |
|
Merge with main. Waiting for pull-ws to be merged to be able to move forward. After that is in, the tests for this should be working again. |
|
It seems like pull-ws is dead. Nobody with commit access seems interested. I'm considering just forking it and using that for my projects, and update multiserver to use the fork. Any thoughts on this? |
|
sounds great to me |
|
I'm in support of forking too @arj03 |
|
Ah right. Had forgotten about this one completely. Thanks for the ping :-) |
|
Alright so I think this is ready. I forked pull-ws as pull-websocket and fixed the tests. With this I have had a quite stable room connection with patchwork. @cryptix shall we test rooms2 tomorrow? :) |
|
Cool, I'd like to merge it but gotta fix some git conflicts first |
|
The conflicts should be fixed, the only thing I'd maybe like to fix is this line. That test was not in the old code so I don't think it's a new problem. |
|
Alrighty part2: I fixed the tests now as well (mainly in pull-websocket). When you call close on a server it will only stop accepting new connections, not actually kill existing. IIRC the argument for that is related to deadlocks and its a common gotcha. |
|
Does your comment @staltz mean that you are okay if I merge this? :) |
|
Yes! |
These changes are necessary after recent changes to pull-stream/pull-ws#26
I also improved test coverage, reusing the same tests for net and ws. @mixmix