-
Notifications
You must be signed in to change notification settings - Fork 333
Conversation
Running this for test, I get this: TypeError: next is not a function
at Server.no_handler (/home/daan/code/scuttlebutt/patchwork/node_modules/ssb-ws/inject.js:20:7)
at Server.emit (events.js:187:15)
at parserOnIncoming (_http_server.js:652:12)
at HTTPParser.parserOnHeadersComplete (_http_common.js:109:17) Not quite sure how to trigger those though, the seem to come in batches of six, but at irregular intervals. |
Thanks! That must be from ssb-ws, I had it pinned to 6.0.0 (I think) because that problem was introduced in 6.1.0, but I *thought* it was resolved in 6.2.0. Guess not.
Feel free to push a commit downgrading ssb-ws or I'll do it in an hour or so. Thanks for the bug hunting!
…On Sat, May 4, 2019, at 07:06, Daan Wynen wrote:
Running this for test, I get this:
TypeError: next is not a function
at Server.no_handler (/home/daan/code/scuttlebutt/patchwork/node_modules/ssb-ws/inject.js:20:7)
at Server.emit (events.js:187:15)
at parserOnIncoming (_http_server.js:652:12)
at HTTPParser.parserOnHeadersComplete (_http_common.js:109:17)
Not quite sure how to trigger those though, the seem to come in batches of six, but at irregular intervals.
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub <#1020 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEDIZERU5VAKIE2WJHYDQLPTWJ5ZANCNFSM4HKXCZBQ>.
|
I thought that the "TypeError: next is not a function" but was resolved in [email protected] but it looks like it still exists. See: ssbc/ssb-ws#19 Note: this commit previously featured a deep update to secret-handshake but it broke the build on Windows. See: auditdrivencrypto/secret-handshake#18
dfedde9
to
ead493c
Compare
okay, no more of those messages, tested on two machines. seems to work for me. |
Cool, I appreciate you testing. Hopefully you're seeing more messages now? I'm seeing normal activity on my end, FWIW. I'm thinking maybe we should release this as a beta, ask for some testing on SSB, and then plan to release 3.12.0 or something on Monday -- what do you think? |
There are so many version changes, bug fixes etc in this release that having some more people test it is definitely a good idea. |
Sweet, once someone has time to approve this PR then I'll merge, release as 3.12.0-beta.0, and ask for folks on SSB to test it out. 👍 Regarding that bug: that's weird, it looks like two instances of atomic-file may be trying to write to |
I can repro this consistently on one machine. I checked and there are no two processes. On the other machine at home it doesn't happen, so it might just be my setup. |
Interesting, I'd love if you could make a bug report so that we can refer to it. I'll merge this and plan to release a beta. The thing with atomic-file is that we use it in a few different places, so I think we could get a race condition even within the same process. Here's a bug I opened upstream a bit ago: ssbc/ssb-server#597
👍 Sounds good to me. |
npm test
via Travis (oops)Anything else we need before a release? Should this be 3.12.0 or 3.12.0-beta?