handle hypercore replication timeouts

[cblgh]

originally suggested by @substack, it seems we have a few connectivity issues for sessions that are open for a long amount of time.

by adding a timeout to the replication stream, we can likely solve the problems we've been seeing

implementation

• upgrade to hypercore@8
• solve the issue with the following hypercore-protocol methods
  • stream.setTimeout(ms, ontimeout)
  • stream.on('timeout', cb)

[telamon]

Um. i've been working on upgrading the replication to proto7 since last week, you should be able to simply upgrade hypercore dependency to v8 once multifeed pr#32 is merged and my working branch replication#proto7 is merged.

edit: core replication tests passed yesterday (yay!)

[cblgh]

ohhh that's great news, yr wonderful @telamon ^_^

[hackergrrl]

hypercore-protocol@6.x.x implements a ping & timeout feature already:

https://github.com/mafintosh/hypercore-protocol/blob/v6.12.0/index.js#L61

Does this not currently work?

[cblgh]

Ah, I missed that.

On Sun, 22 Sep 2019 at 05:55, noffle notifications@github.com wrote:

hypercore-protocol@6.x.x implements a ping & timeout feature already:

https://github.com/mafintosh/hypercore-protocol/blob/v6.12.0/index.js#L61

Does this not currently work?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cabal-club/cabal-core/issues/58?email_source=notifications&email_token=AA5O6WW5SJTHRRIU4AIHW5LQK3UEDA5CNFSM4IY5IYRKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD7I552Q#issuecomment-533847786, or mute the thread https://github.com/notifications/unsubscribe-auth/AA5O6WSJ46LEBDCOWL2A643QK3UEDANCNFSM4IY5IYRA .

[hackergrrl]

I think this ok now that we're on hypercore 8? It sends keep-alive messages every 10s by default.