Updating CI files

[victorb]

This commit updates all CI scripts to the latest version

[victorb]

So upgrading wrtc (to latest 0.0.63) fixes issues on linux, but breaks thing on macOS... However, even with old version, macOS is failing in timeout in one test (even when tests timeout is set to 1 minute [Edit: I see now that that test is overriding the timeout amount...]), and miss-match in number of peers in another...

  1) TCP + WebSockets + WebRTCStar

       nodeAll.dial nodeWStar using PeerInfo:

     Error: Timeout of 10000ms exceeded. For async tests and hooks, ensure "done()" is called; if returning a Promise, ensure it resolves.

  


  2) TCP + WebSockets + WebRTCStar

       nodeAll.hangUp nodeWStar using PeerInfo:


      Uncaught AssertionError: expected [ Array(2) ] to have a length of 3 but got 2

      + expected - actual


      -2

      +3

      

      at parallel (test/nodejs-bundle/tcp+websockets+webrtc-star.js:234:48)

      at node_modules/async/internal/parallel.js:31:39

      at eachOfArrayLike (node_modules/async/eachOf.js:65:9)

      at exports.default (node_modules/async/eachOf.js:9:5)

      at _parallel (node_modules/async/internal/parallel.js:30:5)

      at parallelLimit (node_modules/async/parallel.js:88:26)

      at Timeout.check [as _onTimeout] (test/nodejs-bundle/tcp+websockets+webrtc-star.js:231:9)

[dryajov]

I think we want timeouts per tests instead of global timeouts - should gives us a better idea of what's slow. That's what we've been leaning towards at least.

[victorb]

Thanks! I think setting short timeouts and then having logic for retrying until success is the way to go, then we don't have to fiddle with timeouts.

However, this change here is just made to confirm if things are really broken, or tests are just failing because of timeouts.

[victorb]

Btw, "retrying until success" would look something like this:

js-libp2p/test/nodejs-bundle/tcp+websockets+webrtc-star.js

Lines 197 to 228 in 10357a3

	it('nodeAll.dial nodeWStar using PeerInfo', function (done) {
	nodeAll.dial(nodeWStar.peerInfo, (err) => {
	expect(err).to.not.exist()
	const numberOfTries = 10
	const pauseBetweenTries = 10
	let currentTry = 1
	function check () {
	console.log('checking')
	const peersAll = nodeAll.peerBook.getAll()
	const peersWStar = nodeWStar.peerBook.getAll()
	try {
	expect(Object.keys(peersAll)).to.have.length(3)
	expect(Object.keys(peersWStar)).to.have.length(1)
	expect(Object.keys(nodeAll.swarm.muxedConns)).to.have.length(1)
	console.log('everything ok')
	done()
	} catch (err) {
	if (currentTry >= numberOfTries) {
	console.log('done trying')
	done(err)
	} else {
	currentTry = currentTry + 1
	setTimeout(check, pauseBetweenTries)
	}
	}
	}
	check()
	})
	})

[victorb]

Leaving this for today, back at it tomorrow.

Status:

CircleCI is failing because the machine has to be upgraged, I have not yet done this as it would mean existing PRs and PRs in-between the change is made, and this very PR is merged, would not pass.

Jenkins is having troubles with dialing on windows, ( @richardschneider oh windows god, any ideas? ), log is here:

  1) TCP + WebSockets + WebRTCStar

       nodeAll.dial nodeWStar using PeerInfo:

     Uncaught AssertionError: expected [Error: Could not dial in any of the transports or relays] to not exist

      at nodeAll.dial (test\nodejs-bundle\tcp+websockets+webrtc-star.js:199:26)

      at swarm.dial (src\index.js:267:18)

      at attemptDial (node_modules\libp2p-swarm\src\dial.js:33:20)

      at dialCircuit (node_modules\libp2p-swarm\src\dial.js:96:22)

      at dialCircuit (node_modules\libp2p-swarm\src\dial.js:138:16)

      at nextTransport (node_modules\libp2p-swarm\src\dial.js:94:18)

      at swarm.transport.dial (node_modules\libp2p-swarm\src\dial.js:106:20)

      at dialer.dialMany (node_modules\libp2p-swarm\src\transport.js:55:18)

      at f (node_modules\once\once.js:25:25)

      at map (node_modules\libp2p-swarm\src\limit-dialer\index.js:61:14)

      at node_modules\async\internal\map.js:32:9

      at node_modules\async\internal\once.js:12:16

      at iteratorCallback (node_modules\async\eachOf.js:60:13)

      at node_modules\async\internal\onlyOnce.js:12:16

      at Object.callback (node_modules\async\internal\map.js:29:13)

      at node_modules\async\internal\queue.js:101:31

      at node_modules\async\internal\onlyOnce.js:12:16

      at _dialWithTimeout (node_modules\libp2p-swarm\src\limit-dialer\queue.js:44:16)

      at Timeout.timeoutCallback [as _onTimeout] (node_modules\async\timeout.js:74:13)


  2) TCP + WebSockets + WebRTCStar

       nodeAll.hangUp nodeWStar using PeerInfo:


      Uncaught AssertionError: expected [ Array(2) ] to have a length of 3 but got 2

      + expected - actual


      -2

      +3

      

      at parallel (test\nodejs-bundle\tcp+websockets+webrtc-star.js:239:48)

      at node_modules\async\internal\parallel.js:31:39

      at eachOfArrayLike (node_modules\async\eachOf.js:65:9)

      at exports.default (node_modules\async\eachOf.js:9:5)

      at _parallel (node_modules\async\internal\parallel.js:30:5)

      at parallelLimit (node_modules\async\parallel.js:88:26)

      at Timeout.check [as _onTimeout] (test\nodejs-bundle\tcp+websockets+webrtc-star.js:236:9)

TravisCI seems fine.

[daviddias]

@victorbjelkholm going to merge this PR to have the latest configs and timeouts. Note that:

• I had to revert Circle because the new config was not working
• Jenkins still does not work fully

Mind checking why separately ?