Migration to next AEgir

[daviddias]

Updated deps and applied the same migrations that happen on the test repo with exception to the PubSub ones https://github.com/ipfs/aegir-test-repo/commit/75c82b3775db69a2e42d1005b767ef78e51161ea##commitcomment-23903275

Seeing a lot of errors coming from JEST, 94 tests fail out of 280.

[dignifiedquire]

js-ipfs-api needs to run in --no-parallel mode because of the way test are set up currently.

[daviddias]

Interesting. Ok added that and now I've multiple of one type of error:

Timeout - Async callback was not invoked within timeout specified by jasmine.DEFAULT_TIMEOUT_INTERVAL.

[dignifiedquire]

that means your tests are too slow and you have to increase the timeout for those

[daviddias]

Ok, updated all the timeouts for the tests that had timeouts. Now I've ECONRESET errors.

[daviddias]

Found that these problems also exist with the test repo, see ipfs-inactive/aegir-test-repo#1

[dignifiedquire]

are you using all patches I made to ipfs-factory on the test repo?

[daviddias]

@dignifiedquire see the PR ipfs-inactive/aegir-test-repo#1, it also fails there. I made no changes to the aegir-test-repo

[daviddias]

Seems that in CI most of the test suites are skipped:

VS local

[daviddias]

nvm, seems that test count is correct, they just not get that pretty print

[dignifiedquire]

@diasdavid one lonely test failing on circle :D

[daviddias]

@dignifiedquire gotta catch them all

[dignifiedquire]

But travis is much slower and still timing out on tons of tests, plus it seems you haven't updated all files yet to pass linting

[daviddias]

• update linting

Let's see now on another run :)

No worries here, I'll be following on this one

[daviddias]

@dignifiedquire something is wrong with code coverage (last piece!)

https://travis-ci.org/ipfs/js-ipfs-api/jobs/269735569#L3113-L3166

Running locally I get nothing:

> ./node_modules/.bin/aegir codecov

> ./node_modules/.bin/aegir coveralls

[dignifiedquire]

seems your commands are wrong:

[daviddias]

Seems that the default doesn't default

Also, running aegir coverage shouldn't require a provider if it isn't going to upload.

[dignifiedquire]

@diasdavid I am confused about your statements

• Uploading works fine to both providers as tested here: https://travis-ci.org/ipfs/aegir#L317
• Uploading only happens if you pass the upload flag -u or --upload
• There is no need to pass a provider if you are not using upload or at all if you are happy with the default

[dignifiedquire]

But I think I see that here is a bug with the aliasing, let me publish a release for that fix, that should solve the issue you are seeing

[dignifiedquire]

published 12.0.1 please try that

[daviddias]

Ok, so now tests take over 40 mins to run in Travis (from previously 17 mins) plus they fail because it seems that npm run coverage interprets timeouts differently - https://travis-ci.org/ipfs/js-ipfs-api/jobs/270369029

[dignifiedquire]

@diasdavid well circle ci finally works. And it has much more reasonable build times which could probably be improved with setting up a parallel pipeline. As I mentioned in the community discussion, I think it is time for us to depart ways with travis

[daviddias]

circle is not running Lint nor coverage

running coverage locally also fails and also takes a ton of extra time. I don't think we can blame travis on this one.

[dignifiedquire]

Coverage is much slower than regular runs (always has been) because of the overhead of instrumentation. Not sure if there is some regression that jest is a lot slower for this though.

[dignifiedquire]

@diasdavid looking at travis output, it seems that the errors are actually not timeouts strange enough: https://travis-ci.org/ipfs/js-ipfs-api/jobs/270369040

[daviddias]

It's always Timeout and then a cascade of failures

[dignifiedquire]

@diasdavid there is no timeout here: https://travis-ci.org/ipfs/js-ipfs-api/jobs/270685309#L4821

[dignifiedquire]

@diasdavid coverage works now on circle, slow travis is still slow

[daviddias]

Thanks @dignifiedquire. I've mixed feelings about accepting our tests getting 2x or more slower due to JEST by just switching coverage to circle..

Just squashed commits and rebased master onto this branch, let's see what's left to do.

[daviddias]

Test Node.js
 FAIL  test/interface/swarm.spec.js
  ● Test suite failed to run

    ENOENT: no such file or directory, open '/Users/imp/code/js-ipfs-api/node_modules/interface-ipfs-core/test/fixtures/testfile.txt'

      at Object.fs.openSync (fs.js:652:18)
      at Object.fs.readFileSync (fs.js:553:33)
      at loadFixtures (node_modules/aegir/src/fixtures.js:9:26)
      at Object.<anonymous> (node_modules/interface-ipfs-core/src/pin.js:12:18)
      at Object.<anonymous> (node_modules/interface-ipfs-core/src/index.js:6:15)

 FAIL  test/interface/pubsub.spec.js
  ● Test suite failed to run

    ENOENT: no such file or directory, open '/Users/imp/code/js-ipfs-api/node_modules/interface-ipfs-core/test/fixtures/testfile.txt'

      at Object.fs.openSync (fs.js:652:18)
      at Object.fs.readFileSync (fs.js:553:33)
      at loadFixtures (node_modules/aegir/src/fixtures.js:9:26)
      at Object.<anonymous> (node_modules/interface-ipfs-core/src/pin.js:12:18)
      at Object.<anonymous> (node_modules/interface-ipfs-core/src/index.js:6:15)

Needs: ipfs-inactive/interface-js-ipfs-core#154

[daviddias]

Ok, some progress. Now it just crashes because...

[daviddias]

Travis go home, you are drunk.

[dignifiedquire]

woooooot it is merged 🎉