-
Notifications
You must be signed in to change notification settings - Fork 1.2k
fix build on nodejs stable - update hapi when it is released #541
Comments
Not sure a deprecation warning should break are tests here, but I do agree we should change it. The error is a bit weird though.
Why would we check that the output is "deeply equal" to empty string? Not sure if it's the test cases that are bad, or just printing of the error, but something is wrong here. |
@victorbjelkholm the checks do make sense as it is testing the cli interface and checking stdout is the way to do it. Similar many of the sharness tests will run into this. |
@dignifiedquire checking that there is no output doesn't feel like a very stable way of testing a CLI. Maybe we could turn on verbose/debug output and make assertions on that? |
@victorbjelkholm the other we test it is with sharness, any other way you would recommend? |
I'm trying to see where the |
Fixed but not released in hapijs: hapijs/hapi#3368 |
-> #603 |
* feat: add support for a simpler and async test setup * fix: tweak some tests * fix: fix connecting multiples node in the browser when running with a 'proc' pre-setup we could end up with multiples in-browser setup connecting to each other. whihc isnt a very useful situation. with this the first node gets the pre-configured config and all the others get a JS daemon. * fix: change secondary node to Go * fix: one more to go * chore: update to the latest ctl api * fix: add connect to wantlist tests * fix: fix pubsub tests in firefox * fix: increase timeout findprovs * fix: revert pubsub to go * fix: add more info * chore: bump ipfsd-ctl
os.tmpDir
was deprecated in favor ofos.tmpdir
, resulting in failing tests on CI as we are checking against console output in our tests.see https://travis-ci.org/ipfs/js-ipfs/jobs/170467892
The text was updated successfully, but these errors were encountered: