Message protocol 5 doesn't use 'source' field

[takluyver]

http://jupyter-client.readthedocs.io/en/latest/messaging.html#display-data

[takluyver]

Tests fail on Mac. Not clear why, but it doesn't look like it's related to this change:

* checking tests ... ERROR
  Running ‘testthat.R’ [9s/25s]
Running the tests in ‘tests/testthat.R’ failed.
Last 13 lines of output:
  3. Failure: kernel tests pass --------------------------------------------------
  close(con) not equal to 0.
  1/1 mismatches
  [1] 256 - 0 == 256

[flying-sheep]

Yeah, they’ve been failing for some time, but I have no idea why.

Could you please add a test for protocol validity? (or enable checking this in jupyter-kernel-tests, as i assume it should be in there)?

[takluyver]

It looks like JKT doesn't currently support checking for unexpected extra attributes. I'll work on that, but it's not a five minute job to do it right.

Are you happy if I comment out the Mac tests and open an issue for someone to look at them later? They're slower than the Linux tests, and CI is less useful if we're used to it failing.

[flying-sheep]

yes, very happy, i wanted to do that long ago. no idea why i didn’t…

[kojix2]

Thank you @takluyver @flying-sheep