@t20100 What would need to be done for these tests to be run by travis?

Sorry - I don't understand. These aren't test cases, they're utilities to allow tests w/ units to be written.

Just a short feedback: I wanted to review this PR, but turned away because it's too much effort.

It would be better for review to have the changes separated. There's a lot of whitespace change and hidden in between some functional and py3 changes. At least the github viewer does not really show the difference. Some larger whitespace changes are interpreted as completely new text, and I don't want to recheck everyone of them for a possible py3 or functional change.

This is a special case. As soon as the merge conflict is resolved, I think we should go ahead and merge this.

@timhoffm Have a look at then diff without (most) whitespace changes.

@jklymak yes the exceptions should be removed from pytest.ini.

@QuLogic Thanks, I've learned something new.

Sorry about that. I tried to update my local branch with a rebase and resolve the conflicts and somehow I ended up with a mess showing all the diffs from the original branch point. Let me know what's the best way to resolve this (deleting it and starting over is the only way I know to easily fix this) - my git merge/update skills are clearly lacking...

Hah, I just did that yesterday too:

1. copy the files you modified somewhere else (hopefully finite #)
2. rename the branch you messed up (i.e. git branch -m py3_jpl_units_broke)
3. make a new branch from upstream/master with the same name as the old one (py3_jpl_units)
4. copy the modified files back onto the branch.
5. commit all the modified files (you shouldn't need to rebase onto master because you just pulled from master).
6. push to origin

The trick to rebasing to master is to squash all your own commits first and then rebase onto master.

Hope that helps - feel free to ask if any of the steps not clear...

PS: its possible some of the changes on master were to files you are trying to modify. If thats the case things get a little more difficult...

@jklymak Sorry - that didn't seem to work. There were changes on the same files - that's why there were conflicts in the first place. I tried what you suggested and it tells me the push was rejected it's behind the remote branch (I'm assuming this is because I pushed all those commits so they are already there).

So... perhaps I should just close this and submit two different pull requests, one for the original cmp update and one to update the flake8 warnings. Sound ok?

Well, try a force push first!

git push origin  py3_jpl_units --force

You can't break anything at this point 😉

There seem to be a conflict, please backport manually

Does this still want to be backported to the 2.2.x branch? My guess would be no since it doesn't look like it fixes a regression.