fixes salt-run for kitchen test

[fortunejuggle]

fixes kitchen

this reorders .yml to avoid pillar failed to render error

[aboe76]

@fortunejuggle can you please edit your commit message this will fail:

✖ subject may not be empty [subject-empty] ✖ type may not be empty [type-empty]

see docs: https://github.com/saltstack-formulas/postgres-formula/blob/master/docs/CONTRIBUTING.rst

[myii]

@fortunejuggle Thanks for providing this PR. A suggested commit message you can use when amending the commit:

fix(kitchen): reorder `.yml` to avoid `Pillar failed to render` error

@javierbertoli @daks Can I borrow your Kitchen expertise to discuss if we need to apply this reordering back in our other formulas as well, starting with the template-formula?

[fortunejuggle]

finally - it is done - thanks for your help @myii and thanks for stating it @aboe76

[daks]

I'm surprised it changes anything. I don't think there is any notion of order in kitchen. It could exist for salt (like states ordering in defined top.sls) but not for kitchen.

[javierbertoli]

@fortunejuggle, I fail to see what's the issue you're trying to fix here. An error yaml failed to render means the content of the file is not correct yaml (or that kitchen failed to read it?)

In this case, moving the declarations from the general (default) stanza to the suites stanzas does not change a thing regarding the issue you mention, and should not fix it as far as I know.

Taking into consideration the fact that it 'fixed it', the only difference I see is that, in the stanza, you used the old/deprecated syntax for pillars_from_files with dashes instead of underscores. So I wonder if the "real" issue is either an old kitchen-salt gem in your testing environment or similar thing, which makes kitchen unable to read the pillar because the variable is 'not declared'?

OFC I'm guessing here. As I said, the combination of error and solution you propose don't make sense to me.

Can this be a possibility?

[daks]

I cloned master and tried to run tests on default-debian-9-2019-2-py3 and it fails. Are we sure that tests run correctly on standard salt-call? And how can salt-run be tested?

[fortunejuggle]

@daks -the heading is fixes salt-run for kitchen tests. I always test the tests first - that's why I haven't made any progress yet. However, salt run through the docker container after these changes. I have not yet been able to check whether this is generally the case.

[fortunejuggle]

@javierbertoli - the error was that postgres was not found in environment base - I have checked my gems - you are right - my kitchen-salt is older - I'm really sorry that I've employed five people here now. - i close this pull request

[myii]

Thanks for joining in @daks and @javierbertoli. I'll try to make some time to do some local tests at my end as well. @fortunejuggle Thanks for your patience, it's a good chance for us to review the test setup, to see whether anything else can be done from our end to ease local testing. By the way, can you share how you are calling the tests?

As I'm typing this, you just closed the PR. Is there any reason for that?

[javierbertoli]

@fortunejuggle, the current kitchen.yaml stanzas pass (see https://travis-ci.com/saltstack-formulas/postgres-formula)

I tried the current master of the repo locally and passed, so I can't reproduce your error.

Is it possible you paste the kitchen.yaml and pillar you're using to test, as well as the error you get (the logs of it, I mean)

[fortunejuggle]

@myii - sorry did not mention you - I have used the wrong kitchen-salt gem. All other gems are fine. So it was my mistake - @daks can you please also check it again? @javierbertoli - please refer my previous comment .

I'm trying to be more careful - I'm working a lot with salt right now and I'm trying to help, this time it went wrong.

[myii]

@fortunejuggle Not a problem, welcome to Salt! It's good to know that our testing is working as expected. Thanks to @daks and @javierbertoli for confirming that with their advice.

[myii]

And to @aboe76 for the initial review!