Fix BlogPost canEdit permissions.

[GuySartorelli]

resolves #607
If permissions earlier in the inheritance chain fail, we should not allow users to edit posts.
If permissions earlier in the inheritance chain succeed, we should still go through the checks in this method.

[GuySartorelli]

Had the wrong tab open, meant to close a different PR. Sorry.

[dhensby]

The tests need updating to reflect the changes in logic

[GuySartorelli]

Hey there.
It is starting to look to me as though the tests which are failing should be failing (e.g. BlogTest::testRoles, I think editor should be able to edit that post. The canEdit check seems to be failing up the chain in SilverStripe's checking somewhere, and I can't make heads nor tails of why that would be.

Perhaps we should take a slightly different approach (which is already taken by the Blog class), and do our class specific tests first - then, only if we don't want to explicitly allow the user to edit the post (i.e. they're not listed as an author or editor for the post nor the parent blog), we can kick checks back to the superclass.

[dhensby]

It is starting to look to me as though the tests which are failing should be failing (e.g. BlogTest::testRoles, I think editor should be able to edit that post. The canEdit check seems to be failing up the chain in SilverStripe's checking somewhere, and I can't make heads nor tails of why that would be.

that's right, that's why the tests need updating, to reflect the new intended behaviour

[GuySartorelli]

It is starting to look to me as though the tests which are failing should be failing (e.g. BlogTest::testRoles, I think editor should be able to edit that post. The canEdit check seems to be failing up the chain in SilverStripe's checking somewhere, and I can't make heads nor tails of why that would be.

that's right, that's why the tests need updating, to reflect the new intended behaviour

No, what I meant was that I think the tests as written are still correct, unless I have misunderstood the relationship between the test fixtures. Either I am misreading the fixtures, or my change has revealed a different issue. Either way I would appreciate someone taking a look and seeing what's going on.

[emteknetnz]

@GuySartorelli are you still interested in this one? - note - it's targeting master

[GuySartorelli]

Nope. It's something that should be fixed at some stage but I have no specific interest in diving into the tests to see what's going on there.
If someone else comes across this problem again and makes it known I might pick it up again but until then happy to close.