Unwanted behavior when deleting WC, instead of cancel checkout

[ewohnlich]

Plone 5.0
plone.app.iterate 3.2.0

I ran into an issue where some users were deleting the working copy instead of canceling the checkout. Unfortunately I can't just rely on them using the cancel option and need to find a tech solution. Deleting without canceling causes at least two problems.

1. The baseline content remains locked, in a way that is not even easily unlocked TTP
2. When the baseline is checked out again, the new working copy bizarrely declares itself to be both a working copy and its own baseline

I do not think blocking the delete action is the way to go because that is a high level action and other interfaces may provide other ways of deletion (such as the folder contents). Removing the "Delete objects" permission also won't work because deletion is still a part of the cancel checkout process. Instead, I looked to create a delete event listener that would emulate the cleanup part of policy.cancelCheckout as best as I understand it.

<subscriber
        for="plone.dexterity.interfaces.IDexterityContent
             OFS.interfaces.IObjectWillBeRemovedEvent"
        handler=".events.delete_no_checkout"/>

def delete_no_checkout(ob, event):
    """ In case it is deleted without cancel """
    policy = ICheckinCheckoutPolicy(ob, None)
    if policy is not None:
        # simulate policy.cancelCheckout except for the delete part of course
        baseline = policy._getBaseline()
        notify(CancelCheckoutEvent(ob, baseline))

As far as I can tell, this does what i would expect and makes delete and 'cancel' effectively the same. It does not appear to cause adverse affects when properly doing a cancel. The subscribers will be notified twice, but that seems to be fine. I believe the relevant subscribers are in marker.py and locking.py.