Shopping cart adjusting does not re-consider discount

[larf2k]

I'm having a problem with what I believe to be a group pricing discount. The client who's store I'm working on requires there to be a 30% buyer discount on all shopping carts that exceed $500.

The discount is functional as in when you get to the order total page, in red it displays the 30% discount as a negative number right after the sub-total. This is fine until you edit the shopping cart under a COWOA login. When you edit the shopping cart to be less than $500, the 30% discount still stands, as it should not. It's as if the process between the login page and the order total page is responsible for calculating this discount since when you go to edit the shopping cart and then continue to checkout, it skips the login page as it has already been processed and proceeds to the order total page.

I should also note that the 30% discount seems to be applied to the customer permanently - if you EVER make a purchase over $500, you will always get a discount from that point forward, which is the desired outcome. So in COWOA, I believe it temporarily creates a login then later deletes the user, and in the case explained above, the temporary user gets assigned the $500 purchase attribute (which enables the 30% discount) but this becomes problematic in this particular use-case. It would also seem logical that this would happen if a new customer filled up their shopping cart to $550 then backed it down to $300 they would get the 30% discount undeservedly.

I am using fast and easy checkout as a checkout plugin. As another side-note, I'm noticing that on the login page, the sub-total displayed on this page are incorrect: often times the discount is not applied, or the discount is ADDED rather than subtracted. I'm thinking of just hiding it with CSS as it has become just a nuisance and not really necessary for the login/COWOA page. But it is very strange that it isn't consistent. Suspicious.

Any thoughts?

[BlindSide]

I would take addons out of the equation first, and see if the behavior continues. Start with just one of them, and redo your tests. If the problem disappears, you've found the culprit.