11/29/2018 10:43:24 PM
populate the table when creating invoice
|
|
|
12/4/2018 10:17:28 AM
process all account updates in transactions.
These fixes should patch the high concurrency bug described at https://github.com/killbill/killbill/issues/282.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
12/4/2018 7:34:33 AM
Meyer <pierre@mouraf.org>
|
12/4/2018 5:42:07 AM
Meyer <pierre@mouraf.org>
|
12/3/2018 7:06:34 PM
Pierre-Alexandre Meyer <pierre@mouraf.org>
|
12/3/2018 4:39:25 PM
definition
In case of ACCOUNT alignment, we now return null before the
alignment is computed. We cannot easily trigger a computation
unfortunately, as the bulk of the logic belongs to junction today.
This also fixes the previous patch where the BCD calculation wasn't deterministic.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
12/3/2018 7:59:28 AM
the behavior to set the account BCD even for pure usage based
plans.
The implementation relies on the fact that Subscription#dateOfFirstRecurringNonZeroCharge()
is actually the subscription start date in this case.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
11/30/2018 12:58:23 PM
fixes a couple of other tests.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
11/30/2018 9:22:29 AM
using the first calculated BCD of the first BillingEvent seen.
This was a bit arbitrary, especially across bundles (since we weren't looking at the date).
We now look at the first BillingEvent with a non-$0 recurring price and use
its effective date to get the account BCD. This matches better our BCD definition:
date of first recurring non-$0 charge.
One side effect is that for pure usage based plans, the account BCD isn't set anymore.
For instance, a consumable in arrear usage plan billed on the 1st would have previously
set the account BCD to the 1st. Now, it is not updated (unless a recurring price is set to the plan).
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
