[6.0] Important regression in bank statements numbering

[Stéphane Bidoul (Acsone)]

Hello,

A very disturbing regression has been introduced in 6.0 regarding the numbering of bank statements.

I traced the problem to commit
5120 revid:<email address hidden>
which merges
5055.1.1 revid:<email address hidden>

To reproduce the problem, you need to create a bank statement and populate it from open invoices and manually created statement lines.

Before that commit, the (correct) behaviour was as follow:
- the bank statement received a new number from the associated bank journal sequence (say BNK/2012/N)
- each line in the bank statement generated a move with a number constructed as BNK/2012/N/1, BNK/2012/N/2, ...

After that commit:
- the bank statement received a new number from the associated bank journal sequence (say BNK/2012/N)
- some lines increase generate moves and number them by increasing the journal sequence
- this is behaviour is wrong and result in gaps in bank statments numbering, not mentioning non-contiguous numbering of moves on the bank journal!

Reversing that commit fixes it. However that commit was meant to fix bug lp:931748, so a better fix has to be created.

Best regards,

-sbi

[Ravish(OpenERP) (rmu-openerp)]

Hello Stephan,

Yes, you are absolutely right . I have faced the the same problem on 6.0 due to revision no 5105.
If I am revert up to 5105 all are working as expected .it's clear cut regression on stable 6 that why ,I am assigning it to OPW.

Thank for reporting..!!!