account_id is mandatory, and is set thanks to an onchange of partner_id. But partner_id is not required, therefore, if you do not set a customer/supplier, there is no
account_id set by default.
On the purchase, the account_id was always invisible. Therefore, it was impossible to set the account_id if you did not set any supplier
On the sale, the account_id was invisible if you set the pay_now to pay later. Therefore, it was impossible to set an account_id if you did not set any customer and set pay later.
Anyway, now, account_id is always displayed, so the user can change the value of account_id if the default value provided is not was he wants, whatever the value of the other fields
bzr revid: dle@openerp.com-20140212142737-ox2l9twg3x6ywvqg
Some filters needs to be evaluated with the same
language as their author, e.g. when searching on
translatable field values. When a time-based
action is processed by the scheduled task,
we should thus attempt to reuse the same
language as the author of the corresponding
filter.
bzr revid: odo@openerp.com-20140212113924-77sh4oj6dl2qwqka
The previous behaviour used the uom of the product while it is not necessary the one selected (eg: by default the purchase unit of measure for purchase orders).
This was an issue especially when having a different uom with supplier info lines setting degressive prices. The price should be computed based on selected uom and not the product uom.
bzr revid: mat@openerp.com-20140211152703-twnzco2dwxeqt8hz
The previous behaviour used the uom of product while it could be a different one selected (by default the purchase unit of measure for purchase orders).
This was an issue especially when having different uom with supplier info lines setting degressive prices. The price should be computed based on selected uom and not the product uom.
bzr revid: mat@openerp.com-20140211145703-9uut4hw9aqh7326o
Unescape elements that have been wrongly escaped.
Some elements (eg: 'href' tag in links) are sanitised in templates which prevents the execution of mako templates.
bzr revid: mat@openerp.com-20140210152806-rl5h3twgq8dd203d
Closing previously occured on search request (so that a user wouldn't be able
to select "old" data on new search request), but ``search`` is only triggered
after the search delay. Worked when delay was 0, with it being moved to 250 a
user can get results matching the previous search instead of the current one.
Trigger a closing of the current results list on any ``input`` event, which is
when text is entered in any of the searchview's ``InputView``
bzr revid: dle@openerp.com-20140210140032-06dnlxepcc5ae21f
When the user who validated and closed the pos session was not the same user who created the session, and if this user was not in the same company, it wasnt possible to validate and close the pos session.
This explanation if also valid if the user who created the pos session changed of company in his preferences between the creation and the validation.
bzr revid: dle@openerp.com-20140206163444-ckcmurcwk2vhi5vp
Add followers before super call in mail_thread creation.
This allows to have a value for the _set_followers call and be able to have record rules based on followers at creation.
Set the priority of message_follower_ids fields to -10 (default is 0) to be sure it will be executed before other function fields using these followers (eg: _set_stage_per_user from note.note does a browse which would trigger record rules to read the note)
Context hack to avoid checking read access at creation of the record (check create rules instead)
lp bug: https://launchpad.net/bugs/1259913 fixed
bzr revid: mat@openerp.com-20140206121438-epghqo042ync24v5
Make sure the user is subscribed before the super() call in mail_thread. This allows to have follower based record rules working (eg: note.note)
bzr revid: mat@openerp.com-20140206104714-anpehaik4wajnfrw
On invoice residual amount computing, we ignore reconcile entries line if their invoice are not of the same type of the current invoice
The above revision(s) will be removed
bzr revid: dle@openerp.com-20140205100324-5tyquozmylcp40p1
- tracking unlink does not fail anymore
- do not track virtual fields on res.users model (would for .column call)
- make sure read calls with one id works (launchpad bug lp:1214149)
- select the correct model when tracking x2m fields values
bzr revid: mat@openerp.com-20140204151136-51cm1tbgvcsnlsoe
iter only on fields on __all_columns, this avoids getting inexistant fields such as 'in_group_42' for res.users
make sure read method calls works when passing only one id (lp:1214149)
when itering on change values, if different model (eg: if field is o2m) make sure make name_get on the model of the field and not on the parent model
lp bug: https://launchpad.net/bugs/1214149 fixed
bzr revid: mat@openerp.com-20140204141449-2f922k5awixh1kzt