The start command takes the default dbname from the path.
But when running inside a virtualenv, it is a better default to take the
dbname from the virtualenv's path.
SQL that retrieves data for the "Due Payments" report doesn't include
any filter for excluding AR/AP move lines already reconciled, which
shouldn't be presented in this report.
Courtesy of Pedro Baeza. Was PR #16440
On IE (from 9.0 up to at least IE EDGE 14) we have this behavior
for the method serializeToString of XMLSerializer:
> (new XMLSerializer()).serializeToString($('<b>"</b>')[0])
'<b xmlns="http://www.w3.org/1999/xhtml">"</b>'
> (new XMLSerializer()).serializeToString($('<b>"</b>')[0].firstChild)
'"'
Whilst browser such as chromium or firefox have:
> (new XMLSerializer()).serializeToString($('<b>"</b>')[0])
'<b xmlns="http://www.w3.org/1999/xhtml">"</b>'
> (new XMLSerializer()).serializeToString($('<b>"</b>')[0].firstChild)
'"'
Hence for IE9 and over, if in a `<t t-extend/>` a `t-jquery`
sub-directive (without `t-operation`) is available, we can have
broken javascript if a " is transformed into an " at a unfortunate
location.
This commit favour node.data over XMLSerializer serializeToString to
avoid the possibility of this issue when a text node is processed.
opw-727283
The fields `owner_user_id` used `maintenance` search view to filter on
assigned equipment is overwritten in `hr_maintenance` to be computed
based on employee or department manager linked user.
However employee may not have linked user, considering equipment
unassigned.
Use directly the presence of employee or department to
determine if an equipment is assigned or not.
Different method use self._context.get('lang', 'en_US') to get the lang
value but, the lang in the context can be set to None (eg: copy method).
When the context is used with a lang=None, the method can have a
unexpected behavior or crashed (eg: if you try to format a datetime with a
locale=None)
Leads to the Duplicate link (when editing a blog post in website)
yielding a 500 Internal Error instead of duplicating the blog post.
OPW-740308
Fixes#16511
When reconciliating manually, it may be possible to filter between
`All`, `Customers`, `Suppliers`, `Others`.
When changing this mode, we could end up in a situation such as:
1)
- reconciliation widget X is available
- reconciliation widget line XB is being loaded
- mode is changed -> reconciliation widget X is removed
- reconciliation widget line XB loading finishes
-> error because the parent reconciliation widget exists no more.
2)
having no (or less than the expected) or too few results because
removing reconciliation items (when for example we switch from supplier
to customer mode) can be done asynchronically. But the code expected it
to be done synchronically. Also all elements displayed were removed
instead of only the ones of other types.
thus instead of displaying the expected number of reconciliation items:
MIN({num of record to display}, 10)
the records displayed were:
MIN({num of record to display}, 10 - {num records displayed before})
which could amount to 0 event if there was thousands available ones.
This commit solves these two issues:
1) by not adding line to a reconciliation item without error,
2) by counting and removing displayed items correctly
opw-725796
opw-728309
When deduplicating contacts, the function _process_query doesn't use
the orm to fetch the partners to merge according to the criteria.
So there were some access error in multi company when trying to merge
contacts from a not allowed company.
Now a check is made with the orm before merging the contacts.
opw:708457
When you have push rules on move having a 'production_id', the
'production_id' is copied on the generated moves. So it generate a
traceback when we check the tracking method of 'product_id' on
produce_move.
This issue is caused because we get multiple produce_move instead of
one.
To fix this issue, we don't copy the production_id anymore on moves
generated by push rules.
In function _get_top_level_packages, when checking that all the products
have the same destination, the destinations must be stocked in a set
instead of a list because with different products it could return a list
with duplicated destinations.
opw:725720
When a comma was written in the client_order_ref field of the SO,
the name of the invoice was duplicated because by default the name
of the invoice is client_order_ref in function _create_invoice defined
in model "sale.advance.payment.inv".
opw:724846
- Create 2 companies (A & B)
- Create a product shared between companies
- Add a reordering rule in each company
- Add a supplier for each company
- Run the scheduler (from the cron)
The supplier selected is always the supplier with the smallest sequence,
since it is run in sudo mode.
opw-728346
- Create a SO with a product that is a service and will be invoiced
based in delivery quantities and will be tracked on timesheets and
contract.
- Confirm
- Use a simple user that only has for rights: "Employee"
- Create a timesheet as the employee for the contract created by the sale
- Log some time
- Save it
- Check the SO: ok the number of hours should be "to be invoiced"
- Edit the timesheet and set one day to 0 hours => Access Error
In case of a zero value, the timesheet widget sends an unlink command
for the corresponding entry. In this case, the `sudo` is not set at the
appropropriate moment since the context key `force_so_lines` will resue
lines which are not defined in the `sudo` environment.
opw-728453
Each REPL python shell behave differently depending on the exit method.
Force a rollback at the end to avoid inconsitent behavior.
Thanks to Harry Jollenbeck for the report: https://www.odoo.com/groups/59/26318707
SIPS payment has two limitations:
- A payment reference can only be used once, even if the payment was not
proceeded
- A payment reference can only contain alphanumerical characters
These limitations do not fit the payment process which:
- Reuses existing transactions
- Introduces a `-` character when creating a new transaction for the
same order.
First issue is solved by an ugly hack since there is no link module
between `website_sale` and `payment_sips` in order to override the
behavior of `sale_get_transaction`.
Second issue is solved by modifying the character to `x`. In this case,
we could consider overriding the method `get_next_reference` in
`payment_sips`. However, this method is called before the creation of
the transaction, and we should do a crappy workaround using the context
to get the `acquier_id`.
opw-728209
When you do a group by on field that is not present in the view,
this.fields contains all the model's fields, not only the ones present
in the view.
As we use the method 'has_active_field' to check if the field active is
available in the view, the value is not always correct.
To fix this we take the fields_view fields instead of fields.
This bug has been introduced in rev: https://github.com/odoo/odoo/commit/5b4f7c13f32f36d86805eb7865d8a077746181d8
- Create 3 products (A, B & C) with tracking set to 'serial'
- Create a BOM: A consumes 1 unit of B and C
- Create a MO for A, create the lot A1 for the finished product
- Create another MO for A, use the existing lot A1 for the finished
product
- Try to mark the MO as done => you get an error
- In the finished products tab, change manually the lot, for example by
creating A2
- Mark the MO as done
- Open the Upstream Traceability => under the MO, there is nothing, the
traceability is lost.
The `check_finished_move_lots` method takes care of setting the
`lot_produced_id` on the `move_raw_ids`. However, when changing manually
the finished product lot, this modification is not propagated.
Therefore, when posting the inventory (`post_inventory`), the line:
```
move_raw.move_lot_ids.filtered(lambda x: x.lot_produced_id.id == lot)...
```
doesn't return any lot. Consequently, `consumed_quant_ids` is empty, and
the traceability is lost.
opw-724944
In 599bd509 for inline image in a mail the `Content-Id` of the
attachment was used as the file name to avoir an issue of several inline
images with the same name only displaying the first one.
This solved the issue but with it the file name when downloading the
image were lost as well as the extension which can be an issue.
This changeset refactor the code so the behavior stay the same but the
original file names are kept.
opw-725305
Issue: when a field is computed on a new record, the record is not marked as
"being computed", and it could therefore be invalidated. The fix consists in
using the standard method that marks records before calling the compute method.
This commit prevents from erasing the origin part of the url
when the tour manager does a redirection. This was an issue in
phantomjs if it allows local url as in this case, it will try to
load file://<tour_url> instead of <origin><tour_url>.
Christophe Simonis
