Explicitly remove the cursor of the request. Dropping a database
is not transactional and if overrides of ir_http's dispatch worked
with the database (for example, utm), we are not able to commit
the changes at the end of the request.
It worked nice before rev 40561b0fb6 that changed the use of self.pool
by request.env. Note: a cursor is always attached to an environment. Since
the call to super possibly broke the cursor (for example by dropping the
database), the end of the dispatch operation that try to commit and/or close
the cursor could possibly result in the traceback 'unable to close a closed
cursor' with request.env. With self.pool, it didn't happen because the
code actually called in the utm's dispatch override doesn't use a cr
and so there's nothing to close.
This will improve the onbarding of project and the usability in general.
Project: When creating a project a small modal opens with one or more fields, and you have to click on it just to type the project name.
Usability: one many2one (and others), when clicking on "Create and Edit", the focus is set on the first field.
Currently, when rendering a list view cell with a many2many we would
empty the list of ids, and fill it again once a name_get is resolved.
But in some instance, the code could use the data when it has been
emptied out.
For example, if we set the tax_id field (inside the order_line list view
inside the sale.order form view) as requred, if we modify the order line
and save directly (without clicking outside of the list view) we can get
an incorrect error saying that the "Order Line" is not valid.
It has been reproduced when saving with CTRL + SHIFT + S on google
chrome and firefox, and there have been reports that for some
configuration it also happen when clicking on the "Save" button.
This commit change the behaviour so the value is kept whilst the name_get
is ongoing, and just use a default "false" value for the name during this
interval.
closes#13478
opw-668067
The class `oe-search-options`
has been renamed to `o_search_options`
at the revision
977db823c7
See the full explanation of the bug fix here:
4601b044a6e4dbf162cfca9bf0be5ff2a21dd4f5q
A new preprocess method is added to the template engine in order to do
preprocessing/sanitization of the templates before compilation.
This method can be overloaded in order to add new features such as
translation, ...
Until 9.0 our psycopg2 DSN connection strings do not allow having
spaces within the db name, and passing some can cause duplicate
registries to be loaded.
Stripping spaces is a simple workaround until we actually support
spaces within db names.
Fixes#13078
Wrong class selector and must add '#' to force the browser to reload the picture because the picture is in the cache with the previous cache informations.
The view `sale_team_dashboard` (which extends kanban) has been removed
and replaced by a default kanban view.
To use it in the ViewManager, an option `js_class` is added
on the arch description and this class be used as the View.
This avoids creating a new view type just for one view.
When we drag and drop a record in a folded column, the number of elements
in this column displayed in the column title is not updated.
This issue is present because when a record is droped in a column
its id is not added to the dataset of its new column.
To fix this, we had to add the id of the record in the dataset, and
override the method 'add_ids' in 'DataSetSearch' to have the size
of the dataset equal to the number of ids in it. We have replicated
the behaviour of the function 'remove_ids' in the new function 'add_ids'.
- As of v10, manifest files should be named `__manifest__.py`
- For backwards-compatibility, __openerp__.py manifest files
will still be supported for the time being
- Limited refactoring, to add support for the 2 different
naming conventions
- All textual references to __openerp_.py updated in
documentation and examples
Simon Lejeune