The method removeSrcAttribute is critical, because each changes trigger
this one. this method must be optimized to the maximum to be the fastest.
Before this fix, removeSrcAttribute method search the image and iframe for
all body nodes instead of search only in the inserted nodes, and the method
are very slow.
test_utils.js contains helper functions, which should be stateless
(good practices). This was not the case with patchWindowGetSelection
and unpatchWindowGetSelection, so instead we keep only the first
function which returns the second one.
From the test POV, instead of this:
utils.patchWindowGetSelection();
// ...
utils.unpatchWindowGetSelection();
It now looks like that:
var unpatchWindowGetSelection = utils.patchWindowGetSelection();
// ...
unpatchWindowGetSelection();
Before this commit, phantomjs was crashing when some code manipulates
window.getSelection (such as the composer input in discuss or in the chatter).
This commit lets us mock window.getSelection in tests that need this function
in order to work properly.
Another benefit of this commit is the possibility to simulate the state of
window.getSelection in our tests, which enables tests on the composer that
were not possible before (e.g. making sure that mentions and emojis are
correctly rendered in the input, after the user clicked on them in the dropup
menu or the bottom bar.
Fix over this commit: https://github.com/odoo/odoo/commit/02ec09cb1c3e2d7bc7968f40c18f2208d7f3f498
In mobile, when discuss is installed, there is a traceback.
This is a consequence on an issue with JS services, where some of them
are not registered in the service provider.
For instance, the webclient (the service provider) was instantiated before
chat_manager (a service), so it was not aware of this service.
To solve this issue, service providers now listen on newly registered services.
Also improving deployment of JS services by not relying on topological sort.
In the tests environmnent, we don't want img and iframe nodes to
perform RPCs when they are inserted into the DOM. For img nodes,
we prefixed the src attribute by '#test: ' as soon as they were
inserted into the DOM. Unfortunately, this didn't work as expected:
an RPC was done to fetch route '/web/tests?' (i.e. the tests page),
because everything after '#' in the url isn't sent to the server.
This rev. completely removes the src attribute of img nodes, and
set a 'data-src' attribute with the original src attribute's value,
so that it can still be checked in the tests.
Tests using this attribute have been adapted accordingly.
In the tests environment, we listen to the 'DOMNodeInserted' event
to remove the src attribute of img and iframe nodes as soon as they
are inserted into the DOM, so that they don't perform RPCs.
However, this listener was bound in the create(Async)View, and
createActionManager helpers only. This rev. moves it to the
addMockEnvironment helper which is more global, and basically used
by all tests.
Moreover, we now listen to the 'DOMNodeInserted' event on the body,
instead of on the widget, as it may happen that some content is
inserted into the DOM outside the widget itself (e.g. dialogs are
appended to the body).
Before this rev., it might happen that the parent of BasicModel instances
wasn't correctly set to the Controller instance. It was actually set to the
ActionManager instance, which is never destroyed. So when this happened,
the BasicModel instance wasn't destroyed when it's controller was.
Before this rev., it might happen that the parent of BasicModel instances
wasn't correctly set to the Controller instance. It was actually set to the
ActionManager instance, which is never destroyed. So when this happened,
the BasicModel instance wasn't destroyed when it's controller was.
This commit improves the JS code of the mail module,
which comes maily from the new coding guidelines
and the addition of "JS services".
JS services are important objects that do not fit
well in the component tree, such as chat_manager
or ajax.
The benefits of JS services are improved readability
of the code, reduced coupling, and more testable
modules. In particular, discuss was hard to
Summary of the changes:
- ClientAction has been renamed into Discuss
- Clear instantiation of chatManager
- New coding guidelines in most mail modules
- JS Services can interact with each other
- Chat Manager and Window Manager are services
- Window Manager renamed to Chat Window Manager
- bus.bus is now encapsulated in Bus Service (a service)
- The test infrastructure has been tweaked with JS Services
- Chat Mixin has been removed
A future improvement would be to translate some 'trigger' into 'trigger_up'.
This rev. improves the QuickCreate widget for grouped Kanban views.
It allows specifying in the Kanban arch a form view xmlid, e.g.:
<kanban quick_create_view="some_form_view_xmlid">
...
</kanban>
When this attribute is set, the referenced form view is loaded and
rendered inside the QuickCreate widget. A real FormView widget is
instantiated, meaning that all form view mechanisms are available
in the QuickCreate (modifiers, default_get, onchanges, widgets...).
Task #27475.
In Less, the size_class XXS is defined and used in media queries.
This rev. adds it in JS for sake of consistency.
Code in addons comparing the size_class needed to be adapted due
to the new XXS size_class.
Moreover, we now use the helper 'config.device.isMobile' everywhere
we should, instead of manually comparing the size_class.
Finally, we changed the way the config.device has to be specified
in the test environment. From now on, only the size_class should be
set, and the isMobile flag is computed automatically.
The `additional_context` key was missing from the mock function of `load_action`.
This was causing an issue when testing a use case (the context was correctly
passed in Studio).
This rev. is the first part (out of 3) of the refactoring of the
ActionManager and ViewManager layer of the webclient.
The main changes are:
- there is no ViewManager anymore, its work is now handled by the
ActionManager itself, but isolated in a specific file ; this
eases a lot of things, as the breadcrumbs handling for example.
- the ActionManager code is converted to the new coding principles
and guidelines ; mainly, children widgets communicate with it by
triggering events up, and not by function calls anymore.
- the ActionManager layer is now testable, and a lot of tests have
already been written.
- the code in other addons has been adapted consequently.
What's coming next:
- introduce the 'AbstractAction' Widget, and make client actions
and view controllers inherit from it ; this widget will
implement a common API that could be used uniformly by the
ActionManager (e.g. restore(), canBeLeft(), renderButtons(),
getTitle()...).
- move the ControlPanel handling from the ActionManager to the
AbstractAction.
There is a drawback in 12048bd6, if a kanban view with an
"Undefined" column (exist when a kanban view grouped on a relational
field and some record have this field unset) had a column resequenced a
traceback would happen if more than one column was displayed.
This happen because before we had a "undefined" ID for the "Undefined"
column and after 12048bd6 we get a framework unique ID which was not
managed.
note: this commit also adds `left`/`right` values in the dragAndDrop
test utility function so an element can be dropped at the right/left of
another element.
related to opw-782433
closes#21708
This rev. introduces two helper functions for the tests: patch
and unpatch. Those function allow patching an Odoo Widget (e.g.
AbstractField), and unpatch it at the end of the test with the
guarantee that the Widget's prototype is exactly the same as it
was before the test.
This was necessary because the way it was done before was a little
bit too naive, and could lead to unexpected behavior. Indeed, doing
var someFunction = AbstractField.prototype.someFunction;
// test content
AbstractField.prototype.someFunction = someFunction;
is correct if someFunction is actually define (or overriden) in
AbstractField, as in that case, it really exists on its prototype.
Otherwise, it isn't correct if it is not directly defined on the
prototype before the test, as it will be after..
... to allow them to be called several times without having to
wait for the delays.
This was already done for _.throttle, by default for the whole
test suite, and it couldn't be parameterized in individual tests.
This was enough for _.throttle, but the same solution couldn't be
applied for _.debounce, because some tests actually test that some
handlers are correctly debounced.
Those function are now patched in addMockEnvironment, and two
parameters can be set: 'debounce' (default: true) and 'throttle'
(default: false).
*board,sale_timesheet
This rev. changes the API of load_views, as it now assumes that
the given context has no longer to be evaluated. This is more
convenient that way, as we can now call load_views with contexts
that are dict, and not instances of the Context class.
instead of createView. This allows to log info like RPCs or
uncaught Odoo events even if we are not in a view.
Also remove the logLevel param from MockServer as it wasn't really
used.
Before this commit, it was not possible to reexecute functions using the
_.throttle function in the same test (e.g. drag and drop two elements in the
studio tests).
Before this rev., sorting groups in list views wasn't supported. It
hadn't been implemented in the new views.
This rev. activates the feature, so groups can now be sorted on
their aggregates again. It also implements a basic (first level
only) sorting in mockReadGroup, in the mock server.
opw 781288
Here is a situation where we had a problem:
- a form view with a one2many field, which has no inline views
- the (non inline) list view has a field A, and is not editable
- the (non inline) sub form view has fields A and B, with an onchange on
B which modifies A
In that case, the user could do this:
- go to edit mode
- click on 'add' a new one2many line
- change the value of B in the form view, this changes the value of A
- click on save to close the modal form view
- click on the new o2m record to reopen the modal form view
- rechange B
=> the onchange does not work
The explanation is that when we reopen the modal, we update the known fields
information, but we had to perform a fieldviewget to fetch the list
view, so we have a full knowledge of the fields. However, the code did
not update the fields info (because it uses _.default), which means that
the onchange information contained in the form view is lost.
Note: the test system had to be adapted to more closely simulate what
actually happens. In particular, the onchange flag is no longer added
by the mock server, since it should be done by the data manager, like
'real' code. This change broke the basic model tests, which had to be
modified accordingly, by setting manually the onchange flag.
Generalization of the field 'intercepts' at the
level of the function addMockEnvironment in
/web/.../testutils.js.
Previously, only createAsyncView had support for
'intercepts'. By moving to addMockEnvironment,
we generalize it, so that even mock environment
others than views could intercept bubbling events.
This is desirable in order to test the app_switcher,
so that we can intercept bubbling events like
'hide_app_switcher' to web_client.
Since addMockEnvironment already makes use of some
special intercepts ('call_service', 'load_views', ...),
We must handle the remaining intercepts afterwards,
hence at the very end of the definition of addMockEnvironment
A recent fix in web changed the field FieldImage to make sure it also
loads a __last_update field. This is fine, except that a test in
enterprise (web_clearbit_tests) was using a field image, and failed
because it could not find the __last_update field in the demo data.
With this commit, we make sure that the mock server is always aware of
the __last_update field.
With the new views, we mainly focused on making them work. But we also
need to be able to handle unusual cases, such as some RPC failing. An
example of such a problem was the following:
- a form view with a one2many with an onchange
- edit mode, change a field in the one2many, click elsewhere
- the onchange triggers, calls the server
- it may happen that the server fails with a ValidationError
- the onchange deferred fails, the line in the one2many is restored to
the value known by the list renderer, which is the previous value (but
the new value was applied to the model)
- we now have a 'corrupted' state: clicking on save will send the
invalid value, even though the UI displays the previous one.
We solve this problem in this commit by making sure that the onchange
deferred returned by the _performOnchange always succeed (or stays
pending). If the rpc fails, the deferred will succeed with an empty
dictionary. This means that the list renderer will be updated with the
value from the model.
Note that it may not be intuitive that a failure in the onchange is not
seen as an error from the perspective of the _performOnChange, but I
think that it still makes sense:
- the validation/network errors are not really a concern of the UI. The
UI just display the values entered by the user.
- if an onchange fails (for example, because of a coding error), we do
not really want to prevent the user from working.
Also, when the onchange is performed, the initial change was already
applied to the local data. Rollbacking them would be a more difficult
task.
Note that we also improve the mock server in this commit, to allow
mocking failures.
This commit introduces a new way to hide a column in a x2m list view.
The attribute 'tree_invisible' can be add in the field attrs in the x2m
view definition. This attribute can use the 'parent' key to make a reference
to the parent record (e.g. 'parent.id').
When the mockserver handles an unimplemented route, it just log an
error. This is annoying because it does not break the QUnit suite, so
if the developer does not look at the console, it will look like the
test suite pass. However, the runbot will see that as an error.
With this commit, we throw an error to force the test suite to fail in
that case.
Christophe Matthieu