GitLab

This reverts commit 403aaea3.

Probably some OTP refactoring broke displaying of ITIP invitation.
This patch fixes the problem. It may not fix the root cause of the
problem, but at least it makes ITIP invitations work again.

What happens is that OTP::parseObjectTreeInternal() first calls
Formatter::format() without the asyncResultObserver. Without this
patch it still creates the new memento, but does not connect it to
the observer. Then OTP calls the Formatter again immediatelly after,
but this time with a valid asyncResultObserver. Because the memento
already exists, this only triggers the codepath that checks whether
the memento is finished (which it is not) and exists. When the memento
is finished nothing happens, because it wasn't initially connected
to the observer to force update of the Viewer.

With this patch we don't create the memento if there is no
asyncResultObserver. That means that the next time the Formatter is
called with a valid asyncResultObserver we create the memento and
connect it to the observer. Once the memento is finished, the Viewer
is updated and the invitation is displayed correctly.

Differential Revision: https://phabricator.kde.org/D1512

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

The copy-paste mistake caused that the issues fixed in previous commits
were really only fixed in EventEditTest, not the other two.

The tests accidentally started Akonadi because the tested class
instantiates EventEdit (or Todo/NoteEdit) class, which internally
instantiates CollectionComboBox, which then starts Akonadi. This
can be workaround by passing setting the MessageViewer::_k_fooEditStubModel,
which prevents the CollectionComboBox from using Akonadi models.

Make sure the window is active before checking for a widget focus.
This does not really affect CI, but when running tests locally, the
WM may cause the test window to not be active making focus tests fail.

First make sure the test is not overwriting actual config files
when running locally. Then we can write fake values into the config
to make the check for default pre-selected collection work again.

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"