The flatpak@aarch64 is not allowed on other branches than master but
is allowed on forks. It would be nice to allow manual run in those cases
to save resources, but to allow testing aarch64-specific fixes.

The pages job is allowed on forks currently. It is not desired to run it
there and it is going to fail there anyway. Let's explicitly disable this
job on forks.

The only/except keywords were deprecated in favor of rules. Let's
replace them with rules.

The pages job is not running currently on master. This is because
`only: refs: $CI_DEFAULT_BRANCH` is not supported (e.g. see
https://gitlab.com/gitlab-org/gitlab/-/issues/29591). Let's fix that
using one of the suggestions from that issue.

With non-local drops, when the value is initially NULL, we get
warnings because of some of the checks done on a NULL value.  This was
a regression from 3d17ed25.  Return
action copy before any check dependent on a dropped value.

Peter Eisenmann authored Jan 23, 2023 and Corey Berla committed Jan 23, 2023

Add a check to assert nautilus_files_view_drop_proxy_received_uris is
not called with an empty source uri list.

Make nautilus_dnd_perform_drop return whether the drop event was
actually handled. Adjust callers accordingly.

Fixes: #2787

sunflowerskater authored Jan 20, 2023 and António Fernandes committed Jan 20, 2023

This change makes the rubberband to have rounded corners,
making it consistent with other elements such as buttons and text entries,
which also have rounded corners.

Cheng-Chia Tseng authored Jan 19, 2023 and Administrator committed Jan 19, 2023

(cherry picked from commit 2c2a8394)

Corey Berla authored Jan 15, 2023 and António Fernandes committed Jan 16, 2023

We are still leaking NautilusViewItem in several places. Essentially
every NautilusViewItem (and thus every NautilusFile) ever created is
leaked.  In list-base, the queue in real_add_files is freed, but not
the items themselves. Make sure to also free the results of
g_list_model_get_item() and our macro listitem_get_view_item which now
calls gtk_tree_list_row_get_item.

Corey Berla authored Jan 15, 2023 and António Fernandes committed Jan 16, 2023

view-item holds a reference for the cell and the cell holds a ref
to the item.  In order to break this cycle, let's make item_ui
a weak pointer in view-item.  The cell will get freed by the
list factory.

Corey Berla authored Jan 13, 2023 and Ondrej Holy committed Jan 16, 2023

* Update gpl to match project
* Update authors to match about page (otherwise it says "unknown author")
* Update website so it doesn't show a blank site
* Add logo
* Add dependencies

Corey Berla authored Jan 09, 2023 and Ondrej Holy committed Jan 16, 2023

We are generating the docs, so lets make them accessible online.

See: !941

Corey Berla authored Jan 10, 2023 and Ondrej Holy committed Jan 16, 2023

As part of 42c19589 uncrustify was being
built from source to address a bug that is now fixed in the current release
0.75.

Corey Berla authored Jan 09, 2023 and Ondrej Holy committed Jan 16, 2023

We have a docs options, but we are generating the docs regardless
of if it is set. This was a regression in  d47403bc.
`gi_docgen.found()` is redundant because when 'docs' is set to true
gi_docgen is a requirement.

GtkTreeExpander includes a built-in "expand on drag hover" feature.

But our own "open on drag hover" feature competes (and often wins)
over it, making it hard to reliably expand rows on hover without
opening the folder.

So, separate the content and the expander. Trigger "expand on drag
hover" only if over the expander, and "open on drag hover" only if
over the rest.

No functional change in this commit, but this is going to make it
possible for each view to pick the areas which trigger the hover
timeout, to be done in the next commit.

A property binding is unfit for our use case because directory
loading is async, so the expander arrow is temporarily hidden,
which is wrong and looks like a glitch.

Fixes #2791

When using the collapse shortcut (Left arrow in LTR locales), if the
row is already collapsed, we want to move to the parent row.

Currently we select the parent but don't move focus. Fix that.

They have been set on the cell in commit 728bec15

But they don't work because the focus is never in the cell.

So, set them in the view instead, and trigger them in the capture phase
to prevent GtkListBase's cursor movement shortcuts taking priority.

Toggling sort directories first in list view causes a crash.
This is a regression in commit b9ac68eb

gtk_sort_list_model_set_sorter() first unrefs the old sorter, then refs
the new sorter. The crash happens because the new sorter is the new
sorter. It gets destroyed by the unref, and therefore we crash when
trying to ref a destroyed object. (It's arguably a GTK API bug too.)

The solution (from our side) is for us to hold a temporary reference to
keep the object alive.

When an item is moved, we are still showing it in the old location as a
regression caused by commit 510a4202

With the introduction of tree model, we maintain multiple list stores
(one for each TreeListRow when the TreeListRow is a directory) and the
item is associated with a directory store upon creation by checking
nautilus_file_get_parent(). When removing an item, we again find the
directory store to remove from, by checking directory_reverse_map for
the parent file.

But, when an item has been moved, the parent file is no longer the same
as when it was added. This means we try to remove the file from the new
parent instead of the old one.

Fortunately we already have a solution for this from the old list view:
NautilusFilesView.remove_file() gets passed the directory to remove the
file from. Let's use it instead of the current parent.

Heavily based on diagnosis and commit message from patch by Corey Berla

Fixes #2784

Tooltips are not showing up for long file names that have Pango markup
characters. This is happening because the cell widget label is set using
`set_text`, but is retrieved in the `query-tooltip` signal handler using
`get_label`. While `get_label` can return Pango markup, it is not
guaranteed to return text with markup characters escaped. Avoid this
problem by fetching the file name from label and setting tooltip using
text.

Ondrej Holy authored Jan 11, 2023 and António Fernandes committed Jan 13, 2023

The `org/freedesktop/FileManager1` interface is not currently exported
from the `dbus_register` vfunc. This causes issues for projects (e.g.
desktop-icons extension) that want to use all the Nautilus intefaces
over the `org.gnome.Nautilus` connection. Let's use the already established
connection and export the `FileManager1` interface from the `dbus_register`
vfunc.

https://bugzilla.redhat.com/show_bug.cgi?id=2150894

Ondrej Holy authored Jan 11, 2023 and António Fernandes committed Jan 13, 2023

This reverts commit 3014f694 and part of
the commit a1751d3d. This extra handling
was needed because `NautilusFreedesktopDBus` methods might be called before
establishing its own D-Bus connection. The follow-up commit is going to use
the already established connection from the `dbus_register` vfunc. It
should be ensured that the `dbus_register` vfunc is called before the
`window_added`, `window_removed`, or `startup` vfunc. Consequently, this
extra handling is going to be redundant. Let's remove it as a preparation
for the next commit.