| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
The flatpak devel job uses .pre stage currently.
This seems to be the reason why pipelines are not triggered in some
cases according: https://gitlab.com/gitlab-org/gitlab/issues/198518.
Let's use the unused test stage instead to fix the pipeline and to make
the pipeline output nicer.
(cherry picked from commit e7c2a0182a4ba2d6eb05fe170cc9b4d018a70feb)
|
|
|
|
| |
We consistently forget to update it when releasing.
The intention was good, but this is hard to maintain.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
File creation mask should be honored when creating new files from
templates as it is when creating new folders, or dragging raw data. But
it is not because G_FILE_COPY_NONE flag is specified when creating new
files from templates. Let's use G_FILE_COPY_TARGET_DEFAULT_PERMS flag
to ensure that file creation mask is honored in this case as well.
Just note that this behavior is not wanted when copying in general
(although it is also honored by "cp" cmd in this case) as it might have
some unexpected consequences as discussed on:
https://bugzilla.gnome.org/show_bug.cgi?id=167102
|
|
|
|
|
|
| |
The Select items matching (Ctrl + S) feature allows to select files
which match the pattern. However, it is confusing that current selection
is not cleared when any files don't match the pattern.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The total number of items and their size are shown in Properties dialog.
However, the deep count calculations are currently restarted with each
"changed" event of `NautilusFile` object(s). This is not usually a problem
if only one file is selected, but it is a pretty big issue when more
files are selected. It is common that the calculation never ends. This
is because the "changed" events are emitted in many irrelevant cases
(e.g. free space change) and it totally doesn't make sense to restart
the calculation in most of the cases. The initial idea was to react
on ongoing file operations, however, the calculation currently doesn't
react on file changes deeper in the tree anyway, or on changes, which
happened after the calculation is done. Thus the current result can be
outdated anyway.
Let's ignore `NautilusFile` changes at all when calculating the content
size as it is pretty impossible to implement this properly to dynamically
react on all size changes in the tree.
Fixes: https://gitlab.gnome.org/GNOME/nautilus/issues/363
|
| |
|
|
|
|
|
|
|
|
| |
The flatpak master job fails currently. This is because it tries to rewrite
artifacts from flatpak devel job which fails with permission denied error,
because the runners are not more privileged. Let's use empty dependencies
for flatpak master job to ensure that the artifacts from flatpak devel
job are not passed in.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Nautilus randomly crashes when g_object_bind_property is used to bind
extensions-background-menu and templates-menu properties of
NautilusFilesView and NautilusWindowSlot. Although, we don't have any
reproducer, it seems that this is because NautilusWindowSlot uses
g_object_ref when storing the property, but not g_object_unref, and
NautilusFilesView does the opposite. It seems that NautilusFilesView
can hold invalid references in certain cases because of that. The
both objects should hold its own references to prevent this. Let's
replace the custom codes by g_set_object to achive this behavior.
Fixes: https://gitlab.gnome.org/GNOME/nautilus/issues/813
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
tracker-miners master requires tracker-testutils 2.0, but they are not
provided by tracker 2.3.2 and tracker master provides just version 3.0.
Let's use tracker-miners 2.3.2 to workaround this issue.
|
|
|
|
|
|
| |
Flatpak generation fails currently because tracker-miners still requires
tracker API version 2.0. Let's use tracker 2.3.2, until tracker-miners
will be ported to 3.0.
|
|
|
|
|
|
| |
I don't see any reason to manually specify libdir for tracker as
it seemingly uses the same directory by default. Let's remove that
redundant definition.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
It doesn't actually work, due to https://gitlab.gnome.org/GNOME/gtk/issues/626
So, hide it from the Keyboard Shortcuts dialog.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Empty trash widget is displaying for empty folder inside trash
instead of Empty folder widget.
`eel_uri_is_trash` only checks whether the URI starts with "trash:",
that's why it also returns true for empty folders inside trash.
Create `eel_uri_is_trash_root` to distinguish trash root (trash:///)
from empty folder inside trash (trash:///emptyfolder)
Closes https://gitlab.gnome.org/GNOME/nautilus/issues/1367
|
| |
|
|
|
|
| |
Let's make .1 release as I forgot to bump the version in meson.build.
|
| |
|
|
|
|
|
|
|
|
| |
Currently, hidden files are filtered out when looking for template files.
However, it would be nice to have a way to create hidden files from templates.
Let's offer hidden files as templates, but only when hidden files are shown.
Fixes: https://gitlab.gnome.org/GNOME/nautilus/issues/1320
|