| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
| |
This was introduced in commit 95d42ea01f7f9c60210b415554e7bbf6f4e1b334 and released in 3.5.91.
The error meant that when a network location was disconnected the next time you connected to it you would get the old NautilusFile object that was marked as "gone" and hit an assertion.
https://bugzilla.gnome.org/show_bug.cgi?id=708282
|
| |
|
|
|
|
| |
Signed-off-by: Gheyret Kenji <gheyret@gmail.com>
|
|
|
|
|
|
|
|
| |
While the operation is in progress, since we're returning the result
asynchronously, we need to keep a reference to the invocation, or it
could be invalid when returning later.
Related: https://bugzilla.redhat.com/show_bug.cgi?id=874534
|
| |
|
|
|
|
| |
Thanks to Kamil Paral.
|
|
|
|
|
|
|
|
| |
I.e. don't treat them like thumbnails, and don't add a border.
This was changed in d9ef715ea11e92917414d5d7bddd4dd1487fac1b but it
causes problems for other (more important) use cases.
https://bugzilla.gnome.org/show_bug.cgi?id=688808
|
|
|
|
|
|
|
|
| |
When unmounting from the sidebar with an operation other than eject, we
still need to set up the progress notification, as some drives return
FALSE for can_eject.
https://bugzilla.redhat.com/show_bug.cgi?id=885133
|
| |
|
| |
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=688237
|
| |
|
|
|
|
|
|
|
| |
Like we do in the pathbar, otherwise it could potentially grow
unboundedly large.
https://bugzilla.gnome.org/show_bug.cgi?id=688186
|
|
|
|
|
| |
Make it consistent with the model engine - the mimetypes don't need to
match exactly.
|
|
|
|
|
| |
Or we'll get extra results when applying a mimetype filter when the
model engine is used.
|
|
|
|
| |
We weren't freeing the row list and all the row structs.
|
|
|
|
|
|
|
|
|
|
|
| |
Since we also use nautilus_query_editor_set_query() when the query
editor already has a query, don't unconditionally add mimetype rows from
there, but only when we're being called with a previous query.
Fix a bug where mimetype rows were added when switching between current
location and all files.
https://bugzilla.gnome.org/show_bug.cgi?id=687537
|
| |
|
|
|
|
|
|
|
|
|
| |
We already trigger a nautilus_file_call_when_ready() on the symlink
file, which will include link information.
Invalidating symlink info manually here will make us read invalid
information in the ready callback instead.
https://bugzilla.gnome.org/show_bug.cgi?id=687937
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
When the file watched by the bookmark is trashed or deleted, record the
fact in the object state; later, when checking whether the URI exists,
use that information to avoid connecting to a vanished NautilusFile.
Currently, we will hit an assertion later for unmounted remote volumes;
this bug wasn't present in master and the cherry-pick for commit
231f0b074da973f2c0b69513b743548810d4e716 introduced this regression.
https://bugzilla.gnome.org/show_bug.cgi?id=687518
|
|
|
|
|
|
|
|
| |
We will set the visibility of the action to FALSE when we go to a
location that supports deletion but not trashing, but we fail to restore
it to TRUE when we go back to a directory that supports trashing.
https://bugzilla.gnome.org/show_bug.cgi?id=687619
|
| |
|
| |
|
| |
|
|
|
|
|
| |
This fixes a regression introduced in commit
17f17214a64674bd3079d1a252fed2b6103f9ff4
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The workaround we need to make to focus the entry on-the-fly and
synchronizing the visible state with the toolbar button badly clash with
the focus widget state saving mechanism we have in place.
We end up calling remember_focus_widget() on the newly-focused
search entry in certain circumstances, so we end up restoring the focus
on the (then hidden) search entry, which will then eat key events while
invisible.
Sidestep the issue entirely by just focusing back the view when closing
the query editor.
|
|
|
|
| |
Or we'll always return to $HOME.
|
|
|
|
|
|
| |
Fix a case where entering a directory from the search results would not
take the query editor down, and refactor some code around the state
synchronization between the two objects.
|
| |
|
| |
|
|
|
|
|
|
|
| |
Or the hyphens will be treated as NOT operators instead of word breaks,
which will prevent results from being returned.
https://bugzilla.gnome.org/show_bug.cgi?id=683633
|
|
|
|
|
|
|
|
|
| |
Ensure the current window URI gets selected in the sidebar also when
it's already set on the window at construction.
At the same time, move the sidebar construction from
nautilus_window_show() to nautilus_window_constructed().
https://bugzilla.gnome.org/show_bug.cgi?id=674052
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We used to keep track of the search active state using the state of the
search toggle button; unfortunately, the button is per-window,
so its active state will also be synchronized when switching tabs.
This will cause a non-search slot that becomes active following a tab
switch from a search slot to think its search just became inactive, and
will then change its location to the default ($HOME).
Fix the bug by keeping an additional state per-slot for whether there's
an active search.
https://bugzilla.gnome.org/show_bug.cgi?id=686893
|
|
|
|
|
|
|
| |
Since we don't use a GtkToolButton directly, we need to propagate the
action tooltip manually to the GtkButton.
https://bugzilla.gnome.org/show_bug.cgi?id=686903
|
|
|
|
|
|
|
|
|
|
| |
Add a new flag to signal that we really want the default mount location
to be loaded if possible. This way we can still get to the mount root
e.g. by going up or typing it directly in the location bar, but the
default location between sidebar mounts and using Connect to Server will
be the same.
https://bugzilla.gnome.org/show_bug.cgi?id=522917
|
|
|
|
|
|
|
|
| |
Cleanup and rearrange typedefs to avoid the need for a separate types
header.
Conflicts:
src/nautilus-window.h
|
|
|
|
|
|
|
|
|
|
|
| |
When we detect a deletion, we try to find an existing parent in the
hierarchy to display in the window.
While this is desirable when a file is deleted, we never want to walk a
non-native hierarchy, or to walk up a hierarchy from a mount root
being removed.
This fixes the window redirecting to /run/media/$username/ instead of
$HOME when a mount is removed.
|
| |
|
| |
|
|
|
|
|
|
|
| |
In addition to checking whether the directory is a search, check for the
query editor visibility. This ensures the toggle button state is
consistent when switching tabs with the query editor visible in one of
those.
|
|
|
|
|
|
|
|
|
|
| |
We already do this for all other kind of references, and I think it
makes sense to do it for symbolic links too.
This also ensures applications don't have to deal with e.g. ensuring
backup files are saved relative to the original file and not the
symlink, as reported in
https://bugzilla.gnome.org/show_bug.cgi?id=686465
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Nowadays, we rely on G_FILE_MONITOR_EVENT_UNMOUNTED to be emitted when a
mount disappears, since we already monitor the directory, in order to
switch the view to a different location.
Since non-native mounts very likely won't have file monitoring though,
we're not going to receive such events in that case, and fail to
redirect the view as a consequence.
This patch fixes it by listening to the mount-removed signal on
GVolumeMonitor when a monitor is requested for a non-native mount, and
emulating a file removal in that case.
https://bugzilla.gnome.org/show_bug.cgi?id=684226
|
|
|
|
|
|
| |
We might be refreshing while the mount disappears; in that case, the
pathbar will try to get a symbolic icon from a NULL mount, hitting an
assertion.
|
|
|
|
| |
Didn't mean to remove this line.
|