| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
This was causing problems for Launchpad's translation importer
for Ubuntu.
"Line 7166: Unsupported plural case number."
|
|
|
|
|
|
|
|
|
|
|
|
| |
Nautilus allowed line wrapping for filenames after a . or _
The code made an exception for version numbers, but it
was too rigid and broke things like Ubuntu's 12.04 version
numbering. This relaxes the exception so that the line isn't
wrapped for a . or _ immediately followed by any numbers.
https://launchpad.net/bugs/942539
https://bugzilla.gnome.org/show_bug.cgi?id=781875
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The issue is that when pressing enter in search while the search entry
has the focus, right after the callback to activate the selection is
called, Nautilus also tries to activate the location corresponding to
the base of the search, as the view is still searching.
This occurs because in the activated signal the search is set as not
visible manually, which will make Nautilus check immediatly if it should
go back to the base of the search location, without waiting for the view
to update.
To fix this, don't set the search as not visible in the activated callback,
as when the search-mode-enabled property changes, if necessarry, the
cancel signal will be emitted which will set the search as not visible,
which will happen only after the view is updated. This is the behavior for
activating the selection in other ways, so activation with focus on the
entry should behave the same.
https://bugzilla.gnome.org/show_bug.cgi?id=764981
|
|
|
|
|
|
|
|
| |
This hides the autostart from Ubuntu's Startup Applications
app since we don't want users to easily disable this without
understanding why it's there.
https://bugzilla.gnome.org/show_bug.cgi?id=781874
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Pressing enter with no selection while in search makes Nautilus
crash.
The problem is that Nautilus will try to access the first element
of a list that is NULL.
To avoid this, make sure that the list is not NULL.
https://bugzilla.gnome.org/show_bug.cgi?id=777214
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
If there is a file indexed by tracker that doesn't have nie:url, then
the query will fail everytime, because it will pass null to
uri-is-parent which will then give an error.
To fix this, in the query, check first that the file has nie:url.
https://bugzilla.gnome.org/show_bug.cgi?id=770944
|
|
|
|
|
|
|
|
|
| |
The sparql query will fail if the file name or the parent uri have a
character that needs escaping. An example would be the character '
To fix this, escape the file name and the parent uri in the query.
https://bugzilla.gnome.org/show_bug.cgi?id=770944
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
"Adwaita-dark" is supplied by gnome-themes-standard and has a distinct
name to theme GTK+ 2. For GTK+ 3, this theme just imports gtk-contained-
dark.css from the library, so it's _as if_ Adwaita was the system theme.
But Nautilus only applied its Adwaita-specific CSS if the system theme
name was exactly "Adwaita". So using "Adwaita-dark" meant we got Adwaita
but without Nautilus' additions to it, even though they are applicable.
Fix this by also loading Adwaita.css if the theme name is "Adwaita-dark"
https://bugzilla.gnome.org/show_bug.cgi?id=779266
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When the metadata of the desktop needs to be recalculated or it's
missing, we need to reposition the icons. They have what we call a
"lazy position". In order to place them on the desktop we either
position them by it's saved position if possible, and if not, moving
them as close as possible without overlaping, or by what we call
"auto layout", which is basically a perfect grid similar to a regular
nautilus window. It's clear from this logic that we do either one way
to place them or the other, and both at the same time doesn't make
sense. For that we assert we just apply one of this placements
algorythms. However, we were hitting this assertion if desktop-metadata
was missing (so they have a lazy position) since we need to reorder the
icons using the auto layout algorthm but the code was also trying to do
the "saved position" algorythm.
This issue is introduced by a commit intended to avoid overlapping icons,
with id: 40c79aec2d2fdc860eadab9b59682ae0b04d5131. In the initial
implementation of "lazy position" support, if "auto layout" was chosen,
the icons were repositioned only by the "auto layout" algorithm.
To fix this re-add the check that repositions icons only by the
"auto layout" algorithm if "auto layout" is chosen.
https://bugzilla.gnome.org/show_bug.cgi?id=747662
|
|
|
|
|
|
| |
The name of nautilus_file_is_regular_file had a typo.
https://bugzilla.gnome.org/show_bug.cgi?id=773525
|
|
|
|
|
|
|
|
|
|
|
| |
When the notebook is visible, the pointer intersection with the floating
bar check fails. That is due to how the floating bar bounds are
calculated - the offset of the overlay when tabs are visible is not
taken into account, allowing the upper y bound to become smaller than
the lower bound. That can be fixed by calculating the upper y bound
relative to the lower one.
https://bugzilla.gnome.org/show_bug.cgi?id=773119
|
|
|
|
|
|
| |
And put them together to have a single idfdef, more manageable.
https://bugzilla.gnome.org/show_bug.cgi?id=775935
|
|
|
|
|
|
|
|
|
| |
Batch rename-related source files are conditionally built depending on
whether Tracker support is enabled. Despite that, some code that depends
on batch rename support is included in some core source files without
being conditionally compiled as well.
https://bugzilla.gnome.org/show_bug.cgi?id=775935
|
|
|
|
|
|
| |
Otherwise we might pick up an icon at an invalid size for a given scale.
https://bugzilla.gnome.org/show_bug.cgi?id=776896
|
|
|
|
|
|
|
| |
This is one of the condition that needs to trigger an invalidation to
recompute layout at the correct size.
https://bugzilla.gnome.org/show_bug.cgi?id=776896
|
| |
|
|
|
|
|
|
|
| |
Progress info details are set instead of taken, resulting in a string
leak.
https://bugzilla.gnome.org/show_bug.cgi?id=774232
|
| |
|
|
|
|
|
|
|
|
| |
Currently, when packing the separator for the prompt, it is expanded.
That means that it can grow vertically as the window grows, creating a
visually unappealing effect.
https://bugzilla.gnome.org/show_bug.cgi?id=773211
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The search provider gets unregistered twice when trying to run Nautilus
with a primary instance active, resulting in the infamous warning:
GLib-GIO-CRITICAL **: g_dbus_interface_skeleton_unexport: assertion
'interface_->priv->connections != NULL' failed
The fix is clearing both dbus_manager and search_provider in
nautilus_application_dbus_unregister() as opposed to during
finalization.
https://bugzilla.gnome.org/show_bug.cgi?id=773383
|
|
|
|
|
|
|
|
| |
When the user toggled the Show Hidden Files option on the toolbar menu,
previously the menu disappeared - which isn't the correct behaviour. This
commit changes it so that the menu doesn't disappear.
https://bugzilla.gnome.org/show_bug.cgi?id=766952
|
| |
|
|
|
|
| |
g_file_get_uri needs to return non-null.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
commit 24dae9e0a "file-operations: avoid scanning the same files"
added a way to skip files which have already been processed
in scan_dir(). To do that, the code uses G_FILE_ATTRIBUTE_ID_FILE,
which does not seem to be widely implemented by gvfs modules.
When the file has no ID, it will not be accounted for in the SourceInfo
structure used to track copy progress, resulting in progress not being
shown, and Nautilus saying that it's copying 0 files.
This commit changes the logic to use the file URI instead of the file ID
to check if the file has already been processed.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We were setting the icon position metadata for any operation in any file
due to legacy code where nautilus was able to have a
freedom-of-placement canvas in the regular windows.
However that changed and we can only have freedom-of-placement in the
desktop window.
Setting the metadata is a extremely expensive operation, and was causing
issues, outlined in https://bugzilla.gnome.org/show_bug.cgi?id=757747,
summarizing copy and move operations where really slow when using drag
and drop, operations when we need to store the icon position when using
a freedom-of-placement canvas view.
This patch tries to only set the metadata when necessary, which is when
using the desktop window. However the code is pretty integrated with the
rest of Nautilus, so the patch was able to isolate the part when the
metadata is set and only provide metadata when the target is the desktop.
However, for unsetting the metadata the patch need to check if it's
the desktop inside the file-operations, which is less than ideal.
https://bugzilla.gnome.org/show_bug.cgi?id=775910
|
|
|
|
|
|
|
|
|
|
| |
In the function initial_permission_state_consistent, when the check was
made to see if the files have the same permissions, the check was not
correct.
To fix this, also do & between the first_permissions and the mask.
https://bugzilla.gnome.org/show_bug.cgi?id=698955
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We were scanning the sources as we do for copy and move operations,
and to count the files we were recursing inside directories.
However, the trash operation doesn't recurse at all, so the operation
progress was wrong.
Instead of that, don't recurse for counting the number of files, so the
trash progress is correct. Although is not coherent with the rest of the
operations, recursing for just showing all the files in the progress
would be extremely counterproductive.
https://bugzilla.gnome.org/show_bug.cgi?id=775094
|
|
|
|
|
|
|
|
|
|
|
| |
We were not checking whether the operations was aborted from the user
side when setting the details,so that could make a race and the info
thread setting the details to canceled and the operation thread setting
the details too afterwards before realizing the operations was canceled.
To fix it, always check whether the info progress was canceled or not.
https://bugzilla.gnome.org/show_bug.cgi?id=775094
|
|
|
|
|
|
|
| |
The item list in nautilus_clipboard_is_cut_from_selection_data () is
never freed.
https://bugzilla.gnome.org/show_bug.cgi?id=774232
|
|
|
|
|
|
|
| |
The emblem acquired in apply_emblems_to_icon () is not unreffed as it
should be, since g_emblemed_icon_add_emblem () adds a ref to it.
https://bugzilla.gnome.org/show_bug.cgi?id=774232
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
That means, now we will show the grid icon when the current view is
the list view and the opposite.
This goes in similarity with gnome-boxes.
The main question to decide this is, is the button an action, and
therefore the icon should be the "target" of the action, or is more
about an information of the current state, and therefore the icon should
be about the current view type?
This patch decides the icon represents an action, although would be good
to have some user testing.
https://bugzilla.gnome.org/show_bug.cgi?id=771075
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We were relying on the current view to return the icon to put in the
toolbar, however that requires a view instance and also we cannot
control really what icon we want to show in which circumstances.
We want more control about that, so we need a single entry point where
we can get the icon to show depending on the known view types we
have.
So this patch converts the view property to a static method.
https://bugzilla.gnome.org/show_bug.cgi?id=771075
|
|
|
|
|
|
|
|
| |
nautilus_file_list_filter() creates lists in the wrong order, causing
files activation to occur in the same fashion. This is caused by double
reversal of the file list when filtering files.
https://bugzilla.gnome.org/show_bug.cgi?id=773988
|
|
|
|
|
|
|
|
| |
When renaming a file, the extension is not selected for editing, even if the
file is a directory. In order to fix this, deselect the extension only if the
renamed file is of regular type.
https://bugzilla.gnome.org/show_bug.cgi?id=759718
|
|
|
|
|
|
|
| |
The sort-directories-first gsetting should be consistent
system-wise.
https://bugzilla.gnome.org/show_bug.cgi?id=750996
|
|
|
|
|
|
|
|
|
|
| |
If the metadata has the character "/" the metadata cannot be used.
Rather than showing a conflict when this metadata is used, it should
actually be disabled, because the user will never be able to rename the
files using this metadata.
https://bugzilla.gnome.org/show_bug.cgi?id=771070
|
|
|
|
|
|
|
|
|
| |
The transitions for the rename popover are disabled.
To fix this, use gtk_popover_popdown and gtk_popover_popup instead of
gtk_widget_show and gtk_widget_hide.
https://bugzilla.gnome.org/show_bug.cgi?id=771614
|
|
|
|
|
|
|
|
|
| |
At least for the time being, desktop icons are not supported on
wayland. This may or may not change, but in the meantime we can
still meet user expectations by forcing the supported backend
instead of failing with an unsupported one.
https://bugzilla.gnome.org/show_bug.cgi?id=748347
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
That means, now we will show the grid icon when the current view is
the list view and the opposite.
This goes in similarity with gnome-boxes.
The main question to decide this is, is the button an action, and
therefore the icon should be the "target" of the action, or is more
about an information of the current state, and therefore the icon should
be about the current view type?
This patch decides the icon represents an action, although would be good
to have some user testing.
https://bugzilla.gnome.org/show_bug.cgi?id=771075
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We were relying on the current view to return the icon to put in the
toolbar, however that requires a view instance and also we cannot
control really what icon we want to show in which circumstances.
We want more control about that, so we need a single entry point where
we can get the icon to show depending on the known view types we
have.
So this patch converts the view property to a static method.
https://bugzilla.gnome.org/show_bug.cgi?id=771075
|
| |
|
| |
|