| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
The README was not welcoming enough
Add hello world to the README
Closes https://gitlab.gnome.org/GNOME/nautilus/issues/973
|
| |
|
|
|
| |
The gtk4 branch should have the proper manifest already
|
|
|
| |
Was giving errors in the CI and doesn't really provide much benefits
|
|
|
|
|
| |
Errors are usually old deps or such, and needs someone taking care of those.
It's not much of a benefit upstream, so let's remove them.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
The signal 'trash-state-changed' could be emitted by NautilusTrashMonitor while
a NautilusWindowSlot is still initializing the content view. In this case, don't
show/hide extra location widgets to avoid deferencing invalid memory. They will
be created anyway once the content view has been set.
Closes: https://gitlab.gnome.org/GNOME/nautilus/issues/890
Closes: https://gitlab.gnome.org/GNOME/nautilus/issues/862
|
|
|
|
|
|
| |
I was getting:
error: variable 'string' is used uninitialized whenever 'if' condition is
false [-Werror,-Wsometimes-uninitialized]
|
|
|
|
|
|
|
| |
According to the documentation[1], GFileMonitor object must be freed
when it is not necessary. g_file_monitor_cancel() doesn't free the object.
[1]: https://developer.gnome.org/gio/stable/GFile.html#g-file-monitor-file
|
|
|
|
|
| |
Clean up do_finalize method to use NautilusBookmarkList object
instead of casting to itself every lines.
|
| |
|
| |
|
|
|
|
| |
Generate the json from the yaml file.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
GVfs communicates with daemons over abstract sockets. Currently, synchronous
GIO API doesn't work for GVfs locations, because abstract sockets seem to be
blocked by sandbox permissions. This is not problem in most cases as Nautilus
uses asynchronous API usually. But this breaks e.g. "Empty Trash" functionality
as it is for some reason based on synchronous API.
In order to fix this problems, it is necessary to allow access to sockets
beginning with `unix:abstract=/dbus-vfs-daemon/socket-`, but I am not aware of
better way than adding `--share=network`, which allows access to all abstract
sockets:
http://docs.flatpak.org/en/latest/sandbox-permissions-reference.html#f2
Just a note that it is not clear to me, why it affects only the synchronous
API as the asynchronous API also uses abstract sockets. See the corresponding
Flatpak bug:
https://github.com/flatpak/flatpak/issues/2711
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Signed-off-by: Trần Ngọc Quân <vnwildman@gmail.com>
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
To be honest I don't know when it was used and where, but it's unused
now.
Fixes: https://gitlab.gnome.org/GNOME/nautilus/issues/801
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
* Cleaner separation of phases
* Avoid some races
* Fix some leaks
|
|
|
|
| |
To avoid clogging up the main loop.
|
|
|
|
|
|
| |
With the new icons most of the area was white.
Fixes: https://gitlab.gnome.org/GNOME/nautilus/issues/827
|
| |
|
|
|
|
|
| |
The original linked list can be O(n^2) (n = the number of files)
in the worst case.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
glib now offers a reference counted api.
This allows us to replace a lot of legacy code and get rid of
eel_ref_str. GRefString [1] is almost a drop-in replacement.
nautilus-file-operations.c was indirectly depending upon
eel/eel-string.h from nautilus-file-private.h and thus
a new include was added.
This bumps the minimum version of glib to 2.58
Close #781
https://gitlab.gnome.org/GNOME/nautilus/issues/781
[1] https://developer.gnome.org/glib/stable/glib-Reference-counted-strings.html
|
| |
|
|
|
|
| |
(cherry picked from commit 9b69e5eb8f14b251390fd8dd7c8e20863acbff41)
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When the current cursor's row gets deleted, GTK will move the cursor to
the next row, and when setting the cursor it also selects the new
cursor's row, thereby triggering selection signals. The new cursor will
soon be deleted again and the loop repeats.
Since clear() removes all entries, those selections are useless but they
take up most of the time in clear(). For example, when a search returns
a large list, exiting from the search view would make nautilus hang.
At the time simply removing the cursor solves the problem, but to be
future-proof in case GTK does anything fancy with the current selection,
this commit also removes the selection.
Because GTK internally seeking the cursor takes time, only blocking the
selection signal like everywhere else will not remove that overhead.
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently, the search popover is not dismissed when
clicking outside.
Clicking anywhere other than on popover should close the popover.
This patch fixes this behaviour by making the search-popover
modal.
Closes https://gitlab.gnome.org/GNOME/nautilus/issues/528
|
|
|
|
| |
It could create bad blood to close a recently discussed or updated
issue...
|