| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
I still think it's odd that it's the backend's responsibility to
delete the source file, as opposed to the client.
|
| |
|
|
|
|
| |
Signed-off-by: Gheyret Kenji <gheyret@gmail.com>
|
| |
|
|
|
|
| |
Signed-off-by: Gheyret Kenji <gheyret@gmail.com>
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
New udev/blkid expose more ISO9660 properties, including 'Boot System
ID'. Existance of this property implies that media is bootable. The
libosinfo udev rules that provided the 'OSINFO_BOOTABLE' property will
soon be ditched as a result but we better keep the check for backward
compatibility.
https://bugzilla.gnome.org/show_bug.cgi?id=694007
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Drive test requires root privileges as it uses the scsi_debug kernel module
and running some commands as root, such as injecting a temporary udev rule for
working around some scsi_debug limitations and running udisksd under a mock
polkit daemon.
Add a "gvfs-testbed" script which sets up some unshared tmpdir overlays as a
sandbox (to ensure that the tests don't destroy anything in the real system),
set up a temporary user etc.
This also enables the Sftp.test_unknown_host, as this depends on a particular
client-side configuration and ssh does not allow using a temporary $HOME.
Integrate this into "make installcheck", so that this uses gvfs-testbed when
being called as root.
https://bugzilla.gnome.org/show_bug.cgi?id=691336
|
| |
|
|
|
|
| |
This was just an oversight.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Signed-off-by: Gheyret Kenji <gheyret@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Turned out we can't easily check for 'Mount(0)' in `gvfs-mount -l`
output since that string can be present from a system mount,
either fstab or native volume monitor. Think of a mounted CD-ROM
drive.
We can't disable volume monitors either since there are tests that
require it. The best solution would be to check for the particular
URI but then we need to ensure the strings match (i.e. escaping issues),
bringing more stress to properly formatted URIs.
This fixes false unmount error reports when there was something else
mounted as 'Mount(0)'.
https://bugzilla.gnome.org/show_bug.cgi?id=691336
|
|
|
|
|
|
|
|
| |
This adds more debugging output for both smb and smb-browse backends
and also brings more verbosity for user-visible error messages.
Changes made to be able to debug weird samba issues remotely:
https://bugzilla.gnome.org/show_bug.cgi?id=691863
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
We need to tell systemd to stay out of our fuse mounts before
any data are destroyed.
This patch reuses libsystemd-login detection as long as there is no
general header or pkg-config file in systemd.
https://bugzilla.redhat.com/show_bug.cgi?id=902743
https://bugs.archlinux.org/task/32715
https://bugzilla.gnome.org/show_bug.cgi?id=560658
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
With implementing proper unmount method in the last commit the mount
should be gone when GVfsJobUnmount returns. It's not guaranteed when
the "mount-removed" signal is received to the application.
Let's not wait for the "mount-removed" signal, a return value from
g_file_unmount_mountable_with_operation_finish() should be enough to
indicate success or error.
https://bugzilla.gnome.org/show_bug.cgi?id=691568
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A proper unmount method was missing from the daemon interface as it was
basically not needed. The mount tracker was watching mounts on the bus
and if child disappeared, that event was treated as unmount.
With recent introduction of gvfs test suite we started using
man-in-the-middle program for connecting to private SMB server. However,
certain libsmbclient versions are forking gvfs backends (gvfsd-smb in
this case) which makes d-bus confused and name owner tracking doesn't
work as expected.
Moreover the forked process got stuck due to internal samba LIBSMB_PROG
handling after the original process did its job and unmount was triggered
from user side. The original process exited just fine but the forked one
was blocking the name on the bus probably and gvfs mount tracker was fooled.
Eventually the stuck libsmbclient call times out and exits gracefully.
This happens with samba-3.6.9
This patch adds proper unmount call to indicate the backend is going away
so even if there are leftovers, user is indicated a proper unmount.
https://bugzilla.gnome.org/show_bug.cgi?id=691568
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=686526
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
In load_image(), process all pending main loop events to get an up to date
volume monitor. Not doing that can cause tests to still see the temporarily
added and removed device that was created for loading the image.
|
| |
|
|
|
|
|
|
|
| |
The computer backend needs to use gvfs GVolumeMonitor in order
to track gvfs mounts and local and remote volume monitor changes.
A fallout from commit 79a53dc1b86b35dfbeb377e13f760d6817f7597b
|
| |
|