Package: gvfs / 1.22.2-1

Metadata

Package Version Patches format
gvfs 1.22.2-1 3.0 (quilt)

Patch series

view the series file
Patch File delta Description
01_modules_dir.patch | (download)

configure.ac | 2 1 + 1 - 0 !
1 file changed, 1 insertion(+), 1 deletion(-)

---
04_hurd_path_max.patch | (download)

monitor/gdu/ggduvolumemonitor.c | 33 31 + 2 - 0 !
monitor/hal/ghalvolumemonitor.c | 18 17 + 1 - 0 !
2 files changed, 48 insertions(+), 3 deletions(-)

 [patch] volume monitors: on glibc systems, let realpath() allocate
 its own buffer
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This was originally to support Hurd, which doesn't have a PATH_MAX
at all. It also seems a good idea for systems where PATH_MAX is just a
convenient lie, such as Linux.

Based on a patch by Samuel Thibault, as revised by Martin Pitt,
Raphal Hertzog and Emilio Pozuelo Monfort.

Bug-Debian: http://bugs.debian.org/555055

metadata dont flush null tree.patch | (download)

metadata/metatree.c | 5 4 + 1 - 0 !
1 file changed, 4 insertions(+), 1 deletion(-)

 [patch] check if metatree exists before flushing it

Also change meta_tree_clear () so it sets fd to -1; this will make
meta_tree_exists () work as expected. Bug 598561.

metadata nuke junk data.patch | (download)

metadata/metatree.c | 12 11 + 1 - 0 !
1 file changed, 11 insertions(+), 1 deletion(-)

 [patch] nuke the metadata file if magic blob is wrong


dont crash on null job.patch | (download)

daemon/gvfschannel.c | 5 3 + 2 - 0 !
1 file changed, 3 insertions(+), 2 deletions(-)

 don't try to announce the finish of a null job.
handle inactive vfs.patch | (download)

client/gdaemonvfs.c | 2 1 + 1 - 0 !
1 file changed, 1 insertion(+), 1 deletion(-)

 if the vfs never initialized, don't crash when creating volume monitors
ref jobs in thread.patch | (download)

daemon/gvfschannel.c | 5 3 + 2 - 0 !
daemon/gvfsdaemon.c | 7 5 + 2 - 0 !
2 files changed, 8 insertions(+), 4 deletions(-)

 make sure to keep a ref to jobs while they run in a thread
fix unmount crash.patch | (download)

daemon/gvfsbackend.c | 6 3 + 3 - 0 !
daemon/gvfsbackend.h | 3 2 + 1 - 0 !
daemon/gvfsjobunmount.c | 26 18 + 8 - 0 !
3 files changed, 23 insertions(+), 12 deletions(-)

 gvfsjobunmount: block new requests before calling unmount() on a
 thread

Block new requests before calling unmount() on a separate thread to
prevent a race where new jobs are received and processed while the
unmount() is being executed.

This is not necessary for try_unmount() because all the job handling is
done on the same thread as the try_unmount() method and requests are
blocked when the try_unmount() method completes.

https://bugzilla.gnome.org/show_bug.cgi?id=710986

fix cancellation.patch | (download)

client/gdaemonfile.c | 14 12 + 2 - 0 !
client/gvfsdaemondbus.c | 13 9 + 4 - 0 !
client/gvfsdaemondbus.h | 2 2 + 0 - 0 !
3 files changed, 23 insertions(+), 6 deletions(-)

 client: use the correct serial with file transfer cancellation

When using Copy, Move, Push and Pull with progress callbacks, the
progress callbacks increment the connection serial, so retrieve the
serial immediately after the dbus method call so that the correct serial
is used for cancellation.

https://bugzilla.gnome.org/show_bug.cgi?id=720058