Age | Commit message (Collapse) | Author | Files | Lines |
|
When accessing/storing key info with userprefs, a device uuid is
required that makes it possible to distinguish between different
devices. On execution of lockdownd_client_new, the uuid is queried
via lockdown and now stored in the client struct for later reuse.
This patch also removes the uuid parameter from lockdownd_pair().
|
|
This fixes a bug where lockdown_check_result() might return -1
and lockdownd_pair() still returns success.
Thanks to dborca for spotting this.
|
|
This will change session_id out of the lockdownd_client_int struct
to a pointer instead of using a buffer of fixed size. The session_id is
allocated anyway by libplist when reading it from the plist received
from the device, so why don't just use it?
[#94 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
This removes the session_id parameter from lockdownd_stop_session
because the session_id is stored in the lockdownd_client_int structure
anyway.
|
|
This is required if the device does not send the EnableSessionSSL:true
key-value pair in the answer to the StartSession request.
[#92 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
This makes afc_truncate look more like afc_file_truncate which is also
using uint64_t for the file size.
[#82 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
This patch also adapts all corresponding internal functions.
The buffer lengths are now consistently handled as uint32_t.
|
|
This should make libiphone compatible with big endian machines.
[#85 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
|
|
[#87 state:resolved]
|
|
gnutls_session_t is already a pointer, we don't need to manipulate a
gnutls_session_t*
[#87 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
|
|
[#77 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
|
|
This patch removes the additional buffer that was used inside
afc_dispatch_packet. So instead of 'alloc, copy header, copy data, send,
free' it will now simply do 'send header, send data'. This should reduce
cpu usage.
|
|
|
|
trying to read a negative amount of data doesn't make much sense, and the
returned 'bytes' value will overflow if we try to do that. Just treat
negative length values as an invalid argument. An alternative way of
handling it would be to silently return OK/0 bytes read.
|
|
trying to write a negative amount of data doesn't make much sense, and bad
things will happen if we try to do that. Just treat negative length values
as an invalid argument. An alternative way of handling it would be to
silently return OK/0 bytes written.
|
|
This function allows to set the modification time of a file. The
time value needs to be specified as nanoseconds since epoch.
[#81 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
Signed-off-by: Matt Colyer <matt@colyer.name>
[#77 state:resolved]
|
|
|
|
|
|
|
|
|
|
We should return any underlying error afc_get_device_info returns
so one is able to act properly. Also renamed it to "key" instead
of "field" to be more generic.
|
|
On a system without a previous pairing, libiphone fails
to connect to the device:
userpref_get_host_id: Using (null) as HostID
The code checks if host_id is NULL but it needs to check
if *host_id is NULL, otherwise it won't generate a new HostID.
[#69 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
For use in front-end applications, also serves to document
the available keys.
|
|
Valgrind complains:
==2678== Conditional jump or move depends on uninitialised value(s)
==2678== at 0x3ACD40646C: lockdownd_client_new (lockdown.c:662)
==2678== by 0x405BA6: g_vfs_afc_volume_new (afc-volume.c:113)
==2678== by 0x406541: g_vfs_afc_monitor_device_added (afc-volume-monitor.c:46)
==2678== by 0x4066B1: g_vfs_afc_volume_monitor_constructor (afc-volume-monitor.c:173)
==2678== by 0x3A7FE11A38: g_object_newv (in /lib64/libgobject-2.0.so.0.2000.4)
==2678== by 0x3A7FE12584: g_object_new_valist (in /lib64/libgobject-2.0.so.0.2000.4)
==2678== by 0x3A7FE126DB: g_object_new (in /lib64/libgobject-2.0.so.0.2000.4)
==2678== by 0x408AD5: monitor_try_create (gvfsproxyvolumemonitordaemon.c:1651)
==2678== by 0x408C20: g_vfs_proxy_volume_monitor_daemon_main (gvfsproxyvolumemonitordaemon.c:1694)
==2678== by 0x37E881EA2C: (below main) (in /lib64/libc-2.10.1.so)
==2678== Uninitialised value was created by a stack allocation
==2678== at 0x405B00: g_vfs_afc_volume_new (afc-volume.c:138)
|
|
|
|
[#60 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
[#46 state:resolved]
|
|
add missing includes, add some casts to avoid format warnings,
mark one function as static. There are still warnings about
unused argc/argv
[#61 state:resolved]
Signed-off-by: Matt Colyer <matt@colyer.name>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
non-blocking locks are supported via AFC.
Signed-off-by: Martin Szulecki <opensuse@sukimashita.com>
|