Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
|
|
This change requests the initial backup files more gracefully as
the device might need more time to prepare the backup data for
bigger backups. Otherwise we cause a crash by attempting to parse
a plist message we did not receive yet when the device is busy.
|
|
|
|
|
|
|
|
This commit fixes the problem that the .mddata and .mdinfo files were
not removed due to an incorrect check for the result of stat(). Since
the file data is sent in chunks we have to check for existing files
only when the first chunk of a file is received. That's why we now reset
hunk_index each time a file has been completely received.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
[#114 state:resolved]
|
|
|
|
[#113 state:resolved]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This hammered the device with BackupFileReceived messages and caused the
backup to fail in the middle of the backup process.
|
|
The checked node was already changed within the loop.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This better reflects that a newly allocated plist_t is returned.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
These communication functions operate on an iphone_connection_t, thus their
function name should reflect that.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The lockdown constructor was doing more than needed. Pairing and session
negotiation is now handled by lockdownd_client_new_with_handshake().
|