Technical notes on version 2.3
-New features (buried):
+
+New features:
+Accurate files
TLS authentication
bextract Win32 data
Run afterVSS
Multiple command/console per runscript
-console command
+console command in runscript
Bat API
Plugin API
State file updated on job term
Improved status commands for Bat
Spooling/despooling status
Implement ftruncate for NFS devices
-Add long term statistic job table
+Add long term statistics job table
General:
+03Apr08
+ebl Add statistics prunning with option and menu
+kes Separate unload_dev() from unload_other_device() in autochanger.c
+02Apr08
+kes Apply patch from bug #1069 that corrects spurious error messages
+ when ACLs enabled on SGI but no ACL exists.
+ebl Fix small segfault in bacula-sd when debug level is 500
+kes Modify run_program() and run_program_full_output() to use
+ call by reference for the results string. This corrects a long
+ standing problem where the address of the string may be changed
+ in the subroutine but not in the calling program.
+01Apr08
+kes Re-enable code to remember last volume mounted on a non-tape
+ Autochanger.
+kes Add patch supplied in bug #1068 that fixes a SD crash when using
+ a Virtual autochanger.
+kes Generate correct JobMedia records during spooling/despooling when
+ running concurrent jobs. Thanks to Tom Ivar Helbekkmo
+ <tih@hamartun.priv.no> for excellent analysis and testing.
+31Mar08
+kes Tweak hash algorithm for htable using Martin Simmons idea
+ for doing a circular shift.
+kes Remember what volume is in VTLs (previously turned off for
+ testing).
+kes Analysis of JobMedia records by Tom Ivar Helbekkmo <tih@hamartun.priv.no>
+ pointed out some problems in creating JobMedia records during
+ spooling with concurrent jobs. Fix the problem by updating the
+ start file position before despooling and generate a jobmedia
+ record at the end of despooling.
+28Mar08
+kes Fix handling of JobMedia records during spooling and when a
+ Volume is changed without actually writing data to the Volume.
+27Mar08
+kes Rework certain SD locking based on gdb tracebacks of deadlocks
+ sent by Eric. This code now uses only the volume_lock() rather
+ than the reservations lock when at EOM and acquiring a new
+ volume.
+kes Ensure only one exit point in several subroutines.
+26Mar08
+kes Implement bsock code to permit locking. Use the new code in Verify
+ in the FD to prevent the heartbeat thread from disrupting the Verify
+ data sent to the Dir.
+kes Modify SD locking to eliminate locking the reservations system from
+ outside the system. Use the volume lock when getting media data
+ and reserving a new volume at end of media.
+kes Add Eric's volume_unused() fix to stored.c so that Volumes found
+ when SD starts are not locked into place.
+24Mar08
+kes Fix error compiling runscript code on 64 bit machines.
+kes Tweak Win32 mount point code.
+kes Fix a couple of bugs in the accurate code (bad sscanf, buffer not
+ cleared giving bad file estimate).
+kes Implement BIG_MALLOC in htable code. Runs 2.5 times faster for
+ 5 Million entries.
22Mar08
kes Keep Schedule run lists in order as suggested by
Bastian Friedrich.