Technical notes on version 2.3
General:
+08Jan08
+kes Undo recent reservations changes ... will apply them later
+ when they work better.
+kes Fix bsnprintf for float point numbers. I broke recently when
+ parameterizing some variables.
+06Jan08
+ebl Fixes #1034 which cause mysql to hang the connection after 8h
+ebl Change default statistics target of filename.name and path.path
+ fields for PostgreSQL. This speeds lookup by indexes. Thanks
+ to Marc Cousin.
+kes A few more tweaks to new reservation code. Make sure to clear
+ vol released flag when retaking volume. When reading label and
+ label is bad mark volume unused. When recycling, mark volume
+ unused so it can be renamed.
+05Jan08
+kes Fix reserve_volume() so it doesn't release a volume in use
+ (i.e. a volume entry not marked released). This should be
+ the last part needed to fix bug #1018.
+03Jan08
+kes Move Heartbeat documentation from Job to Director resource.
+ This fixes bug #1033.
+02Jan08
+kes Fix existing switch drive SD code to call autochanger to release
+ any old volume. This must be done to keep the autochanger from
+ releasing subsequently newly reserved volumes in doing a close().
+ This should fix bug #1018.
+kes Fail if attempting to get console input in batch mode. This
+ should help fail RunScript console commands that are incomplete.
+kes First cut implementing switch_drive() in SD (not actually called).
+31Dec07
+kes Implement first cut running console commands in a RunScript.
+29Dec07
+ebl Fixes bug #1028 where "Selection Type" option was not usable
+ with JobDefs.
+ebl Fixes bug #897 and bug #1005 where bacula doesn't display runscript
+ output in error.
24Dec07
+kes Add new free_tls() entry point so that Win32 builds.
+kes Fix compile warning in src/lib/util.c
+kes Fix entry point for edit_job_codes() so that Win32 builds.
kes Fix seg fault Frank Sweetser reports in regression testing
on his systems. The problem was that the original author of
bsnprintf.c did not take into account the side effects of