Technical notes on version 2.1
General:
+24May07
+kes Correct typo in UTF-8 error message as reported by:
+ jhernandez <julianhernandez@gmail.com>
+kes Add additional sm_check() code on debug level 1.
+kes Set default debug level to 1.
+kes Trap SIGABRT in attempt to get traceback.
+23May07
+kes Reduce bat connect timeout from 30 to 15 seconds.
+kes More restructuring and implementing BSOCK class in place of
+ old bnet.c code.
+kes Remove a few unnecessary malloc() tests in crypto code.
+kes Turn off crypto calls for > 128 bits in crypto.c if HAVE_SHA2
+ not defined. Bug reported by Allan Black <Allan.Black@btconnect.com>
+kes Remove duplicate HAVE_CRYPTO definition in config.h.in
+22May07
+kes Fix Verify InitCatalog mysql_escape_string() trashing memory.
+ Make buffer bigger.
+kes Fix Verify InitCatalog. The attributes were not pointing to the
+ correct saved location.
+kes Implement message callback so that GUI is sure to get all
+ daemon messages. Currently displayed in a dialog in bat.
+kes Ignore SIGUSR2 (Bacula timer) in bat.
+kes Apply regress patch from bug #859 to allow regress to work
+ with Solaris compiler.
+21May07
+kes Begin adding TLS support to bat.
+kes Apply UTF-8/16 patch from Yves Orton <demerphq@gmail.com> to
+ clean up lex.c and make it more readable.
+20May07
+kes Move more bnet functions into the BSOCK class.
+kes Fix tray-monitor by not requiring a timer interval in bnet_connect()
+kes Complete change of berrno strerror() method to bstrerror()
+Release: 2.1.10 beta
+18May07
+kes Cleanup incorrect email addresses in bsmtp.
+kes Make bat display initial messages rather than discard them.
+16May07
+kes First cut of adding .mod to the run command.
+kes Implement auto display of messages in bat.
+kes Enhanced preferences to allow all Bacula output to be
+ displayed in bat.
+15May07
+kes Try to make bsmtp date routines more generic.
+kes Fixed bug #856 autochanger documentation.
+kes Rework bsmtp date editing for Win32. This fixes bug #854.
+kes Add new cats entry point so Win32 builds.
+14May07
+kes Attempt to fix bsmtp date editing on OSes without %z editing.
+ Fixes bug #854.
+kes Do better checking for the MySQL 64 bit libraries looking for both
+ .a and .so libraries.
+kes Add a kludge for Ubuntu's non-standard locations for MySQL -- allow
+ /usr/include/mysql/mysql.h AND /usr/lib/... rather than /usr/lib/mysql/...
+kes Require the thread safe version of MySQL during configure. Thanks
+ to Andre Noll for pointing this out.
+kes Ensure that calls to find_jobids_from_mediaid_list() check count before
+ continuing in migration code.
+kes Apply implementation of tray-monitor font derived from gnome-console
+ submitted by Andreas Piesk <a.piesk@gmx.net>
+13May07
+kes Eliminate important memory loss in SQLite
+kes Fixed the batch insert MySQL thread specific data that was not
+ being freed. See cats/mysql.c:db_thread_cleanup().
+kes Fix restore before command.
+kes Convert old hdr.name to name() in a few places.
+kes Implement update jobid command.
+kes Return all time_t dates in db_get_job_record()
+kes Stop watchdog in SD earlier.
+kes Put Pool just after Storage in please mount message.
+kes Fix pointer usage bugs in Verify InitCatalog pointed out by
+ Eric.
+12May07
+kes Update the po files.
+kes Fix Verify InitCatalog by making it cache attributes.
+11May07
+kes Change name bacula-bat to bat for man 1 page to correspond to
+ Makefile.
+kes Check for client and store NULL pointers in status command. This
+ fixes bug #845.
+kes Fix non-ssl build so that it works (add JCR *jcr to crypto structures).
+kes Fix sign extended editing of %p.
+kes Clean out a little old #ifdefing in cats/mysql.c
+10May07
+kes If time diff between DIR and FD is more than 600 seconds convert
+ INFO message to WARNING.
+kes Implement heap size display in status for all daemons.
+kes Attempt to fix MySQL db open memory loss (not successfull).
+kes Don't set SIGKILL and SIGSTOP neither can be trapped.
+kes Eliminate memory loss in MySQL (possibly other SQLs) during
+ Query. This looks like a MySQL but but by doing extra releases,
+ the memory goes away. This should fix bug #847.
09May07
kes When backup fails, cancel SD before waiting for sd termination.
kes If SD is waiting for FD to connect during cancel, wake up the