[pkg-bacula-devel] [pkg-bacula-commits] [SCM] Bacula, a network backup, recovery and verification program branch, migrate-to-ucf, created. debian/5.2.6+dfsg-1-18-g0331c27

[Alexander Golovko]

On Fri, 29 Jun 2012 00:53:39 +0200, Luca Capello wrote:
> Hi Alexandre!
>
> On Thu, 28 Jun 2012 22:54:51 +0200, Alexander Golovko wrote:
>> The branch, migrate-to-ucf has been created
>>         at  0331c27f115b6e9ba51508d7dbba13f41c290d54 (commit)
>
> Another big THANK YOU :-)
>
> Just FTR, I am uploading a new version to unstable before the freeze:
> Other than the two bug fixed (#656912 and #675227), there are other
> improvements (some cosmetic, like the init.d messages, other more
> important, like the -x check for the daemons).  Meanwhile, I am
> finishing checking other "simple" bugs like #602191 or #602431 to see 
> if
> they are still present.  This will be the last upload for wheezy, 
> except
> obviously (any possible) bug fix.
>

Yes, there was a great job! But 602191 is not simple at all and it 
still present in squeeze...
we must fix it in squeeze as soon as possible, due to it mean, that 
there is a count of servers with default bacula-director password.
but it also mean, that we must analyze non-mantainer changes in package 
and add all that changes into git.


> Then we can move to ucf and also to a single build process with all 
> the
> three SQL types :-)
>

yes, you are right.
But what a method for switch between databases types we will preffer?
different packages (as now), /etc/alternatives or something else (for 
example, ask upstream to make next step - choose database type from 
config)?


>> commit e8b89177609328118b991bfc57fd06a399be8160
>> Author: Alexander Golovko <alexandro at ankalagon.ru>
>> Date:   Tue Jun 5 19:11:25 2012 +0400
>>
>>     change template variables names
>>
>>     It's needed for use new substitute() function
>>
>>     @debian_hostname@ -> XXX_HOSTNAME_XXX
>>     XXX_*PASSWORD_XXX -> XXX_*PASSWD_XXX
>
> IMHO we should move to `./configure --with-basename=XXX_HOSTNAME_XXX
> --with-hostname=localhost` instead of manual patching.
>

yes, you are definetelly right, there will be a better choice.


> Is there any reason why you used --with-db-password=XXX_DBPASS_XXX 
> and
> not XXX_DBPASSWD_XXX as in the other *PASSWD?

Only for conjuction with dbconfig-common variables names (dbc_dbuser, 
dbc_dbname, dbc_dbpass).


>
> Thx, bye,
> Gismo / Luca

-- 
with best regards,
Alexander Golovko
email: alexandro at ankalagon.ru
xmpp: alexandro at ankalagon.ru