Hi Kern and list,

I am back with my Unicode problems, but with more information and
try-and-error testing stuff.

First of all, yes, bacula-dir.conf is UTF-8, so it is no problem to
write Unicode strings there. However, bacula-fd or bacula-dir still
ignores them when it comes to Exclude directories or files on bacula-fd
side.

In fact, I didn't mentioned that bacula-fd runs on OS X, and bacula-dir
- on Debian. So there is that catch why my system fails to work.

So, I changed locales in both systems, in OS X it was proper to set
en_US.utf8 to all LC stuff and I put it and exported it in /etc/profile.
In Debian I did in same file, instead of putting en_US.UTF-8, which is
proper label of Unicode system in Linux system. So both system consoles
are UTF-8 enabled.

I also verified bacula-dir.conf file and it is indeed in Unicode and all
characters are fine.

So I want down with debuging and launched bacula-fd (Installed by Fink
on OS X 10.4 up to date) with -vf -d200 flags. Results where interesting
- bacula-fd clearly got UTF-8 chars from bacula-dir.conf, which were
send by bacula-dir (It was shown in bacula-fd debug output). However,
when bacula-fd started to roll it's job, file names which it got from
system where something like this:
/Volumes/RAID/Kopejais/Administrācija/Ilze Ancāne instead of
/Volumes/RAID/Kopejais/Administrācija/Ilze Ancāne

It (propably OS X, because such interpretation I see also when I ls
files in console trough ssh, altough everything is set to Unicode)
converts additional three UTF bytes to something bizzare. What is most
strange, that when I use Terminal from Applications in OS X, it shows
UTF fonts perfectly, but when I log in with ssh, then it is all messed
up.

Anyone has similar problems? Suggestions? Where to look, dig futher?

Peter.

p.s. Haven't tried this with bacula-fd on Linux. If there won't be
problems, then it is OS X problem. Should dig there then. *sights*

On T , 2006-06-21 at 12:14 +0200, Kern Sibbald wrote:
> On Monday 19 June 2006 14:16, Peteris Krisjanis wrote:
> > p.s. I suppose not a first who bothers this list with this question, but
> > as I couldn't find right answer while doing search on list archive, here
> > I go...
> >
> > Hi, my name is Peter and I'm first time Bacula user. At first, I was
> > just confused by system scale first, but now I got lot of things and
> > really love it. However, I have one problem with it's usage in real life
> > in my company and that's when I came to ask help here.
> >
> > What I want:
> > * I want to define FileSet with excluding several files and directories
> > which has UTF-8 symbols there;
> > * I want to Bacula really backup them or exclude them, according to
> > definition :)
> > * I want to restore files with UTF-8 symbols in their names;
> >
> > So far I have tried several combinations with Exclude definition.
> > Problem is that configuration file (bacula-dir.conf) is in standard
> > ASCII/ISO-6689-1, so I tried different combinations to feed UTF-8
> > symbols - escape codes, two bytes - but so far no luck.
> 
> I don't know why you say that bacula-dir.conf is not in UTF-8 because it is 
> *by definition* in UTF-8 format.   Any file that is in US ASCII (as is the 
> case for the default bacula-xx.conf files) is totally indistinguishable from 
> UTF-8, or said another way, "is by default also in UTF-8".
> 
> >
> > My question really is - does Bacula support UTF-8 fully in file names?
> 
> Yes, of course, I test it every day.
> 
> The problem is the interpretation of UTF-8.  If you have not properly defined 
> your system language, you may not be running your system in UTF-8, which is 
> as far as I know, more or less the default for Linux systems.
> 
> For example, (I'm not an expert on this), if your LANG environment variable 
> does not end with .UTF-8, you are very likely NOT running in UTF-8 mode, so 
> none of the system routines that Bacula uses for filename comparisons will 
> work as one would expect.
> 
> For example, here my LANG environment variable is:
> 
> en_US.UTF-8
> 
> 
> 
> > And if not, it could be possible to improve this point (more question to
> > devs)?
> >
> > Just side note - using no UTF-8 file names is no option. And besides, it
> > is just time to get it right - file names with other symbols than ASCII
> > standard is just reality :)
> 
> I admit that the UTF-8 problems and issues are not properly documented in the 
> manual, but I suspect that by starting to get your own house in order will 
> resolve all your problems.
> 
> By the way, all my comments about UTF-8 concern Linux systems (and probably 
> others such as FreeBSD and Solaris).  If you are talking about Win32 systems, 
> it is totally a different story...
> 


Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Reply via email to