Re: Partman

[Anton Zinoviev <anton@lml.bas.bg>]

On  3.IX.2003 at 14:21 Anton Zinoviev wrote:
> 
> The Debconf templates contain rather terse messages.  They contain
> dots instead of spaces because Debconf deletes spaces (and it does
> this also for the non-break spaces U+00A0).

No.  It doesn't do this for non-break spaces, but I happened not to
have properly configured utf-8 environment.  Now U+00A0 works
wonderfuly. :-)

> I wasn't able to test these udebs because I don't know how to satisfy
> the library requirements of partman-parted.

I resolved this.  My library problems were caused by #183143.  It
seams that glibc-people don't fix #183143 only because they don't know
if there is a consensus about the solution proposed by Sebastian
(libc-udeb should be renamed to libc${SONAME}-udeb and provide
libc${SONAME}).  Is there a consensus about this?

Today I was able for first time to start partman from inside d-i.
Then I repartitioned succesfuly my hard disk. [*] The new packages
replaced the old at http://people.debian.org/~zinoviev.

Unfortunately the newt frontend dies with segmentation fault when I
try to change the flags of some partition (the problematic question is
a multiselect).  I used version 0.39 of cdebconf.  I will see if the
same happens with the newer version.  It is possible however that this
problem is caused by incompatability of the libraries I use.  I will
try to recompile them.  

The text frontend works fine in bterm, but in the regular text console
U+00A0 may show as other symbol.

Anton Zinoviev

[*] It is probably harmless to try how different partitioning
    operations work right now if you don't choose the option `Exit and
    commit changes'.  However I would like to give these packages some
    extensive testing and bugfixing before you use them to partition
    hard disks with important data.