[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: partman needs testing

On Mon, Sep 06, 2004 at 04:55:17PM -0400, Joey Hess wrote:
> So this is a call for anyone who can to test partman in the daily
> builds. We need to not only do automated partitioning tests, but manual
> partitioning and try all the little nooks and corners and strange
> combinations that are not often tried and make sure it works at least as
> well as it did in rc1. Thanks.

Unfortunately I can not do this right now.  Here is a list of my
changes in partman since rc1:

1. The conversion long-number -> human-string is changed to round
   correctly numbers so 1990MB is shown as 2.0GB rather than as 1.9GB

2. Do not show the dialog "Write changes?" when the partition table is
   not changed, no partition is to be formatted and partman has
   formatted (in a previous run) at least one partition.  If there are
   not changes and no partition has already been formatted then show the
   dialog "No changes! Really continue?".  Joey did some additional
   changes here

3. Several changes regarding the support of dvh disk labels (SGI). Needs
   testing! (but partman is not default on SGI anyway)

4. Some changes in the recipes of partman-auto, Joey undid the 
   changes regarding mount options

5. Do not sort the available file systems alphabetically so that again
   Ext3 is default instead of Ext2

6. Autoset "noatime" and "quiet" mount options for new partitions, Joey
   undid this

7. For most file systems the dialog "Select mount point" is changed

8. For all file systems the dialog "Select mount options" is changed.
   In rc1 there was the following bug - if the user selects for example 
   "notail" mount option with ReiserFS and then changes ReiserFS to Ext3
   the "notail" mount option is still active despite that Ext3 doesn't
   support "notail"

9. On Ext2/3 allow the user to specify the percentage reserved space for
   superuser and the number of inodes

7 and 8 affect all partman-file_system packages so all they have to be

Anton Zinoviev

Reply to: