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

Re: experimental parted in sarge ...



On Fri, Jan 28, 2005 at 05:38:47AM -0800, Steve Langasek wrote:
> Martin,
> 
> Thank you *very* much for your help in testing this.  This bug has been a
> thorn in our sides for some time now, and now that we have solid
> confirmation that the new parted does fix it, we can move forward.

Ok, i have uploaded both 1.6.20-1 to unstable, and 1.6.21-0.exp.1 to
experimental, as Colin told me that he preferred to wait some before moving to
1.6.21. I am not sure i 100% agree, but whatever, as long as go for the new
API, i am happy, and changes will be easier along the way.

Anyway, there are a couple of points in the 1.6.21 Changelog which are of much
interest, apart from 1.6.21 obsoleting a couple of debian specific patches. I
think the following alone :

  * libparted/linux.c (_device_probe_type): sets the type to
    PED_DEVICE_UNKNOWN on block devices that don't fall into the
    standard categories.  This makes libparted correctly update
    the kernel's view of partition tables, etc.

Is one which would be of interest to Anton, not sure though.

Below is the full blown 1.6.21 upstream Changelog :

2005-01-16  Andrew Clausen <clausen@gnu.org>
* released 1.6.21

2005-01-10  Sven Luther <luther@debian.org>
* libparted/disk.c, doc/parted.texi : renamed prep-boot prep partition name to
  prep.

2005-01-08  Andrew Clausen <clausen@gnu.org>
* libparted/linux.c (linux_disk_commit): if the BLKPG method fails, try
  the BLKRRPART method.

2005-01-06  Andrew Clausen <clausen@gnu.org>
* libparted/disk_dos.c (msdos_write): put in a unique identifier (which
  is misleadingly called the "MBR signature" by Microsoft docs and
  EFI specs).  Based on a patch by Matt Domsch <Matt_Domsch@dell.com>.

2005-01-05  Sven Luther <luther@debian.org>
* include/parted/endian.h : fixed PED_SWAP macros missing (), since it broke
  on 64bit hardware as shown in the PED_CPU_TO_BE32(part ? part_num : LINK_END)
  code snipplet, which applied the uint32 cast to part, and not the value.
* libparted/disk_amiga.c, libparted/fs_amiga/amiga.c : Added some uint32 casts.

2005-01-05  Andrew Clausen <clausen@gnu.org>
* libparted/disk_sun.c (sun_alloc, sun_read, sun_write): fixed
  cylinder recording stuff.  pcylcount is now 2 more, and ncyl is now
  2 less than it was before.
* libparted/filesys.c (ped_file_system_clobber): removed a useless
  (always-thrown) assertion for file system support for clobbering.

2004-12-26  Harley D. Eades III  <hde@grics.net>
* libparted/gnu.c (gnu_new): Added three new variables ro_err, rw_err and
  ispath.  The ispath variables is set to 1 if path is indeed a path and not a
  store type name.  This prevents having to use strchr more then once.  Added
  a check for determining if the inputted path is a path and not a store-type
  name.  If (const char *)path is a path then use store_open (). If (const
  char *)path is a store-type name use store_typed_open ().

2004-12-21  Andrew Clausen <clausen@gnu.org>
* libparted/linux.c (_disk_sync_part_table): removed duplicate error message.

2004-12-19  Andrew Clausen <clausen@gnu.org>
* libparted/linux.c (_device_get_part_path): only treat the device as
  devfs if it has /disc on the end.
* libparted/linux.c (_partition_get_part_dev): added.
* libparted/linux.c (_partition_is_mounted_by_dev ): added.
* libparted/linux.c (_partition_is_mounted): now uses
  _partition_is_mounted_by_dev().  This is more reliable than guessing
  the partition's device name.
* libparted/linux.c (_device_probe_type): sets the type to
  PED_DEVICE_UNKNOWN on block devices that don't fall into the
  standard categories.  This makes libparted correctly update
  the kernel's view of partition tables, etc.

2004-12-18  Andrew Clausen <clausen@gnu.org>
* libparted/disk_dos.c (probe_partition_for_geom): improved heuristic
  for finding broken partition tables: test if any feasible solution
  geometry exists against the end cylinder rather than 1024.  This
  heuristic now has fewer false negatives and no false positives.

Friendly,

Sven Luther



Reply to: