Re: bugreport concerning progsreiserfs

[Szakacsits Szabolcs]

On Thu, 28 Oct 2004, B.Hakvoort wrote:

> Before adding this kind of programs directly to gparted i like to try to
> add them to libparted. 

I wanted to do the same for ntfs over two years ago but because people
relatively often reported problems using Parted, unlike using other
partitioners, I thought I wait until it matures.

Then I realized Parted isn't developed anymore and Andrew doesn't really
have the time to properly maintain it. There was time, not too long ago,
when the situation was really catastrophic and MANY people thrashed their
partition tables by Parted.

> Much like libparted interfaces to progsreiserfs now, 

Corrupting data, known for a year, still not fixed? ;-)
  
        http://qtparted.sourceforge.net/forums/viewtopic.php?t=62
        http://qtparted.sourceforge.net/forums/viewtopic.php?t=175

Or the ext2/ext3 support that's about totally useless for 2-3 years now?

I've even seen several FAT problems, nobody investigated, fixed them.

Thrashing Windows dynamic disks is yet another unsolved problem (not
directly Parted's problem, it just doesn't provide the needed data for 
its users to be able to prevent destructions).

> it maybe possible to let it interface to libntfs.. 

It's not. You can't use anything from libntfs for Parted. Ntfsresize has
fsck and resize, mkntfs has the creation, ntfsclone the fsck again and
efficient copy but they need a lot of very careful modifications for
librarization.

I was working on moving them to libntfs for libparted but I stopped
because basically the invested time is much more than the benefit would be
and still very doubious, hard to support results due to Parted current
shortcomings and status.

        http://mlf.linux.rulez.org/mlf/ezaz/ntfsresize.html#parted

> If this proves too hard/clumsy/time-consuming/whatever i can always
> add it to gparted directly.

After over two years, waiting and working on it, I say it's not worth
unless you _really_ have a lot of time and energy to spend on it and
support the outcome on the long term.

> Think about it, if Graeme succeeds in fixing the ext2/3 code and Yury is
> able to fix progsreiserfs and i manage to add support for ntfs, We give
> our good ol' libparted an enormous boost :)

I doubt. I've seen several great plans on this list without much or any
work done. 

And even if you succeed, how long will they work again? Those who don't
learn from history are doomed to repeat it ;)
 
> Isn't that a bright future? ;) (maybe i'm getting a bit over-
> enthousiastic here :P)

Sorry to be "negative" but I really think you're over enthousiastic 
and don't realize how big task you undertook. 

Several GUI partitioners reached the GParted level or even more in 
the past: PartitionMorpher, PartGUI, GTKParted, Partition Surprise,
Kbpartition, QTParted but unfortunately none of them is developed or
maintained anymore. I don't know why but I suspect one of the reasons 
is the underestimated complexity and time needed to keep it further
developed, or even just maintained.

The only survivals are the commercial distro partitioners, DiskDrake 
and YAST and they exactly do what I suggested: with minimal invested 
time they get the best functionality and quality.

Like many other people, I would also like and wish if you succeeed in 
the future, instead of seeing another attempt being dead after a year.

        Szaka