Re: problems with parted1.4.1

[Andrew Clausen]

Andrew Clausen wrote:
> 
> Hi Joachim/all,
> 
> The actual seg fault was fixed in 1.4.2.  (There was a bug in
> the way it was reporting the error, causing the seg fault)
> 
> Further, it looks like your file system is corrupt (but, it
> probably wasn't corrupted by Parted... I can't see any evidence
> of this)
> 
> (The reason why the corruption didn't cause problems for
> the other resize commands, is no blocks had to be moved.  Some
> blocks needed to be moved for the larger resizes, because the
> group descriptors needed to be grown)

I was wrong.

There is a MAJOR bug in ext2_push_meta().  I'm uploading 1.4.4
now.  (1.4.x makes dodgey group descriptors, which usually
don't get written, because it seg faults on the garbage)

The bug was introduced when I added big-endian support.  I can't
believe no-one noticed!  AND, I can't believe I didn't think of
adding a bigger resize to the regression tests (important for
ext2, because of GDT growth)

Andrew Clausen