Enrico Forestieri wrote: > On Sat, Jan 22, 2011 at 01:53:40AM +0100, Pavel Sanda wrote: > > > this patch should help at least for symlinks. > > Why not restricting the workaround to the case where the file actually > is stored on an ext4 file system?
in principle i'm all for it. i would just rename the method to isOnTheDamnedExt4(). and this patch is not workaround. workaround is the current code, when we move the file to backup instread of copying it. maybe they put some new heuristical rescuing code in the kernel meanwhile and we can get rid of the whole thing. i would be also eager to know whether things are better with btrfs. having ext3 i cant test but if there are ext4/btrfs people with up-to-date kernel i'm happy to provide patch and instruction how to test this bug. pavel
