just to folow up on this. I did some tests with real filesystems on block devices, instead of loopback devices, on an etch system that previously exhibited the bug. (nis 3.17-6, libc6 2.3.6.ds1-13)
While the system still showed the bug with ext3 filesystem images mounted via /dev/loopN, it did not occur with ext3 filesystems on real block devices like /dev/hda8. It appears this issue is confined only to loopback devices and in some circumstances. I could not see any obvious errors in the C code that handles such devices but I only made a brief inspection. Cheers Vince -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
