-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 05 Apr 2003 09:11:49 -0500, David Krider wrote: > > What is written above is misinformation. At least partly. > > So to sum up your correction to my "partial misinformation"... As if I > were going about, purposely telling people wrong things... Not so hostile, please. ;) Nowhere did I claim that you spreaded wrong information by purpose. > > If you don't refer to rare times when "cat /dev/cdrom > image.iso" > > ends too early, you probably refer to run-out sectors at the end of > > a disc. > > I specifically said that I didn't know where the problem lies, only that > it didn't work. Let me quote: "Just using md5sum on /dev/cdrom WILL NOT WORK." It _does_ for many of us, especially when the CDs were manufactured professionally or burnt correctly. Another quote: "The actual burned CD image is different from the iso file, but I don't know why." Whether or not you don't know why, doesn't matter. There is nobody on this world who knows everything. But the ISO image can be found on the CD unmodified. Whether it is possible to read the image back to hard disk with "cat", "cp", "dd" or "readcd", is another matter. > Indeed, I was getting I/O errors at the end of my efforts to `md5sum > /dev/cdrom'. According to your instructions, unless you've specifically > taken care to write the disk in DAO mode (and I didn't), you are going > to run into problems. (Or very likely, even in the case of specifying > "-pad".) Let's put it differently. If and only if (!) you get I/O errors at the end, you know for certain that you've run into a problem. :) I don't claim that burning in TAO mode is a source of problems with _all_ CD-writers out there. It may as well be that some writers don't have problems reading their discs. > The problem is that the original poster specifically said that > the disks he cared about were burnt in Windows, and who knows how that > was done with any number of different burning software packages. Doesn't matter. Important is only that you can actually read the ISO image from the burnt CD. If you can't, the CD was burnt incorrectly. Note that even when you get I/O errors at the end, you may still be able to mount the CD and verify all files. The problem is that if you get I/O errors at the end, you cannot check whether everything else passes the MD5 check. Even if you knew the exact size of the ISO filesystem, you could not always create an image with "dd" due to I/O errors at the end. Padding, on the other hand, allows you to read exactly as many sectors from the CD as necessary to re-create the original ISO image. > So the problem remains. For all your explanation, using `md5sum > /dev/cdrom' STILL doesn't work in the general case. I see what you're aiming at. But "md5sum /dev/cdrom" is the proper way to check correctly burnt CDs. It should give the same MD5 fingerprint than the ISO image does. If it doesn't, the image was burnt incorrectly. > If you can't tell, I really take offense at your characterization of my > post. Especially since it doesn't change anything in regards to the > original subject at hand. Don't take offence. I'm not here to offend people. ;-) - -- -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQE+jvIY0iMVcrivHFQRAl9CAJwOArjEXvFBZb8kjQcaZ2lJb3GpYACfQmly 7kNK6iUUoiPfFWdLxn5exA4= =3eCU -----END PGP SIGNATURE-----
