Subject: Undelete help screen and a sketch of the howto contents Hi all, here the result of my efforts to cut down the length of the help screen donated by Emanuele Cipolla to fewer lines :-) FreeDOS Undelete - (C) Copyright 2001-2002 Eric Auer Undelete is free software; it is distributed under the terms of the GNU General Public License. See COPYING for details. SYNOPSIS: undelete [action] [what] [destination] [optional size] Possible [action]s: syssave Saves the 1st or 2nd copy of the FAT, boot sector or root directory. Select fat1, fat2, boot, or root in [what]. follow Looks for a (possibly deleted) file starting at the cluster [what] and saves data to a file given as [destination]. The output of DIRSAVE helps you to find the right cluster number. dirsave Like FOLLOW, but saves a directory to a file. Directory [what] must be given by absolute path starting with \ OR by cluster number. Also shows a technical directory listing on the screen. Specifying [size] is not needed, but you can override the autodetection by specifying [size] (in clusters for FOLLOW, in sectors for DIRSAVE). [Destination] must be on another drive than the current drive. Data is always read/recovered from the drive from which undelete is invoked. And here a sketch of the topics I would like to see in the HOWTO. Using undelete to "mirror" important drive data (if your filesystem gets completely broken, you can try to write back this data): > Run undelete in syssave mode for all 4 "what" values and keep the > resulting files in a safe place. On normal systems, you can glue > them together in the order "boot fat1 fat2 root" to reconstruct > the first part of your partition, starting with the first sector. > To write back the data, you could use debug (w command). You can > also merge saved and existing data with a hex editor, and so on. > IMPORTANT: While "mirror" data is very helpful to experts that try > to repair your system, please do not try to write it back if you > do not exactly know what you are doing!!! Using undelete to find undeleteable files or directories: > Run undelete in dirsave mode. You will see deleted directory entries > specially marked, and you will see their cluster numbers on the > screen. You can redirect screen output to a file, and you will get > a binary file for further processing by experts (also useful to > "mirror" the status of that directory, if you fear that it may break > soon for some reason or another). Dirsave mode is also useful to get > technically detailled directory listings. They are, however, not very > readable for humans if the directory in question contains long file > names. You can also cascade dirsave runs: Run dirsave on an existing > directory to find the starting cluster of a deleted directory. Then > run dirsave on that cluster to find deleted or dropped files in that > deleted directory, and so on. TAKE CARE: Dirsave does only accept > absolute paths without drive specification. As all undelete modes, > the source drive is implicitly selected as the drive from which you > invoke undelete. So those are wrong [what] values: c:\somewhere > somewhere, .., ..\somewhere, and so on. > A correct [what] value would be \here\there or 1234 (a cluster number). Using undelete to recover a file: > Find the starting cluster of the file using dirsave, as explained > above. Then run undelete follow 1234 x:rescued.bin (for example) to > save the contents of the deleted file starting on cluster 1234 to > the file rescued.bin on drive x: - You can either extract deleted or > existing files. The use of following existing files is that you can > override the size value (in clusters, undelete tells you how big a > cluster on the current drive is when you start undelete), so if you > have accidentally overwritten a long file OLD with a short new file > NEW, you can find the cluster number of NEW, give the size of OLD > and then undelete follow will give you a file that starts with the > contents of NEW but contains the not-overwritten end of OLD as well, > hopefully. Normal use, following simply deleted files, will normally > not need a size to be given, but auto-detect the end of the deleted > file. So far for my ideas on that undelete howto. Feel free to publish the sketch (for example to the list) if you need something for ripcord and do not want to wait for the better and more readable version by Emanuele Cipolla. It would be nice if somebody could provide some sample "screen- shots", for example undelete syssave boot >sample-syssave-screenshot.txt and especially similar files for a practical example of "how do I recover that one file which I have just deleted in directory d:\foo\bar\ to my temp directory e:\temp?". Maybe a short example of how long filenames look in undelete output and which of the lines then contains the 8.3 version of the filename, too? Greetings, Eric PS: Even a "walkthrough" style example of undeleting a single file could be a very useful bonus.