X-Git-Url: https://git.wh0rd.org/?p=dump.git;a=blobdiff_plain;f=TODO;h=61fbfb7356c39545f871d3bdac2b8fe39582ddf9;hp=05e9a818547b44353188eb1b28a4e040560c1f5c;hb=3dd5ecd75e8008104e314726ec16a4b9118f3d8b;hpb=8590d6133f79f8701243d9a9605d5336581ab2e8 diff --git a/TODO b/TODO index 05e9a81..61fbfb7 100644 --- a/TODO +++ b/TODO @@ -1,45 +1,61 @@ -$Id: TODO,v 1.6 1999/11/17 22:51:27 tiniou Exp $ +$Id: TODO,v 1.21 2001/07/25 13:21:31 stelian Exp $ Need to verify: --------------- -1. Verify that dump works correctly on rev 1 ext2fses (need success - reports) - -2. Verify that dump works with kerberos authentification (this +1. Verify that dump works with kerberos authentification (this was ported from *BSD and was not tested - in fact, this wasn't even compiled in!). Need success reports for this. Urgent items (for the next stable version): ------------------------------------------- -1. Make dump honor the no-dump attribute on directories. Currently, it - honours it only on files. This means that a directory flagged with - the no-dump attribute will be dumped if it contains a file without - the no-dump flag. - All others: ----------- -1. Make dump able to backup several directories and/or files in one - invocation, like the SunOS version. - -2. More documentation? Examples, crontab? +1. More documentation? Examples, crontab? -3. Explore and correct dump problems on a 4GB+ active - filesystem (lseek/read negative count). +2. Explore and correct dump problems on active filesystems + (lseek/read negative count) (This should be OK as of 0.4b14. + Unfortunately, this seems to continue for a very few users). -4. Clean-up the code by removing all those #ifdef _BSD, - since nobody would ever re-compile this on a BSD system. - Since the tape format is the same, the native BSD - restore can be used to restore tapes... - -5. Reimplement the ext2 specific code in a "backend" and +3. Reimplement the ext2 specific code in a "backend" and make the dump code more generic. This would allow creation of other backends for other filesystems. Implementing a (v)fat backend should be quite easy, as for BSD ffs (we - already have the code for this). + already have the code for this). The BSD code in traverse.c + (all those #ifdef _BSD) should go into the ffs backend. -6. Implement a DEBUG option which doesn't fork on each tape, making +4. Implement a DEBUG option which doesn't fork on each tape, making it able to debug dump with gdb. +5. Extend the compression patch to use bzip2. + +6. Make a bootable dump tape? I don't know if it is possible... + +7. From Kjetil Torgrim Homme : + a archive_file + Archive file. Archive a dump table-of-contents in + the specified archive_file to be used by + ufsrestore(1M) to determine whether a file is in + the dump file that is being restored. + +8. Modify (or get rid of) rmt in order to work with non Linux + systems (limited succes has been reported). A good version + of rmt to be looked at is part of star + http://www.fokus.gmd.de/research/cc/glone/employees/joerg.schilling/private/star.html + +9. EA/ACL support in dump (requested by Michael Ju. Tokarev + . + +10. Better readline completition in restore (escape spaces etc). + +11. Dump compiled with RedHat's 7.0 gcc seems to hang for some + people. When compiled with kgcc all the problems disappear. + Will the RedHat's 7.1 gcc resolve this ? + +12. Dump uses /dev/tty for reporting some messages. When invoked + from a script, /dev/tty may not exist. Why not using just + stderr ? + +13. -a and -B option to dump are exclusive.