[EMAIL PROTECTED] wrote on 03/16/2007 02:40:38 PM: [...] > > > I'm working on a patch. The hold up is digging through restore to make > > > sure I do this correctly. No point in frogging up dumpinomap, > > > dumpdirmap, and usedinomap if restore doesn't do what you expect. > > > > Why should this interfere with restore? I was thinking of re-reading the > > What do you think is written to the dump device (file, tape, or > otherwise)? The maps make an appearance with the dumped data. Restore > reads those maps back in for it's use. If I screw up the dump, your > screwed on restore. Kind of defeats the purpose of doing the dump in > the first place. ;)
I can understand that, of course, I was just thinking that the only modification required was to tell dump(8) not to descend the tree, but someone on the list made me note that dump(8) in fact does not descend anything but reads the inode table directly. I think I'll leave the patch to whom who's capable of reading the dump source... :-) Bye. D.