> > +* 4 bytes cluster number > > Is that sufficient, or is it possible to have an image larger than 64k*4G that > would overflow?
Well, that is 256TB per image. This is sufficient for us. > > +* 1 byte not used (reserved) > > + > > +We only store non-zero blocks (such block is 4096 bytes). > > + > > +Each archive is marked with an unique uuid. The archive header and > > +all > > s/an unique/a/ (by definition, 'uuid' is an acronym that already means > unique; also, it is 'a' and not 'an' before any 'u' pronounced as 'y', which > is > true for both 'unique' and a spelled-out 'uuid') many thanks for your help - will fix those things. > > + * This work is licensed under the terms of the GNU GPL, version 2. > > + See > > Can you please use GPLv2+ (the 'or later' clause is essential if you want your > work to be reusable in GPLv3[+] projects)? sure, no problem. > I didn't review the code, just the specification. I have to wonder how much > of your work overlaps with Paolo's 'drive-mirror' and NBD server work; and it > seems to me that it is better to use 'drive-mirror' for doing backup work into > existing disk formats, rather than inventing yet another archive format. I already tried to explain why that does not work - see my previous posts. What image format provides that functionality?