> I've been gathering information about mkbfs and learning > about the filesystem in order to write the "mkbfs" > program, which initializes a file or partition with the > BFS. I'm ready to get started with the code but I've > noted that there are various different mechanisms in use > throughout the source for parsing command line arguments. > > Can we adopt a standard? It's an annoying little thing > that is not too difficult to screw up. Especially when > one adds flags that require arguments like this: That's correct but I would suggest for now to stay compatible with the application you're copying. For mkbfs - I don't intend to do it the Be way; they included the whole BFS in the mkbfs command and let it create the file system with the native functions. I don't like that. Since I am currently working on the VFS I have yet to find a way to make it better - for the current BeOS architecture I would let mkbfs write a dummy super block with the flags set to 'INIT' (or something like that) and BFS does the work on first mount. Anyway, I think it's best when I will do that, too, or at least wait until the new VFS is finished (although we could also do a R5 replacement; Bruno mentioned some problems with partitions bigger than 50 GB). In that case, please contact me. Adios... Axel.