Re: [Duplicity-talk] Re: [rdiff-backup-users] Pretty pictures and new version of proposal

[Will Dyson]

On Tue, 2003-09-30 at 17:05, Ben Escoto wrote:

> From Will Dyson's and other's comments, when a file system looks up a
> file, it expects to get its inode number from the directory contents.
> Scanning the index is possible, but may make many operations very slow
> I think.

It is certainly the standard way of doing things in a unix-style
filesystem. It is hard to do better for a real read-write filesytem. But
scanning the index is definatly possible. And at least the Linux
filesystem layer does agressive cacheing of the results of lookup and
readdir operations.

Also, it occured to me this morning that storing the index in sorted
order could partially compensate for the lack of directory contents.
Lookups are binary searches in that case. Reading the contents of (say)
the root directory is still kinda slow, but I think you could make
optimizations even for that case.

-- 
Will Dyson
"Back off man, I'm a scientist!" -Dr. Peter Venkman