[tahoe-dev] "tahoe stats" command

Andrej Falout andrej at falout.org
Wed Feb 11 00:06:01 UTC 2009


Hello Brian,

> Is it possible that there were only 4247 unique files in that directory (and
> 965 duplicates)? 'tahoe backup' counts files on disk, but 'tahoe stats'
> removes duplicate filecaps before counting, so if you have two copies of the
> same data (even under different names), they will only be counted once.

Which is a great feature to have, but for an end user hard to a)
verify and b) use as information. Especially since there is no
"count-files-duplicates" field.

OK, I have 991 duplicates of 850 files, consuming 678.13 mb, according
to DupFinder and to my embarrassment. Still, does not match.

Maybe when displaying information, a human-readable terminology would
be useful. Like "Files" means files that user will see on his hard
disk after restore, and maybe an entry "Duplicate files" or
"Duplicates found" can be added.

> The size-immutable-files has the same property: it only counts the sizes of
> unique files.

Them can you add "size-files" or "size-restored-files" or similar?

> Incidentally, "literal files", in tahoe, refers to files that are small
> enough to contain entirely within the filecap. I think the cutoff is 55
> bytes. Any file that is larger than that will be included under "immutable
> files".

Consider a poor user with a job, trying to backup bunch of files to a
safe place. I do appreciate the developers need for "literal files",
but quite honestly, everybody else will wish they dont have to know
about them.

I would even suggest that 'stats' should print out only information an
(average) user expects: number of files, directories, sizes - as seen
on his file system. Instantly recognizable and verifiable.

The rest can become visible with 'stats --verbose'

-- 
Andrej Falout



More information about the tahoe-dev mailing list