[darcs-users] Converting MoinMoin to Markdown/RST

Trent W. Buck trentbuck at gmail.com
Sat Mar 7 22:51:00 UTC 2009


"Trent W. Buck" <trentbuck at gmail.com> writes:

> On Sat, Mar 07, 2009 at 07:06:58PM +0100, Petr Rockai wrote:
>> Hi,
>> 
>> "Trent W. Buck" <trentbuck at gmail.com> writes:
>> > The resulting repo is 1.2GiB in size!  This sucks, because the
>> > original Moin data directory is only 335MiB.
>> That's still an awful lot. Can we filter out the useless junk? (There's bound
>> to be about 330M of it). Also, how big is the working directory of the repo?
>> As it is, this is going to be pretty useless with darcs, which already can't
>> reasonably handle repos an order of magnitude smaller...
>
> In MiB (rounding up),
>
>     1186    _darcs
>     1153    _darcs/inventories
>     27      _darcs/patches
>     11      DarcsTwo
>     6       _darcs/pristine.hashed
>
> I don't know why inventories should be so large.

OK, so this is better:

    $ darcs get --complete $PWD $PWD.copy
    Copying patches, to get lazy repository hit ctrl-C...
    Finished getting.
    $ du -mx $PWD.copy | sort -nr
    45      /var/tmp/DarcsWiki.darcs.copy
    33      /var/tmp/DarcsWiki.darcs.copy/_darcs
    27      /var/tmp/DarcsWiki.darcs.copy/_darcs/patches
    11      /var/tmp/DarcsWiki.darcs.copy/DarcsTwo
    5       /var/tmp/DarcsWiki.darcs.copy/_darcs/pristine.hashed

So my next question is: why does this particular process generate a
*gigabyte* of unnecessary inventories, and why didn't "darcs optimize"
clean them up?



More information about the darcs-users mailing list