[darcs-users] Easily managing over 100, 000 lines of code with darcs

[Simon Michael]

On 3/30/10 7:17 AM, Mark Stosberg wrote:
> I thought I'd share the metrics on our largest project we manage with
> darcs:
>
>   100,000 lines of Perl code
>   127 Megs of disk storage
>   3983 files
>   5 regular committers
>
> Tools to compute the metrics
> ( I exclude the _darcs directory and third-party libraries in our tree )
> - sloccount project
> - find project -type f | wc -l
> - du -hc project

Hi Mark, thanks for sharing those. Here's a script I just made to get similar stats:

#!/bin/sh
# darcsstat - extract information useful for user support.
# Run in a repo's top-level directory. Requires unix.
darcs show repo
echo "     Repo size:" `du -sh _darcs | cut -f1`
echo "    Repo files:" `darcs show files | wc -l`
echo "Working copy files:" `find . -name _darcs -prune -o -name '*' | wc -l`
echo "          Tags:" `darcs show tags | wc -l`
echo "       Authors:" `darcs show authors | wc -l`

(http://gist.github.com/359114)

Amongst my repos, largest patch count is 2k, largest file count is 3k, largest repo is 30M.

The darcs repo has 524 files totalling 45M in 8k patches from 177 authors over 8 years.

The ghc repo has 1274 files totalling 11M in 22k patches from 186 authors over 14 years (converted from previous vcs), 
and darcs2.4 get --no-cache --lazy fetched it for me in under 3 minutes.

I'd love to see more such stats built in to darcs show repo, making troubleshooting and comparing easier.

Cheers - Simon