On Thu, 14 Feb 2008, Eric Kow wrote:

>>  I notice that this command doesn't sort or perform any sorting, nubbing or
>>  counting.  Is this intended so that users can perform their own counting?
>
> Yep!  Somebody should update the docs to suggest
> darcs show authors | sort | uniq -c | sort -n -r

I happen to try it before I read this thread, and I came here to report
a bug. It spewed about 2000 lines to the screen with no paging, and
ended like this:

[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]

####

That's not the pleasant, intelligent UI from darcs that I've come to expect.
I think this format should be the default:

  darcs show authors | sort | uniq -c | sort -n -r | less

I don't really see any drawbacks

- Sorting the data adds predictability.
- Reporting the count of unique entries is not lossy. You can reconstruct a 
boring
   format like the one above with the counts, if you wanted to.
- paging shows the most interesting values first

To contrast what darcs presented me with above, the result of sorting
and showing the unique counts is much more interesting:

    1451 David Roundy <[EMAIL PROTECTED]>
     828 David Roundy <[EMAIL PROTECTED]>
     507 [EMAIL PROTECTED]
     304 Ian Lynagh <[EMAIL PROTECTED]>
     285 Mark Stosberg <[EMAIL PROTECTED]>
     267 Tommy Pettersson <[EMAIL PROTECTED]>
     217 Juliusz Chroboczek <[EMAIL PROTECTED]>
     168 Eric Kow <[EMAIL PROTECTED]>
     142 Eric Kow <[EMAIL PROTECTED]>
     101 [EMAIL PROTECTED]

Also, the result is easily parsed between the numbers and the authors if you 
care to.

     Mark

-- 
http://mark.stosberg.com/


_______________________________________________
darcs-devel mailing list
darcs-devel@darcs.net
http://lists.osuosl.org/mailman/listinfo/darcs-devel

Reply via email to