[darcs-users] darcs patch: Complain if the user just runs darcs-ben... (and 2 more)

Max Battcher me at worldmaker.net
Sun Jan 24 23:30:15 UTC 2010


Ah-ha! Benchmarks in Windows.  One of these patches was one I was 
contemplating writing myself:

On 1/24/2010 12:49, Eric Kow wrote:
> Sun Jan 24 17:25:25 GMT 2010  Eric Kow<E.Y.Kow at brighton.ac.uk>
>    * Create temporary darcs-stats file in temporary directory.

With this patch (and a minor tweak detailed below) I now see actual 
error-free output from darcs-benchmark (ignore the numbers for now; 
background processing on a busy machine):

darcs-2.3.1 get (full) [darcs]: 1!..2!..                   88.9s, 10.0M
darcs-2.3.1 get (lazy, x10) [darcs]: 1!..2!..              137.0s, 3.0M
darcs-2.3.1 pull 100 [darcs]: 1!..2!..                     15.9s, 17.0M
darcs-2.3.1 annotate [darcs]: 1!..2...                     45.9s, 196.0M
darcs-2.3.1 wh x50 [darcs]: 1...2...                       7.2s, 0.0M
darcs-2.3.1 wh mod x50 [darcs]: 1...2...                   11.8s, 3.0M

With the patch alone I was getting:

darcs-2.3.1 get (full) [darcs]: 1!..2!..                   51.7s, 
darcs-benchmar
k: Prelude.read: no parse

To get the working output, I commented out the line:

                    -- removeFile stats_f

My knowledge of Haskell File I/O doesn't provide me enough of an insight 
into why that was the fix I needed. The decision to comment out the 
removeFile was a gut decision. (I had noticed that the temp files 
weren't being removed anyway, and figured I can delete them manually for 
now anyway.)

So yeah, using multiple temporary files for darcs-stats does seem to be 
the answer to the problems on Windows.

I'll give my system some time to focus on the benchmarks for a full run 
and report back the results when that finishes.

--
--Max Battcher--
http://worldmaker.net


More information about the darcs-users mailing list