[Urwid] Optimize urwid - cutable benchmarks

[Ian Ward]

Rebecca Breu wrote:
> Zitat von Ian Ward <ian at excess.org>:
> 
>> .. or is measuring relative speed something only marketing folks do?
> 
> Ah, you were talking of the improvement of speed. Mmh, at least 22% sounds
> better than 18% !  :)

exactly!

> The attached c-module implements all functions from Ticket #2 except for
> calc_text_pos. The byte encoding is hard coded to "utf8" at the moment.

I've created a branch for your c-module:
   svn co https://excess.org/svn/urwid/urwid/branches/rbreu-cmodule

I haven't modified setup.py to do the compiling yet, so you have to:
   cd source
   make
   mv str_util.so ../urwid
for the moment.

> Happy benchmarking!

I've added a UTF-8 heavy benchmark bench_2.py to the contrib directory. 
  Here are my results for both benchmarks:

time (s)        bench_1   bench_2
       0.9.7.1:  4.367906  42.393004
       0.9.7.2:  1.591115  22.192668
rbreu-cmodule:  1.433236  14.414327

Those are 11% and 54% speed improvements on top of the improvements made 
in 0.9.7.2.  Very nice.

Ian