Re: Hm. ItemFactory vs. gtk2-perl 0.90 ? :)

[Ross McFarland <rwmcfa1 neces com>]

On Tue, 2003-07-22 at 15:30, Chas Owens wrote:
> I am just the sort of person who has to run a benchmark.
>
>          Rate brett   joe  ross
> brett 21427/s    --  -13%  -41%
> joe   24594/s   15%    --  -32%
> ross  36232/s   69%   47%    --
>          Rate brett   joe  ross
> brett 18325/s    --   -3%  -41%
> joe   18976/s    4%    --  -39%
> ross  31211/s   70%   64%    --
>          Rate brett   joe  ross
> brett 26376/s    --  -17%  -52%
> joe   31963/s   21%    --  -42%
> ross  55493/s  110%   74%    --
>          Rate   joe brett  ross
> joe   10175/s    --  -18%  -51%
> brett 12375/s   22%    --  -40%
> ross  20613/s  103%   67%    --
>
> It looks like the simple if based switch cleans the other methods clocks.  That
> being said,  Joe's makes it easy to add new keys and to alias keys (type and
> item_type can both point to $item), but the keys for this function are already
> written in stone (the Gtk API).  I can't see a reason not to use the fastest
> function in this case (since there will be little to no maintenance).

funny you should post this, muppet and myself were discussing this very
issue a few hours ago and had pretty much the same opinions. while the
cool perl magic is just that, cool and you get plenty of geek points,
fancy things normally are much slower than their straightforward counter
parts. that being said the easy addition of new keys (either of the
other two, although it's really trivial regardless) would be a benefit
in many situations, but not so much things one where things are pretty
static. for now what's there isn't broken, and as you said is the
fastest so it will probably stay. 

people are more than welcome to scour the code and try to find places
where they can think of better methods of implementation, at the very
least your stuff will get considered.

-rm