[help-GIFT] address@hidden: Re: Testing GIFT-modifications]

[risc]

Another Gift Hacker. We need more. ;)

I'm just modifying the feature extraction program, so I have a small 
shell script that runs the extractor from upstream, and then diffs
that against the result of 'my' extractor, 40 or so times.

Just to be safe. :)

It then takes the time of the runs, and evens them out, so I know
how much performance I've gained.

Available Upon Request.

Other than that, I don't verify the 'querying' part of the engine
at all. Its written in C++, and i am not a C++ programmer (yet..).

I've religiously avoided C++, as I've seen too many college grads
who think the whole world fits their object model, and who write
*HORRIBLE* performing applications. to say nothing of how they look.

I justify it to myself by knowing I've got leaps and bounds of
improvements to add to the C feature extractor.

After todays commit, CVS should run at 1.42 on my machine,
VS 0.1.14 running at 1.9, and my 'private' branch running at .85!

I'm tired of spending hours waiting on my images to import, how 
about you? ;)

Hope this helps! Feel free to ask for more details!

Julia Longtin <address@hidden>


On Wed, Aug 16, 2006 at 08:13:32AM +0200, Jonas Lindqvist wrote:
> Hi Julia!
> 
> I'm writing to you directly instead of to the help-gift list, since  
> I'm not sure "everyone" is interested...
> 
> I have understood that you are modifying the code to gain  
> performance, right? (I must admit I AM curious to what your  
> approaches are, but that's not what I'm writing for...)
> 
> I myself would like to clean up the code to gain robustness, but I  
> need to find a good way of making sure I haven't broken anything in  
> the process.
> 
> How do you verify that all the vital functions are still in working  
> order after you modify them?
> Are there any automated testing procedures available, written by you  
> or by others? (a "make test" Makefile target would be nice...)
> Or is there a list of tests that should be verified after each  
> modification?
> 
> Wolfgang suggested I'd test "feature extracting and querying", but  
> that's a bit too fuzzy for me. I might miss something important by  
> just testing random stuff.
> 
> Regards,
> 
> Jonas Lindqvist
> 
> 


----- End forwarded message -----