[poppler] General interest: regression testing techniques for pdfs?

[Christian Feuersaenger]

Hi Albert,

I see... perhaps I should start creating a testing repository, since 
your approach appears to work pretty well. Thank you for sharing your 
experience and for the imagemagick calls!

Best regards

Christian

Am 29.10.2010 20:08, schrieb Albert Astals Cid:
> A Divendres, 29 d'octubre de 2010, Christian Feuersaenger va escriure:
>    
>> Dear Developers,
>>
>> I am author of a program which generates pdf output and as such I am
>> thinking about improvements to my regression tests... perhaps I can
>> learn from your very successful testing framework.
>>      
> I wouldn't call it "very successful" nor "framework" as it's a 10 line bash
> script :D
>
>    
>> Is there some sort of automatic system which compares screenshots
>> somehow?
>>      
> diff a.png b.png
>
> If they are equal, good no change happened! If they are different then
>    convert a.png b.png -compose difference -composite -colorspace gray miff:- |
> display
> will give you an image with the differences, and
>    convert a.png b.png -compose difference -composite -colorspace gray miff:- |
> identify -verbose -
> will give you some interesting data and you can have look at the standard
> deviation for example to know how much the rendering has changed
>
> Albert
>
>    
>> Or any other type of automatic comparison function? In that
>> case, I would be interested in using a similar approach for my own
>> work... and I would appreciate any hints.
>>
>> Thanks!
>>
>> Best regards
>>
>> Christian
>> _______________________________________________
>> poppler mailing list
>> poppler at lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/poppler
>>      
> _______________________________________________
> poppler mailing list
> poppler at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/poppler
>