[KPhotoAlbum] More patches

[Robert Krawitz]

On Sat, 15 Aug 2015 11:06:22 -0400, Robert Krawitz wrote:
> On Sat, 15 Aug 2015 16:47:36 +0200, Johannes Zarl-Zierl wrote:
>> ...resending because I was sending from the wrong address before..
>> On Sunday 09 August 2015 14:11:01 Robert Krawitz wrote:
>>> The logic is still not right.  This structuring, in particular, is
>>> incorrect:
>>
>> Yes, it is:
>>
>> for(datum : data)
>> {
>>   if (datum.key == *.LensModel)
>>   {
>>     value = datum.value;
>>     break;
>>   }
>>   if (datum.key == *.LensIDNumber)
>>     value = datum.value;
>>   if (datum.key == *.LensType && value.isEmpty())
>>     value = datum.value;
>> }
>>
>> Ordering A:
>> Exif.LensType = "lt"
>> Exif.LensIDNumber = "id"
>> Exif.LensModel = "lm"
>
> ...
>
> It's more confusing that way, and if we later need to add more
> heuristics, it's going to make the code more complicated.  Even with
> possible multiple loops, kph uses something like 1% CPU time; it's
> spending essentially all of its time waiting for I/O.
>
>>> Also, the canonHack thing isn't the right way to go about it; other
>>> cameras may return bogus values.  Those need to be detected and
>>> ignored.
>>
>> How do you propose we detect arbitrary bogus values?
>
> The most obvious to me is that if LensType looks like an integer with
> no translation, that we keep looking for Lens and LensModel.  Having
> something meaningful is better than not.  Certainly, however, -1 looks
> like a default/error value; it's never going to be translatable.

Have you had a chance to think about these suggestions?
-- 
Robert Krawitz                                     <rlk at alum.mit.edu>

***  MIT Engineers   A Proud Tradition   http://mitathletics.com  ***
Member of the League for Programming Freedom  --  http://ProgFree.org
Project lead for Gutenprint   --    http://gimp-print.sourceforge.net

"Linux doesn't dictate how I work, I dictate how Linux works."
--Eric Crampton