[Digikam-devel] [Bug 296141] Make digikam compile on ARM

[Francesco Riosa]

https://bugs.kde.org/show_bug.cgi?id=296141

--- Comment #9 from Francesco Riosa <francesco+kde at pnpitalia.it> ---
(In reply to comment #8)
> The float precision is important of course, especialy for geolocation data.
> Do you know a way to fix qReal usability problem with DB ?

All the following are guesstimates, take them with a grain of salt:
< cortex-A7 generally don't do floating point (FP)
cortex A7 has VFP extension which should do single and double FP
>= cortex A8 have NEON extensions (A9 optional) which are also suitable to vector processing

So we must assume that ARM _only_support single FP precision. This is baad
because it could taint a database for every other architecture.

Assuming we are not using those values for calculations, they could be stored
the values as strings instead of number, but this has to be done everywere from
EXIV(?) to the insertion in the db.

So really no good ideas here.

Rohan Garg;
   since you have acces to ARM do you have any idea on how a double could be
managed?
What was the actual error the compiler gave you?

Links:
http://en.wikipedia.org/wiki/ARM_architecture
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.faqs/ka9347.html

-- 
You are receiving this mail because:
You are the assignee for the bug.