On Fri, Jan 13, 2012 at 8:30 PM, Youness Alaoui
<[email protected]> wrote:
> On Fri, Jan 13, 2012 at 10:32 AM, Cedric BAIL <[email protected]> wrote:
>> On Fri, Jan 13, 2012 at 8:12 AM, Youness Alaoui
>> <[email protected]> wrote:
>> > I've had an issue recently when I tried to run my app (using edje) on the
>> > PS3, the aspect ratio of all the images were wrong, and it looked really
>> > bad. I investigated the issue and found out that the aspect_preference
>> was
>> > the cause and that when it's set to  'BOTH' for example, the
>> > desc->aspect.prefer value is 50331648 which is.. 0x3000000 .. so it's a
>> big
>> > endian vs. little endian issue since the EDJE_ASPECT_PREFER_BOTH value in
>> > the enum is '3'.
>> > So I figured the reading of the .edj is wrong, so I looked and it seems
>> to
>> > read it as a 'EET_T_CHAR', but the structure contains the enum as type,
>> > which makes it an int.. so what happens is that it stores 1 byte (the
>> char)
>> > in the 32bit variable.. on little endian, it's fine, it works, but on big
>> > endian, it makes the value huge. So I fixed it by changing the
>> declaration
>> > of he 'aspect.prefer' structure to a char instead of the enum it
>> > represents. I tested and it seems to work and not break anything (and
>> fixes
>> > the bug). However, since this seems a bit sensitive, I thought it's best
>> to
>> > send the patch here to make sure I'm not doing something wrong.
>> > Thanks for reviewing this simple one liner patch :
>> http://pastie.org/3176835
>> > I have noticed other structures do the same thing, 'fill mode' for
>> example
>> > is defined as EET_T_UCHAR in the eet data description and as 'unsigned
>> > char' in the structure, that's why I fixed it this way. Note also that
>> this
>> > shouldn't break the .edj file's compatibility or anything.
>>
>> Good, you figured out why it was broken. The fix sounds fine for me.
>> Did you check that the only enum, we are using directly in one of our
>> saved structure or should I check ?
>>
> Ok cool, I will commit it then. I checked and didn't see any other enum
> being used, but I didn't do an extensive check. I will make sure it was the
> only one and fix any other I might see.

Cool, thanks !
-- 
Cedric BAIL

------------------------------------------------------------------------------
RSA(R) Conference 2012
Mar 27 - Feb 2
Save $400 by Jan. 27
Register now!
http://p.sf.net/sfu/rsa-sfdev2dev2
_______________________________________________
enlightenment-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to