ext2_types.h in package e2fslibs-dev is whacked

Phillip Susi psusi at cfl.rr.com
Sat Jan 7 17:59:53 GMT 2006

Sorry, I was up all night and a bit cranky when I wrote that ;)

Now that I'm well rested, I see clearly what the problem is... both long 
and long long are 64 bits on amd64.  The .h.in tests for both to be 8, 
but tests long before long long and goes with the first one it finds.  
This conflicts with the long long definition used in asm/types.h.  This 
should be easy to fix by simply reversing the order of the two checks so 
that long long gets priority.  I guess I should make the fix in the 
package and upload the debdiff to a malone bug. 

Colin Watson wrote:

>Firstly, please relax and turn your flame switch off before posting to
>this list.
>Secondly, ext2_types.h is generated from ext2_types.h.in, and the
>relevant part looks like this:
>#if (@SIZEOF_INT@ == 2)
>typedef int             __s16;
>typedef unsigned int    __u16;
>The header won't say "4" there on all architectures, and that #if branch
>is never relevant to any Ubuntu architectures.

More information about the ubuntu-devel mailing list