LITTLE_ENDIAN conflicts

What is the story of LITTLE_ENDIAN?

Recently a change was made in cpu/samd21 and cpu/saml21. The include files from Atmel contain      #define LITTLE_ENDIAN 1 But it was changed to surround it with #ifndef / #endif

Commits: d2e8efc6, 7275f98 Is it too late to say that I don't like them? What problem was being solved by adding the #ifndef/#endif ?

BTW. I'm about to add a whole bunch of samXYZ.h files that all have that same define set. And I'm also planning to update the existing include files.

And this process of updates and additions may happen again in the (near) future when we want to update these Atmel CMSIS files.

Hi Kees, I think there was some problem with this macro in newlib, that’s why we define it only in certain cases. From the usage point it only seems to be used by sys/quad_math/ and the PCI implementation of the x86 port. Both were introduced by René Kijewski, maybe he can shed some light on this (because for his very well used byteorder lib he used __ORDER_LITTLE_ENDIAN__ which does not seem to have these problems).

Cheers, Martine