"float64 size isn't 8 bytes on this platform", can I ignore?

[Kazade]

I'm trying to compile a library of mine (which uses GLM) for the Sega Dreamcast, and because I'm compiling to an SH4 CPU, I hit this error here:

type_float.hpp:62:2: error: static assertion failed: float64 size isn't 8 bytes on this platform

My question is, am I OK to ignore this / comment out the assertion? Does GLM rely on this being 8 bytes elsewhere? I should be OK if I just don't use float64 right?

Thanks!

[Groovounet]

Hi
What compiler are you using? Does it mean there is no double on your platform? I never really consider this case... It's probably not a big deal but I guess I'll like to figure something smoother for this platform.

[Kazade]

The compiler is GCC 4.7.3 although it's patched for the Dreamcast. I'm guessing that double is the same size as float on this platform.

[Groovounet]

Can you confirm that sizeof(double) is 4 on your Dreamcast?

[Kazade]

Yes confirmed. Double and float are the same size - 4 bytes (it's compiled with --m4-single-only)

[Groovounet]

This has been implement using the define GLM_FORCE_SINGLE_ONLY.

When defined before including GLM, all requirements for 64-bits float is released.

This feature is implemented in master branch for GLM 0.9.9 release.

Thanks for reporting,
Christophe