Don't guess CHAR_BIT

[thetic]

This change requires use-cases without the standard library to provide their appropriate byte size. Given the commitment to maintain support for use without standard C, I think we need to be more prudent about the assumptions we're making about those environments.

[coveralls]

Coverage remained the same at 99.625% when pulling 8c7e75e on thetic:byte-sized into 775e760 on cpputest:master.

[basvodde]

Interesting. Before merging, why did you chose to remove the CPPUTEST_CHAR_BIT? The same could be accomplished by adding the #error to the #else in de #if STDCLIB ? Just curious.

[thetic]

I didn't see much value in the indirection. When the std C library isn't used, we don't have to worry about collisions. A quick scan of the Linux kernel shows some use of (a custom-defined) CHAR_BIT already, so my hope was to avoid impact on such use cases.

[basvodde]

Ok, for me, it feels as a design inconsistency. So far, we've decided to e.g. keep the StdC functions out of the CppUTest code by wrapping them. Alternatively, we could have provided the missing standard C functions when they were not there. There we chose for explicitly wrapping them rather than replacing them.

This one means we are now creating a dependency on the StdC interface in the CppUtest code. It is not a big problem (and I do intent do merge the PR) but it does feel a bit inconsistent with the general design.

[thetic]

Ok. I'll restore use of CPPUTEST_CHAR_BIT, but drop the old "guess."

[basvodde]

Ok. I would have merged it :) Since you've been doing a lot of contributions, I'll try to explain more often the reasoning on why things are the way they are.

[thetic]

Thanks! I don't mind doing rework to keep things consistent.