My two cents: Most variables in EMC correspond to something physical. It might be time, length, velocity, force, etc. There are cases where 32 bits (+/- 2 billion) isn't enough resolution or range, those were specified to be long long or long double.
The physical reality doesn't change when the CPU becomes 64 bit. If 32 bits was enough to represent the value before, it is still enough. If 64 bits was enough before, it is still enough. That is the main reason I prefer types with explicit lengths, like u64 or s32. Code using those types will work the way the author intended regardless of what later generations of CPU and/or compiler do. John Kasunich On Thu, Apr 24, 2014, at 06:10 AM, EBo wrote: > Turns out there are some places where where they specify "long double" > in libnml. Should I convert that to a float64, or a float128 (which it > is on 64 bit machines)? Keeping with the 32bit architecture, defining > double as float32, and long float64 would probably make things > consistent between all the different typs of machines... > > What do people think? > > EBo -- ------------------------------------------------------------------------------ "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE Instantly run your Selenium tests across 300+ browser/OS combos. Get unparalleled scalability from the best Selenium testing platform available. Simple to use. Nothing to install. Get started now for free." http://p.sf.net/sfu/SauceLabs _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
