Several float to int potential rounding errors/inconsistencies

[Lubomir I. Ivanov]

On 11 March 2017 at 07:21, Jérémie Guichard <djeBrest at gmail.com> wrote:
> Hey guys,
>
> I've added the flag conditionally as discussed in previous mail, and updated
> the pull request.

i've tested the changes locally on Windows and it builds fine.
llrint() is also available.

4.9.0 seems to be the version they've added "float-conversation", so
this is correct:

+ if (NOT CMAKE_CXX_COMPILER_VERSION VERSION_LESS "4.9.0")
+ set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wfloat-conversion")
+ endif()

of course, there is no need of lrint() in a lot of locations and
rint() can be used instead, like here:

- sample->depth.mm = depth * FEET * 1000;
+ sample->depth.mm = lrint(depth * FEET * 1000);

"mm" is of type int.

here, t1 is of type int as well:
+ int t1 = lrint(max_d / slope);

but that's not a big issue as long as numbers are not truncated for
64bit builds.
our main Windows build is 32bit, so this is pretty much in effect:
sizeof(long) == sizeof(int).

i haven't tested if the changes break functionality in some way, hopefully not.

lubomir
--