Fix clang warning due to implicit conversion from int to float for the RAND_MAX macro

[Rossmaxx]

A minor warning, and a minor fix. RAND_MAX is an integer macro, which seems to introduce a warning when divided.

[firewall1110]

It seems You have fixed not all such warnings:

/plugins/Organic/Organic.cpp:239:18:
/plugins/Organic/Organic.cpp:241:18:
/plugins/Vibed/VibratingString.h:110:49:
/plugins/Vibed/VibratingString.h:116:49:
/plugins/Vibed/VibratingString.h:127:50:
/plugins/Vibed/VibratingString.h:136:50:
/plugins/Vibed/VibratingString.cpp:81:48:

floatwarnings.txt

P.S.
I am using clang compiler

[Rossmaxx]

@firewall1110 thanks for the catches. I just noted the one case which was giving me trouble. I'll fix the rest later, but I'll take some time off so can't really say when.

[Rossmaxx]

@firewall1110 can you check again?

[Rossmaxx]

I am using clang compiler

Ahh yes, now i remember. I was using clang for iwyu which is when i saw these warnings. After that, I was on gcc, and thought the warnings disappeared.

[firewall1110]

@firewall1110 can you check again?

After ~5 h - now I about to go to rehearsal with vocal ensemble ...

[firewall1110]

There is no more RAND_MAX related warnings.

P.S.
There are more clang different type warnings ...

Compilation stderr:
err_log.txt

Used build script:

export CC=clang
export CXX=clang++
export CFLAGS="-Oz"
export CXXFLAGS="-Oz"

cmake  .. -DCMAKE_INSTALL_PREFIX=../clang_Oz/

make -j6

Invocation command: bash build_clang_Oz.sh 2>err_log.txt

[Rossmaxx]

@firewall1110 you fine if i get done with just the RAND_MAX warnings. I'll look at the other warnings seperately.

[firewall1110]

i get done with just the RAND_MAX warnings. I'll look at the other warnings seperately.

File in P.S. is only for complete information, all others warnings should not be fixed here.

[Edited:] You respect me too much (Thank You for this, but ... )

[Rossmaxx]

@sakertooth @messmerd can i get one approval and get this merged real quick?

[messmerd]

I wonder if it would be worth adding a range version for rand like @rubiefawn did for fastRand:

lmms/include/lmms_math.h

Lines 88 to 99 in 3417dfe

	template<std::floating_point T>
	inline auto fastRand(T range) noexcept
	{
	constexpr T FAST_RAND_RATIO = static_cast<T>(1.0 / 32767);
	return fastRand() * range * FAST_RAND_RATIO;
	}
	template<std::floating_point T>
	inline auto fastRand(T from, T to) noexcept
	{
	return from + fastRand(to - from);
	}

[Rossmaxx]

That might make sense too, but I am focusing more on #7677 rn so this has to wait a bit, if i'm refactoring

[rubiefawn]

There's a lot of places in the code where rand()/std::rand() is used in this way, which we may want to change to just use fastRand(). I would argue that belongs in a separate PR since it is beyond the scope of "fix compiler warnings".

Edit: See #7741