16 Nov
2014
16 Nov
'14
12:36 a.m.
To me it's a matter of readability/semantics.
Even though a BOOL is still an integer, if a variable is clearly meant for
the purpose of a bool, such as "if (bEnableThis)", then using != FALSE is
redundant.
If the variable doesn't have an implicit "boolean-ness" in its name, then
it may be best to write != FALSE to clarify, although in such case it may
be a better choice to rename the variable itself rather than use the
verbose comparison.
For pointers, it's a similar matter: I understand "if (ptrSomething)" to
mean != NULL. But for numbers of any type, be it int, float, or otherwise,
I strongly prefer using != 0.
It may sound silly, but I prefer it that way because in my head "if
(myNumber)" reads as "if there is a number in this variable", and of
course
there is a number, it just happens to be 0, which is still a number. So
although I understand the way it works, it just feels better for me to
compare with != 0 and make my intention clear.
On 15 November 2014 23:22, Timo Kreuzer <timo.kreuzer(a)web.de> wrote:
Am 14.11.2014 15:32, schrieb Love Nystrom:
On 2014-11-14 00.41, Alex Ionescu wrote:
I would much rather see if (f != FALSE) instead of if ( f ).
Well, it's certainly a valid option, and C++ compilers seem to generate
the same "CMP boolRm, 0" in both cases (though the latter could actually
generate the potentially faster and better "OR boolRm, boolRm" instead).
"if (f != FALSE)" and "if (f)" are exactly 100% the same for the
compiler.
"if (f)" is nothing but a synonym for "if (f != 0)" and FALSE is 0.
Whether
it creates a CMP, or an OR or a TEST is all up to the compiler and it will
be the same in both cases. If it was not the same, there is something ...
well not neccessarily "wrong" but at least very strange with the compiler.
I expect some people will use the former.
Personally I much prefer the latter.
I think both have their pros and cons. So I don't really care.
_______________________________________________
Ros-dev mailing list
Ros-dev(a)reactos.org
http://www.reactos.org/mailman/listinfo/ros-dev