How would you do your If and else statements

[Little Bobby Tables]

Personally I don't think anyone on here could explain there whole style because its just so many rules to adapt to.

its not too difficult to explain a style...

[3hM2Nwbp]

I usually just type it out however, and then format my entire project 100% perfectly and accurately within a few seconds by using my IDE's automatic functionality.

This works for all style formats, even custom ones (assuming people use a modern, functional IDE).

[IWishIKnew]

switch(...)
{
    default:
    {
    }
    break;
}

That's how I do my if-else, cause it's way more efficient... (try if/else and switch a million times and see which one's better)

[3hM2Nwbp]

@IWishIKnew -- I wouldn't be so quick to make that claim. Often times, compilers may opt to generate a switch-like structure in the resulting object code from an if/elseif chain. Likewise the same may be said about a compiler generating a simple branch if a switch has few paths. I would suspect that switches are a little bit more tricky to get right with branch-prediction optimizations like simple branches are.

Optimizations are also often times compiler-specific. One compiler might spit out drastically different binary than another.

* Just to be clear, I make no claims at being an optimization guru. I have next to no experience in any assembler level object code other than x86 (that is to say, no experience from technology in the past 10 years).

[Stewbond]

I can't rely on my IDE, if I don't format it as I type, I risk mis-matching a brace and that's a problem I don't have time for.

I use auto-formatting when importing code from someone else (another developer, not an external library).

For switches I like this:

switch(...)
{
default:
    code here
    break;
}

[Little Bobby Tables]

also, switches only work the same as if chains if and only if each operator is ==. if you want to compare using < > <= then you would need to use if blocks. also, you can only use chars, ints, bools, and floats (i think. never tested it)

[EwCjE3v7]

@Little Bobby Tables
No you cannot use floats, only integral values. Well you can use float but it will be turned into a int by taking away the decimal part

[Little Bobby Tables]

hmm i thought as much. never really had a need for it though so i assumed yes

[S G H]

float test = 0.4f;
switch(test > 0.39f && test < 0.41f)
{
    case 1:
        break;
    default:
        break;
}

_{(That's the nearest you can get)}

[Drrockso]

What is preferable this:
for (int i = 0; i < 5; i++)
Or this:

int i;
for (i = 0; i < 5; i++)

I always initialize i within the for statement.

[Little Bobby Tables]

it depends. in the first one the i doesnt exist past the for statement. in the second it has a a longer scope. if you are going to be using it multiple times then the bottom, otherwise the first

[BHX]

Don't forget, with the new standard some prefer this version of the first method:

for (auto i = 0; i < 5; i++)

[Drrockso]

i doesnt exist past the for statement

That is only in some compilers. I was just wondering because my tutor told me that he usually prefers the second one.

[BHX]

That is only in some compilers.

What?! If compilers let you access i outside of the scope of for (int i = 0; i < 5; i++){} then never touch those compilers as it should go out of scope and not be accessible.

For example:

for(int i = 0; i < 5; i++)
{
      std::cout << i << std::endl;
}

std::cout << i << std::endl; // should return a compile error for being out of scope 

g++ -g -pg -std=c++11 -Wall -c "ex2_33.cpp"  (in directory: /home/equinox/project/tests/src)
ex2_33.cpp: In function ‘int main()’:
ex2_33.cpp:22:23: error: ‘i’ was not declared in this scope
  std::cout << "\n" << i << std::endl;
                       ^
Compilation failed.

[cire]

Drrockso wrote:

That is only in some compliant compilers.

Fixed.

Drrockso wrote:

I was just wondering because my tutor told me that he usually prefers the second one.

I would view said tutor's advice with suspicion.

I generally use another form of loop when I must access the counter afterwards. If I have to access the value after the loop to find out what it was, the condition for continuing the loop obviously doesn't just depend on the counter's value.

[IWishIKnew]

Keep in mind: the tutors ae students themselves. You should call into question their credibility if they say somthing questionable. That doesn't mean that the tutor is wrong for prefering the latter example, but the latter example is somwhat..... un-conventional? In any case, I wouldn't declare the for-loop's variable outside of it unless I was doing somthing really special with it.

[Lowest0ne]

Scope should be kept as limited as possible.

Edit: Before C99, you couldn't declare a variable from inside the for loop.

[LB]

Before C99, I believe all variables had to be declared before any statements. Thank goodness that absurd restriction got removed.

[Lowest0ne]

warning: ISO C90 forbids mixed declarations and code [-Wpedantic]

... which is what you're talking about. ( guess I'm not using that :) )