I'm reading some lecture notes of my C++ lecturer and he wrote the following:
- Use Indentation // OK
- Never rely on operator precedence - Always use parentheses // OK
- Always use a { } block - even for a single line // not OK, why ???
- Const object on left side of comparison // OK
- Use unsigned for variables that are >= 0 // nice trick
- Set Pointer to NULL after deletion - Double delete protection // not bad
The 3rd technique is not clear to me: what would I gain by placing one line in
a { ... }
?
For example, take this weird code:
int j = 0;
for (int i = 0 ; i < 100 ; ++i)
{
if (i % 2 == 0)
{
j++;
}
}
and replace it with:
int j = 0;
for (int i = 0 ; i < 100 ; ++i)
if (i % 2 == 0)
j++;
What's the benefit of using the 1st version?
Let's attempt to also modify i
when we increment j
:
int j = 0;
for (int i = 0 ; i < 100 ; ++i)
if (i % 2 == 0)
j++;
i++;
Oh no! Coming from Python, this looks ok, but in fact it isn't, as it's equivalent to:
int j = 0;
for (int i = 0 ; i < 100 ; ++i)
if (i % 2 == 0)
j++;
i++;
Of course, this is a silly mistake, but one that even an experienced programmer could make.
Another very good reason is pointed out in ta.speot.is's answer.
A third one I can think of is nested if
's:
if (cond1)
if (cond2)
doSomething();
Now, assume you now want to doSomethingElse()
when cond1
is not met (new feature). So:
if (cond1)
if (cond2)
doSomething();
else
doSomethingElse();
which is obviously wrong, since the else
associates with the inner if
.
Edit: Since this is getting some attention, I'll clarify my view. The question I was answering is:
What's the benefit of using the 1st version?
Which I have described. There are some benefits. But, IMO, "always" rules don't always apply. So I don't wholly support
Always use a { } block - even for a single line // not OK, why ???
I'm not saying always use a {}
block. If it's a simple enough condition & behavior, don't. If you suspect someone might come in later & change your code to add functionality, do.