I agree with the requirement for the parenthesis, simply because it helps separate the various conditions that are being tested for.  Even looking at the code without braces, you should be questioning exactly what is going on in there since you generally don't do assignments inside if statements.  The fact that the compiler failed to catch it because of the parenthesis doesn't excuse a programmer from not seeing the problem.  I personally would never assume that a dev intended to do an assignment just because he or she put parenthesis around such a statement, I would automatically assume they messed up typing and left off a = in the check.

2009/2/16 Timo Kreuzer <timo.kreuzer@web.de>
I agree with Dmitry on that matter. The brackets are bad. They only hide possible bugs. I don't see any advantage. They don't add anything for the compiler nor do they add anything to readability. IMO it's easier to read with less braces.

Timo

Aleksey Bragin schrieb:
Good, let's discuss it developers-wide then, and add a result into  
the (to become official) Coding Style Guidelines for kernel-mode code.
The only thing I'm bothered with is to keep the style consistent  
across kernel (and drivers, where possible). And right now bracing  
arguments of && and || prevails.

I'm eager to listen to other dev's opinion on that topic. You just  
showed one disadvantage of such method. Are there any advantages?


WBR,
Aleksey.

On Feb 16, 2009, at 8:47 PM, Dmitry Gorbachev wrote:

  
No, it's just a matter of coding style used in the kernel.
      
What is a logical basis for adopting this particular rule?  
Sometimes it is
not just a matter of preference.

These parentheses are unnecessary for humans, and if they are not  
needed for
compilers or some other tools, they are bad! Without them, the bug  
could be
noticed earlier.

Cheers,
Dmitry
    
_______________________________________________
Ros-dev mailing list
Ros-dev@reactos.org
http://www.reactos.org/mailman/listinfo/ros-dev

  


_______________________________________________
Ros-dev mailing list
Ros-dev@reactos.org
http://www.reactos.org/mailman/listinfo/ros-dev