This is a mirror of official site: http://jasper-net.blogspot.com/

Danger, Will Robinson!

| Friday, March 4, 2011
As long-time readers of this blog know, I am often asked why a particular hunk of bad-smelling code does not produce a compiler warning.

"Why not?" questions are inherently hard to answer because they turn causation on its head; normally we ask what caused a particular thing to happen, not what caused a particular thing to not happen. Therefore, rather than attack that question directly I like to rephrase the question into questions about the proposed feature. (A warning is of course a feature like any other.) By answering these questions we can see whether the compiler team is likely to spend its limited budget on the feature rather than prioritizing something else. Some questions I think about when considering "why did you not implement this warning?" are:

Did someone think of it?

If no one on the compiler team ever thinks of the possibility of a particular warning then obviously it won't happen. As an example of that, I was recently asked why:

using(null) statement;

does not produce a warning. That code is legal; it is equivalent to

IDisposable temp = null;
try
{
 statement;
}
finally
{
 if (temp != null) temp.Dispose();
}

Of course the code is completely pointless; all it does is introduce a try-finally block that adds nothing useful to the program. Why don't we give a warning for this useless code? Well, the primary reason is that (to my knowledge) no one on the compiler team ever thought that a customer would do such a thing, and therefore, never thought to design, implement, test and document a compiler feature to deal with the situation.

Read more: Fabulous Adventures In Coding

Posted via email from Jasper-net

0 comments: