X

Aggregator

CodeSOD: Would a Function by Any Other Name Still be WTF?

The Daily WTF
1 month ago

"Don't use exception handling for normal flow control," is generally good advice. But Andy's lead had a PhD in computer science, and with that kind of education, wasn't about to let good advice or best practices tell them what to do. That's why, when they needed to validate inputs, they wrote code C# like this:

public static bool IsDecimal(string theValue) { try { Convert.ToDouble(theValue); return true; } catch { return false; } }

They attempt to convert, and if they succeed, great, return true. If they fail, an exception gets caught, and they return false. What could be simpler?

Well, using the built in TryParse function would be simpler. Despite its name, actually avoids throwing an exception, even internally, because exceptions are expensive in .NET. And it is already implemented, so you don't have to do this.

Also, Decimal is a type in C#- a 16-byte floating point value. Now, I know they didn't actually mean Decimal, just "a value with 0 or more digits behind the decimal point", but pedantry is the root of clarity, and the naming convention makes this bad code unclear about its intent and purpose. Per the docs there are Single and Double values which can't be represented as Decimal and trigger an OverflowException. And conversely, Decimal loses precision if converted to Double. This means a value that would be represented as Decimal might not pass this function, and a value that can't be represented as Decimal might, and none of this actually matters but the name of the function is bad.

[Advertisement] ProGet’s got you covered with security and access controls on your NuGet feeds. Learn more.
Remy Porter