I found this article about switch statements pretty interesting:
http://741mhz.com/switch/
The author goes into detail about how the C switch statement ends up as assembly on a couple of implementations (Clang and GCC) and a variety of situations (e.g. Different numbers of cases, continuous values vs. sparse values, default case vs. non-default case).
http://741mhz.com/switch/
The author goes into detail about how the C switch statement ends up as assembly on a couple of implementations (Clang and GCC) and a variety of situations (e.g. Different numbers of cases, continuous values vs. sparse values, default case vs. non-default case).
Very interesting article. Goes to show that you should never just assume that your mental model of how your code will be translated to assembly/machine code is correct.
Totally tangentially, one of the features that I enjoy in D (the programming language) is "final switch" - http://dlang.org/statement.html#FinalSwitchStatement
When used with D's enumerated type, all possible values must exist as case statements, and no default case is allowed. That way if you extend an enumerated type with a new value, whenever you have used "final switch", the compiler will tell you where you need to handle this value.
When used with D's enumerated type, all possible values must exist as case statements, and no default case is allowed. That way if you extend an enumerated type with a new value, whenever you have used "final switch", the compiler will tell you where you need to handle this value.
As a sidenote, compilers like LLVM also allow you do to this - if you use an enum for the switch, and omit a default case, it will warn you if not all enum values are handled.
- Casey
- Casey