Why don't Microsoft implement some sort of "typeof" for properties, fields,
methods etc?

For example, lets suppose we have "propertyof" keyword, then we can do the
following:

PropertyInfo pi = propertyof(MyClass.MyProperty)

Is there any problems that prevents Microsoft from implementing this?

I would appreciate any thoughts.
Thanks in advance.

In article <FAB852C2-A5E8-48A5-807A-9F14DCC8B***@microsoft.com>,
A***@discussions.microsoft.com says...
What are you trying to accomplish?  Do you want to get a PropertyInfo
without using reflection?

I'll try to explain. Letts suppose we have the following method to fetch data
from the database:

BizEntity[] e = bll.ExecuteQuery("type of the business object", " property
name of the business object", ....);

It takes as first parameter name of the business logic entity which objects
I want to get, and other parameters are names of the properties I want to be
loaded of the fetched objects (suppose we have a delay loaded business
objects).
I can rewrite this snippet:

BizEntity[] e = bll.ExecuteQuery(typeof(BizCustomer), " property name of the
BizCustomer", ....);

Now if for some reason business entity "BizCustomer" will be deleted or
renamed the error will be show at compile time. This is really good ability,
espessially for the large systems with lots of business objects.
Now suppose we have "propertyof" keyword, then the snippet will be like this:

BizEntity[] e = bll.ExecuteQuery(typeof(BizCustomer),
propertyof(BizCustomer.Name), ....);

And the compiler will show us an error even if the property "Name" of the
"BizCustomer" object will be renamed or deleted.

As you see I cannot use reflection because I will need to use string names
of the properties anyway so it will not help.

And I cannot see why Microsoft havn't provided such possibility.

The suggestion has come up before, and I believe the C# team has said
that it's on their list of things to consider for future versions. I
can certainly see some challenges though. For example, you need to
find a way to deal with overload resolution for methods and indexers.
I'm sure they could implement it, but I don't know if the feature is
asked for often enough to justify it.

Personally I'd settle for an operator that "string-ifies" (i.e.
compiles down to a string literal) an identifier while still verifying
that the identifier is valid at compile time. It would be very useful
when throwing ArgumentExceptions and the like.


Mattias

Document mangment in .net
Debugger Already attached, help
Global Semaphore problems
Scientific Notation when serializing doubles
Large text file - in memory ( > 60mb)