On Mon, 28 Nov 2011 18:21:03 +0200, Max Samukha <max...@spambox.com> wrote:
Could you clarify what is the most important part?
As i tried in the above post, header files are specs, a contract between
library writer and the user.
A dll itself is not a spec, it is the implementation.
If you want to use a library, you would need to read its docs anyway.
The interface files won't be of much help (D's ones will, since dmd
liberally pours implementations into them). If the docs are unavailable
or outdated, you can always extract the declarations from the library:
monop -a -r:lib.dll
lib
Version=0.0.0.0
Culture=neutral
PublicKeyToken=null
public class Lib {
public Lib ();
public static void hello ();
}
Creating separate interface files, taking trouble to push them around
along with the library and then feeding them to the compiler are
needlessly redundant steps. Or what am I missing?
Well if you think that is redundant, you should neither comment nor
document your code,
Because i think they are less important than what header files trying to
solve, they don't even give you any guaranties.
The fact is that apps with tight real-time requirements cannot be
developed in C# for obvious reasons, and compilation models are
irrelevant to the fact.
They are mostly libraries so it is more than relevant.