Why do I listen to my sleep coding self?
-
In bed last night, I dreamt a wonderful improvement to the design of some software I'm working on. It would clean it up, move the code into it's appropriate classes, make it all beautiful. So I have just coded it. Or at least, I coded 3/4 of it. That was the point when I thought "Hang on a moment, you can't have
abstract staticmethods in C#". :doh: Bugger. Note to self: don't listen to me when I'm sleepy. Anyone else done this? Constructed a massive edifice of beautiful code in your mind, only to find when you have got sufficient caffeine in your system that it has a fatal, fundamental, flaw?Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952) Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
Usually the code I dream about is good but occasionally I waste time on a bomb, depends what I had been drinking the night before? :)
If first you don't succeed, hide all evidence you ever tried!
-
You're thinking too much about the C++ style implementation. There's actually no reason why you can't have vtable entries for static methods, though, even using that model – they'd just get passed nothing for the instance parameter. You just have to choose a different way of determining the type that should get dispatched, instead of using the instance's dynamic type, for example the declaration type of the method which was originally called. Delphi goes some way towards this but it doesn't really work well because it doesn't sort out the dispatch properly.
BobJanova wrote:
There's actually no reason why you can't have vtable entries for static methods
Ok fair enough, obviously you can put anything in there, including random bytes, but that then brings us to:
BobJanova wrote:
You just have to choose a different way of determining the type that should get dispatched, instead of using the instance's dynamic type,
I don't really see any options there, but that might be because I don't understand exactly what you mean by this:
BobJanova wrote:
the declaration type of the method which was originally called.
-
Usually the code I dream about is good but occasionally I waste time on a bomb, depends what I had been drinking the night before? :)
If first you don't succeed, hide all evidence you ever tried!
-
OriginalGriff wrote:
That means I don't get "fixated" on "this must be why" and my subconscious is free to "bing" me and say "Try this, not that".
And that's how I defend smoking a pack a day :laugh:
The console is a black place
That's what I used to do - take a cigarette break every hour. When I gave up, I realized I wasn't getting as much work done, because I wasn't taking a break! :laugh: So rather than take up smoking again (and having seen the price they are charging for them these days, I can think of much, much better things to do with the money) I take a break without the cancer-stick.
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952) Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
-
Quote:
I waste time on a bomb
Perhaps this is an indication of a flatulence problem in your sleep? :laugh:
Kenneth Haugland wrote:
Perhaps this is an indication of a flatulence problem in your sleep?
When I can't sleep I don't count sheep I count... :)
If first you don't succeed, hide all evidence you ever tried!
-
BobJanova wrote:
There's actually no reason why you can't have vtable entries for static methods
Ok fair enough, obviously you can put anything in there, including random bytes, but that then brings us to:
BobJanova wrote:
You just have to choose a different way of determining the type that should get dispatched, instead of using the instance's dynamic type,
I don't really see any options there, but that might be because I don't understand exactly what you mean by this:
BobJanova wrote:
the declaration type of the method which was originally called.
Ok so let's have an example.
public static class BaseService {
public static void LoadData() {
Console.WriteLine(GetDataSource());
}protected virtual static string GetDataSource();
}public static class FileService : BaseService {
protected override static string GetDataSource() { return "file.dat"; }
}public static class DatabaseService : BaseService {
protected override static string GetDataSource() { return "database"; }
}... and calling code
void Run() {
DatabaseService.LoadData();
FileService.LoadData();
}This would print "database" and "file.dat", because the class used for the vtable lookups would be taken from the data type used for the call (i.e. DatabaseService.LoadData). That would be resolved at compile time and based on the static declaration type used in the call, not dynamic object type lookups as with instance dispatch. A related good idea is the ability to pass service classes around and dispatch off them:
void Run(Class<BaseService> service) {
service.LoadData();
}... called like Run(DatabaseService) which would bake in the type for dispatch.
-
Ok so let's have an example.
public static class BaseService {
public static void LoadData() {
Console.WriteLine(GetDataSource());
}protected virtual static string GetDataSource();
}public static class FileService : BaseService {
protected override static string GetDataSource() { return "file.dat"; }
}public static class DatabaseService : BaseService {
protected override static string GetDataSource() { return "database"; }
}... and calling code
void Run() {
DatabaseService.LoadData();
FileService.LoadData();
}This would print "database" and "file.dat", because the class used for the vtable lookups would be taken from the data type used for the call (i.e. DatabaseService.LoadData). That would be resolved at compile time and based on the static declaration type used in the call, not dynamic object type lookups as with instance dispatch. A related good idea is the ability to pass service classes around and dispatch off them:
void Run(Class<BaseService> service) {
service.LoadData();
}... called like Run(DatabaseService) which would bake in the type for dispatch.
-
:laugh: I work for myself, so I give myself special permission to "goof off". :laugh:
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952) Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
Lucky you! Allowed to give yourself some special permissions :laugh:
-
Lucky you! Allowed to give yourself some special permissions :laugh:
The down side is the Christmas party is a little lonely... ;)
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952) Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
-
The down side is the Christmas party is a little lonely... ;)
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952) Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
Started working here the week after they had their Christmas party so I can't give you any feedback on that one, yet! :-D
-
0 times. I've wished that interfaces could contain static methods so that T.someFunction could exist for generics (constrained to that interface), but that's a completely separate issue. (that can't work in C# of course) How is "static virtual" even a reasonable concept? No instance = no vptr = no dynamic dispatch.
harold aptroot wrote:
no vptr
Don't get hung up on how current languages are implemented. I too would like a language that allows static members in interfaces and abstracts. If a Bright compiler developer wanted to make this happen, I'm sure he could, it would likely have to be very different from current implementations. (It's way beyond my skills of course.)
You'll never get very far if all you do is follow instructions.
-
In bed last night, I dreamt a wonderful improvement to the design of some software I'm working on. It would clean it up, move the code into it's appropriate classes, make it all beautiful. So I have just coded it. Or at least, I coded 3/4 of it. That was the point when I thought "Hang on a moment, you can't have
abstract staticmethods in C#". :doh: Bugger. Note to self: don't listen to me when I'm sleepy. Anyone else done this? Constructed a massive edifice of beautiful code in your mind, only to find when you have got sufficient caffeine in your system that it has a fatal, fundamental, flaw?Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952) Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
OriginalGriff wrote:
abstract static
OriginalGriff wrote:
Anyone else done this?
Yes, for exactly the same thing. Was such a perfect solution to the problem. Except that it wasn't. Obviously. I even still have the code in place, safely commented out, just to remind me.
cheers Chris Maunder
-
In bed last night, I dreamt a wonderful improvement to the design of some software I'm working on. It would clean it up, move the code into it's appropriate classes, make it all beautiful. So I have just coded it. Or at least, I coded 3/4 of it. That was the point when I thought "Hang on a moment, you can't have
abstract staticmethods in C#". :doh: Bugger. Note to self: don't listen to me when I'm sleepy. Anyone else done this? Constructed a massive edifice of beautiful code in your mind, only to find when you have got sufficient caffeine in your system that it has a fatal, fundamental, flaw?Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952) Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
Make it more dynamic. Instead of static abstract, make it a static property of a delegate type.
-
Delphi allowed virtual static methods... so I think it probably allowed abstract static methods too.
-
In bed last night, I dreamt a wonderful improvement to the design of some software I'm working on. It would clean it up, move the code into it's appropriate classes, make it all beautiful. So I have just coded it. Or at least, I coded 3/4 of it. That was the point when I thought "Hang on a moment, you can't have
abstract staticmethods in C#". :doh: Bugger. Note to self: don't listen to me when I'm sleepy. Anyone else done this? Constructed a massive edifice of beautiful code in your mind, only to find when you have got sufficient caffeine in your system that it has a fatal, fundamental, flaw?Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952) Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
I also made that mistake once.... :laugh: :laugh: :laugh:
-
If you started to dream of coding you have only two options - doctor or vacation...
I'm not questioning your powers of observation; I'm merely remarking upon the paradox of asking a masked man who he is. (V)
-
harold aptroot wrote:
no vptr
Don't get hung up on how current languages are implemented. I too would like a language that allows static members in interfaces and abstracts. If a Bright compiler developer wanted to make this happen, I'm sure he could, it would likely have to be very different from current implementations. (It's way beyond my skills of course.)
You'll never get very far if all you do is follow instructions.
very different from current implementations[^] :-D
**_Once you lose your pride the rest is easy.
I would agree with you but then we both would be wrong._**
The report of my death was an exaggeration - Mark Twain Simply Elegant Designs JimmyRopes Designs
I'm on-line therefore I am. JimmyRopes -
0 times. I've wished that interfaces could contain static methods so that T.someFunction could exist for generics (constrained to that interface), but that's a completely separate issue. (that can't work in C# of course) How is "static virtual" even a reasonable concept? No instance = no vptr = no dynamic dispatch.
-
I once coded a class function in C++ that calls realloc() or malloc(), based on whether "this" points to null. It then deletes this and returns This.
-
Not only that, it calls _msize() on the this pointer and checks against another size parameter for the realloc. The last member in the class was a 0-byte array, for which the _msize() of the pointer minus size of other class members gave the array size.
