Could anybody enlighten me on how to make the 'SetLength' procedure into an equivalent function

I have many instances of
   SetLength(xxx, Length(xxx)+1);
   Num := High(xxx);
I would like to have Num :=  SetLength(xxx, Length(xxx)+1);
  or even something like Num := IncLength(xxx, 1);

These are on varying types of array, so not really feasible to create a function for each one.
Hoping it is possible within pascal itself :)

Thanks - SteveG

Steve Gatenby wrote

Could anybody enlighten me on how to make the 'SetLength' procedure into an equivalent function I have many instances of SetLength(xxx, Length(xxx)+1); Num := High(xxx);

What is the problem with this code?
Why force these 2 commands into a function?
Just to save a little bit of typing?
When *reading* this code it is much
clearer what it does as when reading

Num := IncLength(xxx, 1);

It is only a personal preference, not a serious problem.

I would find code in the form of
   ArrayLen := IncLength(ArrayName, 1);

to be much more readable then

    SetLength(ArrayName, Length(ArrayName)+1);
    ArrayLen:= High(ArrayName);

I was just hoping somebody may know how its done.

Thanks - SteveG

Without support for generic function, currently you have to use generic class method:

type
  generic TDynArrayUtil<T> = class
  public
    class function SetLength(var a: T; const n: Integer): Integer;
  end;
class function TDynArrayUtil.SetLength(var a: T; const n: Integer): Integer;
begin
  System.SetLength(a,Length(a) + n);
  Result := High(a);
end;

and use it like:

var
  a: TIntegerDynArray;
...
maxn := specialize TDynArrayUtil<TIntegerDynArray>.SetLength(a,1);

note that you have to instantiate with the correct type.

Thanks for that - I will give it a try :)

Why? It is not clear which index ArrayLen will receive from IncLength. You need to look up (or remember) what exactly this function does while in the two-line version it is all clear and unambiguous. Especially, if you not only add 1 element but more as in

   ArrayLen := IncLength(ArrayName,2);
    SetLength(ArrayName, Length(ArrayName)+2);
    ArrayLen:= High(ArrayName);

In the first version, which value will ArrayLen be set to? High(ArrayName) or High(ArrayName)-1?
This is very clear in the second version.

IMO it is not worth investing time into writing a function that obscures the code. If you look at it months later you may no longer know what IncLength does exactly, not to talk about if someone else has/wants to read your code.

I very seldom comment on this, or any, forum unless I have a good answer and no one else seem to have one. This case is different. I have noticed quit often here that instead of answering the question people try to reason why in the world the author of the question would even want to do this. Would it not be better to tell him how to do it or at the most say "I don't know". I know there are times when you want to prevent someone from making a serious mistake, but this is not the case here. Sometimes the question is placed by someone new to programming and it is important to learn from your own work, mistakes and all. Often the question also comes from people migrating from, or knowledgeable with, another language where a similar functionality exits. This type of argument will only discourage them away from our beloved Pascal.

Nur mein Senf.

Andreas Berger wrote

I have noticed quit often here that instead of answering the question people try to reason why in the world the author of the question would even want to do this. Would it not be better to tell him how to do it or at the most say "I don't know". I know there are times when you want to prevent someone from making a serious mistake, but this is not the case here.

I know what you mean but in this case I find it worth mentioning that the 2 line code *is* easier to read. And for me it would be wasting time to somehow find a solution for his problem instead of simply using the 2 line code and use the saved time for real problems. The 2 line code is definitely more in the spirit of Pascal than obscuring what happens with a (not so clear) function.

But of course, everybody can do what he wants.
It was just meant in the sense of an advice.

Andreas Berger wrote

I have noticed quit often here that instead of answering the question people try to reason why in the world the author of the question would even want to do this.

I agree that this is an annoying pattern on this list.

I would suggest everyone at least wait until someone tries to respond with an actual answer to the question before diverting discussion in this way.