Does MATLAB call the double constructor on a=a+1?
1 view (last 30 days)
Show older comments
Jan Siegmund
on 22 Apr 2020
Commented: Jan Siegmund
on 25 Apr 2020
I want to overload the double class and add some properties to it.
And i also have to override the plus function. A similar code could
look like this:
classdef myclass < double
properties (Access = protected)
foo
end
methods (Access = public)
function obj = myclass(val,foo)
obj = obj@double(val);
obj.foo = foo;
end
function out = plus(obj,in)
out = myclass(plus@double(obj,in),obj.foo);
end
end
end
However, if out = in on plus, I could just obj.foo = in.foo and call plus@double(obj,in), could I?
How does MATLAB itself determine, if a plus is in fact an increment, so out = in?
0 Comments
Accepted Answer
Walter Roberson
on 22 Apr 2020
How does MATLAB itself determine, if a plus is in fact an increment, so out = in?
Reference counting.
I do not know about class methods in this regard.
Outside of classes, in-place operations are only done inside a function of the form
function Variable = FunctionName(arguments that include Variable)
and only if the reference count for Variable is 1, and only if the outside output variable name is the same as the input variable name
In-place operations are not even done for simple code like
count = 0;
for K = 1 : 5
count = count + 1;
end
it would have to be something like
count = 0;
for K = 1 : 5
count = PlusPlus(count);
end
function var = PlusPlus(var)
var = var + 1;
end
For more discussion of in-place operations, I recommend searching postings from James Tursa
3 Comments
Walter Roberson
on 24 Apr 2020
Mind you, exactly what the Execution Engine optimizes to in the case where it detects built-in arithmetic types and non-overloaded operations is unknown.
More Answers (0)
See Also
Categories
Find more on Class File Organization in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!