I thought I was improving my code by adding ::type to all the _self in my OOP code. As this is one main advantage in programming in Maple, which is being able to attach types to all name and variables. Make code more robust.

But it turned out Maple is not very happy now and gives that drearded error

    Error, static procedure ... refers to non-static local or export ... in surrounding scope

Only place I found that adding ::type_name is to _self is allowed, is on the constructor signature.

But in no other method of the object module. local or export method, it does not matter. 

My question is, why is that?

So I went and removed all those _self::type_name and made it just _self to make Maple happy.

I also noticed this happes regardless of having kernelopts('assertlevel'=2): there or not.  Attached  is the worksheet.

Just tryting to understand the logic, that is all. 

This is using Maple 2023.2.1

Download why_adding_type_on_self_gives_error.mw

WHen using _self in object, help says

"As of Maple 2021, if the method has a formal parameter named _self, references
to its object's local or exported variables may be written without prefixing them.
That is, _self:-variable may be written as just variable. Maple will
add the self:- prefix internally when the method is simplified."

And the above is true for all static methods inside an object module, except for ModuleCopy which is also static.

It seems ModuleCopy is special. But why?  I am thinking it is because at time this is called, the object itself does not yet exist, but for all other methods, the object by then is fully constructed. But wanted to be sure.

Here is an example

restart;

person := module()
option object;
local  m_name::string;
local  m_age::integer;
export ModuleCopy :: static := proc(_self :: person
                                   , proto :: person, name::string, age::integer,$
                                )
         m_age := age;
         m_name := name;
    end proc;

export name::static:= proc(_self, $);
    RETURN(m_name); #no need to write _self:-m_name;
    end proc;

export age::static:= proc(_self, $);
    RETURN(m_age); #no need to write _self:-m_age;
    end proc;

export set_name::static:= proc(_self, name, $);
    m_name := name;   #no need to write _self:-m_name := name
    NULL;
    end proc;

export process::static:=proc(_self,$)
   print("name is ",m_name," age is ",m_age);
end proc;

end module:

And now

o:=Object(person,"me",10);

Error, static procedure `ModuleCopy` refers to non-static local or export `m_age::integer` in surrounding scope

Changing ModuleCopy to

export ModuleCopy :: static := proc(_self :: person
                                   , proto :: person, name::string, age::integer,$
                                )
         _self:-m_age := age;
         _self:-m_name := name;
    end proc;

Now there is no error. But notice that in all other static methods, I can write  m_name directly without using _self:-m_name;

I looked at help page for ModuleCopy but see no mention of this.

Why ModuleCopy is different?

Maple 2023.2 on windows 10

I am really confused by this behavior. Should the solution be different when calling dsolve(ode) vs. dsolve(ode,[x(t),y(t)]) where here ode is list of two coupled first order ode's?

Maple shows the solutions in reverse order in one case vs. the other. I do not see why adding [x(t),y(t)] should make x(t) solution now y(t)'s and y(t) solution now x(t)'s.  Here is an example

Here is worksheet. I just find the out changing very confusing. Should not the same solution for x(t) and y(t) remain the same regardless of adding the second option as list of not? What is the logic behaind this change or order, and how it one to know which is the correct one as they look different.

Download why_solution_changes_nov_27_2023.mw

In my code, without knowing what the expression is, other than it has RootOf, the code called allvalues and got internal error 

Error, (in SolveTools:-Basis) invalid input: igcd received 5/7, which is not valid for its 2nd argument

Is this to be expected depending on the input, or is this some internal problem I need to report?

restart;
expr:=RootOf(R^4*b+R^2*a*_Z+2*_Z^2-exp(RootOf(tanh(1/2*(a^2-8*b)^(1/2)*(4*S-_Z)/a)^2*R^4*a^2-8*tanh(1/2*(a^2-8*b)^(1/2)*(4*S-_Z)/a)^2*R^4*b-R^4*a^2+8*R^4*b-8*exp(_Z))))

allvalues(expr)

Error, (in SolveTools:-Basis) invalid input: igcd received 5/7, which is not valid for its 2nd argument

Maple 2023.2 on windows 10

ps.  Reported to Maplesoft

I wonder if this a bug or Am I overlooking something?

I set infolevel[dsolve]:=5; and first time calling dsolve(ode), it works as expected. It prints on the screen the trace and steps it did.  But when I repeat the command dsolve, now it only prints one line and the rest of info that were printed before no longer show.

Should not the same information be printed each time? This is what happens on another example I tried. It seems infolevel does not work the same depending on the ode being solved?  Here is a worksheet attached to show the above.

Is there a workaround to make it display same information each time?

Maple 2023.2 on windows 10.

On a side note. the first line of the display is hard to read. Any one know what it is supposed to mean by the space between the 1 and y(x) at the end there?

Download infolevel_stops_working.mw

Another example where infolevel changes on second call to dsolve: The first call gives more information which is lost in the second call. There does not seem to be a way to fix this at user level.

Download another_example_info_level_changes.mw