is it possible for a string such as "res:=boo()" to be evaluated after calling parse, and have res have the value that the function boo() returned?

Now, it does not work. I must do   res:=eval(parse("boo()")) , meaning the result of the call to boo() is outside the string.

Here is an example of what I mean

boo:=proc()
   return 99;
end proc;  

foo:=proc()
  local s,res,i;
  s:="res:=boo()";
  eval(parse(s)):
  print(res);
end proc;

foo();

The above does not work. It displays res but this works

boo:=proc()
   return 99;
end proc;  

foo:=proc()
  local s,res,i;
  s:="boo()";
  res:=eval(parse(s)):
  print(res);
end proc;

foo();

The above works, and prints 99.

The reason I am asking, is that I was thinking of storing in a data base the complete Maple command as string, along with the LHS of the call, and just load the string the evaluate it, and it will have res automatically set.

It is no problem if this can't work, I could always just eval the command itself, without the assignment being there. Just thought to ask if there is a way.

I have about 30 modules under one root common module.  The total code is about 25,000 lines.

Currently during development, when I change a line of code and want to run a test, I just do

interface(warnlevel=4);
kernelopts('assertlevel'=2):
read   "my_root_module.mpl";
#run some test 

From a worksheet. This works fine, but it is very slow. it takes now about 8-10 minutes to finish each time.

So each time I make small change to the source code, which all sit in separate .mpl plain text files, I have to wait and wait for the read to complete.

This is way too slow. 

The thing is, my my_root_module.mpl files includes child modules (using the $include directive), and each child module might also have $include to pull in its own child modules, if any. Each module is onbe separate .mpl file.

So at the end my_root_module.mpl will end up reading all the code, which is now about 25,000 lines over 30 or so modules. Here is an example of the layout I have

my_root_module.mpl
=======
export my_root_module:=module()
$include  "B.mpl"   
$include  "C.mpl"                            
$include  "D.mpl"   
   export foo:-proc()....     B:-foo().... end proc;
end module:

B.mpl
=======
local B:=module()
$include  "B1.mpl"   
$include  "B2.mpl"                            

   export foo:=proc()..... end proc;
end module:

C.mpl
=======
local C:=module()
$include  "C1.mpl"   

   export foo:=proc()..... end proc;
end module:

etc...

There must be faster way to do this. Having to wait about 10 minutes each time to test one small code change each time is not practical.

What do others do when they develope large code in Maple? Is this a typical time?

I do not want to break my main modules into separate name spaces again in order to  use separate mla for each module, so that my changes only affects one mla file.  This will speed things, as I would then only have to rebuild one small modules, the one I changed.

But I want to put all my modules under one common module, which acts just as a common name space, and have only one mla file.

Are there ways to speed  reading mpl file?

Edit

Thanks to all the answers. It turned out indeed to be the slow Maple GUI again. The option interface(prettyprint= 0) did not help in the worksheet to speed the GUI.

I wrote the commands to read the mpl file and build the mla in a file, and used the DOS command line, and turned off the warning messages using DOS option, and now it finished almost instantly.

"C:\Program Files\Maple 2021\bin.X86_64_WINDOWS\cmaple.exe" UPDATE_MLA.mpl 1> nul

    memory used=18.6MB, alloc=73.3MB, time=0.52

Notice, in the above 1>nul will not even display the warning messages on the terminal. 2>nul does not work, since these are not error messages. Here are the different redirection options from the net

But even without using this redirection, and keeping the messages scroll on the terminal, it was still very very fast from the command line

I complained about how slow the Maple Java GUI when it comes to scrolling and how it can can cause  slow down many times before, so will not complain again about. May be one day Maplesoft will fix this.

I noticed PDEtools:-Solve hangs sometimes on some input where solve does not. When I change the solver for PDEtools:-Solve to use solve, it did not hang.

This means PDEtools:-Solve default solver is something else. I did not see what it is by looking at help.

Here is an example

restart;
eq:=[1/3*(-334*I*(-4807763*I+27000*I*2^(1/2)+60*(1746675+72116445*2^(1/2))^(1/2))^(1/3)-2*(-4807763*I+27000*I*2^(1/2)+60*(1746675+72116445*2^(1/2))^(1/2))^(2/3)+56978)/(-4807763*I+27000*I*2^(1/2)+60*(1746675+72116445*2^(1/2))^(1/2))^(1/3)*v[1]-20*I*2^(1/2)*v[2] = 0, (-18000*I*2^(1/2)+33400*I-40*(1746675+72116445*2^(1/2))^(1/2)-40*(-167*I*(1746675+72116445*2^(1/2))^(1/2)+75150*2^(1/2)+10*(-4807763*I+27000*I*2^(1/2)+60*(1746675+72116445*2^(1/2))^(1/2))^(2/3)+145445)/(-4807763*I+27000*I*2^(1/2)+60*(1746675+72116445*2^(1/2))^(1/2))^(1/3))/(167*I*(-4807763*I+27000*I*2^(1/2)+60*(1746675+72116445*2^(1/2))^(1/2))^(1/3)+(-4807763*I+27000*I*2^(1/2)+60*(1746675+72116445*2^(1/2))^(1/2))^(2/3)-28489)*v[2]-20*I*2^(1/2)*t = 0];

solve(eq,[v[1],v[2]]); #OK
PDEtools:-Solve(eq,[v[1],v[2]],solver=solve); #OK
PDEtools:-Solve(eq,[v[1],v[2]]); #hangs

Maple 2021

I am merging two separate modules  B,C to be child modules inside one main A module.

Now, I found the code breaks, because uses does not work any more.

Before merging, C module did uses B. But now this gives an error when both B,C sit inside one parent module.  An example make this easier to explain

restart;
interface(warnlevel=4);
kernelopts('assertlevel'=2):

A:=module()

   export B:=module()
     export foo:=proc()
         print("In B:-foo()");
     end proc;
   end module;

   export C:=module()
      uses B; #this cases problem
      export boo:=proc()
           foo();
      end proc;
   end module;
end module;

Error, (in A:-C) no bindings were specified or implied

Changing uses B; to uses A:-B; does not help. I get error Error, (in A:-C) `A` does not evaluate to a module

I also tried uses :-A:-B; , now this does not give error, but it does not work. i.e. when doing  A:-C:-boo() Maple does not end up calling foo() inside module B as expected.

One way to avoid all this, is not to use uses B inside the module C and do this instead

A:=module()

   export B:=module()
     export foo:=proc()
         print("In B:-foo()");
     end proc;
   end module;

   export C:=module()      
      export boo:=proc()
           B:-foo(); 
      end proc;
   end module;
end module;

But it means I have to now change lots of code inside the C module, and add an explicit B:- everywhere

This is how it was before the merging

B:=module()
   export foo:=proc()
       print("In B:-foo()");
   end proc;
end module;

C:=module()  
  uses B;    #no problem now
  export boo:=proc()
      foo();  #this now uses B:-foo() automatically due to uses.
  end proc;
end module;

The above works. Now I can do C:-boo() and it works as expected.

The question is: How to make it work after moving both C and B inside one parent module so I do not have to change lots of code? I tried many things, but can't get it to work.

Maple 2021 on windows 10

Edit

Thanks for the answers below. I think I have to change my code then either way to add a prefix to the call. I actually always use the long form of the call for everything as in module:-function() but for one function, which I use so much everywhere, using the fully qualified name would make things hard to read.  This function converts Maple expressions to Latex after some filtering. Here is an example

cat(",toX(y),"' = f_0(",toX(f),")",toX(y),"+f_1(",toX(x),")",toX(g),"^n \\tag{2}, etc.....")

Now the function toX comes from one module, the one I had uses for it. Now I have to change the above to becomes

",module_name:-toX(f),"' = f_0(",module_name:-toX(x),")",module_name:-toX(y),"+f_1(",module_name:-toX(x),")",module_name:-toX(g),"^n \\tag{2}

Since my strings are very long as it is, (program generates Latex on the fly as it runs), this will make them even longer and harder to read.

But I can change all this in the editor, using global search and replace.

I was just hoping I do not have to just because I moved the modules all into one main module. I still do not understand why Maple does not allow uses when moving the modules inside one bigger module, but I guess this is by design.

Edit

I found the problem. Let me rephrase it. I keep original question below. 

This causes an error

restart;
interface(warnlevel=4);
kernelopts('assertlevel'=2):
A:=module()
   export module_A1;
   local person_type;

   module person_type()
       option object;
       export name::string; 
       export age::string; 
    end module;

   module_A1:=module()
        export boo;
        boo:=proc()
          local X::A:-person_type;  #THIS LINE CAUSES ERROR
          #local X::person_type;    #THIS LINE ALSO CAUSES ERROR
          #local X;                 #THIS WORKS. No error
          X:=Object(person_type);
          return X;
       end proc;
    end module:
end module:

A:-module_A1:-boo();
           Error, (in boo) module does not export `person_type`

If I change  local X::person_type;  to just local X;  then no error is generated:

restart;
interface(warnlevel=4);
kernelopts('assertlevel'=2):

A:=module()
   export module_A1;
   local person_type;

   module person_type()
       option object;
       export name::string; 
       export age::string; 
    end module;

   module_A1:=module()
        export boo;
        boo:=proc()
          local X;  
          X:=Object(person_type);  #now this works
          return X;
       end proc;
    end module:

end module:

A:-module_A1:-boo();

              `Object<<person_type,1822742359104>>`

So the problem was in the variable declaration. Maple for some reason does not like   local X::A:-person_type;  and does not like local X::person_type; but things seem to work by just removing the ::type altogother. 

So this solves this problem for me for now, I can keep person module local, and still use it in other child modules. Even though I do not understand why Maple complains about it when I use ::

original question. Can be ignored now.

I have one main module, with 2 submodules (child modules) A1 and A2. The child modules are all exported so they can be called from outside the main module.

But I want to have one additional module X as object, to share among the child modules A1 and A2, but without having to also export module X.

But Maple wants me to also export module X in order to use it inside the child modules A1, and A2.

Is there a way to share X  among child modules, without exporting it? I only want X be visible to child modules (i.e. modules inside the main module)

The module X has option object and I want to use it as just a type basically (instead of using Record).

Here is an example

restart;
interface(warnlevel=4);
kernelopts('assertlevel'=2):
A:=module()
   export module_A1,module_A2;
   local person_type;

   module person_type()
       option object;
       export name::string; 
       export age::string; 
    end module;

   module_A1:=module()
        export boo;
        boo:=proc()
          local X::A:-person_type;
          X:=Object(A:-person_type);
          return 1;
       end proc;
    end module:

   module_A2:=module()
        export boo;
        boo:=proc()
          local X::A:-person_type;
          X:=Object(A:-person_type);
          return 1;
       end proc;
    end module:

end module:

Now the call A:-module_A1:-boo(); fail with Error, (in boo) module does not export `person_type`

This can be fixed by changing local person_type;  to export person_type;  but this means person_type can now be seen from outside the main package.

In Ada, it is possible to have private data type shared among child packages. How to do the same in Maple?