Interesting problem I found

ID:264653
 
Feb 13 2010, 6:28 am
So I was going through my code, coding my dna structuring system, when all of a sudden a few procedures stopped working. I inspected my set up and found an interesting solution to why it wasn't working.

First of all I have two code files, dna.dm comes first before inc_dna.dm does in this order:
dna.dm
inc_dna.dm


Now in dna.dm I have this:
#include "inc_dna.dm"

generate_structures()
    world<<"generated"

I then have in my inc_dna.dm,
proc/generate_structures()


I learned that it wasn't calling the procedure in my dna.dm file but instead going to inc_dna.dm to call it, thus having no effect when I call the procedure. First of all dna.dm comes first, but calls in the code from inc_dna.dm defining the generate procedure. I then define the generate_structures procedure after the include line but I think it is then reading from the next code file; inc_dna.dm thus having a reversing effect on the procedure definition.

This is my theory, but what I find interesting is I have the same setup under multiple procedures only they are defined in datums and are not global. They work perfectly...? Has anyone else run into this problem? I mean I can easily get around it but I just find these circumstances interesting.

Feb 13 2010, 10:58 am
If you override a proc in the exact same scope, in separate files, then the behavior is as good as unspecified. The solution is not to do that.
Feb 13 2010, 11:05 am
In response to Garthor
Yes but I do that with other procs only the procs are defined in datums, and they work perfectly... I'm just wondering why that is?
Feb 13 2010, 11:19 am
In response to Smokey Joe
Because you aren't doing "that" in other procs. I doubt I can explain this to you in words, so I'll just give an example:

// OKAY:
mob
  proc/dance()
    src << "You dance"
  dancer
    dance()
      src << "You dance really well"

// NOT OKAY:
proc/blah()
  // stuff

blah()
  // different stuff


On its own, it's poor design. When you start mixing them up in different source files, it starts messing things up.
Feb 13 2010, 11:27 am
In response to Garthor
Yes but I AM doing that in other procs, here look at this.
The file body.dm precedes inc_body.dm

This is body.dm:
#include "inc_body.dm"

datum
    
    body
        create(nam)
            src.name = nam
            dna = get_dna()
            update_dna()
            
            who = ckey(nam)
            density = 1
        
        get_dna(list/L,src.name)
            var/datum/dna/d = new()
            d.create_dna(L)
                    
            return d


Then in inc_body.dm I have this:
datum


    body        //the body datum

        proc/create(name,species)       //creates the body
        proc/destroy()      //destroys body


        proc/tickle()
        proc/get_dna(list/L)        //grabs a new generated dna, L is anything specific
        proc/update_dna()   //update body status based on dna code
        proc/change_dna()


Same setup as my dna files, yet this works while my dna procedures failed.
Feb 13 2010, 11:30 am
In response to Smokey Joe
Does it have anything to do with being global? The only difference between these two I can see is the dna procedures were global, these are not.
Feb 13 2010, 11:55 am
In response to Garthor
 
mob
 proc
  abc()
   world << 1

 abc()
  world << 2
  ..()

 abc()
  world << 3
  ..()


You can override the datum function a million times in the same scope and it will work.

proc
 abc()
  world << 1

abc()
 world << 2
 ..()

abc()
 world << 3
 ..()


This however doesn't.
Feb 13 2010, 12:46 pm
In response to Keeth
I just was wondering why that is? Maybe its not that big of a deal but things like this just bug me.
Feb 13 2010, 12:50 pm
In response to Smokey Joe
I think the only guys who can answer this are the guys who work on the source.