[gecode-users] Debugging branching

[Mikael Zayenz Lagerkvist]

First of all, I meant the copy constructor in your custom branching.

Secondly, I re-read your originally posted code, and saw that you use
the status/description/commit interface erroneously. From the
documentation of description [1]:
    "This member function can only be called after the member
     function Space::status on the same space has been called
     and in between no non-const member function has been
     called on this space."

Cheers,
Mikael

[1] http://www.gecode.org/gecode-doc-latest/group__TaskSearch.html


On Wed, Jan 21, 2009 at 8:12 AM, Malcolm Ryan <malcolmr at cse.unsw.edu.au> wrote:
> On 21/01/2009, at 5:36 PM, Mikael Zayenz Lagerkvist wrote:
>
>> Hi,
>>
>> While I'm not sure about the exact reason for your failure (especially
>> without seeing your code), it could be that you call the wrong
>> Branching-constructor in your copy constructor. See the example custom
>> branchings in examples/black-hole.cc and examples/queen-armies.cc for
>> guidance.
>
> Yes, I was following those examples. My code for the space is:
>
>
> ConcretePlanSpace::ConcretePlanSpace(Graph *graph, int nRobots, int
> maxLength) :
>        Space(), myPlan(this, graph, nRobots, maxLength, true) {
>
> }
>
> ConcretePlanSpace::ConcretePlanSpace(bool share, ConcretePlanSpace
> &space) :
>        Space(share, space) {
>        myPlan.update(this, share, space.myPlan);
> }
>
> ConcretePlanSpace::~ConcretePlanSpace() {
>        // TODO Auto-generated destructor stub
> }
>
> Space* ConcretePlanSpace::copy (bool share) {
>        return new ConcretePlanSpace(share, *this);
> }
>
> CompletePlanVar& ConcretePlanSpace::getPlan() {
>        return myPlan;
> }
>
>
>
>>
>>
>> Cheers,
>> Mikael
>>
>> On Wed, Jan 21, 2009 at 7:21 AM, Malcolm Ryan <malcolmr at cse.unsw.edu.au
>> > wrote:
>>> I've written a custom branching and I am trying to write some unit
>>> tests to make sure it is working properly. As such, I am doing the
>>> branching and propagation process step-by-step, but there is
>>> something
>>> wrong with my code. I have:
>>>
>>>       ConcretePlanSpace *space = new ConcretePlanSpace(graph, 1, 10);
>>>       ConcretePlanBranching *branching = new (space)
>>> ConcretePlanBranching(space);
>>>
>>>       assert(space->status() == SS_BRANCH);
>>>
>>>       const BranchingDesc* desc = space->description();
>>>       space->commit(desc, 0);
>>>
>>>       assert(space->status() == SS_BRANCH);
>>>
>>> That much of the code works fine, but I want to create a clone of the
>>> space to test out different alternatives:
>>>
>>>       Space* copy = space->clone(false);
>>>       desc = copy->description();
>>>       copy->commit(desc, 0);
>>>
>>> When it gets to commit() the code fails. GDB shows:
>>>
>>> Program received signal EXC_BAD_ACCESS, Could not access memory.
>>> Reason: KERN_PROTECTION_FAILURE at address: 0x0000000b
>>> 0x001d1e0b in Gecode::Space::commit (this=0xf17460, d=0x7, a=0) at
>>> gecode/kernel/core.cc:330
>>> 330         while ((b_commit != Branching::cast(&a_actors)) &&
>>>
>>> Any ideas what I am doing wrong?
>>>
>>> Thanks,
>>>
>>> Malcolm
>>>
>>> _______________________________________________
>>> Gecode users mailing list
>>> users at gecode.org
>>> https://www.gecode.org/mailman/listinfo/gecode-users
>>>
>>
>>
>>
>> --
>> Mikael Zayenz Lagerkvist, http://www.ict.kth.se/~zayenz/
>
>
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users
>



-- 
Mikael Zayenz Lagerkvist, http://www.ict.kth.se/~zayenz/