Re: [reSIProcate-users] Why push mChain.size and mAddress in?

[Byron Campen <docfaraday@xxxxxxxxx>]

One of the properties of the processor chain is that it can perform filtering etc asynchronously; the chain needs to be able to resume at the same point later by using a ProcessorMessage. In addition, you can have chains within chains, so the addressing scheme needs to be recursive. Those vector<short> are basically recording something like "chain x, subchain y, subsubchain z, processor a". I could be wrong, since I wrote that stuff a long time ago when I was a n00b, but the reason this was not handled as a simple linked list, with the address simply being a Processor*, was to allow processors to be removed without leaving dangling pointers around. I bet this code could (should?) be made simpler...

Best regards,
Byron Campen

On Jul 5, 2013 12:18 AM, <jiangjungeneral@xxxxxxxxxx> wrote:

Hi,

    I am working on repro example,these two lines in the function addProcessor of class ProcessorChain confused me (in file ProcessorChain.cxx,line 55):

 

void
repro::ProcessorChain::addProcessor(auto_ptr<Processor> rp)
{
   DebugLog(<< "Adding new " << mName << " to chain: " << *(rp.get()));
   rp->pushAddress((short)mChain.size());
   rp->pushAddress(mAddress);
   rp->setChainType(mType);
   mChain.push_back(rp.release());
}

    So why rp(class Processor) put a vector size (mChai.size) and a vector (mAddress) in it's vector (mAddress),mAddress is defined in class Processor (in file Processor.hxx,line 49):

   protected:
      std::vector<short> mAddress;

and the two functions pushAddress are (in file Processor.cxx,line 23):

void
Processor::pushAddress(const std::vector<short>& address)
{
   for(std::vector<short>::const_iterator i=address.begin();i!=address.end();++i)
   {
      mAddress.push_back(*i);
   }
}

void
Processor::pushAddress(const short address)
{
   mAddress.push_back(address);
}

Thanks.

_______________________________________________
resiprocate-users mailing list
resiprocate-users@xxxxxxxxxxxxxxx
List Archive: http://list.resiprocate.org/archive/resiprocate-users/