Hi,

I've moved the bitcoin-dev list to bcc:, as this question is better suited to forums dedicated to Bitcoin Core implementation specifics, rather than the general bitcoin development list.  

Please feel free in the future to ask questions like this on the bitcoin-core-dev mailing list (https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-core-dev) or on the #bitcoin-core-dev freenode IRC channel.

The work limit (that was put in place in https://github.com/bitcoin/bitcoin/pull/6654, when the concept of "dirty" entries was introduced) was removed in https://github.com/bitcoin/bitcoin/pull/7594, in preparation for ancestor-feerate-mining.  So those comments should have been cleaned up to match the new code.

Please feel free to file an issue or open a PR to update those comments at https://github.com/bitcoin/bitcoin.

Thanks,
Suhas
 

On Mon, May 8, 2017 at 5:38 AM, DJ Bitcoin via bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org> wrote:
Hi Guys,
 
I have a question about the use of txmempool. find attached the code in txmempool.h
 
 
======================================================
/* Adding transactions from a disconnected block can be very time consuming,
 * because we don't have a way to limit the number of in-mempool descendants.
 * To bound CPU processing, we limit the amount of work we're willing to do
 * to properly update the descendant information for a tx being added from
 * a disconnected block.  If we would exceed the limit, then we instead mark
 * the entry as "dirty", and set the feerate for sorting purposes to be equal
 * the feerate of the transaction without any descendants. */

class CTxMemPoolEntry
{
   private:
   // ...    
   // Information about descendants of this transaction that are in the
   // mempool; if we remove this transaction we must remove all of these
   // descendants as well. if nCountWithDescendants is 0, treat this entry as
   // dirty, and nSizeWithDescendants and nModFeesWithDescendants will not be
   // correct.
   
   int64_t nCountWithDescendants; //!< number of descendant transactions
   // ...
======================================================
 
 
Now, the only place where nCountWithDescendants is modified is the following (txmempool.cpp):
 
 
======================================================
void CTxMemPoolEntry::UpdateDescendantState(int64_t modifySize, CAmount modifyFee, int64_t modifyCount)
{
    nSizeWithDescendants += modifySize;
    assert(int64_t(nSizeWithDescendants) > 0);
    nModFeesWithDescendants += modifyFee;
    nCountWithDescendants += modifyCount;
    assert(int64_t(nCountWithDescendants) > 0);
}
======================================================
 
 
Therefore, nCountWithDescendants is never zero.
Am i missing something? Where is this concept of "dirty" defined?
 
Thanks a lot,
DJ
 
 
 
 
 

_______________________________________________
bitcoin-dev mailing list
bitcoin-dev@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev