getblocktemplate: longpolling support (v2) #4503

1. 
   laanwj commented at 9:07 am on July 10, 2014: member

   A bit reworked version of #2834 ‘getblocktemplate: longpolling support’. Just a final push in trying to get it merged.

   • Add a function IsRPCRunning()
   • Remove unnecessary added includes
   • Use our own CWaitableCriticalSection instead of boost::mutex (amounts to the same, ours is a wrapper)
   • Add our own typedef CConditionVariable for boost::condition_variable
   • Make it compile and work in disable-wallet builds, as well as when running in runtime disable-wallet mode

   Note: currently untested! I still need to find something to test this with @luke-jr.

2. 
   luke-jr commented at 3:38 pm on July 10, 2014: member

   See #2834 (comment) for a test plan.
3. in src/rpcmining.cpp: in 7ce99d1b74 outdated

   387+            while (chainActive.Tip()->GetBlockHash() == hashWatchedChain && IsRPCRunning())
   388+            {
   389+                if (!cvBlockChange.timed_wait(lock, checktxtime))
   390+                {
   391+                    // Timeout: Check transactions for update
   392+                    if (nTransactionsUpdatedLast != nTransactionsUpdatedLastLP)
   

   ---

   ---

   

   luke-jr commented at 9:01 pm on July 10, 2014:

   This should use mempool.GetTransactionsUpdated() instead of nTransactionsUpdatedLast, otherwise it will never be triggered unless another getblocktemplate call updates it. :(

   ---

   

   laanwj commented at 5:55 am on July 11, 2014:

   So to be clear, this should be?

   0if (mempool.GetTransactionsUpdated() != nTransactionsUpdatedLastLP)
   

   Or,

   0if (mempool.GetTransactionsUpdated() != nTransactionsUpdatedLast)
   

   ---

   

   luke-jr commented at 6:03 am on July 11, 2014:

   0if (mempool.GetTransactionsUpdated() != nTransactionsUpdatedLastLP)
   

4. 
   luke-jr commented at 9:02 pm on July 10, 2014: member

   ACK except for the 1 bug noted inline for transaction updates.
5. 
   laanwj commented at 6:17 am on July 11, 2014: member

   Ok, updated, will look at the test plan next.
6. getblocktemplate: longpolling support ff6a7af154
7. Add test for getblocktemplate longpolling b45a6e8394
8. 
   laanwj commented at 12:51 pm on July 11, 2014: member

   Ok, I added tests (qa/rpc-tests/getblocktemplate.py):

   • Test 1: test that the longpolling waits if we do nothing else
   • Test 2: test that longpoll will terminate if another node generates a block
   • Test 3: test that longpoll will terminate if we generate a block ourselves
   • Test 4: test that introducing a new transaction into the mempool will terminate the longpoll

   I consider this sufficiently tested and unless someone screams bloody murder, I will merge this later today.

9. 
   BitcoinPullTester commented at 1:05 pm on July 11, 2014: none

   Automatic sanity-testing: PASSED, see http://jenkins.bluematt.me/pull-tester/p4503_b45a6e8394e8c9e2886ae8b9aa0734996448ff37/ for binaries and test log. This test script verifies pulls every time they are updated. It, however, dies sometimes and fails to test properly. If you are waiting on a test, please check timestamps to verify that the test.log is moving at http://jenkins.bluematt.me/pull-tester/current/ Contact BlueMatt on freenode if something looks broken.
10. laanwj merged this on Jul 14, 2014
11. laanwj closed this on Jul 14, 2014

    ---

12. laanwj referenced this in commit 3554df9b99 on Jul 14, 2014
13. MarcoFalke locked this on Sep 8, 2021

Contributors
laanwj luke-jr BitcoinPullTester