RanchiMall/node-stratum-pool
3
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fixed duplicate shares on vardiff difficulty update. Added more comments.

Browse Source
This commit is contained in:
Matt 2014-02-19 13:52:43 -07:00
parent 897dadbf11
commit 0dcccbd56e
1 changed files with 8 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
lib/pool.js
View File

@ -54,20 +54,21 @@ var pool = module.exports = function pool(options, authorizeFn){
client.sendDifficulty(newDiff); client.sendDifficulty(newDiff);
/* /*
Disabled this line of code as it will because it will force the miner Disabled this line of code because it will force the miner
to restart and submit duplicate shares. The stratum-python sends out a to restart and submit duplicate shares. Stratum-python sends out a
mining.notify but rolls the jobID and sets "clean jobs" to false. mining.notify but rolls the jobID and sets "clean jobs" to false.
Meaning that the worker will only start on the new work once it Meaning that the worker will only start on the new work once it
exhausts its current nonce range. But if the miner were to start the exhausts its current nonce range. But if the miner were to start the
new job, the shares would be invalidated since stratum-python doesn't new job, the shares would be invalidated since stratum-python doesn't
insert the new jobID that the share-limiter generated into the jobs insert the new jobID that the share-limiter generated into the jobs
array. Also, since the new work is only sent with a new jobID but with array. Even worse, since the new work is only sent with a new jobID
the same extranonce as the last job, the shares will be duplicate but with the same extranonce and other job parameters as the last job,
anyway. Perhaps this bug has gone unnoticed because of how likely it the shares will be duplicate. Perhaps this bug has gone unnoticed
is for a miner to exhaust the nonce range before new work is sent. because of how likely it is for a miner to exhaust the nonce range
before new work is sent.
So lets only send a new difficulty, and the worker will use it when So lets only send a new difficulty, and the worker will use it when
it receives a new job when the block template updates. it receives a new job from when the block template updates.
*/ */
//client.sendMiningJob(_this.jobManager.currentJob.getJobParams()); //client.sendMiningJob(_this.jobManager.currentJob.getJobParams());