Topic: [DVC]DevCoin - Official Thread - Moderated - page 194. (Read 1058949 times)

Is there on devtome a howto add merged mining capability for sha256D bitcoin clones?

ok so https://github.com/sidhujag/devcoin/commit/e646ca6740862d6d5914d83b7e69bf31f95c25dc the initial 0.8.5 commit seems to be working ok. The daemon loaded (after a little compile coaxing), got all the blocks up to date, and p2pool seems to have stopped complaining that devcoind isn't playing with him.

Moving up the commit list now.

As a side note, from this point onwards: https://github.com/sidhujag/devcoin/commit/36d595726636255a55c62780a671386c0cc7c505 (Updated unix makefile to match mingw)

make -f makefile.unix clean always gives the following result:

so far I've mostly just been ignoring it, but I just discovered it was working without errors before this commit. Previous commit to the one above:


I don't know if that's actually important though.

Possibly more important, possibly not, while compiling the dec23 commit I noticed this warning:


I've seen that in the newest one too, could p2pool be sending one kind and the daemon expecting the other? The error from devcoind was "ERROR: CTransaction::CheckTransaction() : txout.nValue negative", so could there be a buffer overflowing somewhere? Did you declare any ints which needed to be unsigned by any chance?

EDIT:

hmm, pretty much everything I try compiling after that point is giving me a weird problem (until the most latest builds). The devcoind that's produced is...not there. I mean it compiles and puts a file there, but when I try to run it (as root or user, and no matter where its placed), it just says "bash: ./devcoind: No such file or directory"...

The old version is like 0.3.x new one is 0.8.5.. good idea though.. see if it haooens in old one.

Also its a good idea to cooy the data dir so you can recopy it over when you change versions.

Hmm it might be related to this.

I remember something about what I was doing yesterday - I had actually set the user as Devcoinrpc and the pool was trying to log in as devcoinrpc...devcoind may not have been crashing initially because p2pool wasn't even connecting. I changed it when I started looking at the p2pool logs, and noticed it was saying "Authorization failure" every time it tried to access devcoind. That may actually have been when it started terminating process, so I'm thinking it's less likely server problem atm.

I'm going back to initial 0.8.5 and trying that now

I think so, Im new to git I use HG at work and used SVN before that. but its pretty easy I use the windows client and it shows nicely the latest changes and diffs in a GUI and I can go rollback to any version.... and using the online github you can get the checkout id and go back just the way you describe.

When you do go back I would do a make -f makefile.unix clean and also get rid of the obj files in the src/obj folder... i dont think clean does it.

During my tests I set up a mining proxy and was able to call getauxblock in the node using the test in a box approach. I will try it again to see if it crashes this time.

On your end if you try on a few older versions, if it crashes the same way then its probably the port over to bitcoin 0.8.5 somewhere, or it could still be the server config since you said it was fine before... seems like when p2p connects it tries to talk and it crashes...

I had a similar issue when the android wallet would connect the node would crash and that was because it would try to deserialize a aux block that was null, and deferencing null crashes. So I added a null check there and then everything worked... but i was never sure why its getting null there, why would it try to pull up a block from the disk, deserialize it and find that the merged-mine section was null? Anyways I think beginning with commit 13cf0dd190e59b3ebf83f638872db6b2b9a5bc7f on jan 13 I tried to fix this issue, so maybe first try before this commit, I think it was stable on Dec 31. Before that we started to play with min fees and changed the dust fee definition.... so that might be an issue... before that we have initial implmentation.. so we have a few to try out and it would be interesting if as we go back that it works.

Ah it was there earlier, I had to restart the pool to detach devcoin merging, and I think it must have wiped the two blocks p2pool found from local memory, so they're not on the local list any more - they'd still be on the p2pool network.

EDIT: hmm just had a look myself, couldn't find it...hmm...I wonder if it got bumped.

EDIT: If the payment doesn't arrive and you should have received it, worry not, I'll cover it until I know the server's working properly (and I get the devcoind thing sorted out)

while I can see the share in 'payout' page, but cannot find in the 'last blocks', what's the problem? or did I missing something?

Ok I'm a bit git nooby, this is what I'm doing to get a previous version:

git checkout 016bd1e7f8c289b05158e8005a69ff33bb262573 (whichever branch I want to use)
cd ~/devcoin/src && make -f makefile.unix

that then compiles what I think is the branch version, whose executable I can use to run. Is this the correct way to do it?

It's too late to get into round 32, you'll go into round 33. We need about a week to approve new writers, each writer is checked by three people.

It is less than half the work. If blocks can be mined with your code you will also get the pledges. Given that the shares this month will be less than 130,000, the pledges of at least 6 million are worth at least 46 additional shares, in total probably around a hundred. So 24 would be about a quarter of that.


Sidhujag is getting admin salary for developing the devcoin source. On top of that for ongoing open source development, he also qualifies for the bitcoin share list. Sidhujag, please contact Hunterbunter to go on that list:
http://www.devtome.com/doku.php?id=earn_devcoins_by_developing

I made fee changes that may have broke it.. maybe its a good idea to work backwards from the first stable release like 1.0.3 or whatenot looking at history.

What's your devtome username again?

I agree with you, and I'll split the bounty with you if we can figure this out together, and the pool manages to get a raw devcoin block. Whatever amount you think is fair. It may not even be a problem with the source at all, and just a problem with my server, which another pool might not have. The server seemed to have been working fine up until I announced it yesterday, and then my isp replaced faulty ram and it died a horrible death in many ways. Everything got back up except devcoind, which didn't have a problem up until then. It's probably not a source problem, but I need source advice to try and identify a potential problem with my server. Either way, if you can help me fix it, everyone wins.

On a side note, I do also agree with emfox's suggestion that the devcoin wallet/daemon developers need to be well taken care of. Our whole network runs off it! One could easily spend 40 hours a week on something like the source and make it awesome + features.


I'm not sure if it's when it tries to create a block, it looks like it's failing when it's trying to check the transactions. There might be something incompatible with what p2pool thinks its asking and what devcoind is expecting. Have you (or anyone else) made any changes to the transaction code after the fork?

What looks like is happening is the daemon starts up fine, gets the new blocks fine, and then as soon as p2pool throws what I linked above at it (merged getauxblock), it refuses connections until it terminates. We need to figure out what's causing this:

whenever p2pool tries to getauxblock on it. (and more specifically:  p2pool.util.jsonrpc.NarrowError: -1 CreateNewBlock() : ConnectBlock failed) source for p2pool is https://github.com/forrestv/p2pool/

Thanks, glad someone is reading them haha

I tried searching myself and nothing came up so I guess noone has put me on the list yet.

I just signed up four or five new writers. If you messaged me in the past two days and still haven't gotten a reply, please message me again as that means I missed you.

I'm getting enough requests that I'm processing them in batches once a day or every other day, so if you don't hear back right away, that's why.

Welcome aboard to all the new writers. It appears word about the Devtome opportunity is getting out.

Would that not fall under the share for being a full time Open Source developer? Assuming there were shares given out for the original development? Please don't consider this a vote pro or con... just a question.

- Nova

Do the references / links have to be to outside sources or can it be my previous articles on devtome?

Lot of my articles posted recently are for trading systems which are either created by me or my modifications of trading systems I've found from publicly accessible threads on trading forums. Referencing this would be an issue as the source is basically my mind? An added problem is that even for the strategies I've modified, for a lot I can't even remember where I got the original idea from. As there is no copyright on trading systems I never really bothered to keep track of where I borrowed ideas from.

I plan to change up the type of content I post starting from the next round so I don't expect this to be an issue going forward but obviously I wouldn't want to get stuck with a bad rating in the meantime.

Calm down man, I think it's more hard work developing wallet than testing it, but having mining pool is also important.

I think I'd like to propose bounty of 12 shares every round for sidhujag, after his wallet made official and as long as he continuously maintain it.

Seems it throws a fit when creating a block? Doesntthis get called when doenloading blocks from the chain? whats the difference. Is it the transaction stuff tht happens after checkblock is called?

Seems there is an issue in the creation of a new block... Hmm how would i debug this one?

Good thing is it is consistent.. so the p2ppool is sending the command right away amd I can put a bunch of console output statements to see what is going on.. I can add these to github ill tell you when to rebuild.

You really think its half the amount of work to test a pool than it is to develop the source? So does this mean he would have to dive into the source to figure out whats wrong if that os the case? Seems most of the hard work is in the code since that would be the problem and not the setup of a generic p2pool.