Bitcoin Forum

Update Nov 13, 2014:
https://github.com/litecoin-project/bitcoinomg
0.9.3 plus many backports including watch-only.  Can you help to test especially the watch-only parts?

Bitcoin OMG is an unofficial branch of Bitcoin 0.9.x plus well-tested features backported from Bitcoin master.

It is the intent for Bitcoin OMG to be relatively conservative and usable after wide-spread testing, but please be careful and make backups of wallet.dat.

Historically this branch has helped to find bugs in Bitcoin master before they go into the next official release.

Feature Highlights - stuff backported from Bitcoin master

• Watch-only
• and lots more! read the commits in the source below

Git Source
https://github.com/litecoin-project/bitcoinomg/commits/OMG-0.9

Binary Downloads
http://download1.rpmfusion.org/~warren/bitcoin-0.9.3-OMG30/
Please use GPG to verify the integrity of each download.

Gitian Build Verification
https://github.com/litecoin-project/gitian.sigs.omg
You can follow the directions here to compare your gitian builds to better ensure that the build was not compromised.

Please :)

Hundreds of Litecoin users and a few developers have been testing Coin Control many many times for the past several months.  We suggested improvements to the Coin Control patch and I am quite satisfied with it right now.  We are shipping it in the next version of Litecoin and it is also in this unofficial Bitcoin branch.  We do not control what goes into the official Bitcoin source.  The best we can do is to give things more test exposure, which has a tendency to shake out bugs and improve things faster than sitting in a source review queue forever.

Awesome.

Just installed this one one of my EC2 nodes.

No-wallet mode is going to be nice for that application.

Thanks, @wtogami - well done!

Oh My God  :D

Well done!

Very nice! ::)

Wow you are either DPR or the FBI LOL

This is great - and it be merged into Bitcoin 0.9. Double win.

I think wtogami just got tired waiting for 0.9 with all the crap and decided to take the development into his own hands, putting in the useful features.
And that's exactly what this community needs.

@wtogami, make your own 0.9 and then 1.0 - people need it. Don't look behind.

p.s.
Will you pull my gitian sigs if I make some?

Nice Balance!  8)

I can't speak for Warren, but the general idea with gitian sigs is the more the better. We want lots of people from the community to use gitian - pretty much no matter who you are a gitian sig from you will increase someone's confidence in the build.

ok - I'm on it..

thx warren, good work. that was alot of work i guess  :) !

cant wait to see that for Litecoin.

all I can tell you ATM, before going to bed, is that my win32 sigs are different
https://github.com/piotrnar/gitian.sigs.omg/tree/master/0.8.5-OMG1-win32/piotrnar

though I took the HEAD - might that be the problem?
I would not think so, since its the same commit id as the tag: 5b474dacd8ae0d32f8c76dead7b7e9905d868ec3

EDIT:
Please see this post: https://bitcointalk.org/index.php?topic=320695.msg3441007#msg3441007

I'll redo my dep sha256sums when I return from an appointment today.  We need to get this right.

http://blog.litecoin.org/2013/10/litecoin-v0852-rc2-release-notes.html
Could you please verify if that bug exists in Litecoin v0.8.5.2-rc2 too?  We copied nearly all the patches from our Litecoin branch, so I'm curious if it needs to be fixed there too.

http://download1.rpmfusion.org/~warren/litecoin-0.8.5.2-rc2/
It was linked from that previous URL.  Please use GPG to verify the download.

Thanks.  We'll look into fixing this in the next build.  If you can find a patch to fix this it would help us out.  I personally don't have a Mac, and our Mac developer is very busy right now.

I rebuilt all deps from scratch and the resulting .exe's came out identical to my previous gitian sigs.  I'm asking others to join gitian build tests to figure out what is going on here.

Update: pooler's gitian sigs were identical to me.  There might be something wrong with your build environment.  Do you get equal sigs for standard Bitcoin giitan sigs (https://github.com/bitcoin/gitian.sigs)?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

0a95da7567d8a9dabeadffe5b74e4813072e4a2c6fc02c130bb383e949c78d22  bitcoin-0.8.5-win32-setup.exe
7b06dc9534b1e12849780be0481c629a8cc9d925ca8849dadc4d593a5d9e4f6f  bitcoin-qt.exe
dfeac5941ea70141d447c1d79c84d8c74c84126c6df49f2b08c6bff34c07a7b6  daemon/bitcoind.exe
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.15 (GNU/Linux)
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=
=HmNj
-----END PGP SIGNATURE-----

I wouldn't say the build environment is acting wrongly, isn't this just due to the builds not being deterministic ?

If anyone is wondering what I'm talking about, please check https://blog.torproject.org/blog/deterministic-builds-part-one-cyberwar-and-global-compromise or others related.

https://github.com/litecoin-project/gitian.sigs.omg
Two people independently built identical binaries.  It is deterministic.

It's fine that you are using gitian, but did you check the link I pasted (most likely the part two of that) ? Tor uses gitian too, but that link points out that gitian does not actually guarantee deterministic builds, and two people getting the same binary is not a counter-argument for that. It might be that this package is not affected by the issues pointed there, no matter the platform, but I don't know that.

One of the issues pointed in that link: On Windows builds, something mysterious causes 3 bytes to randomly vary in the binary.

Can you guarantee that this is not the case for this package ?

I make no guarantees about anything.  Identical build results from Gitian suggests that I didn't add anything secretly to this build.  It could also mean it is possible that all builds are equally compromised.  If that's true, then the standard Bitcoin gitian builds are also compromised.

If you get different results than these builds then there is something different about your build environment.  Are you using Gitian kvm or lxc?

Do you know if BItcoin master currently has this issue for Mac?  I am unable to test Mac.

If we reintroduced this problem by pulling a 0.9 patch, it's possible master has this issue now, or they subsequently fixed it with another patch.

Watch Only introduced some kind of coin selection issue.  It was removed for now until I can figure it out.  The remaining patches are very close to what has been shipping in Litecoin since mid-September.

Now I see what I did wrong.
I did not set the VERSION var properly (0.8.5 instead of 0.8.5-OMG1) and fetched wrong sources.

Shouldn't be doing it so late and announcing such a message just before going to sleep - my apologies.

I'm redoing it right now.

Yep, the binaries are all right.
I've also done OMG2: https://github.com/piotrnar/gitian.sigs.omg/tree/master/

Sorry again for messing up.

As a LTC dev shouldn't you have to create this for LTC first?

He did, basically he took the latest litecoin client and made it a bitcoin client.

Went to check out the Binary downloads and the link in the original post is just showing blank directories for each OS version:
http://download1.rpmfusion.org/~warren/bitcoin-0.8.5-OMG3/

Are the binaries going to be available again anytime soon?  Has anyone tested the Node Only/Disabled Wallet functionality with a local P2Pool node?

Oops.  Linux binary copied into the correct folder now.  Mac builds are done by another dev and hopefully will be available today.  I acquired a Macbook to do it myself but need to figure out the build environment.

Are you one of the rare users who regularly experiences the data corruption with Bitcoin-Qt on Mac?  We need testing of this build which includes two fsync patches and an upgrade to leveldb-1.13.  Does it make the corruption problems go away?  If you still experience corruption with OMG3, please report:

• MacOS version?
• Mac hardware?
• Brand/model of hard drive or SSD, especially if you replaced it with an after market drive

http://download1.rpmfusion.org/~warren/bitcoin-0.8.5-OMG3/macosx/
Binary download, please verify the GPG signature.

https://github.com/litecoin-project/bitcoinomg/commits/bitcoin-omg-0.8
Source

I usually encountered the data corruption issue on starting up with Bitcoin-Qt only on reboots, which necessitated a rebuild of the local copy of the blockchain so eager to try this beta... but...

Downloaded the OMG3 version, and already encountered a problem.

I tried running it with disablewallet=1 in the bitcoin.conf, but it crashed on startup with the following crashlog:
 (crashlog was too long to insert here even in code tags)
https://dl.dropboxusercontent.com/u/11139673/crashlogBTCOMG3.rtf

I'm running it on a Mac Pro, 16GB Ram on OSX 10.9 Mavericks.

I'm also running concurrently with a P2Pool node, namecoind for merged mining, and BFGMiner.

When I comment out the disablewallet=1 in the bitcoin.conf it starts up fine, so I'll keep it this way for now, but would definitely prefer to run it with the wallet disabled.

I have noticed that this happens to me only when my MacBook goes to sleep with Bitcoin-qt open. If I shut down Bitcoin-qt manually, everything works fine.

Hmm, disablewallet=1 mode is meant only for bitcoind.  bitcoin-qt is a wallet by definition.  It should probably refuse to run with an error message instead of crashing.  Thanks for the report.

Please report:

• MacOS version?
• Mac hardware?
• Brand/model of hard drive or SSD, especially if you replaced it with an after market drive

Wait a minute, you experienced the data corruption issue with OMG3 specifically?

No, previously with the current non OMG Bitcoin-qt beta 0.8.5.  

Haven't rebooted yet to see if it happens with your version.

Also the crash log linked has detailed system info for my Mac (system details are at the bottom of the crash log) and I mentioned in the post above it was a Mac Pro with 16GB ram on OSX 10.9 Mavericks.  

One note about my Mac Pro that probably has no bearing but I'll list here anyway, it's actually an original Mac Pro 1,1 (first intel ones) that was hacked with Chameleon to have an EFI64 boot loader and smbios.plist to look like a Mac Pro 3,1 so I could install Mavericks. It still runs circles around many current MacBook Pro's and iMacs.

So I take it if I wanted to run with disablewallet=1 I'll have to build bitcoind from your git.
No binaries available for bitcoind for the Mac?

bitcoin.org does not distribute Mac bitcoind.  32bit Mac bitcoind 0.8 does not even build.  64bit does build but nobody distributes it.

The primary reason to use disablewallet is to reduce RAM use, which is mainly useful on small headless servers or VM's where you would use bitcoind instead of bitcoin-qt.

Gotcha.  I'll stick with the OMG3 version of Bitcoin-qt then.   My reason for looking for bitcoind for Mac was to get as high an efficiency rating as possible with my P2Pool node.  Your version of Bitcoin-Qt has already reduced GetBlockTemplate Latency and has improved efficiency for my node noticeably.

gweedo,

Thank you for your report.  We are unable to reproduce this bug on 10.6.8, 10.8.x, 10.9.  Are you using our builds of Bitcoin-Qt OMG3 or building it yourself? 

https://github.com/litecoin-project/litecoin/issues/82
Yes thank you. Our Mac dev is looking into this.

https://github.com/bitcoin/bitcoin/issues/2770#issuecomment-27669022
It appears that there may be more than one MacOS X corruption issue.  It is difficult to know if any of them have been fixed when other issues still happen.  Any detailed reports of corruption with Bitcoin 0.8.5 OMG3 specifically would be helpful to Bitcoin dev.

https://github.com/litecoin-project/bitcoinomg/commits/0.8.5-OMG4
The source tree was cleaned up in preparation for the next release of Bitcoin OMG.

https://github.com/litecoin-project/bitcoinomg/tree/0.8.5-OMG3
The previous release source was moved to this branch.

https://github.com/litecoin-project/bitcoinomg/commits/0.8.5-OMG4
Added a tiny patch requested by BlueMatt for the pulltester.

What features do you want backported to Bitcoin OMG?  Please submit pull requests or just ask here.

https://bitcointalk.org/index.php?topic=337294
A few of you commented that you experience continued corruption even with all the patches in Bitcoin OMG3.  Gavin Andresen says it happens in Bitcoin master too.  A bounty is now funded by Gavin and Litecoin Dev.  Mac users, please consider adding to the bounty to increase incentive of someone to fix this sooner.

https://github.com/litecoin-project/bitcoinomg/commits/0.8.5-OMG4
This branch is very close to what will be in the next build of Bitcoin OMG.  Please test and report back if you find any problems.

Any requests to include any other particular patches in OMG?

https://bitcointalk.org/index.php?topic=337294.msg3718821#msg3718821
If you previously experienced corruption with MacOS X Bitcoin or LItecoin, please test these MacOS X binaries and report back!

Please see the OP for Bitcoin 0.8.5-OMG8 (https://bitcointalk.org/index.php?topic=320695.0) with potential fixes for the long-standing MacOS X corruption issue (https://bitcointalk.org/index.php?topic=337294.0;all).

Good work Warren. I've been impressed with your work for Litecoin, I didn't know you were doing things for Bitcoin too! I've been pretty busy, but I will try to help you test things out more often in the future (I despise Apple though, so I can't help you there.) You are the best thing to ever happen to Litecoin, seriously. I hope that the Bitcoiners appreciate your work here too.

Thanks!

Hey Warren, with the release of 0.8.6rc1, does the main branch include any your changes from the OMG releases? I see their changelog mentions a fix for the Mac Corruption issues too.

Alternately, if features like "disable wallet" and the optimizations aren't included, will there be an effort to make an OMG version of 0.8.6 merging the fixes and features like the increased mining block size from the main branch?

Features will never end up in official 0.8.x releases, as these are fix-only stable releases. "Disable wallet" and such will be in 0.9.0.
Of course -OMG is a bit different here as they attempt to backport features so they're a kind of hybrid stable unstable version.

OMG10 already has all bug fixes in 0.8.6rc1.  The minor fee related rule changes will be included in 0.8.6-OMG after 0.8.6 is released.

Status Update: OMG10 already has all bug fixes that were included in 0.8.6.  OMG will be rebased on "0.8.6" only after we have substantive fixes or new features to make a new release worthwhile.

How are the .9 private keys different?

Thanks for this OMG client with Coin Control.

But I wonder how much can I rely on the tx size/fee estimation in Coin Control?
I just noticed that tx size estimated is 300 bytes less than actual tx size in some cases.
Maybe that's a minor bug?

The underestimation in tx size occurs when multisignature address is involved.
This tx
https://blockchain.info/tx/d84613726465c27ce641bd20a4385dfada887effee99d77b01548e13403066d2
contains 5 inputs and 2 outputs.
The address starting with 33 is a 3-of-3 multisignature address.
The actual size is 1264 bytes, but in Coin Control it is reported as 946 bytes.

Coin Control does not check the address type, I guess.
Multisig address is larger than normal address, and ignoring address type leads to this underestimation in tx size.

Hope this could be fixed soon, and we can expect a reliable fee estimation. :D

Very good advice, listen to gweedo.

Thanks for suggestions. I understand the risk using coin control. I will double check when using it.

So, for fees, I still need to estimate the fee myself even when using coin control?

I am interested in using CoinControl and am wondering what direction to go...

I'm running standard Bitcoin-Qt ver 0.8.6 at the moment.

Can I just run an installation of win32  0.8.5-OMG10 and overlay onto my existing  0.8.6 version of bitcoin-QT?

or should I do a seperate install of 0.8.5-OMG10 into its own directory, then copy over existing blockchain and wallet.dat directories?

Which method would least troublesome?


I would of course do full backup of existing sys before trying this.

Bitcoin OMG10 has all the bug fixes that were in Bitcoin-Qt 0.8.6.  It is fully compatible with the data and wallet of the standard Bitcoin-Qt.  That being said you should always make backups of wallet.dat on a regular basis.

https://github.com/litecoin-project/bitcoinomg/commits/0.8.7-OMG12
tag v0.8.7-OMG12 is similar to Luke-Jr's 0.8.7rc1 with the following differences.

* Removed the three OP_RETURN pruning patches.  It seemed to have made it incompatible with indexes from 0.8.x.
* Reduced the relay fee by 10x to match the behavior of Bitcoin 0.9.0.
* Added the getnormalizedtxid RPC that was not accepted in Bitcoin Core.  This is an experiment.
* Many 0.9 features backported and carried forward from earlier versions of Bitcoin OMG.

You really should be using Bitcoin 0.9.  But if for some reason you can't, this version should be fully compatible with data from Bitcoin 0.8.x while including the important bug and security fixes from 0.9.  Most of the credit for this round of patches goes to Luke-Jr.

Status Update March 26th, 2014:
v0.8.7-OMG12 contains all of the security and bug fix workarounds pertaining to transaction malleability that were included in Bitcoin 0.9.  You really should be using Bitcoin 0.9.  Use this only if you need 0.8.x for some reason and you understand how to make backups.  At the moment this branch is only lightly tested, so please report back if you run into any issues.

well, you just saved 380MB of my memory :D thanks for disablewallet=1 :P

before adding it to litecoin, better have a careful test.

Update Oct 1st, 2014:
https://github.com/litecoin-project/bitcoinomg
0.9.3 plus many backports including watch-only.  Can you help to test especially the watch-only parts?

Bitcoin-0.9.3-OMG27
https://github.com/litecoin-project/bitcoinomg/commits/0.9.3-OMG27
http://download1.rpmfusion.org/~warren/bitcoin-0.9.3-OMG27/linux/
Please use GPG to verify the integrity of each download.

These builds are not well tested.  Please report any problems in a Github ticket.

Bitcoin-0.9.3-OMG30
https://github.com/litecoin-project/bitcoinomg/commits/OMG-0.9
http://download1.rpmfusion.org/~warren/bitcoin-0.9.3-OMG30/
Please use GPG to verify the integrity of each download.

Major update!