Compiling NameCoin on RedHat and CentOS

Post Reply
dolphin
Posts: 16
Joined: Wed Jul 02, 2014 6:52 pm
os: windows

Compiling NameCoin on RedHat and CentOS

Post by dolphin » Mon Oct 06, 2014 11:52 pm

A while back I've submitted a pull request to enable compilation and operation of NameCoin software on RedHat derivatives (CentOS, Fedora etc.). As some of you, who followed that request, may recall, the key issue with these platforms is that their OpenSSL version is missing the Secp256K1 algorithm, used by BitCoin.

My initial patch attempt relied on custom, 3-rd party implementation of the Secp256K1. However, as core developers pointed, that was not as desirable as porting similar implementation in BitCoin, which relies on libsecp256k1 that is part of BitCoin tree. I took away to work on that, which I did not have a chance to do until a few days ago.

While working on the recommended option I've ran into a question that has broad implications and seems worth community discussion and guidance.

Key related functionality in BitCoin evolved since the NameCoin fork. BitCoin added support for so-called BIP32 -Hierarchical Deterministic Wallets before addition of libsec256k1 integration. Hence libsecp256k1 changes in BitCoin are not quite compatible with the key related functionality in NameCoin - relatively intrusive changes are necessary to make NameCoin compile and run with BitCoin changes for libsecp256k1, i.e. changes that enable BitCoin to compile and run on RedHad based Linuxes. I could certainly do the changes and publish another pull request, but before I do that, I thought to ask the community on other available options that may be even better. So, here is what can be done:

a. Port not only libsec256k1 changes but also BIP32 implementation. This will trigger bigger changes but code-wise will be quite close to BitCoin. May even be identical for the most part.

b. Port just libsecp256k1 - this is likely to translate to new code that is specific to NameCoin but based on BitCoin ideas. Changes will touch lesser files/lines than the option (a) but will not match those, in BitCoin tree

c. Go back to the initial proposal that utilizes a custom made Secp256K1 implementation, borrowed from discussions on BitCoin forums.

Any other options that I've missed?
NameCoin and Security News
Chrome extension for Gmail encryption - protect your identity and E-mails using NameCoin

biolizard89
Posts: 1971
Joined: Tue Jun 05, 2012 6:25 am
os: linux

Re: Compiling NameCoin on RedHat and CentOS

Post by biolizard89 » Thu Oct 09, 2014 2:30 am

In my opinion, rebasing on current Bitcoin is a better way to achieve this without wasted effort. I'm curious what Daniel and Ryan think about this though.
Jeremy Rand, Lead Namecoin Application Engineer
NameID: id/jeremy
DyName: Dynamic DNS update client for .bit domains.

Donations: BTC 1EcUWRa9H6ZuWPkF3BDj6k4k1vCgv41ab8 ; NMC NFqbaS7ReiQ9MBmsowwcDSmp4iDznjmEh5

dolphin
Posts: 16
Joined: Wed Jul 02, 2014 6:52 pm
os: windows

Re: Compiling NameCoin on RedHat and CentOS

Post by dolphin » Thu Oct 09, 2014 6:21 pm

I agree, that would be ideal. That being said, it feels like there is no defined timeline for rebased NameCoin release... at least, I was unable to find date. It also feels that it won't happen soon and all that time RHEL/CentOS/Fedora users will have to jump through hoops which will artificially restrict the growth of the community.
NameCoin and Security News
Chrome extension for Gmail encryption - protect your identity and E-mails using NameCoin

Post Reply