Page 6 of 6

Re: Officially releasing Namecore

Posted: Wed Aug 19, 2015 7:33 pm
by phelix
fyi: namehistory for me made a difference of about 175MiB (@ height ~ 244737)

Re: Officially releasing Namecore

Posted: Thu Nov 19, 2015 5:07 pm
by phelix
Do I understand the current status correctly that there are not Linux binaries yet?

Re: Officially releasing Namecore

Posted: Fri Nov 20, 2015 3:22 am
by josephbisch
phelix wrote:Do I understand the current status correctly that there are not Linux binaries yet?
Sort of correct. There were just the binaries that were made for the OpenSSL consensus issue, but none were made since then, and Domob did a lot during that time. So new binaries should be made.

We could use a tagged version on GitHub, so that we have some consistent commit to work off of for not only the various platforms, but also for the various builders building reproducibly.

Re: Officially releasing Namecore

Posted: Fri Nov 20, 2015 1:37 pm
by domob
josephbisch wrote:We could use a tagged version on GitHub, so that we have some consistent commit to work off of for not only the various platforms, but also for the various builders building reproducibly.
I'm open for tagged commits (signed by me), but I think we need a "plan" to decide which ones to tag. I'm planning on branching off a "stable" version when Bitcoin does that for 0.12. For now, we could simply tag "one random" commit. We could also create a "0.11.9" (or something) branch for a release before that, but then people need to commit time testing that particular release. Should we do that?

Re: Officially releasing Namecore

Posted: Fri Nov 20, 2015 4:23 pm
by erkan
>but then people need to commit time testing that particular release. Should we do that?
Are there existing test plans/cases ?

Re: Officially releasing Namecore

Posted: Sat Nov 21, 2015 10:25 am
by domob
erkan wrote:>but then people need to commit time testing that particular release. Should we do that?
Are there existing test plans/cases ?
Of course, there are both unit and regression tests in place. I actually run them before even pushing any changes, so that all published commits "should" be good with respect to them. For an official release, I would expect that actual users try the software out in the wild in addition. (Otherwise, everyone can just take the newest commit and run the unit tests instead of asking for releases.)

Re: Officially releasing Namecore

Posted: Sun Nov 22, 2015 11:06 pm
by biolizard89
domob wrote:
josephbisch wrote:We could use a tagged version on GitHub, so that we have some consistent commit to work off of for not only the various platforms, but also for the various builders building reproducibly.
I'm open for tagged commits (signed by me), but I think we need a "plan" to decide which ones to tag. I'm planning on branching off a "stable" version when Bitcoin does that for 0.12. For now, we could simply tag "one random" commit. We could also create a "0.11.9" (or something) branch for a release before that, but then people need to commit time testing that particular release. Should we do that?
My opinion is that releasing official Namecoin binaries that are not based on official Bitcoin release tags is sufficiently risky that we should avoid it. We did so for the OpenSSL consensus bug because it was an emergency, but we later found some nasty bugs in those binaries (wizkid from Eligius ran into problems with the wallet code and mining code, which Daniel traced to a bug in upstream Bitcoin master Git).

I would suggest doing what Daniel suggests: branch off of a Bitcoin stable release for our stable releases. The Bitcoin Core devs do an absolutely phenomenal job at QA testing, which we simply do not have the resources to match.

Re: Officially releasing Namecore

Posted: Tue Nov 24, 2015 9:18 am
by phelix
I just needed a binary for the API server. Getting the existing binary from Google was a huge pita so I put a copy up here: https://namecoin.org/files/namecoin-0.1 ... x32.tar.gz

Works fine on 64bit Debian so far.