Releases: blog post & release notes for 0.15.0#422
Conversation
When Travis CI builds a test version of the site, it doesn't have the /bin directory and so all links to that directory will be reported as broken (failing the build). This commit ignores all links to /bin so the site can build successfully.
|
|
||
| - **40% to 50% faster validation of blocks consisting of previously-seen transactions** as the result of repeating fewer validation steps when a previously-verified mempool transaction is later received in a block. | ||
|
|
||
| - **Moderate performance gains on some platforms** as the result of using hardware acceleration for some operations, such as support on modern computer processors for single-SHA256 hashing which results in about a 50% improvement each time one of Bitcoin's double-SHA256 hashes needs to be computed. This mainly benefits users of 64-bit Intel and AMD processors produced in 2008 or later. |
There was a problem hiding this comment.
Hmm, this is misleading due to the fact that this is disabled-by-default (and in the release binaries). Maybe mention it is just "in testing" or so?
There was a problem hiding this comment.
I don't feel comfortable saying something is in testing in a post about a release, so in the edit commit I'm about to push, I changed the example to mention the LevelDB CRC acceleration instead. It's less of a cool example IMO, but thanks for not letting me get away with misleading readers.
|
|
||
| We decided to take the later option, but we're also not going to wait the normal six months before the next major update. Instead, our next release will be a feature release that generates segwit-compatible addresses by default. This will be made available as soon as it has been written and thoroughly tested. | ||
|
|
||
| For those of you interested in technical details, our plan is to use P2SH-wrapped segwit addresses that are compatible with nearly all other wallets on the network. Bech32 native segwit addresses will either be implemented in a subsequent release (assuming no problems arise) or will only be available through expert options for now. |
There was a problem hiding this comment.
I believe the plan is to support Bech32 native segwit addresses for sending only until the rest of the community updates.
|
|
||
| The timing of segwit lock in and activation meant that we had to choose between either delaying the planned release of 0.15.0 and all its features described above or shipping 0.15.0 without a user interface defaulting to segwit. | ||
|
|
||
| We decided to take the later option, but we're also not going to wait the normal six months before the next major update. Instead, our next release will be a feature release that generates segwit-compatible addresses by default. This will be made available as soon as it has been written and thoroughly tested. |
There was a problem hiding this comment.
Maybe "our next feature release will generate..." instead? It is possible (though hopefully unlikely) that we have to do a bugfix release between now and then.
|
ACK, thanks for writing this |
|
Looks like this just needs the edate to be set, release notes synced, etc? |
|
Date updated and hashes from https://bitcoincore.org/bin/bitcoin-core-0.15.0/SHA256SUMS.asc added. |
|
Oh, forgot about the release notes. Those have now been updated as well. |
|
ACK. |
17104fd 0.15.0 release: update release notes (David A. Harding) e4fde30 0.15.0 release: add hashes for verification (David A. Harding) a5f3214 0.15.0 release: blog post: use final release date (David A. Harding) 66380fb 0.15.0 release post: edits suggested by BlueMatt (David A. Harding) 2673013 Releases: 0.15.0 blog post + release notes (David A. Harding) adb4f33 Travis: allow linking to bitcoincore.org/bin (David A. Harding)
5 participants
Preview: http://dg0.dtrt.org/en/2017/09/01/release-0.15.0/
Todo:
I'd especially appreciate review of this part about the plans for upcoming default segwit address support in the wallet.