Update libsecp256k1 and use RFC6979 extra entropy for test cases #5952
Conversation
They may not contain all necessary characters for a language
The main thread spends time waiting for the DetectShutdownThread. So why not just run this waiting loop function in the main thread? One thread-stack less saves 4MB of virtual memory on 32-bit, and 8MB on 64-bit.
…uilds Also increase temp dmg filesize to account for a bigger background image
Also do a bit of cleanup: - Make the background name a variable so it's easier to change - Add proper make dependencies
f0172bf osx: bump build sdk to 10.9 (Cory Fields)
1d84aea Coin Control: Use U+2248 "ALMOST EQUAL TO" rather than a simple tilde (which may be mistaken for a negative sign) (Luke Dashjr)
2ce63d3 MOVEONLY: Move struct CBlockTemplate to miner.h (from main.h) (Luke Dashjr)
No longer necessary since bitcoin#5161 / 845c86d.
e7cfcc8 Remove custom pkg.m4 script. (randy-waterhouse)
…SSharedFileListItemCopyResolvedURL() instead
0eade74 fix crash: CoinControl "space" bug (fsb4000)
0cc0d8d Get rid of the internal miner's hashmeter (jtimon)
…nstants for database keys
2fa9a8e Make empty byte arrays pass CheckSignatureEncoding() (Peter Todd)
44bc988 [Wallet] Do not flush the wallet in AddToWalletIfInvolvingMe(..) (Cozz Lovan)
ff09e31 sleep-wait on genesis block during init with -reindex (Matt Corallo)
785bb81 [Qt] remove size grip to get rid of the right margin (Jonas Schnelli)
7b782f5 RPCWallet: Notate all account stuff as deprecated (Luke Dashjr)
5fdc5b0 depends: latest config.guess and config.sub (Michael Ford)
d698ef6 Consensus: Refactor: Decouple pow.o from chainparams.o (Jorge Timón) bd00611 Consensus: Refactor: Introduce Consensus::Params class (Jorge Timón)
fc72020 don't trickle for whitelisted nodes (Ruben de Vries)
5983a4e Add a NODE_GETUTXO service bit and document NODE_NETWORK. Stop translating the NODE_* names as they are technical and cannot be translated. (Mike Hearn)
Instead of manually tweaking the deterministic nonce post-generation, pass the test case number in as extra entropy to RFC6979.
|
ACK. I was uncomfortable with (and complained about) the "test nonce" approach used previously as it used seriously insecure linearly related nonces... this was fine for tests but it seemed like a matter of time before someone wanted to use it to derandomize regular signing (and, indeed, someone recently asked about that). |
There was a problem hiding this comment.
Happy to get rid of this function.
|
ACK to bitcoin core changes, haven't been able to review all secp256k1 changes. |
|
Would it be interesting to have a more automated way for verifying that the contents of the subtree directory matches the contents of the listed commit of the subtree repository? |
|
Yes, a script that compares our secp256k1 directory with an upstream repository would be useful. |
|
upstream secp256k1 commit is 1897b8e
|
|
Another way is checking the git tree id: (in master) (in secp256k1, with 50cc6ab the commit id of the last secp256k1 import) |
|
Checked the subtree using #5965's ACK |
|
I did a first pass review of all changes, but the number of changes is huge! Most of them are simply the removal of a variable declaration from its first use. However, many reviewers of this commit would be helpful. |
|
@SergioDemianLerner This is a subtree merge from the libsecp256k1 repository ( https://github.com/bitcoin/secp256k1/ ), merging in months of work; you can see the individual changes there (and the review process there). The big stirring you're reffering to was the change of the codebase to be strict standards conformant C89. It's much easier to review from its actual repository because e.g. with the changes split out you can verify various changes that shouldn't have changed the binaries didn't and such. |
|
In addition, with the script added in #5965 you can verify that the subtree
corresponds to the result of the commits in the external repository.
|
18 participants
Included libsecp256k1 changes (multiple authors):