< bitcoin-git> [bitcoin] dongcarl opened pull request #18478: build: Automatically include both `git`-tracked and bootstrapped files. (master...2020-03-adventures-in-reconciling-automake-and-git) https://github.com/bitcoin/bitcoin/pull/18478
< bitcoin-git> [bitcoin] amitiuttarwar closed pull request #16698: Mempool: rework rebroadcast logic to improve privacy (master...rebroadcast) https://github.com/bitcoin/bitcoin/pull/16698
< ryanofsky> gwillen: where is the option on github commit pages to "create a branch from it"?
< ryanofsky> the only way i've found to download orphan commits on github not on a known tag or branch is to use web api and reconstruct tree and commit objects locally
< gwillen> ryanofsky: yeah it's slightly tricky, I'm not sure whether or how you can easily reach it directly from the GUI
< gwillen> but e.g. https://github.com/bitcoin/bitcoin/commit/208b9f00140e0243707e98e1f58a0fcd07f427ce is a commit that's been force-pushed over on a PR, and this is the page that you get to if you click the commit ID
< gwillen> then if you change the "commit" in the URL to "commits", you get a different view, like, the branch view from that tip
< gwillen> then if you click "Tree: NNNNNNN" you should be able to commit a branch, if you have that permission on the repo you're looking at
< gwillen> er, create* a branch
< kallewoof> I can't restart travis jobs for some reason.
< sipa> kallewoof: try logging out and back in
< kallewoof> sipa: that fixed it. thanks!
< sipa> i've experienced it as well since a few days
< sipa> i don't know what causes it
< fanquake> sipa: can you block Domel1985
< sipa> again?
< fanquake> In the bitcoin/bitcoin repo now
< sipa> ah, from different repo
< sipa> Done.
< bitcoin-git> [bitcoin] luke-jr opened pull request #18479: RPC: Show fee in results for signrawtransaction* for segwit inputs (master...rpc_sign_show_fees) https://github.com/bitcoin/bitcoin/pull/18479
< wumpus> vasild: yup tarballs work fine for 'orphaned' commits, downloading patches/diffs also works, but never found a trick to be able to do it with git itself
< wumpus> (or alternatively, to download a tarball w/ .git directory, or say, a git bundle for a commit range)
< vasild> https://stackoverflow.com/questions/28958327/does-github-remember-commit-ids/28958418#28958418 is relevant. It's like gwillen's way, but using the github api. Requires write permissions though, to create a new branch from the orphaned commit.
< gribble> https://github.com/bitcoin/bitcoin/issues/28958418 | HTTP Error 404: Not Found
< vasild> I have already forked the repo in my github account, so I cannot fork it a second time to check whether the orphaned commits will be copied into my fork (where I would have write permissions)
< wumpus> oh interesting! that's at least something, though, I'd most frequently wanted it for other people's commits and don't tend to have write access to them
< lucaferr> How does peer protocol command "notfound" behave when a "getdata" request only partially matches. Will it return a partial result plus "notfound", or will it only return "notfound"?
< wumpus> haaaah fork their repository first, of course
< wumpus> lucaferr: it either returns a (partial) result or "notfound" never both IIRC
< lucaferr> wumpus: great, thank you
< vasild_> wumpus: but when I go to https://github.com/practicalswift/bitcoin and try to "Fork" it into my account I am redirected to vasild/bitcoin (nothing gets forked)
< jonasschnelli> Can we consider #18160 a bugfix? I'd like to get that into 0.20?
< gribble> https://github.com/bitcoin/bitcoin/issues/18160 | gui: Avoid Wallet::GetBalance in WalletModel::pollBalanceChanged by promag · Pull Request #18160 · bitcoin/bitcoin · GitHub
< sipa> lucaferr: it wul
< sipa> lucaferr: it will respond with tx/blocks for the found things, and a notfound for the rest
< lucaferr> sipa: ah, ok. So the arity will match between the requested inventory vectors and the notfound inventory plus data messages (tx or block)?
< wumpus> sipa: oh you're right, I was confused with how 'reject' messages used to be handled
< lucaferr> wumpus, sipa: thanks to both for replying :)
< bitcoin-git> [bitcoin] theStack opened pull request #18481: test: add BIP37 'filterclear' test to p2p_filter.py (master...20200330-test-add-BIP37-filterclear-message) https://github.com/bitcoin/bitcoin/pull/18481
< wumpus> jonasschnelli: yes, I think it's a fix of some kind
< wumpus> jonasschnelli: I was holding it off because of the idea to merge #17905 first, which isn't merged yet, seems too late to merge both now (and do a last-minute rebase of the latter)
< gribble> https://github.com/bitcoin/bitcoin/issues/17905 | gui: Avoid redundant tx status updates by ryanofsky · Pull Request #17905 · bitcoin/bitcoin · GitHub
< wumpus> so I don't know what's best now
< promag> #18338 #18160
< gribble> https://github.com/bitcoin/bitcoin/issues/18338 | Fix wallet unload race condition by promag · Pull Request #18338 · bitcoin/bitcoin · GitHub
< gribble> https://github.com/bitcoin/bitcoin/issues/18160 | gui: Avoid Wallet::GetBalance in WalletModel::pollBalanceChanged by promag · Pull Request #18160 · bitcoin/bitcoin · GitHub
< promag> these should be ready to merge
< wumpus> ok!
< promag> #16923 another fix that could also go
< gribble> https://github.com/bitcoin/bitcoin/issues/16923 | wallet: Handle duplicate fileid exception by promag · Pull Request #16923 · bitcoin/bitcoin · GitHub
< promag> #18160 is easer to backport to 0.19 without 17905
< gribble> https://github.com/bitcoin/bitcoin/issues/18160 | gui: Avoid Wallet::GetBalance in WalletModel::pollBalanceChanged by promag · Pull Request #18160 · bitcoin/bitcoin · GitHub
< wumpus> looking at them now, doesn't help that my main computer's case fan decided to die today
< luke-jr> :|
< wumpus> ordered a new one and it'll probably get here tomorrow, but who knows these days
< bitcoin-git> [bitcoin] laanwj pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/965c0c37d54c...f2880e21eff2
< bitcoin-git> bitcoin/master 0933a37 João Barbosa: gui: Avoid Wallet::GetBalance in WalletModel::pollBalanceChanged
< bitcoin-git> bitcoin/master f2880e2 Wladimir J. van der Laan: Merge #18160: gui: Avoid Wallet::GetBalance in WalletModel::pollBalanceCha...
< bitcoin-git> [bitcoin] laanwj merged pull request #18160: gui: Avoid Wallet::GetBalance in WalletModel::pollBalanceChanged (master...2020-02-avoid-getbalance) https://github.com/bitcoin/bitcoin/pull/18160
< luke-jr> even though it's internal to the PC, I'd take precautions to sanitize it somehow
< wumpus> yeah
< bitcoin-git> [bitcoin] laanwj pushed 3 commits to master: https://github.com/bitcoin/bitcoin/compare/f2880e21eff2...9a2b5f22c1f6
< bitcoin-git> bitcoin/master ab31b9d Russell Yanofsky: Fix wallet unload race condition
< bitcoin-git> bitcoin/master 41b0baf João Barbosa: gui: Handle WalletModel::unload asynchronous
< bitcoin-git> bitcoin/master 9a2b5f2 Wladimir J. van der Laan: Merge #18338: Fix wallet unload race condition
< bitcoin-git> [bitcoin] laanwj merged pull request #18338: Fix wallet unload race condition (master...notify-shared) https://github.com/bitcoin/bitcoin/pull/18338
< bitcoin-git> [bitcoin] MarcoFalke pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/9a2b5f22c1f6...d52ba21dfff9
< bitcoin-git> bitcoin/master 0055922 Sebastian Falbesoner: test: add BIP37 'filterclear' test to p2p_filter.py
< bitcoin-git> bitcoin/master d52ba21 MarcoFalke: Merge #18481: test: add BIP37 'filterclear' test to p2p_filter.py
< bitcoin-git> [bitcoin] MarcoFalke merged pull request #18481: test: add BIP37 'filterclear' test to p2p_filter.py (master...20200330-test-add-BIP37-filterclear-message) https://github.com/bitcoin/bitcoin/pull/18481
< emilengler> By the way, is someone working on the switch from C++11 to C++17 right now? Like #17904? Otherwise I would try to pick it up
< gribble> https://github.com/bitcoin/bitcoin/issues/17904 | [wip] build: Switch from C++11 to C++17 by practicalswift · Pull Request #17904 · bitcoin/bitcoin · GitHub
< provoostenator> emilengler: usually we have to wait for ancient distros to support it, luke-jr tends to be up to date on those...
< aj> i thought we were just deferring 'til 0.20 was done?
< emilengler> aj: It has the milestone of 0.21
< aj> #16684 has info about distros
< gribble> https://github.com/bitcoin/bitcoin/issues/16684 | Discussion: upgrading to C++17 · Issue #16684 · bitcoin/bitcoin · GitHub
< emilengler> The only problem was CentOS 7 then but CentOS 8 is already out
< promag> so how about #16923?
< gribble> https://github.com/bitcoin/bitcoin/issues/16923 | wallet: Handle duplicate fileid exception by promag · Pull Request #16923 · bitcoin/bitcoin · GitHub
< luke-jr> dongcarl: for #16684, note Gentoo is at 9.2.0
< gribble> https://github.com/bitcoin/bitcoin/issues/16684 | Discussion: upgrading to C++17 · Issue #16684 · bitcoin/bitcoin · GitHub
< jonatack> promag: re-acked #16923. lone travis failure (mempool_reorg.py) is unrelated
< gribble> https://github.com/bitcoin/bitcoin/issues/16923 | wallet: Handle duplicate fileid exception by promag · Pull Request #16923 · bitcoin/bitcoin · GitHub
< wumpus> yes maybe with 0.21 we can switch to c++17, I hope so
< wumpus> restarted travis run for 16923
< bitcoin-git> [bitcoin] achow101 opened pull request #18484: Correctly compute redeemScript from witnessScript for signrawtransaction (master...signrawtx-p2pkh-p2wsh) https://github.com/bitcoin/bitcoin/pull/18484
< bitcoin-git> [bitcoin] hebasto opened pull request #18485: test: Add mempool_updatefromblock.py (master...20200331-test-mempool) https://github.com/bitcoin/bitcoin/pull/18485
< bitcoin-git> [bitcoin] MarcoFalke pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/d52ba21dfff9...72ef36656439
< bitcoin-git> bitcoin/master fa3cc0b MarcoFalke: test: Remove unsafe BOOST_TEST_MESSAGE
< bitcoin-git> bitcoin/master 72ef366 MarcoFalke: Merge #18472: test: Remove unsafe BOOST_TEST_MESSAGE
< bitcoin-git> [bitcoin] MarcoFalke merged pull request #18472: test: Remove unsafe BOOST_TEST_MESSAGE (master...2003-testNoBoost) https://github.com/bitcoin/bitcoin/pull/18472
< hebasto> travis did not handle #18485. Could reopening help? Or just wait?
< gribble> https://github.com/bitcoin/bitcoin/issues/18485 | test: Add mempool_updatefromblock.py by hebasto · Pull Request #18485 · bitcoin/bitcoin · GitHub
< jeremyrubin> hebasto: good work on that test!
< hebasto> jeremyrubin: how large tx graph do you suggest?
< jeremyrubin> Up to you -- You can see the code in the ComplexMempool benchmark for "generate a large tx graph" idiom.
< jeremyrubin> I think the question is less so largeness but the 'weirdness' of the structure
< jeremyrubin> You want it to have a lot of stuff where a child is also a grandchild
< hebasto> "child is also a grandchild" included already :)
< jeremyrubin> Ah let me re-check
< jeremyrubin> Maybe some ascii art :)
< jeremyrubin> ok so TX 0 pays into tx 1 and into tx2 and tx3
< jeremyrubin> Tx 1 pays into tx 2
< jeremyrubin> gotcha
< hebasto> what are our rules: an arrow from a parent to a child or opposite?
< jeremyrubin> I see
< jeremyrubin> For ascii? I think there are arguments for both arrow directions
< hebasto> if don't mind, let skip ascii :)
< jeremyrubin> Personally I've always preferred arrow to parent conceptually, but arrow to child makes more sense to people intuitively
< hebasto> it seems I tend to agree with your point
< jeremyrubin> yeah arrow to parent makes sense because it's a pointer and the parent is immutable
< jeremyrubin> But value "flows" from parent to child
< hebasto> sure
< jeremyrubin> I'll think if there's cases that wouldn't be caught here.
< hebasto> maybe just add a note that a complete graph is used?
< jeremyrubin> I think making it bigger (e.g., 50 txns) would be useful
< jeremyrubin> because then you will trip over the mempool descendants and ancestors limits
< jeremyrubin> So you check that they aren't being enforced during reorg
< hebasto> yeah, limits...
< jeremyrubin> can be done as follow up work tho
< jeremyrubin> test is fine as is IMO
< hebasto> "they aren't being enforced during reorg" mind pointing me to code?
< jeremyrubin> UpdateTransactionsFromBlock
< hebasto> ty
< hebasto> oh, I know that part of code ;)
< jeremyrubin> yeah we're just not checking things like "how many descendants does a tx have" here
< hebasto> I see
< jeremyrubin> So by doing your test with 50 things you would confirm that we aren't eviciting stuff even if not re-mined
< jeremyrubin> (which can be a potential problem, and behavior we'd want to change, but... let's skip that for now)
< bitcoin-git> [bitcoin] MarcoFalke opened pull request #18486: doc: Explain new test logging (master...2004-docTestLogging) https://github.com/bitcoin/bitcoin/pull/18486
< bitcoin-git> [bitcoin] MarcoFalke pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/72ef36656439...d2db25233cf0
< bitcoin-git> bitcoin/master fac3716 MarcoFalke: test: check that peer is connected when calling sync_*
< bitcoin-git> bitcoin/master d2db252 MarcoFalke: Merge #18474: test: check that peer is connected when calling sync_*
< bitcoin-git> [bitcoin] MarcoFalke merged pull request #18474: test: check that peer is connected when calling sync_* (master...2003-qaCheckConnSync) https://github.com/bitcoin/bitcoin/pull/18474
< bitcoin-git> [bitcoin] promag opened pull request #18487: wip: rpc: Fix rpcRunLater race in walletpassphrase (master...2020-04-fix-rpcrunlater-race) https://github.com/bitcoin/bitcoin/pull/18487
< promag> I'd like to consider #18487 for 0.20
< gribble> https://github.com/bitcoin/bitcoin/issues/18487 | wip: rpc: Fix rpcRunLater race in walletpassphrase by promag · Pull Request #18487 · bitcoin/bitcoin · GitHub
< promag> at worst case it doesn't fix the issue and only reduces the lock scopes