< luke-jr>
achow101: well, if I backport #17204, the test fails apparently because of this
< gribble>
https://github.com/bitcoin/bitcoin/issues/17204 | wallet: Do not turn OP_1NEGATE in scriptSig into 0x0181 in signing code (sipa) by meshcollider · Pull Request #17204 · bitcoin/bitcoin · GitHub
< achow101>
luke-jr: interesting, I'll have a look
< achow101>
luke-jr: my expectation is that 0.20 and earlier would attempt to sign non-wallet inputs and fail silently. Now what's happening is that it's attempting to sign but failing loudly
< achow101>
erm got that backwards
< luke-jr>
achow101: master is succeeding..
< achow101>
this is looking more like a bug in 0.20 to me rather than a bug in master
< achow101>
luke-jr: the test passes on 0.19
< luke-jr>
XD
< sipa>
it was my expectectation too that that should work
< sipa>
if yiu agree to signing a tx, there is no reason why you wouldn't sign it for all inputs/keys that you can
< luke-jr>
achow101: did you bisect it already?
< achow101>
luke-jr: no, but I don't need to
< achow101>
I already know what introduced the bug
< luke-jr>
?
< achow101>
#18115
< gribble>
https://github.com/bitcoin/bitcoin/issues/18115 | wallet: Pass in transactions and messages for signing instead of exporting the private keys by achow101 · Pull Request #18115 · bitcoin/bitcoin · GitHub
< sipa>
so, is there a simple backport we could use in 0.20.2 ?
< achow101>
pretty sure no
< sipa>
a less than simple one? that doesn't involve all of descriptor wallets...?
< achow101>
I don't think there's a commit we can directly backport
< achow101>
but it shouldn't be hard to write a fix
< sipa>
ok
< sipa>
i don't think this is all that worrisome, though it's good practice to fix discovered issues
< luke-jr>
especially while 0.20 is our latest
< instagibbs>
it's conceivable you'd want to not necessarily sign all inputs in a tx(payjoin is an example) but you'd want a specific api to exclude inputs
< instagibbs>
not important to discussion per se, just saying
< sipa>
instagibbs: yes, agree, though that's a much more specialized situation
< instagibbs>
agreed!
< luke-jr>
probably just a PSBT use case though?
< instagibbs>
anyways, not really something pertaining to this issue
< luke-jr>
achow101: if it's trivial to fix correctly, I'd like to include it in Knots 0.20.1 if possible
< bitcoin-git>
[bitcoin] achow101 opened pull request #19740: [0.20] wallet: Simplify and fix CWallet::SignTransaction (0.20...backport-17204-fix) https://github.com/bitcoin/bitcoin/pull/19740
< achow101>
luke-jr: ^ Turns out copy and paste from master works
< luke-jr>
XD
< luke-jr>
achow101: is it really safe to drop all the error checking there?
< achow101>
I think so. See #19568. The errors should already be coming from the called functions
< bitcoin-git>
bitcoin/master b82067b Karl-Johan Alm: wallet: try -avoidpartialspends mode and use its result if fees are below ...
< bitcoin-git>
bitcoin/master 7f13dfb Karl-Johan Alm: test: test the implicit avoid partial spends functionality
< bitcoin-git>
bitcoin/master c831e10 Samuel Dobson: Merge #14582: wallet: always do avoid partial spends if fees are within a ...
< bitcoin-git>
[bitcoin] meshcollider merged pull request #14582: wallet: always do avoid partial spends if fees are within a specified range (master...181026-try-avoidpartialspends) https://github.com/bitcoin/bitcoin/pull/14582
< bitcoin-git>
[bitcoin] hebasto opened pull request #19745: build: Drop back compatibility code for 32-bit x86 Linux binaries (master...200817-i686) https://github.com/bitcoin/bitcoin/pull/19745
< bitcoin-git>
[bitcoin] hebasto closed pull request #19745: build: Drop back compatibility code for 32-bit x86 Linux binaries (master...200817-i686) https://github.com/bitcoin/bitcoin/pull/19745
< bitcoin-git>
[bitcoin] slmtpz opened pull request #19752: test: Update wait_until usage in tests not to use the one from utils (master...master) https://github.com/bitcoin/bitcoin/pull/19752
< achow101>
Why are there both RemoveWallet and UnloadWallet? It seems like unloadwallet calls both, but the GUI only calls RemoveWallet?
< bitcoin-git>
[bitcoin] achow101 opened pull request #19754: wallet: Move UpdateWalletSetting to wallet module and use from GUI interface (master...load-on-start-gui) https://github.com/bitcoin/bitcoin/pull/19754
< ryanofsky>
difference between gui and rpc unloading sequences is that gui sequence is nonblocking while rpc blocks, but that code is a mess