< jtimon> what's the point of if (g_txindex) if g_txindex is never null ?
< jtimon> or I'll rephrase, what's the new way of "if (!fTxIndex)" ?
< sipa> ?
< jtimon> sorry, I'll figure it out
< aj> g_txindex should be null without -txindex?
< sipa> indeed
< sipa> jtimon: sorry, i didn't understand your question by missing that your second line was about fTxIndex and not g_txindex
< jtimon> I just got a ‘fTxIndex’ was not declared in this scope, greped and since I'm quite sure I didn't invented fTxIndex, I thought to ask, but I copied that fTxIndex from somewhere else and now I want ot know where I copied it from and what happened to that line. but git has all the anwers I'm looking for
< sipa> in any case, g_txindex should be null when there is no index
< Kanmanee> [012f6fd7@gateway/web/freenode/ip.] has joined #bitcoin-core-dev
< bitcoin-git> [bitcoin] witty67 opened pull request #13102: for our kids (master...master) https://github.com/bitcoin/bitcoin/pull/13102
< bitcoin-git> [bitcoin] fanquake closed pull request #13102: for our kids (master...master) https://github.com/bitcoin/bitcoin/pull/13102
< wumpus> how... how long is verify-commits.sh supposed to take?
< wumpus> I thought I'd post a local performance comparison in #13066 but while the python script finished in 5 minutes, the old shell script seems to be taking forever
< gribble> https://github.com/bitcoin/bitcoin/issues/13066 | Migrate verify-commits script to python, run in travis by ken2812221 · Pull Request #13066 · bitcoin/bitcoin · GitHub
< wumpus> 5min for the python script, 32min+ for the shell script
< promag> that's a huge difference
< wumpus> this is not a very fast computer (probably slower than what travis uses), but still
< promag> wumpus: re #13028 it's no big deal, the new mutex is used in a couple of places, nothing special
< gribble> https://github.com/bitcoin/bitcoin/issues/13028 | Make vpwallets usage thread safe by promag · Pull Request #13028 · bitcoin/bitcoin · GitHub
< promag> looks like I'm creating a chain of PR's..
< sdaftuar> wumpus: i think it's funny that our project has a slow commit verification script, you'd think someone here would be very good at writing a fast one. :)
< sdaftuar> anyway i just timed it locally, i've never run it before so not sure if i'm doing it right, but it took 5m30seconds but finished with an error
< sdaftuar> this is on a very fast computer though
< wumpus> what commit does it get to?
< sdaftuar> here's my output: https://0bin.net/paste/bcwIMpfNjPlH7JKh#1dF-PaM+KUv4wIDmz1rfrd7PzMrNp5fXmLq8NoFOktU
< wumpus> I managed to make it quite also after 5mXX but it complained about an unclean commit which was not very deep
< gribble> https://github.com/bitcoin/bitcoin/issues/1 | JSON-RPC support for mobile devices ("ultra-lightweight" clients) · Issue #1 · bitcoin/bitcoin · GitHub
< wumpus> the python script verifies 2627 commits in 5m36.844s
< sdaftuar> i'll give that a try
< wumpus> that one gets to commit c17f11f7b43ad3bd9e242c67db1f3679558a0581, that's depth 479
< promag> MarcoFalke: I don't have windows atm to find the problem
< wumpus> 5.5/479*2627 -> would have taken ~30 minutes to go through them all, that's faster than my 32min, but still
< sdaftuar> the python script finished in 3m21seconds and didn't produce an error
< wumpus> I wonder what causes the difference in results
< ken2812221> It runs over 1.5 hours on my PC
< wumpus> ken2812221: hey you're here! any idea why commit c17f11f7b43ad3bd9e242c67db1f3679558a0581 fails the clean merge test with the .sh script, but passes with the .py script?
< ken2812221> sipa says we can only check unclean merge in one month, so I don't check if it commit before one month
< wumpus> oh right
< wumpus> it seems to check merge up to 082e26c08bb0a803cf9cf6b015c08ee57f31e623
< wumpus> and that commit is deeper, so it won't be checked
< wumpus> (well, the subtree check, but not the merge check)
< wumpus> eh no, not that either, just the gpg check? your script only checks the sha512sum treehash for the top commit?
< wumpus> or is that the same for the .sh, I don't remember
< ken2812221> it only check if you has set --tree-checks, same in shell script
< wumpus> I see, just figured out --tree-checks
< wumpus> you're correct, the shell script has the same behavior
< wumpus> ken2812221: so to be clear: the only expected difference in behavior between your python script and the shell script is that your script only verifies merges up to a certain depth?
< wumpus> (which is a sensible change, as the shell script throws an error there, not allowing it to verify the gpg signatures deeper either)
< ken2812221> Yes.
< wumpus> ok
< bitcoin-git> [bitcoin] laanwj pushed 4 new commits to master: https://github.com/bitcoin/bitcoin/compare/487dcbe80c20...17266a13067c
< bitcoin-git> bitcoin/master 2ae705d Andrew Chow: Remove Safe mode
< bitcoin-git> bitcoin/master 7da3b0a Wladimir J. van der Laan: rpc: Move RPC_FORBIDDEN_BY_SAFE_MODE code to reserved section...
< bitcoin-git> bitcoin/master d8e9a2a Wladimir J. van der Laan: Remove "rpc" category from GetWarnings...
< bitcoin-git> [bitcoin] laanwj closed pull request #13090: Remove Safe mode (achow101) (master...2018_04_remove_safemode_rebased) https://github.com/bitcoin/bitcoin/pull/13090
< bitcoin-git> [bitcoin] jamesob opened pull request #13105: [qa] Add --failfast option to functional test runner (master...2018-04-testrunner-failfast) https://github.com/bitcoin/bitcoin/pull/13105
< bitcoin-git> [bitcoin] TheBlueMatt opened pull request #13106: Simplify semantics of ChainStateFlushed callback (master...2018-04-wallet-flush-better-criteria) https://github.com/bitcoin/bitcoin/pull/13106
< jamesob> I'm really down the rabbithole on this qt functional test deadlock stuff -- do we have a way of programmatically capturing stacktraces? any aversion to adding libunwind as a dep when --enable-debug?
< MarcoFalke> jamesob: You could debug locally in a trusty vm?
< jamesob> MarcoFalke: yep, am at the moment
< Master_> Hello
< instagibbs> can someone point me to where .cookie file went on master? Must have missed it.
< sipa> it's there for me
< sipa> $DATADIR/.cookie
< sipa> do you have rpcuser/rpcpassword/rpcauth set?
< instagibbs> ok, probably an issue between chair and keyboard