03:41 < luke-jr> NicolasDorier: stop using Windows kthx

03:41 < luke-jr> test/functional/p2p_permissions.py is missing +x

08:18 < bitcoin-git> [bitcoin] fanquake pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/7d6f63cc2c2b...6519be605480

08:18 < bitcoin-git> bitcoin/master 9b92538 Matt Corallo: Remove unused fScriptChecks parameter from CheckInputs

08:18 < bitcoin-git> bitcoin/master 6519be6 fanquake: Merge #13868: Remove unused fScriptChecks parameter from CheckInputs

08:18 < bitcoin-git> [bitcoin] fanquake merged pull request #13868: Remove unused fScriptChecks parameter from CheckInputs (master...check-inputs-script-checks) https://github.com/bitcoin/bitcoin/pull/13868

08:36 < wumpus> emilengler: fanquake: done

08:39 < fanquake> wumpus: thanks

08:48 < meshcollider> sipa: Q about the bech32 code in bitcoin. Shouldn't the encode function error if the HRP it is given is capitalised?

08:49 < meshcollider> i think encoding currently just uses it as-is which means the checksum it creates will be over a mixed case address and will always be invalid, even if you manually change capitalisation afterwards

08:49 < meshcollider> Compare with the C implementation which explicitly checks this https://github.com/sipa/bech32/blob/master/ref/c/segwit_addr.c#L59

08:49 < meshcollider> Just wanted to confirm with you before I PR

08:53 < meshcollider> the go implementation just converts to lowercase automatically

10:12 < bitcoin-git> [bitcoin] darosior opened pull request #16787: Human readable network services (master...services_for_humans) https://github.com/bitcoin/bitcoin/pull/16787

11:32 < bitcoin-git> [bitcoin] meshcollider pushed 4 commits to master: https://github.com/bitcoin/bitcoin/compare/6519be605480...33f9750b1b86

11:32 < bitcoin-git> bitcoin/master 7f3bb24 darosior: gettransaction: add an argument to decode the transaction

11:32 < bitcoin-git> bitcoin/master b8b3f04 darosior: tests: Add a new functional test for gettransaction

11:32 < bitcoin-git> bitcoin/master 9965940 darosior: doc: Add release note for the new gettransaction argument

11:33 < bitcoin-git> [bitcoin] meshcollider merged pull request #16185: gettransaction: add an argument to decode the transaction (master...decode_tx_gettransaction) https://github.com/bitcoin/bitcoin/pull/16185

11:43 < bitcoin-git> [bitcoin] laanwj opened pull request #16788: gui: Update transifex slug for 0.19 (master...2019_10_tx_slug_0_19) https://github.com/bitcoin/bitcoin/pull/16788

12:07 < emilengler> wumpus: You meant the ban for the spammer? if so, thanks

14:42 < bitcoin-git> [bitcoin] GChuf opened pull request #16789: Update headers and copyright_header.py script (master...headers) https://github.com/bitcoin/bitcoin/pull/16789

17:05 < bitcoin-git> [bitcoin] HashUnlimited opened pull request #16790: break loop early on undefined version bits (master...patch-4) https://github.com/bitcoin/bitcoin/pull/16790

17:33 < bitcoin-git> [bitcoin] HashUnlimited closed pull request #16790: break loop early on undefined version bits (master...patch-4) https://github.com/bitcoin/bitcoin/pull/16790

19:24 < warren> wumpus: everything good with the transifex account now?

19:52 < bitcoin-git> [bitcoin] fanquake closed pull request #16789: Update headers and copyright_header.py script (master...headers) https://github.com/bitcoin/bitcoin/pull/16789

22:14 < elichai2> Playing a bit with clangs `-flto=thin` and first the binaries are down from ~220-280mb to 16-30mb. curious to see if there are also any speed imporvements

22:21 < sipa> elichai2: compared to what?

22:22 < sipa> that kind of difference sounds more like you disabled or removed debug symbols

22:22 < sipa> lto can have a bit of an impact on binary size, but nlt

22:22 < sipa> lto can have a bit of an impact on binary size, but not that much

22:22 < elichai2> sipa: the only difference is gcc vs clang+lto, might be the different compiler, but other than that didn't change anything

22:23 < sipa> my non-lto bitcoind is 6 MB btw, after stripping debug symbols

22:23 < elichai2> didn't touch the file, i'll try stripping in both cases, and try to stick to clang for with and without

22:25 < elichai2> altough i'm explictly doing `--disable-debug` and manually setting CFLAGS and CXXFLAGS so in theory there shouldn't be really any debug symbols (no `-g`)

22:34 < elichai2> after stripping them both you're right, there is no meaningful difference, but without striping there's a big difference, not sure why

22:35 < sipa> poasibly because the debug symbol generation for lto is broken/incomplete :p

22:36 < warren> IMO building with debug symbols and stripping before distribution is superior to building without debug symbols. This is because gdb lets you run the stripped binary while looking at the unstripped put in an optional directory to the side if you need to debug.

22:47 < midnightmagic> What warren said. Although at least in the past this could cause performance issues in rare cases.

22:48 < midnightmagic> Plus with a stripped binary you can rebuild with the same build env and recreate an image which can be used to debug coredumps from stripped bins. Which is pretty neat.