--- Log opened Thu Oct 24 00:00:43 2019 00:44 -!- achow101 [~achow101@unaffiliated/achow101] has joined #bitcoin-builds 00:57 -!- achow101 [~achow101@unaffiliated/achow101] has quit [Ping timeout: 276 seconds] 01:01 -!- achow101 [~achow101@unaffiliated/achow101] has joined #bitcoin-builds 02:29 -!- provoostenator [~quassel@provoostenator.sprovoost.nl] has left #bitcoin-builds ["https://quassel-irc.org - Chat comfortably. Anywhere."] 02:54 -!- jonatack [~jon@2a01:e35:8aba:8220:6627:dad:d967:649d] has quit [Ping timeout: 252 seconds] 03:42 -!- jonatack [~jon@54.76.13.109.rev.sfr.net] has joined #bitcoin-builds 04:15 -!- jonatack [~jon@54.76.13.109.rev.sfr.net] has quit [Quit: jonatack] 04:16 -!- jonatack_ [~jon@213.152.161.249] has joined #bitcoin-builds 04:16 -!- jonatack_ [~jon@213.152.161.249] has quit [Client Quit] 04:16 -!- jonatack_ [~jon@213.152.161.249] has joined #bitcoin-builds 04:18 -!- jonatack_ [~jon@213.152.161.249] has quit [Client Quit] 04:18 -!- jonatack_ [~jon@213.152.161.249] has joined #bitcoin-builds 04:41 -!- Bitcoinera [d9d8574e@217.216.87.78.dyn.user.ono.com] has joined #bitcoin-builds 04:47 -!- Bitcoinera [d9d8574e@217.216.87.78.dyn.user.ono.com] has quit [Remote host closed the connection] 05:24 < MarcoFalke> We can't cross compile depends from an arm64 host to any guest? 06:25 < dongcarl> MarcoFalke: That's probably a path that hasn't been tested much... 06:31 -!- jonatack_ [~jon@213.152.161.249] has quit [Ping timeout: 246 seconds] 07:17 -!- jonatack_ [~jon@2a01:e35:8aba:8220:6627:dad:d967:649d] has joined #bitcoin-builds 07:26 -!- jonatack_ [~jon@2a01:e35:8aba:8220:6627:dad:d967:649d] has quit [Quit: jonatack_] 07:27 -!- jonatack [~jon@2a01:e35:8aba:8220:6627:dad:d967:649d] has joined #bitcoin-builds 11:19 < wumpus> that's interesting to try, let me see if ubuntu aarch64 has a x86 toolchain 11:23 < wumpus> g++-x86-64-linux-gnu... 11:33 < wumpus> ok, local cross-build running, might take a while as it's a Odroid C2 with a SD card 11:52 < dongcarl> cfields fanquake: sync up after meeting? or tomorrow? 11:53 < fanquake> dongcarl: I'll be around for either 12:02 < MarcoFalke> sd cards don't seem to be great with parallel compilation 13:18 < wumpus> FWIW the depends build succeeds, but configure fails on: configure: error: Could not link against boost_filesystem-mt-x64 ! 13:18 < fanquake> wumpus building master ? 13:19 < wumpus> fanquake: yes, building master on aarch64, cross compiling to g++-x86-64-linux 13:19 < wumpus> (without qt, for now) 13:20 < wumpus> config.log is really unhelpful 13:21 < fanquake> heh, some more info was going to be my next question 13:22 < wumpus> "checking whether to build Bitcoin Core GUI... no (Qt)" wait, that was fixed right? maybe I still need to re-run autogen.sh 13:23 < fanquake> It was just changed from Qt5 to Qt iirc 13:23 < wumpus> I thought it was removed for the no case 13:24 < fanquake> hmm you are correct 13:24 < fanquake> maybe need to autogen as you said 13:24 < fanquake> https://github.com/bitcoin/bitcoin/pull/17005 13:26 < wumpus> phew, it gets past the dreaded boost checks now 13:27 < fanquake> ? 13:27 < wumpus> pretty funny to see "checking for x86_64 assembly availability... yes" on an ARM system 13:27 < MarcoFalke> I am still compiling depends :sweat-smile: 13:28 < wumpus> MarcoFalke: I suppose you enabled qt :) 13:28 < MarcoFalke> yeah, qmake takes a trillion years to compile 13:28 < fanquake> even after years of continual slimming down heh 13:29 < wumpus> everything else in depends is super fast nowadays (boost's extraction takes much longer than building it) 13:29 < wumpus> fanquake: yup, despite that 13:29 < wumpus> which means, it *could* be much worse 13:30 < fanquake> At least there's now a possibility of being able to throw away Protobuf, and potentially OpenSSL entirely for 0.20.0 :o 13:30 < wumpus> which would be nice for many reasons, but hardly a win in compile time :) 13:30 < fanquake> Also I see you just looked Marco, but if builds people want to look at something: https://github.com/bitcoin/bitcoin/pull/17231 13:31 < fanquake> wumpus heh yes, a slight saving in depends compile time is just a bonus 13:32 < fanquake> Clangs -dumpversion has been hardcoded to 4.2.1 for a loooong time 13:32 < wumpus> GCC 4.2.1, July 18, 2007 13:33 < fanquake> heh, not quite that long. I became like that in 2011 for GCC back compat 13:33 < wumpus> kind of funny given that Apple hardly cares about backwards compat in general, let alone so long 13:34 < wumpus> wasn't gcc 4.2.1 also the version of gcc that OpenBSD/FreeBSD stuck with for a long time? because of some license change? 13:35 < fanquake> I'm sure I heard that OpenBSD is now actually going to move away from Clang, due to another licensing issue? 13:35 < wumpus> " 13:35 < wumpus> Because of GCC's move to the GPL v3, FreeBSD was forced to remain using GCC 4.2.1 (GPL v2), 13:35 < wumpus> yep 13:35 < fanquake> So I assume back to GCC now? 13:35 < wumpus> moving *away* from clang?!? 13:35 < wumpus> nooo, OpenBSD is moving to clang afaik 13:37 < fanquake> Apparently they are just going to stay with Clang 8 13:37 < dongcarl> I think they're moving more and more towards clang 13:37 < fanquake> and cant move to 9 because of licensing issues 13:37 < fanquake> https://lists.llvm.org/pipermail/llvm-dev/2017-April/112300.html 13:37 < wumpus> it took them somewhat longer because they had a patch stack with various security improvments on top of gcc (like ROP mitigation), which they've been porting to clang 13:37 < wumpus> lol, from one licensing issue to the next 13:38 < fanquake> Some more discussion here: https://news.ycombinator.com/item?id=21020024 13:38 < dongcarl> Ah I see... 13:39 < fanquake> I'm glad that as a project we seem to avoid most of this sort of controversy 13:40 < wumpus> poor them, after having spent 10 years maintaining an old compiler because of license reasons, just when it looked like their problems were over, now they now have to do the same again 13:41 < wumpus> at least not *that* sort of controversy 13:42 < fanquake> heh true 13:44 < wumpus> ah they're going to do the "8 in the base system, 9+ in ports" thing again ... that's going to be a mess when the C++ ABI's inevitably change at some point, deja vu 13:46 < wumpus> it's good to know in advance how doc/build-openbsd.md has to change 13:46 < fanquake> Someone can give it it's semi-yearly update 13:59 < wumpus> it's still building... 14:00 < fanquake> what does the io look like 14:03 < wumpus> dunno, somehow iotop reports negative numbers, must be really bad :) 14:06 < wumpus> it looks like the compilation is CPU bound mostly 14:07 < wumpus> it has four cores but only 2GB of memory, at -j4 it pretty much freezes 14:09 < wumpus> luckily this is only to prove a point and not to do all development builds from now on :) 14:15 < MarcoFalke> it might be faster with -j2 idk 14:20 < wumpus> it is 14:20 < wumpus> nah, not sure 14:22 < wumpus> seems to depend on what is being compiled 14:51 < MarcoFalke> ok, so depends/riscv64-linux-gnu passed on aarch64 14:52 < dongcarl> MarcoFalke: Nice, riscv64-targeting toolchains have never given me any trouble, very solid 14:53 < MarcoFalke> don't say that too early xD 14:54 < dongcarl> True... Knock on...... gounding mat... 14:56 * dongcarl is anxiously awaiting a build that's been haunting his dreams for days 14:57 * MarcoFalke types `make` in the other tab :eyes: 15:06 < MarcoFalke> https://github.com/bitcoin/bitcoin/issues/17244 15:06 < MarcoFalke> :sob: 15:07 < fanquake> dongcarl is that the guix macOS cross-compile :p 15:10 < dongcarl> fanquake: Nah, just fixing up some of my fancy Guix stuff 15:11 < fanquake> noice 15:12 < dongcarl> fanquake: gunna get started on the Mac stuff soon tho... it's all coming together.gif 15:12 < fanquake> Hopefully we'll get the macOS toolchain stuff in shortly ? 15:13 < dongcarl> fanquake: Yeah true, I need https://github.com/bitcoin/bitcoin/pull/17099 at least 15:13 < wumpus> cross-build to x86_64 is still going 15:14 < fanquake> dongcarl soon heh 15:14 < fanquake> a lot of build related things going on at the moment 15:15 < wumpus> MarcoFalke: this went too fast, this can't be right, either there were objects left behind or ccache is doing something strange 15:15 < MarcoFalke> jup 15:15 * MarcoFalke facepalm and make clean 15:16 < MarcoFalke> I guess it is only an issue when cross compiling with the gui enabled 15:17 < midnightmagic> fanquake: I would trade OpenBSD controversy for the crap even I have to endure in BitcoinLand in an instant. :-/ 15:18 < midnightmagic> wumpus: Do you have the unleashed expansion board? 15:20 < wumpus> MarcoFalke: hm, so maybe it chooses the wrong compiler for native gcc in that case? 15:20 < wumpus> so it builds tools like uic with the cross-compiler instead of the native compiler 15:20 < wumpus> midnightmagic: nope, unfortunately not 15:21 < wumpus> I'll do a qt cross-depends build while I sleep to check 15:22 < wumpus> afaik the NO_QT build doesn't build any native tools as part of depends 15:22 < fanquake> not any more, with Protobuf disabled 15:23 < wumpus> protobuf is also only built with qt enabled, right? 15:23 < fanquake> oops right, misread your comment 15:23 < fanquake> and yes, protobuf is now opt-in as well 15:23 < fanquake> PROTOBUF=1 15:25 < wumpus> can I do PROTOBUF=1 with NO_QT=1? 15:25 < wumpus> (just to build some native tool without building qt...) 15:25 < fanquake> Yea pretty sure you can 15:25 < fanquake> Protobuf doesn't need qt 15:26 < fanquake> yep you can 15:26 < fanquake> make -C depends/ NO_QT=1 PROTOBUF=1 15:26 < fanquake> copying packages: native_biplist native_ds_store native_mac_alias native_protobuf boost openssl libevent bdb miniupnpc protobuf zeromq 15:26 < wumpus> nice! trying 15:27 < wumpus> bitcoind cross-compile to x86_64 succesfully completed 15:27 < wumpus> including tests, benches, etc 15:53 < wumpus> $ file x86_64-linux-gnu/native/bin/protoc 15:53 < wumpus> x86_64-linux-gnu/native/bin/protoc: ELF 64-bit LSB shared object, ARM aarch64, version 1 (SYSV), dynamically linked, interpreter /lib/ld-, for GNU/Linux 3.7.0, BuildID[sha1]=a6b4a421938ded93233a53f7195a5004d11f850b, stripped 15:53 < wumpus> at least protoc looks as expected 15:54 < wumpus> if this is a qt issue it's really specific to qt --- Log closed Fri Oct 25 00:00:45 2019