--- Log opened Fri Mar 05 00:00:47 2021 00:34 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Quit: ZNC - http://znc.sourceforge.net] 00:34 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has joined #bitcoin-builds 00:42 -!- meshcollider [meshcollid@gateway/shell/ircnow/x-vtakzzjunnrxyedh] has quit [Ping timeout: 264 seconds] 00:44 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Quit: ZNC - http://znc.sourceforge.net] 00:46 -!- meshcollider [meshcollid@gateway/shell/ircnow/x-ttsnhkjhgyervsge] has joined #bitcoin-builds 00:46 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has joined #bitcoin-builds 00:55 -!- jonatack_ [~jon@37.167.227.177] has joined #bitcoin-builds 01:25 -!- jonatack_ [~jon@37.167.227.177] has quit [Quit: jonatack_] 01:25 -!- jonatack [~jon@37.167.227.177] has joined #bitcoin-builds 02:20 < wumpus> elichai2: it might be interesting to benchmark different allocators but fwiw, glibc does thread-local allocation of some kind, we ran into problems with this on 32-bit platforms where the amount of virtual memory reserved for these pools got out of hand 02:21 < wumpus> there's still a recommendation about setting an environment flag somewhere for low-memory systems let me see, ah yes: https://github.com/bitcoin/bitcoin/blob/master/doc/reduce-memory.md#linux-specific 02:38 < fanquake> wumpus: "guix builds seem to take really long" 1h28m to build for all targets on the 8 core machine I've been using 02:43 < wumpus> fanquake: does it build qt and boost for every platform every time there ? 02:45 < wumpus> (i haven't actually timed it, will do so next time) 02:47 < hebasto> real 237m47,159s / user 514m29,652s / sys 45m14,712s -- https://github.com/bitcoin/bitcoin/pull/21298#issuecomment-788819296 02:47 < fanquake> wumpus: no, that’s just for the binaries. Possible it was a bit slow, as I had some gitian builds and other things running at the same time on that machine 02:49 < wumpus> how do you make it only build the binaries and not the rest? 02:51 < fanquake> That was building #21304 with: "BASE_CACHE="/guix/base_cache" SOURCES_PATH="/guix/sources" SDK_PATH="/guix/SDKs" ./contrib/guix/guix-build.sh" 02:51 < gribble> https://github.com/bitcoin/bitcoin/issues/21304 | guix: Various user workflow improvements by dongcarl · Pull Request #21304 · bitcoin/bitcoin · GitHub 02:52 < fanquake> So built once from scratch, used the guix clean script, then rebuilt with the same "BASE_CACHE="/guix/base_cache" SOURCES_PATH="/guix/sources" SDK_PATH="/guix/SDKs" ./contrib/guix/guix-build.sh". The second time it only built the binaries. 02:52 < wumpus> what do you run between builds? 02:52 < wumpus> (to clean) 02:52 < fanquake> Will post some stuff in that PR 02:52 < fanquake> I ran contrib/guix/guix-clean, which is new in that PR 02:53 < wumpus> it seems that the BASE_CACHE doesn't get the compiled libraries here or i'm missing something 02:53 < fanquake> previously I'd remove distsrc-*, output & windeploy manually 02:53 < wumpus> oh it does, it has ./riscv64-linux-gnu/qt/qt-5.9.8-02ce96fd3f1.tar.gz etc, i'm confused now 02:55 < wumpus> i'm sure i saw qt and boost compiling just now, but don't have a log or screenshot (unless the guix build output goes somewhere else than the terminal) 02:56 < fanquake> No all the build output should end up in your terminal 02:56 < fanquake> guix setup, depends builds, binary builds etc 02:57 < wumpus> i'll tee it to a file next time 03:31 < wumpus> > compiling painting/qtextureglyphcache.cpp 03:32 < wumpus> and im building #21209 not a qt bump or such 03:32 < gribble> https://github.com/bitcoin/bitcoin/issues/21209 | build: use newer source for libnatpmp by fanquake · Pull Request #21209 · bitcoin/bitcoin · GitHub 03:32 < wumpus> so a) it is caching qt 5.9.8 into a tar.gz after build b) it doesn't seem to be using it 03:41 < jonasschnelli> what is the reason for "guix time-machine: error: all build users are currently in use; consider creating additional users and adding them to the `guixbuild' group"? 03:42 < fanquake> You've probably set a max jobs larger than than number of builder users available to guix 03:43 < fanquake> I ran into something similar once, and have since bumped the builder user count to 32, so that I'll likely never run into it 03:43 < fanquake> See the setup here: https://github.com/fanquake/core-review/blob/master/guix/Dockerfile#L36 03:44 < jonasschnelli> fanquake: I'm using ./contrib/guix/guix-build.sh vanilla on a physical host (no VM, no docker) 03:44 < fanquake> How many guix builder users did you create when you setup guix? 03:44 < jonasschnelli> no idea... :) 03:45 < fanquake> heh, no worries. Take a quick look at https://guix.gnu.org/manual/en/html_node/Build-Environment-Setup.html#Build-Environment-Setup 03:45 < jonasschnelli> ah. 10 03:45 < jonasschnelli> Shall I set the MAX_JOBS env variable? 03:46 < fanquake> Yea. Although, what is the output before it fails? I assume it's trying to set max-jobs > 10 03:47 < fanquake> You could also just add more users to the group, so that you can use what I assume are a large # of cores 03:48 < jonasschnelli> Yes. Let me do the later, Thanks 04:11 < jonasschnelli> hmm... 120seconds RPC timeout is not enough for feature_nulldummy.py on bitcoinbuils? 04:12 < jonasschnelli> https://bitcoinbuilds.org/index.php?ansilog=20e8f792-99e6-4371-a0be-f3436448b648.log#l2551 05:13 -!- jonatack [~jon@37.167.227.177] has quit [Ping timeout: 265 seconds] 05:54 -!- gribble [~gribble@unaffiliated/nanotube/bot/gribble] has quit [Remote host closed the connection] 06:05 -!- gribble [~gribble@unaffiliated/nanotube/bot/gribble] has joined #bitcoin-builds 07:00 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Quit: ZNC - http://znc.sourceforge.net] 07:01 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has joined #bitcoin-builds 07:12 < dongcarl> jonasschnelli: I have a local branch that clarifies the max-jobs vs cores debacle, will push sometime soon. For future things, feel free to reach out to me for any help during daytime US Eastern. Even the smallest complaints help me determine what user improvements to work on! 07:15 < dongcarl> wumpus: I usually debug with: `> outerr 2>&1` 07:16 < dongcarl> wumpus: what was your particular invocation? I'd like to make sure that caching is working correctly. 07:21 -!- jonatack [~jon@37.167.227.177] has joined #bitcoin-builds 07:24 -!- mol [~mol@unaffiliated/molly] has joined #bitcoin-builds 07:26 -!- mol_ [~mol@unaffiliated/molly] has quit [Ping timeout: 245 seconds] 07:26 < wumpus> dongcarl: SOURCES_PATH="${HOME}/sources" BASE_CACHE="${HOME}/base-cache" SDK_PATH="${HOME}/sdk-path" contrib/guix/guix-build.sh on commit 7af25024e9563241c72797d4eeabdf660e548f53 (#21209) 07:26 < gribble> https://github.com/bitcoin/bitcoin/issues/21209 | build: use newer source for libnatpmp by fanquake · Pull Request #21209 · bitcoin/bitcoin · GitHub 07:26 < wumpus> i'll repeat the same and see if it happens again (and save the log) 07:32 < dongcarl> Yeah thanks that’d be useful. BTW time-machine/manifest changes will cause cache to be invalidated (as they should) 07:47 < wumpus> dongcarl: it seems to be okay now, it starts out building bitcoin not the dependencies, must have been that then 07:48 < dongcarl> wumpus: Great! In the future I'll try to note in PR descriptions if I expect the cache to be invalidated so people know what to expect 10:21 < wumpus> looks ok now, for a build of all platforms (on 8-core machine) real 107m21.361s / user 507m49.312s / sys 27m0.605s 10:22 < dongcarl> Great :-) 10:59 -!- gribble [~gribble@unaffiliated/nanotube/bot/gribble] has quit [Remote host closed the connection] 11:12 -!- gribble [~gribble@unaffiliated/nanotube/bot/gribble] has joined #bitcoin-builds 11:45 -!- jonatack_ [~jon@37.170.241.200] has joined #bitcoin-builds 11:48 -!- jonatack [~jon@37.167.227.177] has quit [Ping timeout: 260 seconds] 12:47 -!- jonatack_ [~jon@37.170.241.200] has quit [Quit: jonatack_] 12:47 -!- jonatack [~jon@37.170.241.200] has joined #bitcoin-builds 13:07 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Quit: ZNC - http://znc.sourceforge.net] 13:10 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has joined #bitcoin-builds 14:51 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Quit: ZNC - http://znc.sourceforge.net] 14:51 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has joined #bitcoin-builds 15:41 -!- jonatack [~jon@37.170.241.200] has quit [Read error: Connection reset by peer] 17:25 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Read error: Connection reset by peer] 17:27 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has joined #bitcoin-builds 19:06 -!- belcher_ [~belcher@unaffiliated/belcher] has joined #bitcoin-builds 19:06 -!- belcher_ [~belcher@unaffiliated/belcher] has quit [Read error: error:1408F10B:SSL routines:ssl3_get_record:wrong version number] 19:06 -!- belcher_ [~belcher@unaffiliated/belcher] has joined #bitcoin-builds 19:09 -!- belcher [~belcher@unaffiliated/belcher] has quit [Ping timeout: 260 seconds] 20:15 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Read error: Connection reset by peer] 20:15 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has joined #bitcoin-builds --- Log closed Sat Mar 06 00:00:48 2021