--- Log opened Mon Jun 21 00:00:44 2021
01:45 -!- luke-jr [~luke-jr@user/luke-jr] has quit [Quit: ZNC - http://znc.sourceforge.net]
01:46 -!- luke-jr [~luke-jr@user/luke-jr] has joined #bitcoin-core-builds
02:07 -!- belcher_ is now known as belcher
02:27 -!- ghost43 [~ghost43@gateway/tor-sasl/ghost43] has quit [Remote host closed the connection]
02:30 -!- ghost43 [~ghost43@gateway/tor-sasl/ghost43] has joined #bitcoin-core-builds
04:14 -!- orionwl [~laanwj@user/laanwj] has joined #bitcoin-core-builds
04:16 -!- laanwj [~laanwj@user/laanwj] has quit [Ping timeout: 268 seconds]
07:55 -!- orionwl is now known as laanwj
08:06 -!- luke-jr [~luke-jr@user/luke-jr] has quit [Quit: ZNC - http://znc.sourceforge.net]
08:07 -!- luke-jr [~luke-jr@user/luke-jr] has joined #bitcoin-core-builds
08:16 -!- laanwj [~laanwj@user/laanwj] has quit [Ping timeout: 265 seconds]
08:22 -!- laanwj [~laanwj@user/laanwj] has joined #bitcoin-core-builds
10:15 < dongcarl> hebasto: thinking...
10:21 < dongcarl> The obvious way to do this is to just have Guix build toolchains for non-riscv64 against glibc 2.17 and toolchains for riscv64 against glibc 2.27... However, I'm wondering if there'd be fixes between 2.17 and 2.27 that would not change the symbol version, but is still important, especially because 2.17 is EOL whereas the 2.27 release branch is
10:21 < dongcarl> still maintained.
10:27 < dongcarl> What I'm saying is that there are 2 ways:
10:27 < dongcarl> 1. Build against whatever MAX_VERSION we have for GLIBC (pros: eliminate the possibility for version incompatibilities, cons: 1. missing fixes btw 2.17 and 2.27, 2. Guix will be the only build pipeline we can do this on (possibility for subtle bugs?))
10:27 < dongcarl> 2. Build against 2.27 for all and utilize the glibc backcompat (pros: includes fixes btw 2.17 and 2.27, Guix pipeline will align with bionic builders, cons: reliance on the correctness of glibc backcompat and symbol check)
10:29 -!- jamesob [sid180710@id-180710.brockwell.irccloud.com] has joined #bitcoin-core-builds
10:29 < dongcarl> I don't really care if we do either way here, just want to make a decision
10:29 < dongcarl> Also, why 2.24 and not 2.17?
10:45 -!- gribble [~gribble@bitcoin/bot/gribble] has joined #bitcoin-core-builds
11:32 < dongcarl> some musl info: https://github.com/bitcoin/bitcoin/issues/18110#issuecomment-865253852
11:32 < sipa> what is the oldest libc on common platforms we want to support running on?
11:32 < sipa> *glibc, that is
11:48 < dongcarl> sipa: 2.17 I think
11:49 < dongcarl> See: https://github.com/bitcoin/bitcoin/pull/22244/files#diff-e85bcfd32a45808ce584aec5a4935a524045bb1166291156857e2880bcd45cc6R44-R51
12:55 -!- luke-jr [~luke-jr@user/luke-jr] has quit [Quit: ZNC - http://znc.sourceforge.net]
12:57 -!- luke-jr [~luke-jr@user/luke-jr] has joined #bitcoin-core-builds
15:12 < hebasto> dongcarl: re "20:29:58 <dongcarl> Also, why 2.24 and not 2.17?" - as you wrote, no version change in the used symbols plus bug and security fixes
17:07 -!- belcher_ [~belcher@user/belcher] has joined #bitcoin-core-builds
17:10 -!- belcher [~belcher@user/belcher] has quit [Ping timeout: 258 seconds]
17:16 < sipa> dongcarl: we don't actually statically link glibc, right? so even building against an old version doesn't mean we're stuck with security issues that release has?
17:17 < sipa> so the concern is really just about headers-side problems in glibc, which is a very thin layer only i think
18:17 < hebasto> dongcarl: `HOSTS=i686-pc-linux-gnu contrib/guix/guix-build` builds nothing
19:45 -!- gribble [~gribble@bitcoin/bot/gribble] has quit [Remote host closed the connection]
19:50 -!- gribble [~gribble@bitcoin/bot/gribble] has joined #bitcoin-core-builds
20:57 -!- robertspigler [~robertspi@2001:470:69fc:105::2d53] has joined #bitcoin-core-builds
22:15 -!- nanotube [~nanotube@user/nanotube] has quit [Ping timeout: 240 seconds]
22:16 -!- nanotube [~nanotube@user/nanotube] has joined #bitcoin-core-builds
--- Log closed Tue Jun 22 00:00:45 2021