github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
Voldenet sena_kun: I doubt it's a good solution though – musl implements dl_iterate_phdr, but I've had no idea how to detect that it's available and the resulting solution is low-effort fix that makes tests moderately happy 00:12
00:27 lucasb left 01:34 Altai-man_ joined 01:36 sena_kun left 03:35 sena_kun joined 03:36 Altai-man_ left 03:52 Util left, Util joined 05:34 Altai-man_ joined 05:36 sena_kun left 06:01 sena_kun joined 06:02 Altai-man_ left 06:24 Altai-man_ joined 06:26 sena_kun left
nwc10 good *, #moarvm 07:04
07:35 sena_kun joined 07:36 Altai-man_ left 08:27 Geth left 08:28 Geth joined 08:56 zakharyas joined 09:34 Altai-man_ joined 09:36 sena_kun left 10:19 xlat joined
Altai-man_ (cross from #raku-dev) Anyone here objecting me doing moarvm releases routine (signing tags, updating website etc, except changelog, I think) which samcv usually does? 11:15
11:34 sena_kun joined 11:36 Altai-man_ left 11:53 zakharyas left 11:56 zakharyas joined 12:24 zakharyas left 13:34 Altai-man_ joined 13:36 sena_kun left 13:39 zakharyas joined 15:19 AlexDaniel left, AlexDaniel joined, AlexDaniel left, AlexDaniel joined 15:27 harrow left 15:28 bonsaikitten left, harrow joined 15:29 xiaomiao joined 15:35 sena_kun joined 15:37 Altai-man_ left 16:57 xlat left 17:34 Altai-man_ joined 17:36 sena_kun left 18:07 zakharyas left
samcv no objections from me. though. 18:09
Geth MoarVM: c05c268b8e | (Samantha McVey)++ | tools/release.sh
Make tar builds reproducible

Taking pointers from wiki.debian.org/ReproducibleBuilds/Howto
  * Set modification time to the tag's time
  * Set the locale so that is consistent
  * Make sure umask differences are not propagated to the tar
  * Sort the output of git ls-files
18:10
samcv Altai-man_, what OS do you have?
Altai-man_ samcv, gnu/linux box, is it a problem?
or, what the requirement is? 18:11
samcv well. not really. though i'm curious if you checkout 2020.02 and run make release=2020.02; and then tell me the md5sum of the output
Altai-man_ samcv, latest? 18:12
samcv no, checkout 2020.02 tag
Altai-man_ oh, 2020.02, sorry
samcv this isn't a requirement (reproducable builds) but I think it would be pretty neat to do
Altai-man_ samcv, md5 of the whole output?
samcv uh. no the resulting file 18:13
Altai-man_ oki
samcv oh wait. uh.
nevermind. let's do the latest. i forgot my changes are not going to be in 2020.02 :D 18:14
Altai-man_ well, yes. :)
samcv so: make release VERSION=2020.02 && md5sum MoarVM-2020.02.tar.gz 18:15
Altai-man_ it dies for me with `bash -c 'if [[ $(git status --untracked=no --porcelain) ]]; then echo "Dirty work trees will make broken releases; maybe git stash?"; exit 1; fi'` because `MANIFEST` file is created, but it is created by `make release VERSION=2020.02` 18:16
I wonder what I am doing wrong
samcv does "git status" show anything? 18:18
Altai-man_ only MANIFEST file in root. if I'll remove that, on a first run of make release it is created and detected as unclean state?
maybe I should just run tools.release.sh? 18:19
s!tools.release!tools/release!
samcv, 9def3e4c486c90698c069eb0884db6e7 MoarVM-2020.02.tar.gz
samcv Altai-man_, can you try this: 18:27
git clone [email@hidden.address] && cd /tmp/MoarVM && perl Configure.pl && make release VERSION=2020.02 && md5sum MoarVM-2020.02.tar.gz
err wait
yeah. do that. in /tmp or something so we can make sure it's the same exactly
cd /tmp && git clone [email@hidden.address] && cd /tmp/MoarVM && perl Configure.pl && make release VERSION=2020.02 && md5sum MoarVM-2020.02.tar.gz
I get 6d04e9404919e597a1f32ff0d320e3e1 18:28
Altai-man_ let's see 18:33
samcv, totally different for me. :( 18:37
samcv if you do 'make release VERSION=2020.02' again, is the md5sum the same? 18:38
Altai-man_ 9def3e4c486c90698c069eb0884db6e7, I can upload my build to investigate what's different
m: say '9def3e4c486c90698c069eb0884db6e7' eq '9def3e4c486c90698c069eb0884db6e7' 18:39
evalable6 True
Altai-man_ samcv, ^
samcv Altai-man_, what is your 'tar --version' and 'gzip --version'?
Altai-man_ tar (GNU tar) 1.32, gzip 1.10
samcv Altai-man_, yeah, can you upload it somewhere? 18:41
Altai-man_ samcv, drive.google.com/file/d/1tJR-jrQSH...sp=sharing 18:43
19:12 lucasb joined
Geth MoarVM: fee45b9b42 | (Samantha McVey)++ | tools/release.sh
release.sh: enzure gzip doesn't save dates

Also ensure that the locale is exported for all programs, since sort command needs this (minimum).
19:18
samcv Altai-man_, your manifest had a different order. pushed something to try and fix that 19:19
Can you try this: rm -rf /tmp/MoarVM && cd /tmp && git clone [email@hidden.address] && cd /tmp/MoarVM && perl Configure.pl && make release VERSION=2020.02 && md5sum MoarVM-2020.02.tar.gz 19:20
Altai-man_ does
19:35 sena_kun joined 19:37 Altai-man_ left
nine samcv++ # reproducible builds are cool! 20:01
sena_kun samcv, I got f4d0e8201ed2a35a43d38c5a7e971a21 20:03
(sorry for the delay)
and a lot of `perl: warning: Falling back to a fallback locale ("en_US.UTF-8").` 20:04
lizmat you might want to google that last one :-) 20:25
samcv sena_kun, that's what i have too :) 20:40
sena_kun, i don't get that message if i try 'LC_ALL=C.UTF-8 perl -E 'say "hi"' weird 20:42
21:34 Altai-man_ joined 21:36 sena_kun left
samcv 3rdparty/sha1 is not included in the archive. is that intentional? 22:10
23:16 leont joined 23:22 lucasb left 23:35 sena_kun joined 23:36 Altai-man_ left 23:44 AlexDaniel left