El_Che AlexDaniel: claudio.ulyssis.be/var/tmp/output_...fix.tar.gz 00:40
AlexDaniel El_Che: ok that's a bit more believable 01:20
still no Unicode::GCB though :S 01:21
but less modules are “OK” 01:22
El_Che didn't look,at the rapport
more modules were tested?
AlexDaniel less 01:23
I mean… I'm not sure
El_Che weird
AlexDaniel ===> Searching for: Acme::Cow 01:25
===> Found: Acme::Cow6:ver<0.0.1>:auth<cpan:ELIZABETH> [via Zef::Repository::LocalCache]
that's not right?
or is it? I'm not sure 01:26
El_Che: anyway, that's what I wanted to see, thank you!
El_Che ping me if more tests are needed or changes to the img 01:27
AlexDaniel thanks! 01:29
El_Che no, thx you ;) 01:38
02:44 squashable6 left 02:45 squashable6 joined 03:32 redable joined 04:06 redable left 05:29 squashable6 left 05:32 squashable6 joined 06:32 squashable6 left 06:34 squashable6 joined 08:10 sena_kun joined 09:02 ilogger2_ joined 09:05 ilogger2 left 09:07 Geth joined 09:10 samcv joined 09:43 Altai-man_ joined
tyil does anyone have an email address of clarkema, the r* release manager for the macos .dmg packages? 09:47
10:38 jmerelo joined
jmerelo Hey 10:42
Old docs are for the time being deployed in docs-stage.perl6.org 10:43
Every primary page has a link to the new page in docs.raku.org in the footer.
It's still in stage, just in case someone wants to do it in some other way, maybe with a more prominent deprecation notice or something.
This is the relevant issue: github.com/perl6/doc/issues/3046 Please check it out and use it to discuss and suggest improvements. 10:44
releasable6: status 10:45
releasable6 jmerelo, Next release will happen when it's ready. There are no known blockers. 248 out of 454 commits logged (⚠ 9 warnings)
jmerelo, Details: gist.github.com/4ed4887ef08aba169f...7e4d7fb8f2
[Tux] 10501.992 10:53
Rakudo version 2019.07.1-454-g7e7aa2626 - MoarVM version 2019.07.1-312-g7ecfcc9ba
csv-ip5xs0.717 - 0.724
csv-ip5xs-206.661 - 6.842
csv-parser21.688 - 22.074
csv-test-xs-200.421 - 0.426
test7.474 - 7.695
test-t1.764 - 1.883
test-t --race0.885 - 0.889
test-t-2029.925 - 31.040
test-t-20 --race8.995 - 9.128
11:02 jmerelo left
rba jmerelo: Two things. If you like to use the new server as well for docs.perl6.org, there is almost all ready. Just let me know and if you like to get notified while you‘re offline in irc I create a user for you on znc.p6c.org. 11:06
tellable6 rba, I'll pass your message to jmerelo
11:43 sena_kun joined 11:45 Altai-man_ left 13:06 donaldh joined 13:14 lucasb joined 13:33 ggoebel joined 13:43 Altai-man_ joined 13:46 sena_kun left 13:49 donaldh left 14:01 donaldh joined 14:02 donaldh left 14:03 donaldh joined 14:26 donaldh left 15:44 sena_kun joined 15:45 Altai-man_ left 15:48 cfa joined
MasterDuke huh. this gist gist.github.com/MasterDuke17/e3ad9...57ca13e4b1 has a patch i tried and the new/changed error it causes. `Q:qq:cc/\1/` used to cause 'Unrecognized backslash sequence: '\1'. Did you mean $0?', but now causes 'This representation (P6int) does not support positional access (for type BOOTInt)' 17:15
17:44 Altai-man_ joined 17:46 sena_kun left 18:35 jmerelo joined 19:24 jmerelo left
El_Che AlexDaniel: Added a file with the binary dependencies so it can be used as the configuration for the image and the documentation of the needed pkgs 19:40
AlexDaniel: github.com/nxadm/raku-blin-docker/...pendencies
if something is missing or new modules with binary depenencies need to be tested, that's the place to add them 19:41
19:45 Altai-man_ left
AlexDaniel El_Che: that's awesome 19:47
El_Che: that's really really good 19:48
El_Che: any chance we can get that back into the main repo? github.com/perl6/Blin
El_Che: also, can you have comments in that file?
lucasb while trying some changes to core code, I got the error "This type (List) does not support positional operations". It doesn't make sense right? 19:56
moritz weird stuff happens when you screw up core :D 19:57
this could be a type not being fully composed yet, or a missing decont
or an actually bogus error message 19:58
lucasb ah :( I won't have the knowledge to hunt that down 19:59
timotimo it does make sense
you have to getattr the $!reified or something out of the List object
lucasb: ^ 20:00
lucasb the only thing I changed was Capture.perl. The error comes from regex tests... because Match is a Capture. So maybe the problem isn't my changes fault :) 20:01
but my change made the bug visible
timotimo m: say Match.^mro
camelia ((Match) (Capture) (Cool) (Any) (Mu))
timotimo oh, i see
so some code was expecting Capture to hold low-level arrays inside the list rpoperty? 20:02
lucasb hope someone can take a look :) github.com/rakudo/rakudo/issues/32...-549173390 20:08
the intention is to make \('a'=>1, :a(2)) output as-is, not as \(:a(1), :a(2)) 20:09
timotimo right
i remember that
20:11 cfa left
timotimo the existing code should have already tripped over List objects if this is the spot it happens on 20:12
is there a helpful backtrace with --ll-exception? does it just end in the Capture.perl method?
oh, lucasb, since methods have a *% parameter by default, maybe just always pass :arglist there, independent of whether it's a Pair or not 20:15
lucasb that was my original thought. So maybe it's ok to drop the Pair test 20:16
ha! sorry, my rakudo wasn't up to date github.com/rakudo/rakudo/commit/ef...8c143d0791 20:24
error was comming from this samemark stuff 20:25
El_Che AlexDaniel: comments? What type? Skip lines that start with #? Something else? 20:27
AlexDaniel El_Che: I'm thinking about mentioning which modules depend on which packages, like here: github.com/perl6/ecosystem-unbitrot/wiki 20:28
El_Che AlexDaniel: I just use a oneliner; depending on the format a lot is possible with sed/awk/perl/raku 20:29
now is just "give the contents of the file to apt-get" :)
AlexDaniel: columns, splits and so on are easy to do: awk -F\; '{print $2}' (print the 2nd column of a ";" separated file 20:31
AlexDaniel El_Che: right 20:39
El_Che: it's kinda unfortunate that there's no better way to do it
ideally this info should be in META6.json file in some form
because right now adding to this list helps blin but doesn't improve the user experience 20:40
that is, the user still has to figure out which dependencies they need to install
El_Che a suggest system per OS would be nice
AlexDaniel El_Che: right, but a suggest system based on *what* 20:41
El_Che you mean install at run time? (possible but less evident)
AlexDaniel like we can have zef say something like “Native dependencies are missing, you can install them by running `apt install libwhatever-whatever`” 20:42
El_Che "debian_suggest": [ "libfoo-dev", "foobar2" ], centos_sugget: "[...]"
rba btw. maettu made Blin running in 164m25.818s. is this good to go?
AlexDaniel rba: PERFECT.
El_Che AlexDaniel: however, if you want the user to add dependencies at run time, I'll need to adapt the image a little 20:43
we can use the -e (environment switch) 20:44
docker run -ti -e PKGS="libfoo libbar3 libzork" rakudo/blin
the dblin wrapper could wrap that (read env vars, or use switches etc) 20:45
AlexDaniel: about merging, sure, but maybe best when we figure out what we need and want
AlexDaniel rba: I think you're aware of it but it doesn't hurt mentioning it again. Blin installs all modules in the ecosystem, meaning that it takes one idiot to include some destructive garbage in their code to take down parts of your system. Running it in a docker image isn't too bad, but still be aware that angry code can explore your local network, abuse kernel bugs, etc. 20:52
El_Che: what you did is exactly what I wanted :)
rba Yes. I told to maettu to make sure the firewall does block all except the internet. 20:53
El_Che rba: have a look running it with podman instead of docker
rba: is does not requires root
rba El_Che: got it and proxied it to maettu 20:54
El_Che rba: great 20:55
Geth rakudo: tbrowder self-assigned =defn block parsing is not complete github.com/rakudo/rakudo/issues/3275
lucasbuchala++ created pull request #3276: Call .perl(:arglist) for Capture's list elements
El_Che AlexDaniel: for now it skips lines starting with # in the pkg-dependencies 21:05
MasterDuke i thought META6.json already had a section specced for native dependencies? 21:16
El_Che AlexDaniel: re: github.com/perl6/ecosystem-unbitrot/wiki . It looks like 2 colum setup. We can do a 2 column file saparated by ";" (or something else, like white space), with a header that will be skipped 21:22
pkg; module info 21:23
t META6.json already had a section specced for
AlexDaniel: something like this? gist.github.com/nxadm/8fa41561d184...2d055ebd23 21:25
21:48 AlexDani` joined 21:54 tbrowder joined 21:55 ggoebel left 23:33 lucasb left 23:40 AlexDani` is now known as AlexDaniel, AlexDaniel left, AlexDaniel joined
AlexDaniel "runtime": { "requires": [ "python2.7:from<native>" ] } 23:42
El_Che: I'm not entirely sure what it means. Should we then have a squashathon for PR-ing this section to modules that have native dependencies? 23:45
El_Che: then we can write a script that uses that info to figure out which debian dependencies need to be install, so that it generates that file automatically 23:46
El_Che Going to bed, but I am proposing a format that we can easily parse without extra libs 23:48
the file itself can be autogrnerated
23:54 sena_kun joined
El_Che maybe it makes more sense to generaye q simple pkg file from whatever complex format we have for doc 23:55
see ypubtomorrow
AlexDaniel o/