Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
03:26 MasterDuke left 04:16 Blue^NeXuS left 04:25 Blue^NeXuS joined 04:30 Blue^NeXuS left 04:42 Blue^NeXuS joined 04:58 Blue^NeXuS left 05:00 Blue^NeXuS joined 05:48 epony left 06:05 Blue^NeXuS left 07:01 kanopis joined 07:45 kanopis left 08:18 patrickb joined
patrickb .tell vrurg Did you see colabti.org/irclogger/irclogger_lo...-06-16#l75 08:25
yoleaux patrickb: I'll pass your message to vrurg.
09:08 patrickb left 09:26 squashable6 left, pmurias joined
pmurias nine: I'm creating a CompUnit::Repository::Node 09:26
paste.debian.net/1088166 09:27
nine: but the problem is that modules in node.js export an arbitary object
09:27 squashable6 joined 09:50 patrickb joined
AlexDaniel jnthn: hello, any good reason why rakudo repo is not under perl6 organization? 10:05
for example, a disadvantage of that is that it's impossible to move tickets from rakudo repo to problem-solving 10:06
arguably a github limitation, but still
kawaii I assume because Rakudo is an implementation of Perl 6, not Perl 6 itself AlexDaniel 10:08
AlexDaniel kawaii: but why a perl 6 implementation can't live in perl6 organization? 10:09
kawaii So if more Perl 6 compilers were born, would they too live under the Perl 6 org, alongside Rakudo? 🤔 10:10
AlexDaniel if they want to, why not?
the base permission in perl6 org is currently set to `read` only, so it's not about privileges I assume
kawaii I don't think it matters either way, but might be nice to consolidate if there are no objections 10:11
Though, it may give the impression that Rakudo is the 'official' Perl 6 compiler
AlexDaniel IMO same goes for MoarVM/ org which doesn't even have a logo :) 10:12
kawaii Poor MoarVM, someone make those guys a logo please
I _think_ it makes more sense for moar to move under the perl6 org, at least since Perl 6 is the only language that runs under that backend 10:13
AlexDaniel moarvm.org repo is in MoarVM org, but rakudo.org is in perl6/ :)
10:14 squashable6 left
kawaii MoarVM is in sense, a first-class citizen in Perl 6 world, whereas Rakudo is more like a second/third-party component? 10:14
I don't know, too early to engage brain this much
10:15 squashable6 joined, ChanServ sets mode: +v squashable6 10:51 pmurias left 10:52 pmurias joined
releasable6 Next release in ≈3 days and ≈7 hours. 3 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 11:00
11:02 pmurias left 11:03 pmurias joined
pmurias AlexDaniel: I think different sets of commiters are the reason 11:15
AlexDaniel pmurias: but this can be done using Teams, right?
kawaii I doubt the list of committers between rakudo and nqp etc... are that different surely? 11:19
AlexDaniel kawaii: yeah, but committing to rakudo requires signing a CLA 11:21
that's the only difference I think
lizmat Files=1275, Tests=108113, 209 wallclock secs (26.30 usr 7.33 sys + 2905.29 cusr 271.75 csys = 3210.67 CPU) 11:22
11:29 pmurias left
|Tux| Rakudo version 2019.03.1-624-g6f8442ada - MoarVM version 2019.05-55-g8c85b4a46
csv-ip5xs1.101 - 1.109
csv-ip5xs-2057.248 - 57.514
csv-parser21.770 - 22.446
csv-test-xs-200.433 - 0.442
test6.053 - 6.727
test-t1.685 - 1.757
test-t --race0.759 - 0.844
test-t-2027.118 - 27.226
test-t-20 --race8.363 - 9.242
 11:48
patrickb Is there any real reasoning why rakudo would require a CLA and NQP or MoarVM don't? 11:52
Especially because NQP is located under the "perl6/" umbrella and thus seems to have a tighter bound to TPF... 11:53
kawaii Yes I wonder the same thing, Rakudo from what I can tell does not have any official attachment to TPF 11:54
11:55 AlexDaniel left, AlexDaniel joined 11:57 Guest77572 joined
Guest77572 only three blockers, impressive 11:57
12:02 ggoebel joined
vrurg .tell nine Please, when you complain about a build problem – tell me how you got into it. Best – make a ticket. 12:16
yoleaux 08:25Z <patrickb> vrurg: Did you see colabti.org/irclogger/irclogger_lo...-06-16#l75
vrurg: I'll pass your message to nine.
12:17 vrurg left
patrickb vrurg++ 12:18
AlexDaniel, kawaii: What's your thoughts on the release planed in 3 days? 12:22
lizmat patrickb: I understand why MoarVM doesn't, as it is something that jnthn started on his own
patrickb: re nqp, I guess that's a similar situation, but that's more shrouded in the mists of time for me 12:23
kawaii AlexDaniel: of the 3 blockers, I think this one at the very least should be dealt with github.com/rakudo/rakudo/issues/2759
patrickb s/planed/planned/
kawaii AlexDaniel: also schedule wise, I am fine with 3 days time doing a release alongside you, are you also fine with this day? 12:24
the other two blockers are... not as serious IMO
at least not enough to prevent a release
though I certainly need to be able to run Blin again, and can't currently until rba has got that cert in place 12:25
12:30 Ulti left 12:39 Ulti joined
AlexDaniel realistically? I don't think it's going to happen in exactly three days 12:48
we don't even have Blin results yet
I'm busy till tomorrow, then I'm mostly free 12:49
AppVeyor is still broken, is there a ticket about it somewhere? 12:51
patrickb AlexDaniel: It's one of the three blockers 12:52
AlexDaniel ah yeah, marked as blocker
ok 12:53
patrickb AlexDaniel: Did you notice the autocomplete in rakudobrew? :-)
AlexDaniel not great, not terrible
timotimo nqp was initially more a part of parrot, right? did parrot have a cla?
AlexDaniel patrickb: no, but I did notice that I had to manually choose Bash in the config, and was pissed off a bit :) 12:54
patrickb: but now that you said it, it's pretty awesome!
12:58 vrurg joined
patrickb vrurg: o/ 13:00
AlexDaniel patrickb: btw why can't we use $SHELL to figure out which shell it is?
vrurg Hi! 13:01
patrickb: sitting at TPC now. Would try to be useful if anything is needed.
patrickb: do you know what went wrong for nine? 13:02
patrickb vrurg: Which TPC?
timotimo in riga?
vrurg Pittsburg 13:03
timotimo ah, merica
patrickb AlexDaniel: I'm not sure anymore. It was quite an adventure before I arrived at the current solution. I think there is no reliably way to determine the shell currently running. With $SHELL my memory is foggy. Might have been a problem with that variable not being reset when spawning subshells.
AlexDaniel kawaii: the question is whether we can indeed make sure that the current state of rakudo doesn't have additional serious bugs that the previous release didn't have 13:05
kawaii: without the info from Blin I really can't tell
but AppVeyor tests should definitely pass before we make a new release
releasable6: status
patrickb vrurg: Sadly I don't. The piece of the console log he pasted contained "bash config.status".
releasable6 AlexDaniel, Next release in ≈3 days and ≈5 hours. 3 blockers. 221 out of 624 commits logged (⚠ 1 warnings)
AlexDaniel, Details: gist.github.com/5d1b217b07cb0ec551...f9d1968824
kawaii AlexDaniel: yes I know, but Blin is currently inoperable due to the TLS issue, hopefully I can trigger one tonight 13:06
patrickb vrurg: Do you live in the US? I somehow assumed you'd live in the Netherlands... 13:07
AlexDaniel kawaii: yeah, so 3.6 blockers :)
vrurg patrickb: wish I could see what's in that config.status.How come it came up with unix/nqp-m template whereas there is no such thing whatsoever?
AlexDaniel until we get data from Blin
vrurg patrickb: No, haven't been to Europe for almost 5 years now. :) 13:08
patrickb :-)
nine vrurg: I'm pretty sure it was: /usr/bin/perl Configure.pl "--git-protocol=https" "--prefix=/home/nine/rakudo/install" "--backends=moar" "--expand=/home/nine/rakudo/nqp/tools/templates/unix/nqp-m" "--out=nqp-m" $* 13:15
yoleaux 12:16Z <vrurg> nine: Please, when you complain about a build problem – tell me how you got into it. Best – make a ticket.
nine The --expand and --out part was what caused the issue. I got it working by running Configure.pl manually.
vrurg nine: That's fixed already. Perhaps you didn't run your Configure for a couple of weeks. 13:16
Yes, it was erroneously saving config.status with --expand
nine vrurg: well I _was_ trying to run Configure.pl :) 13:17
So what about camelia's build issue? File 'unix/nqp-m-' not found in base directory /home/camelia/rakudo-m-2/nqp/tools/templates in file /home/camelia/rakudo-m-2/nqp/tools/templates/moar/Makefile.in at /home/camelia/rakudo-m-2/nqp/3rdparty/nqp-configure/lib/NQP/Config.pm line 862. at /home/camelia/rakudo-m-2/nqp/3rdparty/nqp-configure/lib/NQP/Config.pm line 34. 13:19
vrurg nine: Can you update submodule or run Configure && git pull? 13:21
13:21 lucasb joined
vrurg It's due to a recent change by patrickb and happens if submodule is not new. 13:21
Or wipe nqp dir if nqp build is initiated by rakudo build. 13:24
14:06 ggoebel left 14:12 vrurg left 14:18 vrurg joined 14:43 skids joined 14:50 vrurg left, vrurg joined 14:51 vrurg left 15:01 vrurg joined
nine vrurg: submodule update seems to have fixed it. Thanks! 15:03
Why doesn't it do that automatically?
vrurg It does. You just didn't run Configure for long on that setup. 15:04
nine m: say $*PERL.compiler.version 15:05
camelia v2019.03.1.624.g.6.f.8442.ada
nine But if it depends on me running Configure.pl it's not really automatic, isn't it?
vrurg nine: Just once, to let Configure set git's submodules.recurse 15:06
After that every git pull will update 15:07
nine Also it looks like camelia's build script does run Configure
github.com/perl6/evalbot/blob/mast...oar.pl#L56 15:08
vrurg Hard to say. Unless somehow submodule.recurse gets reset in between, should be ok. I mean, it could fail once but should do right on next iteration. 15:11
nine I guess we'll find out ;) 15:12
vrurg Nah, ignore! Rakudo will be ok, but nqp – unlikely.
Don't remember if git clean -xdf cleans up subdirs too? Because if it don't then nqp wouldn't have its configure run. 15:13
"If an untracked directory is managed by a different Git repository, it is not removed" – so, nqp dir is left alone. Perhaps we need a kind of --force key to re-run nqp/moar configures forcingly when needed. 15:15
github.com/rakudo/rakudo/issues/2993 – do we really need release-number and codename attributes in Compiler class? 15:21
15:33 vrurg left 15:37 patrickb left
[Coke] .seen mst? 15:41
yoleaux I haven't seen mst? around.
[Coke] .seen mst
yoleaux I saw mst 10 Jun 2019 18:43Z in #perl6: <mst> Rayler: and if you want to play around, volunteering to be a crash test dummy for people is always appreciated
ugexe vrurg: no 15:43
at least codename anyway 15:45
release-number might have some use, although i don't mind seeing that go either 15:46
timotimo i seem to recall we spoke about this at some point in the past, but does anybody know what "nqpevent" is for? it's in the HLL compiler, passing --nqpevent=blah will put it in the adverbs, and the backend class has a nqpevent method that is passed whatever positionals (but not nameds?) the compiler's nqpevent method is called with 16:13
my irclogs don't actually mention it except for this line i just wrote, huh 16:14
17:10 lucasb left
timotimo geth is down again, eh? but i've got to AFKBBL 17:15
github.com/rakudo/rakudo/commit/34...3cd477c195 - i pushed this, but there's no tests yet
17:21 vrurg joined 17:24 vrurg left 17:27 vrurg joined 17:55 lucasb joined, vrurg left, vrurg joined 18:02 vrurg left 18:07 vrurg joined 18:24 kawaii left 18:27 kawaii joined 18:35 vrurg left 18:39 vrurg joined 18:40 vrurg left 18:42 vrurg joined 18:46 discord6 left 18:48 discord6 joined, discord6 left, discord6 joined, Geth joined 19:12 MasterDuke joined, MasterDuke left, MasterDuke joined
[Coke] .seen nwc10 19:20
yoleaux I saw nwc10 13 Jun 2019 06:35Z in #moarvm: <nwc10> good *, #moarvm
[Coke] .seen util 19:21
yoleaux I saw Util 24 Jan 2019 23:10Z in #perl6: <Util> m: my @a = <abc def ghi jkl>; .say for @a.rotor(2);
Geth roast: ea484809fb | usev6++ | 2 files
Update location of spectest.data

That file has been moved from Rakudo to roast with
  github.com/rakudo/rakudo/commit/ed6d91c085 and
  github.com/perl6/roast/commit/ae818a10e7 .		 19:36
19:43 vrurg left
timotimo i don't wanna put a rebootstrap in :( :( 19:44
in that case, i guess rakudo has to mixin the functionality at run time :P 19:45
why don't we have a #?stage2
actually, wouldn't i be able to just put that in and things will be fine? 19:46
19:49 AlexDaniel left 19:58 vrurg joined
Geth rakudo: b1fac3d6f1 | usev6++ | tools/templates/jvm/Makefile.in
[JVM] Use --release 9 instead of -source 1.9

This avoids a warning about -bootclasspath being needed; similiar to
  github.com/perl6/nqp/commit/104a5ce4d4		 19:58
19:59 ggoebel joined
timotimo heck yeah 20:06
i got an nqp with --confprog support but didn't have to rebootstrap
Geth nqp: 268f985ccf | (Timo Paulssen)++ | 5 files
Allow #?if stage1/stage2, add confprog loading support via cmdline flag
 20:09
timotimo check it out!
MasterDuke cool 20:12
20:13 travis-ci joined
travis-ci NQP build failed. Timo Paulssen 'Allow #?if stage1/stage2, add confprog loading support via cmdline flag' 20:13
travis-ci.org/perl6/nqp/builds/546932222 github.com/perl6/nqp/compare/85775...8f985ccf61
20:13 travis-ci left
timotimo oh crap 20:14
ooooh whoops 20:15
of course
installconfprog is a too-new op
it's not even on moarvm's master branch
silly me, sorry about that
Geth nqp: cdc9e773ac | (Timo Paulssen)++ | 2 files
back out installconfprog until it's in moarvm master
 20:16
bartolin timotimo: also, there seems to be a missing question mark in the added endif in Compiler.nqp (line 32). 20:21
timotimo but then how does it work at all o_O 20:22
20:22 travis-ci joined
travis-ci NQP build failed. Timo Paulssen 'back out installconfprog until it's in moarvm master' 20:22
travis-ci.org/perl6/nqp/builds/546935207 github.com/perl6/nqp/compare/268f9...c9e773acbd
20:22 travis-ci left
bartolin ^^ that's the error on the other backends. I guess on moar the #?if is interpreted you just get an additional comment instead of an interpreted #?endif 20:24
I mean 'interpreted by the preprocessor'
timotimo pushed a fix
Geth nqp: ab21c2decc | (Timo Paulssen)++ | src/HLL/Compiler.nqp
add missing ?, thanks bartolin++
bartolin \o/ travis++, timotimo++ 20:25
timotimo any specific reason we don't run the js backend for nqp regularly? 20:26
AFKBBL 20:27
bartolin m: say DateTime.new("2013-12-23T12:34:36Z").later(minute => 1) 20:43
camelia Cannot look up attributes in a Date type object
in block <unit> at <tmp> line 1
20:44 travis-ci joined
travis-ci NQP build passed. Timo Paulssen 'add missing ?, thanks bartolin++' 20:44
travis-ci.org/perl6/nqp/builds/546938940 github.com/perl6/nqp/compare/cdc9e...21c2decca2
20:44 travis-ci left
bartolin ^^ is that a known regression? I didn't find a recent issue but saw a failure in S32-temporal/DateTime.t. It seems to work with latest rakudo star (2019.03.1). 20:47
bisectable6: bad=HEAD good=2019.03.01 say DateTime.new("2013-12-23T12:34:36Z").later(minute => 1) 20:49
bisectable6 bartolin, Cannot find revision “2019.03.01” (did you mean “2019.03.1”?)
bartolin bisectable6: bad=HEAD good=2019.03.1 say DateTime.new("2013-12-23T12:34:36Z").later(minute => 1)
bisectable6 bartolin, Bisecting by exit code (old=2019.03.1 new=b1fac3d). Old exit code: 0
bartolin, bisect log: gist.github.com/9c169693d4900ade4e...eb43c34e29
bartolin, (2019-06-17) github.com/rakudo/rakudo/commit/34...3cd477c195
bartolin oh, it's from today. 20:50
20:51 AlexDaniel joined
bartolin actually there is a similiar failure in S03-operators/misc.t 20:54
lizmat 34c0e892ca9649b9b seems the cause ? 20:56
bisectable6: say DateTime.new("2013-12-23T12:34:36Z").later(minute => 1)
bisectable6 lizmat, Bisecting by exit code (old=2015.12 new=b1fac3d). Old exit code: 0
lizmat, bisect log: gist.github.com/e71abf917e597a88a4...6bbb6ae4a6
lizmat, (2019-06-17) github.com/rakudo/rakudo/commit/34...3cd477c195
bartolin yes (see the other output from bisectable6 above :)) 20:57
lizmat sorry... was still looking at the initial report 20:58
20:58 vrurg left
bartolin np. geth was absent when it was pushed: colabti.org/irclogger/irclogger_lo...06-17#l173 20:58
20:59 patrickb joined, travis-ci joined
travis-ci Rakudo build failed. usev6 '[JVM] Use --release 9 instead of -source 1.9 20:59
travis-ci.org/rakudo/rakudo/builds/546928294 github.com/rakudo/rakudo/compare/3...fac3d6f112
20:59 travis-ci left
bartolin hmm, well, that travis failure is not related to my change. it looks like it was build within the few minutes that nqp HEAD was broken. 21:02
lizmat Geth has been awol for a few days already :-( 21:03
nobody seems to care :-(
bartolin but he is back now! 21:04
lizmat aaah...
lizmat is probably too tired to do this sort of thjng
or too distracted doing the P6W
21:05 skids left 21:06 vrurg joined
timotimo oh damn 21:07
i b0rked things again :) :)
lizmat commit on its way
[Coke] waves to lizmat
Geth rakudo: 5bdeedcbe7 | (Elizabeth Mattijsen)++ | src/core/Date.pm6
You cannot use attributes on type objects

Maybe "new-from-daycount" should become a multi, that also takes instance objects to create new objects from? At least this fixes a very clear execution error.		 21:08
timotimo thanks for picking up after me, lizmat++ 21:09
MasterDuke lizmat: can you share a repro for that busted profile you showed me a little while ago? 21:13
timotimo any reason to make the proto for new-from-daycount any more specific than (|) {*}? 21:15
lizmat introspection ? 21:16
reportable6: 2019-06-10T12:00:00 2019-06-17T00:00:00Z 21:21
reportable6 lizmat, OK, working on it! This may take up to 40 seconds
lizmat notable6: weekly
notable6 lizmat, 9 notes: gist.github.com/8595388760584cf046...afccb52539
lizmat notable6: weekly reset 21:22
notable6 lizmat, Moved existing notes to “weekly_2019-06-17T21:22:08Z”
reportable6 lizmat, gist.github.com/a5b28c410734c096be...20135f3787
timotimo i mean if i constrain it to "exactly one positional" or constrain the return type, it'll reduce what end users can extend it for 21:27
lizmat timotimo: that's a dilemma that applies to many subs in the core.. at one point we changed the proto's of a lot of subs for that reason 21:29
timotimo what direction did we change them in?
lizmat from (|) to ($,$) e.g.
timotimo ah, ok, so i'll make it ($) 21:33
21:54 ggoebel left 22:03 vrurg left
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2019/06/17/...es-of-int/ 22:09
22:23 travis-ci joined
travis-ci Rakudo build passed. Elizabeth Mattijsen 'You cannot use attributes on type objects 22:23
travis-ci.org/rakudo/rakudo/builds/546955749 github.com/rakudo/rakudo/compare/b...deedcbe77e
22:23 travis-ci left
Geth rakudo: f05b1b0803 | (Timo Paulssen)++ | src/core/Date.pm6
only use attribute access for definite objects
 22:30
22:59 patrickb left
timotimo lizmat++ 23:04
23:06 epony joined 23:46 Geth_ joined, Geth_ left 23:48 vrurg joined 23:56 vrurg left

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!