IOninja I stresstest with it installed yesterday... 00:00
IOninja tries the latest and greatest version
TimToady might've been residual precomp stuff from braids...
no, trying to reinstall Inline::Perl5 then 00:02
nope, fails
gets failures like: 00:03
# Failed test 'parse :from<Perl5> syntax'
# at t/spec/S01-perl-5-integration/import.t line 17
# Error: Type check failed in binding to $handle; expected CompUnit::Handle but got Sub (sub EXPORT (*@args) {...)
nine: ^^ does this ring any bells, other than alarm? :)
IOninja oh... I think I recall someone having that issue in #perl6 :o 00:07
TimToady I confess, I'm glad that it didn't turn out to be my braids branches, but this is probably a release showstopper
IOninja Yeah. I should've thought to test a fresh clone in pre-release preparations >_< 00:08
TimToady if you look at the code, the .release method is definitely trying to return a sub, not a CompUnit
so the binder has a valid complaint 00:09
IOninja well, don't see any I::Perl5 fails, but t/spec/S17-channel/stress.t hung again. Waited 3 minutes. When run alone it usually completes 11-40secs 00:19
grrrr... and it finishes just as I finish typing that -_-
All tests successful.
Files=1225, Tests=132898, 313 wallclock secs (21.33 usr 3.30 sys + 5553.67 cusr 325.51 csys = 5903.81 CPU)
Result: PASS
TimToady: ^ I didn't get any failures with a fresh clone of everything 00:20
TimToady using zef?
IOninja No, panda. Just copy-pasting step 12 of release guide: github.com/rakudo/rakudo/blob/nom/..._guide.pod
TimToady maybe I should try a realclean
or panda...
which version does panda install? zeg gets 0.13 00:23
*zef
IOninja Not sure. I've just nuked everything and gonna try zef. I'd guess it installs latest 00:24
TimToady heh, panda fails to install I::P5 because LibraryMake is already install, which is a bit antisocial 00:27
would be nice if panda mentioned which version it is installing... 00:29
TimToady notes that the last entry in 'git log' in .zef's I::P5 directory has a date of Sep 19 00:31
now where does panda keep stuff?...
IOninja no idea 00:32
TimToady btw, panda-installed I::P5 seems to work better
so looks like zef is just not installing the latest, for some reason 00:34
now if only I could figure out a way to get Inline::Perl5 to actually tell me what version it is... 00:38
IOninja installed 0.23 I::P5 with zef... All tests pass. Files=1225, Tests=132898, 137 wallclock secs (20.93 usr 3.03 sys + 2797.51 cusr 278.02 csys = 3099.49 CPU) 00:41
Pretty sure there's a Comp::Unit incantation that let's you look at meta file 00:42
`zef info Inline::Perl5` tells the version 00:44
TimToady zef install 0.13, panda installs 0.23 00:47
slight difference there...
okay, so maybe I can test I::P5 with braids now, yay 00:49
Geth geth: 63cfd48bac | (Zoffix Znet)++ | bin/geth.p6
Update URL
00:54
nqp/braids: 11 commits pushed by (Pawel Murias)++, (Moritz Lenz)++, TimToady++
review: github.com/perl6/nqp/compare/d0d0b...d77f4fcc4e
01:17
rakudo/braids: 4 commits pushed by (Zoffix Znet)++, (Moritz Lenz)++, TimToady++
geekosaur re stupid optimizer tricks from the other day... plus.google.com/u/0/10220845651992...ZsZ6mvA4T6 01:31
ugexe TimToady: `zef update && zef install Inline::Perl5` (like apt-get), or `zef install "Inline::Perl5:ver<0.22>"` `zef install "github:niner:Inline-Perl5:0.22"` should work 03:24
yoleaux2 16 Feb 2017 21:23Z <nine> ugexe: your patch looks fine. It uncovered an interesting riddle though
ugexe m: say CompUnit::DependencySpecification.new(:short-name("xxx")).version-matcher; # otherwise the cache ends up matching Inline::Perl5 as Inline::Perl5:ver<*> 03:26
camelia rakudo-moar ffae3f: OUTPUT«True␤»
ugexe star: say $*REPO.need(CompUnit::DependencySpecification.new(:short-name("JSON::Fast"))).distribution.meta<ver>; # doesn't work on anything not installed to CURI though 03:32
camelia star-m 2016.10: OUTPUT«v0.4␤»
IOninja huggable: version :is: Lookup version of *installed* module: star: say $*REPO.need(CompUnit::DependencySp​ecification.new(:short-name("JSON:​:Fast"))).distribution.meta<ver>; 03:35
huggable IOninja, Added version as Lookup version of *installed* module: star: say $*REPO.need(CompUnit::DependencySp​ecification.new(:short-name("JSON:​:Fast"))).distribution.meta<ver>;
ugexe technically that should work on -I modules if they provide a META6.json as well (it just doesn't yet). This is its advantage over calling .resolve on each CURI 03:36
IOninja huggable: version :is: Lookup version of *installed* module: star: say $*REPO.need(CompUnit::DependencySpecification.new(:short-name("JSON::Fast"))).distribution.meta<ver>; 03:37
huggable IOninja, Added version as Lookup version of *installed* module: star: say $*REPO.need(CompUnit::DependencySpecification.new(:short-name("JSON::Fast"))).distribution.meta<ver>;
IOninja
.oO( log site adding some sort of invisible breaker char? )
IOninja & bed
ugexe m: say map { .meta<name ver> }, grep { .so }, map { .installed.Slip }, grep { $_ ~~ CompUnit::Repository::Installation }, $*REPO.repo-chain; 03:47
camelia rakudo-moar ffae3f: OUTPUT«((Inline::Perl5 v0.22) (CORE v6.c))␤»
ugexe star: say map { .meta<name ver> }, grep { .so }, map { .installed.Slip }, grep { $_ ~~ CompUnit::Repository::Installation }, $*REPO.repo-chain;
camelia star-m 2016.10: OUTPUT«((File::Find v0.1) (Shell::Command v*) (JSON::Fast v0.4) (File::Which v*) (panda v2016.02) (XML::Writer v*) (SVG v*) (SVG::Plot v*) (Terminal::ANSIColor v*) (Test::Mock v*) (Grammar::Profiler::Simple v0.01) (Grammar::Debugger v*) (MIME::Base64 v1.2) (URI vv.0.…»
nine TimToady: it does 07:37
ugexe: I've been bitten by a stale zef cache a couple of times, too and the same is probably true for many people. What do you think about auto-update if the cache is older than a day? 07:40
[Tux] This is Rakudo version 2017.01-223-gffae3ffb8 built on MoarVM version 2017.01-72-g542baec7 08:23
csv-ip5xs 3.203
test 13.324
test-t 4.980 - 5.130
csv-parser 13.958
samcv NeuralAnomaly, status 08:49
NeuralAnomaly samcv, [✘] Next release is today. Since last release, there are 48 new still-open tickets (1 unreviewed and 0 blockers) and 13 unreviewed commits. See perl6.fail/release/stats for details
nine How on earth is it possible that two type objects from two different precompiled modules share the same SC object? 13:19
But only if they have the same short name
What the? Though it's the type objects that share the SC, the exported subs seem to be the cause of the sharing. 14:17
When I give the subs different names, the type objects suddenly get different SCs.
Talking about gist.github.com/ugexe/b8ac60d26a17...62e9716d96 btw. In case anyone's actually reading this ;) 14:18
IOninja well, I'm reading but only in the context of "does impact today's release?" :) 14:23
nine Ooooh....it's not actually the name of the exported sub. It's that both modules have identical source. Even adding a comment to one of them fixes the issue.
IOninja ooh :)
Geth rakudo/nom: ca102c51ae | (Zoffix Znet)++ | src/core/List.pm
Make flat() propagate is-lazy if possible

The **@foo slurpy ditches the is-lazy flag, so we fix that by adding an Interable candidate that will preserve it when just an Iterable is given to flat().
15:23
roast: 1c7470f84e | (Zoffix Znet)++ | S02-types/array.t
Test flat() preserves .is-lazy when possible

Rakudo fix: github.com/rakudo/rakudo/commit/ca102c51ae
15:24
rakudo/nom: 673f06bc8d | (Zoffix Znet)++ | src/core/Mu.pm
De-break round-trip of some circular structures

The fix[^1] to fix .perl.EVAL roundtripping of Array and Hash circular accidentally broke compilation of .perl of all the other types of structures.
Even though they're still broken in the sense that their EVAL does not produce the correct values, this commit at least unbreaks their compilation.
  [1] github.com/rakudo/rakudo/commit/67...35189bcc70
16:04
timotimo Interator %) 16:17
DrForr Interossiter? 16:19
IOninja ? 16:22
DrForr /This Island Earth/ ref. No worries. 16:24
IOninja What's with Iterator %)? 16:26
geekosaur it-erator 16:27
TimToady so, maybe there's a good reason to bump the version number in the source, as well as the metadata... 17:39
Geth rakudo/nom: 8519794d22 | (Zoffix Znet)++ | docs/ChangeLog
log all changes to date

Logs commits: 11d005e 4b85db6 673f06b 67aeefa 8f53a6f dc69daf
17:45
IOninja NeuralAnomaly: status 17:54
NeuralAnomaly IOninja, [✔] Next release is today. Since last release, there are 47 new still-open tickets (0 unreviewed and 0 blockers) and 0 unreviewed commits. See perl6.fail/release/stats for details
IOninja NeuralAnomaly: cut the release 17:55
ummm
what the hell was the command...
Robot, do stuff!
Ohhhhh....
Zoffix NeuralAnomaly, cut the release 17:56
NeuralAnomaly Zoffix, Will do! If you're feeling particularly naughty, you can watch me at perl6.fail/release/progress or go look at some cats icanhas.cheezburger.com/lolcats
Zoffix, ♥♥♥♥♥♥ Prep done
Geth nqp: 1b5595eabd | (Zoffix Znet)++ | tools/build/MOAR_REVISION
bump MoarVM version to 2017.02
nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2......2017.02
9619781073 | (Zoffix Znet)++ | VERSION
nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2......2017.02
d7314032d7 | (Zoffix Znet)++ | t/serialization/02-types.t
IOninja looks like perl6.fail/release/progress is busted by HTTPs as well -_- 18:00
IOninja can see the log in bot's terminal tho
ah, tis a 1-letter fix. s{ws://}{wss://} 18:03
moritz
.oO( whitspace over SSL )
18:04
NeuralAnomaly Zoffix, ☠☠☠☠☠☠ NQP: test 18:05
Zoffix, ☠☠☠☠☠☠☠☠☠☠ ABNORMAL EXIT!
IOninja -_- 18:06
geekosaur .oO { not quite perfect }
timotimo has no clue about that 18:11
timotimo also doesn't usually build and test nqp-j
IOninja I'd guess it's the commits that went in the past day or so, 'cause I tested everything not long ago 18:13
yup: github.com/perl6/nqp/commit/31eab9...dd37d240a5 18:20
IOninja .tell pmurias FYI, I fudged one of the tests you added 'cause it fails on JVM. No idea why: github.com/perl6/nqp/commit/d7314032d7 18:22
yoleaux2 IOninja: I'll pass your message to pmurias.
Zoffix NeuralAnomaly, steps
NeuralAnomaly Zoffix, all pre nqp r post pre-r6 pre-blank-slate nqp-clone nqp-bump-vers nqp-build nqp-test nqp-tar nqp-tar-build nqp-tag nqp-tar-sign nqp-tar-copy r-clone r-prep-ann r-bump-vers r-build r-p5 r-stress r-stress-v6c r-tar r-tar-build r-tar-p5 r-tar-stress r-tag r-tar-sign r-tar-copy post-scp
Zoffix NeuralAnomaly, cut the release
NeuralAnomaly Zoffix, Will do! If you're feeling particularly naughty, you can watch me at perl6.fail/release/progress or go look at some cats icanhas.cheezburger.com/lolcats
Zoffix, ♥♥♥♥♥♥ Prep done
Zoffix, ♥♥♥♥♥♥ NQP: MoarVM version appears to be already bumped
IOninja god... the lolcats on that site are awful. 18:28
NeuralAnomaly: you should pick up some nicer material for next time
geekosaur site jumped the shark years ago
timotimo files.mastodon.social/media_attach...1487371245 - look at this cutie 18:30
how do i get to see the progress again? 18:31
Zoffix timotimo, perl6.fail/release/progress
timotimo it still gives me the failed versio
Zoffix press F5 18:32
NeuralAnomaly Zoffix, ♥♥♥♥♥♥ nqp tests OK
Zoffix .... a few times...
timotimo i pressed F5, and also ctrl-shift-r
Zoffix If I press it sometimes I get the old copy with failed stuff but eventually it swaps to updated on
No idea why
timotimo huh
Zoffix Did it work? 18:33
timotimo no
there it is!
IOninja I restarted the web app 18:34
timotimo ok
it doesn't seem to put anything new in there, though
IOninja it's thinking...
timotimo ok
IOninja outputs stuff slowly 18:35
timotimo still a quite darn handy tool 18:36
oh, i F5'd and now i see it spit out new lines 18:40
NeuralAnomaly Zoffix, ♥♥♥♥♥♥ nqp release tarball tests OK 18:41
Zoffix, ♥♥♥♥♥♥ nqp release DONE 18:42
Geth rakudo/nom: 4e06fc3ab6 | (Zoffix Znet)++ | docs/announce/2017.02.md
Generate release announcement for 2017.02
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017.......2017.02
ca45e72b06 | (Zoffix Znet)++ | tools/build/NQP_REVISION
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017.......2017.02
aa94ffc337 | (Zoffix Znet)++ | VERSION
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017.......2017.02
ba246371ba | (Zoffix Znet)++ | docs/announce/2017.02.md
NeuralAnomaly Zoffix, ♥♥♥♥♥♥ Rakudo stresstest (master) OK 18:50
Zoffix, ♥♥♥♥♥♥ Rakudo stresstest (6.c-errata) OK 18:52
IOninja why do our emails say "Announce:" instead of "Announcement:" ? 18:55
NeuralAnomaly Zoffix, ♥♥♥♥♥♥ Rakudo release DONE 18:59
Zoffix, 🎺🎺🎺📯📯📯📯📯📯🌈🌈🌈📦📦📦
Zoffix, ♥♥♥♥♥♥ Post: upload tarballs to rakudo.org and rakudo.perl6.org
Zoffix, The release of **Rakudo #108 2017.02** has now been completed
Zoffix, 🎺🎺🎺📯📯📯📯📯📯🌈🌈🌈📦📦📦
NeuralAnomaly celebrates with an appropriate amount of fun
Zoffix NeuralAnomaly, thanks, baby
NeuralAnomaly Zoffix, any time, buddy!
rakudo/nom: 366d1cade9 | (Zoffix Znet)++ | docs/ChangeLog
Fix glitched list indent
rakudo/nom: 20c8995090 | (Zoffix Znet)++ | docs/release_guide.pod
2017.02 is now in the past
19:04
IOninja Nice. Today's release is on the 18th and the next release is also on the 18th :)
timotimo cool
IOninja NeuralAnomaly: status 19:05
NeuralAnomaly IOninja, [✘] Next release will be in 3 weeks and 6 days. Since last release, there are 0 new still-open tickets (0 unreviewed and 0 blockers) and 3 unreviewed commits. See perl6.fail/release/stats for details
IOninja a'ight. Job's done. 19:07
IOninja leaves to blow off steam in Witcher 3
timotimo \o/
thanks IOninja
Geth roast/master: 5 commits pushed by (Ronald Schmidt)++ 19:29
TimToady now that the release is out, I'm going to merge braids, so expect a bit of instability and/or warnings from slang modules 19:46
(just doing a final spectest first)
IOninja \o/
TimToady if you have any code that depends on %*PRAGMAS or $*ACTIONS, it'll break 19:47
if you have code that sets %*LANG from a use, it'll emulate and warn
since that's mostly all that slang modules do
so, for instance, expect Slang::Tuxic to issue a warning about use of %*LANG 19:48
Geth nqp/braids: 4 commits pushed by (Zoffix Znet)++, TimToady++ 19:55
rakudo/braids: 10 commits pushed by (Zoffix Znet)++, TimToady++
review: github.com/rakudo/rakudo/compare/0...4b18a8b7e3
19:56
ugexe zef auto-update (a default setting to update the ecosystem list each run) is fixed now. It was a poorly written conditional at fault, not the cache for once... github.com/ugexe/zef/commit/65e970...1c0b32b898 19:57
it *was* auto updating... but only if the package list file didn't exist yet 19:58
TimToady \o/ 19:59
Geth nqp/master: 21 commits pushed by TimToady++
review: github.com/perl6/nqp/compare/d7314...f414161e9c
20:06
¦ nqp/master: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...2-g542baec
rakudo/nom: 25 commits pushed by TimToady++
review: github.com/rakudo/rakudo/compare/2...e1df52ecab
20:07
¦ rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....1-gf414161
TimToady that's the braids merge
IOninja \o/ 20:08
TimToady note that pre/post-braids nqp will not interoperate with post/pre-braids rakudo, so you gotta update both 20:10
also note this is not primarily an optimization upgrade, but a semantic hygiene upgrade, so speed differences will be close to a wash 20:11
it is likely to make some optimizations easier in the future however, and puts the responsibility for adjudicating a lot of the slang changes into the cursor 20:12
(there are still a few things that might go into the braid, such as %*HOW) 20:13
it'll also run slightly faster once we strip out the old %*LANG apparatus 20:15
jnthn TimToady: I figure there's a migration path for things that were playing around with %*LANG? 20:42
TimToady: Seems there's a left-behind printf in HLL::Compiler in the NQP merge 20:45
TimToady if a slang module also does the equivalent $*LANG.define_slang, it should suppress the warning. obviously, would have to be in a try to support backward compat 20:48
jnthn *nod* 20:49
Yeah, think I can see how the new way works :)
I think my only module that uses %*LANG was Grammar::Generative, which busted long ago in a different way anyway
TimToady well, it's just one of those hang-on-the-right-peg things to me, where parallel dynvars are a smell, much like parallel arrays smell compared to OO 20:50
jnthn Yeah, it feels like an improvement to me :) 20:51
TimToady I'm a little unhappy to duplicate LANG as LEAF part of the time, but $*W doesn't always get the current cursor
and sometimes it has to be the current cursor rather than the current "official" language 20:52
Geth nqp: 2685b8c71a | TimToady++ | src/HLL/Compiler.nqp
comment out residual debugging code

Can likely be deleted once braids settle down.
20:55
TimToady oh, I should mention $*PACKAGE is in the braid now too, in case anyone runs into that, though we still have an epiphenomenal $*PACKAGE for reading at the momemt. $?PACKAGE is unaffected, of course, since that's user API 20:58
and $?LANG is no longer actually poked into the lexical scope, cuz that's what caused it to try to serialize BOOTContext 21:01
instead it's treated more en-passant-ly as $?LINE is
travis-ci Rakudo build failed. Zoffix Znet '2017.02 is now in the past' 21:02
travis-ci.org/rakudo/rakudo/builds/202993681 github.com/rakudo/rakudo/compare/3...c8995090f6
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
jnthn m: { sub infix:<abc>($a, $b) { "$b$a" }; EVAL 'say 2 abc 4' } 21:03
camelia (Perl6::Grammar+{Oper} with Perl6::Actions
42
jnthn m: { sub infix:<abc>($a, $b) { "$b$a" }; EVAL 'say 2 abc 4' }; EVAL 'say 2 abc 4' 21:04
camelia (Perl6::Grammar+{Oper} with Perl6::Actions
42
5===SORRY!5=== Error while compiling /home/camelia/EVAL_1
Two terms in a row
at /home/camelia/EVAL_1:1
------> 3say 27⏏5 abc 4
expecting any of:
infix
infix sto…
TimToady hmm
jnthn Seems there's a bebug leftover there too
*debug
But it gets the right languages in the EVAL
So I guess the %?LANG got replaced with something appropriate to mkae that happen 21:05
*make
TimToady I think it's still limping along with $?LANG for that 21:06
jnthn %?LANG it seems 21:07
TimToady probably needs some further tweaking when we get rid of %*LANG
jnthn I thought that's what you'd meant we didn't poke into the scope any more, though
TimToady %?LANG, I mean, is still there for the moment 21:08
$?LANG doesn't go into the scope
and EVAL is still using %?LANG; I misspoke when I said $?LANG
that needs fixing up, eventually 21:09
jnthn *nod*
TimToady I think I didn't have $*LEAF yet when I last visited that code, which might be what's needed there
but I was also considering making EVAL a bit macroier, and take an implicit snapshot of $*LANG at the point of the call 21:11
just didn't get there yet
we have to treat EVAL specially in terms of optimization anyway, so it's not like we don't recognize it at compile time 21:12
geezer nap & 21:15
jnthn That could also work, yeah. 21:17
nap well :)