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.
releasable6 Next release in ≈2 days and ≈11 hours. R6 is down. At least 7 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 07:00
lizmat described her commits in the ChangeLog-Draft 07:41
(it weren't that many)
Geth_ rakudo/newer-sprintf: 8a91739468 | (Elizabeth Mattijsen)++ | src/core.e/Rakudo/Internals/Sprintf.pm6
Yet another try, yet another branch

I sincerely hope this one sticks		 07:48
lizmat wow
AlexDaniel lizmat: thanks!
Geth_ roast/new-sprintf: eec150db1e | (Elizabeth Mattijsen)++ | S32-str/sprintf-d.t
WIP on %d tests
 08:04
rakudo: 4bd25bf53d | (Elizabeth Mattijsen)++ | 2 files
Add core.e/Rakudo/Internals/Sprintf.pm6 stub

To facilitate switching between "newer-sprintf" and "master" branches without having to do a full re-configure.		 08:14
|Tux| Rakudo version 2019.07.1-260-g9dd67cb37 - MoarVM version 2019.07.1-230-gea9b29e85
csv-ip5xs0.666 - 0.687
csv-ip5xs-204.952 - 4.990
csv-parser21.850 - 22.678
csv-test-xs-200.424 - 0.426
test7.392 - 7.489
test-t1.751 - 1.831
test-t --race0.776 - 0.797
test-t-2030.175 - 30.492
test-t-20 --race8.988 - 9.337
 10:28
pmurias vrurg: the truffle branch isn't a gradual performance improvemnt, it's a rebuild using a completely different technology 12:25
vrurg: truffle generates a custom JIT from a interpreter using partical evaluation dark magic rather then relying on generating JVM bytecode 12:26
vrurg pmurias: does it provide other advantages besides of speed up? 13:28
pmurias vrurg: super fast interop with other languages that have a truffle implementation 13:42
(we don't have that yet, as the truffle backend was in early stages) 13:43
vrurg: they had plans to add language-protocol-server integration to truffle but I'm not sure what's the state of that 13:47
vrurg Interesting! 13:48
However much I don't like Java. ;) 13:49
pmurias vrurg: for the truffle backend a lot of the usual Java boilerplate was solved by generating code from annotations for a lot of the repetive stuff 13:50
but it's still written in Java :/
vrurg Whatever. I don't like it for two reasons: it was hellishly slow back then which is not really a case anymore. And they dictate me how to program, but I'm lucky to choose the language and can avoid it. Otherwise I don't care. 13:53
But the idea is interesting, no doubt.
pmurias: BTW, more actual thing. I've started reworking NQP makefiles with the same purpose as for Rakudo. I would try to manage JS without you. But still might ask for your help. 13:56
pmurias vrurg: I'll try to help, but getting nqp.js to work should be a lot easier now that rakudo.js works 13:57
vrurg I hope for this too. Though I already see it's not doing a few things moar and jvm do – like it doesn't have OPS and ASTNodes parts. I would require conditionals in templates which I try to avoid as much as possible. 13:59
By ASTNodes I mean MAST and JAST. 14:00
pmurias vrurg: you want to reorganize the js code to have it the same way? 14:01
vrurg: also a fundamental difference is that the js backend bootstraps using the moar one instead of having a stage0 in js 14:02
vrurg Reorganazing wasn't my intention. Let's see how it goes. I'll have moar and jvm done. Then it'll be more clear what is needed. 14:04
So, for now – just ignore it. :)
I don't think I'd have something working before the next week. 14:05
lizmat vrurg: perhaps already known? bugs.debian.org/cgi-bin/bugreport....bug=940776 15:53
kawaii AlexDaniel: seems like I should run another ecosystem roast this weekend 16:08
vrurg lizmat: Thanks! Though it's more for patrickb. It's like he decides where to put files and I provide the means for this. :) 16:30
lizmat well, I hope you or patrickb will keep the debian folks informed 16:33
vrurg lizmat: where did you receive the message from? 16:41
lizmat twitter.com/DebianBug/status/11747...7198780417 16:42
vrurg .tell patrickb bugs.debian.org/cgi-bin/bugreport....bug=940776 – also related to R#3093 for which, I guess, you have a solution. 16:43
synopsebot R#3093 [open]: github.com/rakudo/rakudo/issues/3093 [BLOCKER] Installer is broken with --libdir=/usr/lib
tellable6 vrurg, I'll pass your message to patrickb
vrurg AlexDaniel: I can't recall if bisect is working on mac. 16:55
patrickb o/ 17:21
tellable6 2019-09-13T19:08:18Z #perl6 <Xliff> patrickb Thanks for the additional info! I'll keep that in mind.
2019-09-16T14:33:30Z #perl6-dev <sortiz> patrickb Yep, already merged (DBDish::ODBC PR 5)
2019-09-19T16:43:19Z #perl6-dev <vrurg> patrickb bugs.debian.org/cgi-bin/bugreport....bug=940776 – also related to R#3093 for which, I guess, you have a solution.
synopsebot R#3093 [open]: github.com/rakudo/rakudo/issues/3093 [BLOCKER] Installer is broken with --libdir=/usr/lib
AlexDaniel kawaii: yes please, maybe even now 18:47
vrurg: `git bisect` does :P :P :P
MasterDuke AlexDaniel: i think we need a new nqp/rakudo bump. to pull in the revert of the template commits and nine++'s fixes 18:48
Geth_ nqp: 2ccade7319 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/templates/MOAR_REVISION
[MoarVM Bump] Brings 14 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...g270f1b0be 270f1b0be Merge pull request #1132 from Kaiepi/cstr 329f669c3 Really fix race condition with exited threads still in the thread list ea9b29e85 Fix destroyed thread context being accessable through the thread list bcc785d23 Fix overflow objects still belonging to exiting thread ... (10 more lines)
rakudo: 565b2ac219 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/templates/NQP_REVISION
[NQP Bump] 2ccade731 [MoarVM Bump] Brings 14 c […]

NQP bump brought: github.com/perl6/nqp/compare/2019....g2ccade731
vrurg AlexDaniel: if you say it also compiles dependant moar/nqp and rakudo itself, I'd praise its brilliance. ;)
AlexDaniel vrurg: yes if you write a script to do that :D
vrurg "Damn you, Addams!" (c) ;p
AlexDaniel don't forget to handle unbuildable commits correctly!
and make sure your exit codes are right :)
tobs: ↑ speaking of which, there are two really needed things we should implement eventually 18:50
vrurg Ok, will get back to it later. Won't have time until finish with nqp makefiles anyway.
AlexDaniel vrurg: honestly, it's more reliable to do it manually. There should be just 11 steps or so
that's about an hour of your attention
tobs: one is a local use of the bots, preferably without running an irc server. So that you can `bisect: say 42` from a command line and it works 18:51
tobs: another is building rakudo on demand. Currently it attempts to get an archive from the main server, and if that fails it just skips it. Instead, if there's nothing on the main server, it should build it on the current machine 18:52
vrurg Perhaps. The problem is that makefiles require me to be very scrupulous and I can't parallelize. NP, it can wait, I think. Or somebody would locate the problem in the meanwhile. 18:53
Thanks!
vrurg is diving back into the world of rules, receipes and incompatibe implementations...

Powered by Cro and the Raku® Programming Language.

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

Thank you!