01:08 lichtkind left 01:11 lucasb left 01:26 Xliff left 01:50 raiph left 04:52 moon-child left 04:53 moon-child joined 07:43 JJMerelo joined
JJMerelo Could maybe someone take a look at old issues? github.com/Raku/old-issue-tracker/ Since there's still some time left for the new release, it would be great if some of them could be solved. 08:12
tellable6 2020-05-29T11:50:47Z #raku <uzl[m]> jmerelo I think some doc issues that were addressed by respective PRs and could thus be closed. From the top of my head: github.com/Raku/doc/issues/3389 , github.com/Raku/doc/issues/3308
JJMerelo .tell uzl[m] OK, done, but please include closing statements in the PRs and/or commits you're doing for them in the future. 08:15
tellable6 JJMerelo, I'll pass your message to uzl[m]
JJMerelo m: class N { method text { return Proxy.new​: FETCH => method () { say 'fetch​: ', ++$ },STORE => method ($new) { say "$new​: ", ++$ } } }; N.text; N.text = 'blah'; 08:38
camelia 5===SORRY!5=== Error while compiling <tmp>
Bogus postfix
at <tmp>:1
------> 3class N { method text { return Proxy.new7⏏5​: FETCH => method () { say 'fetch​: ',
expecting any of:
infix stopper
JJMerelo m: class N { method text { return Proxy.new: FETCH => method () { say 'fetch​: ', ++$ },STORE => method ($new) { say "$new​: ", ++$ } } }; N.text; N.text = 'blah'; 08:40
camelia fetch​: 1
Cannot assign to a readonly variable or a value
in block <unit> at <tmp> line 1

fetch​: 1
JJMerelo m: class N { method text { return-rw Proxy.new: FETCH => method () { say 'fetch​: ', ++$ },STORE => method ($new) { say "$new​: ", ++$ } } }; N.text; N.text = 'blah';
camelia fetch​: 1
fetch​: 2
fetch​: 3
blah​: 1
10:32 sena_kun joined
tbrowder hi, just filed an issue on spectest 10:52
if roast on perl6 and raku aren't synced behind the scenes aren't we asking for trouble? 10:53
or am i missing something? 10:54
some github magic maybe? 10:55
sena_kun tbrowder, I think the repo is moved, no? I mean, there are no two copies, but only a single one with symlinks. 10:58
tbrowder ok, i hope so, but the tooling still needs updating: raku instead of perl6 10:59
sena_kun yes
Geth ¦ rakudo: JJ self-assigned "make spectest" pulls from github:perl6 instead of github:raku github.com/rakudo/rakudo/issues/3729 11:19
¦ rakudo: JJ self-unassigned "make spectest" pulls from github:perl6 instead of github:raku github.com/rakudo/rakudo/issues/3729 11:27
JJMerelo tbrowder sena_kun after looking at this for a while, can't figure out where that is set 11:32
Geth rakudo/set_equality: dbd69ba726 | (Elizabeth Mattijsen)++ | 2 files
Simplify Setty/Baggy.ACCEPTS

Now that we have a set equality op, we can use that.
Also fix up the Setty/Baggy infix:<eqv> check on object identity: the should always decont before checking.
JJMerelo tbrowder perl6/roast is for the time being automatically redirected to Raku/roast, which is why it works.
JJMerelo tbrowder but beyond that, there's a labyrinth of Perl code, Makefile.in macros and git configuration where I am lost. I've taken notes in the issue. 11:34
lizmat vrurg patrickb would probably know better where to look / fix 11:35
JJMerelo AlexDaniel` found it here: github.com/Raku/nqp-configure/blob...#L153-L156 11:36
Really I had no idea we had to look there too... 11:37
tbrowder yep, that's why i gave up on fixing tooling years ago, hat's off to vrurg and others for improving it! 11:38
*hats 11:39
ShimmerFairy The clear and obvious solution is to simply switch over to CMake. Then nothing would be complicated ever again. 11:40
tbrowder it would be cool to gradually move away from reliance on perl, too
ShimmerFairy Speaking of build issues, 'make install' builds all of Raku's modules, not the 'make' step like you'd expect. This wouldn't be much of an issue except that the created '.precomp' folder and all its contents belong to root when you do 'sudo make install', making it a pain to deal with when you reconfigure (since configure tries to delete .precomp) 11:42
tbrowder second the motion on CMake, but that's a steep learning curve, too 11:43
ShimmerFairy cmake is cool, but boy howdy are they allergic to clear documentation sometimes. 11:44
tbrowder yep, i even bought the book...
ShimmerFairy And it's worth keeping in mind that CMake (or any other tooling system) would require a lot of custom plumbing, since they're geared more towards major languages. 11:46
tbrowder the brlcad.org project move to cmake a few years ago and i think it was done too early before the top devs understood it well enough. i think you need to really prepare well before the transition.
11:50 JJMerelo left
[Tux] Rakudo version 2020.05.1-207-g6ead4ce16 - MoarVM version 2020.05-18-g157387214
csv-ip5xs0.816 - 0.821
csv-ip5xs-208.322 - 8.605
csv-parser25.961 - 26.197
csv-test-xs-200.376 - 0.377
test7.230 - 7.658
test-t1.919 - 1.926
test-t --race0.883 - 0.914
test-t-2030.696 - 30.718
test-t-20 --race8.933 - 9.077
tbrowder thnx for clarification on roast...bye 12:04
12:27 Altai-man_ joined 12:30 sena_kun left 14:15 lucasb joined 14:20 maggotbrain left 14:25 Xliff joined 14:28 sena_kun joined 14:30 Altai-man_ left 14:56 patrickb joined 15:40 domidumont joined 15:41 finsternis left 15:46 finsternis joined 16:27 Altai-man_ joined 16:28 domidumont left 16:30 sena_kun left
Geth nqp: a2878ef2f5 | (Patrick Böker)++ | tools/templates/MOAR_REVISION
Bump MoarVM

This pulls in a change that fixes a test in t\nqp/019-file-ops.t
rakudo: sumanstats++ created pull request #3730:
Perl6 to Raku and many more
17:49 JJMerelo joined
Geth nqp: sumanstats++ created pull request #641:
Perl6 to raku + fix links
18:10 travis-ci joined
travis-ci NQP build passed. sumanstats 'Fix the links 18:10
travis-ci.com/sumanstats/nqp/builds/168863588 github.com/sumanstats/nqp/compare/...2865a7c621
18:10 travis-ci left 18:12 lichtkind joined 18:28 sena_kun joined 18:30 Altai-man_ left
patrickb I want to use `nqp::execname()` in an nqp test. That opcode does not exist on JVM. Using an `if` to guard that op does not work as the compiler will still trip on the non existing op. What can I do? 18:38
MasterDuke how easy is it to eval in nqp? 18:43
18:44 JJMerelo left
MasterDuke npq: my $comp := nqp::getcomp('nqp'); $comp.eval('say(nqp::execname())') 18:47
nqp: my $comp := nqp::getcomp('nqp'); $comp.eval('say(nqp::execname())')
camelia /home/camelia/rakudo-m-inst-1/bin/nqp-m
patrickb That could actually work! 18:48
MasterDuke np
Geth rakudo: c9e94623bc | (Sylvain Colinet)++ | 2 files
Make the 'is' routine in Test not throwing an error on failure when the type can't be stringified, falling back on .raku instead.

A good example is the Buf type that has a Str routine that fail to encourage the user to use encode instead.
is Buf.new(42), Buf.new(43)
This will fail with the error : Cannot use a Buf as a string, but you called the Stringy method on it instead of the regular got, expected message.
This fix issue #3535
linkable6 RAKUDO#3535 [closed]: github.com/rakudo/rakudo/issues/3535 Comparing Buf with the is routine from Test does not behave well
Geth rakudo: 9a7d0e6c62 | (Vadim Belman)++ (committed using GitHub Web editor) | 2 files
Merge pull request #3613 from Skarsnik/master

Fix for 'Comparing Buf with the is routine from Test does not behave well', closes #3535
patrickb MasterDuke: How would I go about retrieving the result of the eval? 19:53
Geth nqp-configure: 889df0206a | (Vadim Belman)++ | lib/NQP/Config.pm
Change default github community from perl6 to Raku

Fixes git URLs for roast and NQP. Closes rakudo/rakudo#3729
linkable6 RAKUDO#3729 [closed]: github.com/rakudo/rakudo/issues/3729 "make spectest" pulls from github:perl6 instead of github:raku
MasterDuke patrickb: don't know. dirty hack, but maybe just pass in the test, dying if it succeeds and use dies-ok 20:07
patrickb MasterDuke: I think I don't understand. 20:08
MasterDuke well, there's a `sub dies-ok` defined in a couple of the nqp tests. at least once it's given a block that includes a $comp.eval. so either define a `sub lives-ok` or just negate the dies-ok 20:11
something like `#if !jvm lives-ok({ $comp.eval('<whatever your test is>') })` 20:12
patrickb So you are suggesting putting the entire test in a giant eval?
MasterDuke yeah. it's a hack, but gets around the jvm backend compiler complaining about not having nqp::execname 20:13
patrickb MasterDuke can you have a look at github.com/Raku/nqp/blob/master/t/...od-panic.t <- Do you think putting everything in an eval a viable solution in that case? 20:17
MasterDuke run-command is what does the nqp::execname? 20:20
patrickb MasterDuke: github.com/patrickbkr/nqp/blob/azu...od-panic.t <- Sorry, wrong link. That's the version including the execname 20:21
MasterDuke heh. would be easier if $cmd could just get the result of an eval... 20:23
patrickb it seems that's actually possible on the JS backend, but not on the others. 20:24
MasterDuke nqp: my $comp := nqp::getcomp("nqp"); my $e := $comp.eval("say(nqp::execname())"); say($e)
camelia /home/camelia/rakudo-m-inst-1/bin/nqp-m
MasterDuke nqp: my $comp := nqp::getcomp("nqp"); my $e := $comp.eval("say(\"execname = \" ~ nqp::execname())"); print($e) 20:25
camelia execname = /home/camelia/rakudo-m-inst-1/bin/nqp-m
MasterDuke you just have to say or print what you want the result to be in the eval 20:26
oh hm. wait
20:27 Altai-man_ joined 20:30 sena_kun left
patrickb Would just moving it to t/moar be a sane solution? 20:30
MasterDuke well, copying with a note in each that the other version is the same except for how it get the name? 20:31
patrickb The test is skipped on JVM already...
MasterDuke nqp: my $comp := nqp::getcomp("nqp"); my $e := try { $comp.eval("nqp::die(nqp::execname())"); CATCH { say("from eval: " ~ $_) } } 20:37
camelia from eval: /home/camelia/rakudo-m-inst-1/bin/nqp-m
MasterDuke patrickb: ^^^ even dirtier hack, but looks like it would work 20:38
patrickb I'm giving it a try
MasterDuke: It's looking good so far. I have triggered a CI run to see how it goes on JVM. Thanks for your help! 20:53
patrickb is booting into Windows to find another test failure 20:54
20:54 patrickb left
MasterDuke np 20:54
20:58 patrickb joined 21:05 lichtkind left 21:06 lichtkind joined 21:36 patrickb left 21:37 patrickb joined
Geth nqp: 476642f34b | (Patrick Böker)++ | t/nqp/019-file-ops.t
Correct a wrong skip count on Windows

Also remove a bogus check for backends that don't exist. There currently are only moar, js and jvm.
22:28 sena_kun joined
Geth nqp: 35267b4abd | (Patrick Böker)++ | tools/templates/MOAR_REVISION
Bump MoarVM

Pulls in another fix for dir handling on Windows.
22:30 Altai-man_ left
Geth nqp: c4fcc6438d | (Patrick Böker)++ | t/nqp/114-pod-panic.t
Fix t/nqp/114-pod-panic with a relocated `nqp` executable

As `nqp::execname` is an op code not available on JVM and thus resulting in a compile time error, a quite dirty workaround is used. We hide the op in an eval and retrieve the value via `die`.
Also there is no sane way to retrieve the current nqp executable name available on the JVM, so skip that test there.
nqp: a5f90ae55e | (Patrick Böker)++ (committed using GitHub Web editor) | t/nqp/114-pod-panic.t
Merge pull request #640 from patrickbkr/fix-114-pod-panic-with-reloc

Fix t/nqp/114-pod-panic with a relocated `nqp` executable
23:02 Kaiepi left 23:05 Kaiepi joined 23:06 patrickz joined 23:10 patrickb left 23:11 sena_kun left 23:12 lichtkind left 23:35 patrickz left 23:45 lucasb left