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.
00:12 stmuk joined 00:13 p6bannerbot sets mode: +v stmuk 00:15 stmuk_ left 00:32 fake_space_whale joined 00:33 p6bannerbot sets mode: +v fake_space_whale 01:43 lizmat left 02:01 stmuk_ joined, p6bannerbot sets mode: +v stmuk_ 02:04 stmuk left 02:33 stmuk joined 02:34 p6bannerbot sets mode: +v stmuk 02:35 stmuk_ left 03:15 fake_space_whale left 05:37 brrt joined 05:38 p6bannerbot sets mode: +v brrt 05:50 ZzZombo joined, p6bannerbot sets mode: +v ZzZombo 05:51 stmuk_ joined 05:52 p6bannerbot sets mode: +v stmuk_ 05:54 stmuk left 06:10 robertle joined, p6bannerbot sets mode: +v robertle 06:13 ZzZombo left 06:17 stmuk joined 06:18 p6bannerbot sets mode: +v stmuk 06:20 stmuk_ left 06:22 Ven` joined 06:23 p6bannerbot sets mode: +v Ven` 06:27 Ven` left 06:44 ZzZombo joined, p6bannerbot sets mode: +v ZzZombo 07:04 brrt left 07:16 [TuxCM] left 07:53 jkramer left 08:00 jkramer joined 08:01 p6bannerbot sets mode: +v jkramer 08:31 [TuxCM] joined, p6bannerbot sets mode: +v [TuxCM] 08:33 Ven` joined 08:34 p6bannerbot sets mode: +v Ven` 09:32 Ven` left 09:33 Ven` joined 09:34 p6bannerbot sets mode: +v Ven` 09:50 ZzZombo left 09:52 AlexDaniel joined, p6bannerbot sets mode: +v AlexDaniel 10:29 lizmat joined 10:30 p6bannerbot sets mode: +v lizmat 10:45 ufobat joined 10:46 p6bannerbot sets mode: +v ufobat 10:47 robertle left 11:01 ufobat left 11:18 stmuk_ joined 11:19 p6bannerbot sets mode: +v stmuk_ 11:20 stmuk left 11:21 stmuk joined, p6bannerbot sets mode: +v stmuk 11:24 stmuk_ left 11:28 stmuk_ joined 11:29 p6bannerbot sets mode: +v stmuk_ 11:30 stmuk left 11:38 robertle joined 11:39 p6bannerbot sets mode: +v robertle
tbrowder_ lizmat: don’t know how you do it but the p6 weekly is always well-done and very informative, thanks for the hard but very useful work! 11:50
lizmat tbrowder_: thank you 11:51
MasterDuke you know i was just thinking yesterday that the only two open source communities i've ever got involved in in any way are perl 6 and kde. both in no small part because they have good regular updates easily available 11:55
|Tux| Rakudo version 2018.06-490-g82755848c - MoarVM version 2018.06-437-g5bfad12bb
csv-ip5xs0.957 - 0.960
csv-ip5xs-207.773 - 8.512
csv-parser23.302 - 24.005
csv-test-xs-200.441 - 0.462
test9.325 - 9.753
test-t2.155 - 2.189
test-t --race0.914 - 0.919
test-t-2036.690 - 37.260
test-t-20 --race11.861 - 12.009
12:00
Geth nqp: 09bf927e61 | MasterDuke17++ (committed using GitHub Web editor) | 3 files
Remove some unused JVM interop ops (#498)
12:11
MasterDuke lizmat, et al.: any further thoughts on github.com/rakudo/rakudo/pull/2228 ? 12:18
lizmat MasterDuke: believe it or not, but I'm again about to be on the road until later tonight 12:20
been too busy with other things
timotimo lizmat: i don't fully agree with your suggestion to use the existing constant folding
lizmat timotimo: why ? 12:21
timotimo because it'll turn the result into a WVal containing a Range object
lizmat yes
MasterDuke lizmat: no worries
timotimo though i suppose you could then call its methods to find out what the range endpoints are
i expect there was some reason for not wanting to fiddle with the WVal Range when that was implemented, but i don't remember :|
lizmat ah, ok, I see that "my $a = 1..10" codegens to infix:<..>(1,10) 12:22
but that also means that "my $a = 1..10 + 1" could be codegenned into infix:<..>(2,11) , right 12:23
instead of codegenning it to a constant Range 12:24
timotimo that'd require an extra opt pass somewhere 12:25
it'd look a little like the junction folding opt
12:26 travis-ci joined, p6bannerbot sets mode: +v travis-ci
travis-ci NQP build failed. MasterDuke17 'Remove some unused JVM interop ops (#498)' 12:26
travis-ci.org/perl6/nqp/builds/421539217 github.com/perl6/nqp/compare/c43d1...bf927e6104
12:26 travis-ci left 12:27 stmuk joined 12:28 p6bannerbot sets mode: +v stmuk 12:30 stmuk_ left
lizmat so why don't we codegen "Range.new(1,10)" into a single WVal, instead of calling Range.new with 2 WVals ? 12:34
is it because it wouldn't make any ses
sense in the precomp case?
commuting& 12:35
12:35 lizmat left 12:52 Kaiepi left 12:55 Kaiepi joined 12:56 p6bannerbot sets mode: +v Kaiepi 13:10 travis-ci joined, p6bannerbot sets mode: +v travis-ci
travis-ci NQP build failed. MasterDuke17 'Remove some unused JVM interop ops (#498)' 13:10
travis-ci.org/perl6/nqp/builds/421539217 github.com/perl6/nqp/compare/c43d1...bf927e6104
13:10 travis-ci left 13:41 Ven` left 14:04 skids joined 14:05 p6bannerbot sets mode: +v skids 14:43 j3nnn1 joined, p6bannerbot sets mode: +v j3nnn1 14:48 Ven` joined 14:49 p6bannerbot sets mode: +v Ven` 14:53 Ven` left 15:04 brrt joined 15:05 p6bannerbot sets mode: +v brrt 15:09 robertle left 15:10 Ven` joined, DrForr_ is now known as DrFor, DrFor is now known as DrForr, p6bannerbot sets mode: +v Ven` 15:33 brrt left 15:36 fake_space_whale joined 15:37 p6bannerbot sets mode: +v fake_space_whale 16:02 robertle joined 16:03 p6bannerbot sets mode: +v robertle 16:06 brrt joined 16:07 p6bannerbot sets mode: +v brrt, [TuxCM] left 16:10 stmuk_ joined 16:11 p6bannerbot sets mode: +v stmuk_ 16:12 stmuk left 16:31 brrt left 16:34 fake_space_whale left 16:42 epony left 16:43 Ven` left 16:46 Ven` joined 16:47 p6bannerbot sets mode: +v Ven`
Geth nqp/truffle: d8d27a8d5c | (Daniel Green)++ | 4 files
[truffle] Implement nqp::<bitor bitand bitxor>_s
16:47
16:47 epony joined, p6bannerbot sets mode: +v epony 16:50 Ven` left 17:26 [TuxCM] joined 17:27 p6bannerbot sets mode: +v [TuxCM] 18:17 ExtraCrispy joined, p6bannerbot sets mode: +v ExtraCrispy
Geth rakudo: jstuder-gh++ created pull request #2240:
Make native arrays throw when storing lazy Seq
18:34
19:56 lizmat joined 19:57 p6bannerbot sets mode: +v lizmat
lizmat . 20:09
Geth nqp: tbrowder++ created pull request #499:
add some more nqp examples
20:58
21:08 skids left 21:11 travis-ci joined, p6bannerbot sets mode: +v travis-ci
travis-ci NQP build failed. MasterDuke17 'Remove some unused JVM interop ops (#498)' 21:11
travis-ci.org/tbrowder/nqp/builds/421761779 github.com/tbrowder/nqp/compare/db...bf927e6104
21:11 travis-ci left
Geth rakudo: 0f05d27c36 | (Elizabeth Mattijsen)++ | src/core/BagHash.pm6
Prevent confusion with Iterator role
21:17
tbrowder_ anyone: is the travis jvm failure a show-stopper for my nqp PR? moarvm is fine 21:22
i didn't change nqp code, just added a few examples 21:23
i'll merge later unless i hear a nay 21:24
lizmat tbrowder_: it was MasterDuke's last commit that caused the Travis noise 21:25
please merge, looks ok to me from casual observation
tbrowder_ ok, they work fine on last stable release, thnx 21:26
Geth nqp: 096670f0a7 | (Tom Browder)++ | 4 files
add some more nqp examples
21:28
nqp: da885a2bda | (Tom Browder)++ (committed using GitHub Web editor) | 4 files
Merge pull request #499 from tbrowder/examples

add some more nqp examples
¦ roast: tbrowder self-assigned Rename module suffixes from .pm to .pm6 github.com/perl6/roast/issues/464 21:32
lizmat is too knackered to look at code 21:33
21:33 lizmat left
tbrowder_ i'm warning all that i'm changing roast ".pm" files to ".pm6" per roast issue GH#464 (2 up-votes plus an ok by @JJ). it will be in a PR, but please stop me early if you don't think it is wise to continue. the PR will be tested against the current master branch. 21:35
synopsebot GH#464 [closed]: github.com/rakudo/rakudo/pull/464 Add can-ok subroutine
21:43 travis-ci joined, p6bannerbot sets mode: +v travis-ci
travis-ci NQP build failed. Tom Browder 'Merge pull request #499 from tbrowder/examples 21:43
travis-ci.org/perl6/nqp/builds/421775675 github.com/perl6/nqp/compare/09bf9...885a2bdade
21:43 travis-ci left 21:44 robertle left 22:03 Kaypie joined, Kaiepi left, p6bannerbot sets mode: +v Kaypie
AlexDaniel tbrowder_: you mean NQP#464 22:51
synopsebot NQP#464 [open]: github.com/perl6/nqp/pull/464 Fix issue with building nqp-js on Windows
AlexDaniel no…
tbrowder_: you mean ROAST#464
hmm
tbrowder_: you mean SPEC#464
synopsebot SPEC#464 [open]: github.com/perl6/roast/issues/464 Rename module suffixes from .pm to .pm6
AlexDaniel there it is!! :)
tbrowder_ yes, sorry 22:52
www.irccloud.com/pastebin/MQ0IRVam/ 22:54
AlexDaniel .tell pmurias this test does not pass on jvm github.com/perl6/nqp/commit/f0f384...28755fa1f6 22:56
yoleaux AlexDaniel: I'll pass your message to pmurias.
AlexDaniel tbrowder_: hm, are you sure that's on rakudo HEAD? 23:01
tbrowder_: go ahead and rename .pm files, I see no problem with that 23:02
yes it will make git history a tiny bit more complex, but that's ok 23:03
tbrowder_ well, i just checked out master branch, fetched and pulled from upstream, and configured for a local build. i did notice during the build that there were some msgs about checkng out certain branches of nqp, etc. 23:07
the usual
AlexDaniel tbrowder_: that's very interesting… 23:09
because just a few commits ago it was clean for me (for the release)
tbrowder_ ok, i’ll use the release and see what happens... 23:10
MasterDuke hm, npq passed `make j-test` locally when i made that commit 23:11
yoleaux 19:31Z <uzl> MasterDuke: `if ($number ~~ s/^E//) {`, I don't remember the error correctly but I was getting something related to globbing, which was fixed after adding the parenthesis. It was probably not related though.
19:32Z <uzl> MasterDuke: I usually put the spaces around the variables in the hash accesses to make them easier to read. Are there occasions when these spaces make a difference?
19:33Z <uzl> MasterDuke: `grep: { $_ }`. The thing is that the array shouln't be a number but an Assignment object. `%!assignments<Homework>[1..*]` is meant to skip the zeroth element (a whole number) which keeps tracked of the number of assignments. I didn't know about `.skip` so I'll look into it.
19:34Z <uzl> MasterDuke: I don't know why but I always get ARGS and ARGV (which I think isn't a thing in P6) confused. Anyway...Thanks for taking the time to proofread it. Really appreciate it. I've updated the post to reflect your corrections.
AlexDaniel tbrowder_: git checkout release-2018.08
tbrowder_: in rakudo repo
tbrowder_: then later use `git checkout master` to go back 23:12
tbrowder_ ok, thnx
AlexDaniel MasterDuke: I'm also trying to figure out the meaning of this github.com/rakudo/rakudo/commit/82...0e6a025bb1
MasterDuke: any comments?
I *think* everything looked good for the release, but maybe not? 23:13
I just tried buliding r-j on the release branch and there were some errors…
MasterDuke there is no Ops.istype_nodecont, it's Ops.istype_nd
AlexDaniel MasterDuke: so how did it work in the past 23:15
?
tbrowder_ git checkout release-2018.08
MasterDuke dunno. rebuilding everything now to see what's up 23:16
tbrowder_ duh, wrong window :D
MasterDuke but i think the *_nd stuff is pretty new
jnthn The various _nd (no-decont) forms appeared as part of rescalar
I'm pretty sure there weren't any before that 23:17
Geth rakudo/release-2018.08: 11f19025ca | (Daniel Green)++ (committed by Aleks-Daniel Jakimenko-Aleksejev) | src/vm/jvm/runtime/org/perl6/rakudo/Binder.java
Unbreak the JVM build
23:19
AlexDaniel MasterDuke: thanks 23:20
MasterDuke github.com/perl6/nqp/commit/47271b...fdc24aeae7 23:23
so it was renamed to istype_nd in nqp ~2 weeks ago, but that change didn't make it to rakudo until my recent commit 23:25
huh. i do now get `t/nqp/081-radix.t (Wstat: 0 Tests: 208 Failed: 1) Failed test: 203` as well as `t/qregex/01-qregex.t (Wstat: 0 Tests: 782 Failed: 0) TODO passed: 663` (before i just got the 01-qregex.t TODO passed 23:35
AlexDaniel MasterDuke: yeah but that's github.com/perl6/nqp/commit/f0f384...28755fa1f6 23:41
maybe it flaps or something?
Geth nqp: b335d22f8b | (Daniel Green)++ | t/qregex/rx_modifiers
Unfudge now passing test for JVM
23:50