Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2017-06-05

| Channels | #perl6-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:01 d4l3k_ joined #perl6-dev
00:01 pmurias joined #perl6-dev
00:31 samcv joined #perl6-dev
01:01 samcv joined #perl6-dev
01:09 astj joined #perl6-dev
01:26 MasterDuke joined #perl6-dev
01:32 TimToady joined #perl6-dev
04:42 Zoffix perl -wlnE  'say "$.:$_"; close ARGV if eof' foo bar # Perl 5
04:42 Zoffix perl6 -MLN -ne 'say "$*LN:$_"' foo bar               # Perl 6
04:42 Zoffix :) step aside, Perl 5 :)
04:44 Zoffix Wonder if we can pass options along with `-MSomeModule` like you can in p5
05:00 samcv Unicode Grant Status Update 1 is up https://cry.nu/perl6/grant-status-update-1/
05:08 Zoffix samcv++
05:19 BenGoldberg samcv, There's a broken link in the doc, which goes to "perl6/unicode-property-names" but should go to "/perl6/unicode-property-names"
05:20 samcv thanks!
05:20 samcv BenGoldberg, can you check the link now? should work
05:21 BenGoldberg Yup, looks good
06:01 astj joined #perl6-dev
06:02 astj joined #perl6-dev
06:50 dalek joined #perl6-dev
06:51 SmokeMachine joined #perl6-dev
06:56 ilmari[m] joined #perl6-dev
07:19 Zoffix $ p6 -Ilib -MLN -e 'say $*ARGFILES.ln'
07:19 Zoffix (Int:D)
07:19 Zoffix lulz? A :D that ain't a :D :S
07:22 Zoffix m: role Foo { has Int:D $.ln = 0 }; (42.^mixin: Foo).ln.say; (42 does Foo).ln.say
07:22 camelia rakudo-moar 36bc41: OUTPUT: «(Int:D)␤0␤»
07:23 SourceBaby joined #perl6-dev
07:24 Zoffix Ah; `does` also does `obj.^mixin($role).BUILD_LEAST_DERIVED({});`
08:03 [Tux] This is Rakudo version 2017.05-353-g36bc41022 built on MoarVM version 2017.05-25-g62bc54e9
08:03 [Tux] csv-ip5xs        2.830
08:03 [Tux] test            12.950
08:03 [Tux] test-t           4.406 - 4.515
08:03 [Tux] csv-parser      12.827
08:09 Zoffix m: say '' ~~ <0>
08:09 camelia rakudo-moar 36bc41: OUTPUT: «True␤»
08:09 Zoffix m: say '' ~~ '0'
08:09 camelia rakudo-moar 36bc41: OUTPUT: «False␤»
08:09 Zoffix Kinda annoying. Gonna have to do a bunch of special-casing for this in my code :/
08:15 Zoffix I think I'm too deep in the yak shaving :) I started this weekend wanting to finish the Toaster and I haven't even started on the actual Toaster :/
08:54 Zoffix Benched two pieces of code and surprised just how faster one version is than the other: https://gist.github.com/zoffixznet/30ba2fbc0244a9d19395b4c9486bb157
08:55 Zoffix I guess istype op is so cheap assignment and addition dwarfs it
08:55 Zoffix ZofBot: not sure if istype is too faster or assignment is too slow
08:55 ZofBot Zoffix, my class X::Bind::ZenSlice is X::Bind::Slice { "Cannot bind to {$
08:57 Zoffix well, s/assignment/binding/;
09:02 lizmat joined #perl6-dev
09:18 Zoffix .tell AlexDaniel Present:  dd <00>   vs.   dd <00.0+0i>   vs.   dd <00+0i>
09:18 yoleaux Zoffix: I'll pass your message to AlexDaniel.
09:22 Zoffix ZFOVM: Files=1253, Tests=138338, 124 wallclock secs (23.57 usr  3.75 sys + 2551.96 cusr 157.57 csys = 2736.85 CPU)
09:22 Zoffix ZOFVM: Files=1253, Tests=138338, 124 wallclock secs (23.57 usr  3.75 sys + 2551.96 cusr 157.57 csys = 2736.85 CPU)
09:22 Zoffix Wonder what commit made stuff slow :/
09:22 Zoffix buggable: speed 12
09:22 buggable Zoffix, ▁▁▄▄↑▃▂▂▆▆▂▃ data for 2017-05-31–2017-06-05; range: 4.364s–4.807s; 3% slower
09:22 Zoffix buggable: speed 40
09:22 buggable Zoffix, ▂▃▂▂▁▁▆▇▂▂▁▁▃▂▃▄▂▆▇▄▆▅▃▃▄▃▄▇▃▃▆▆↑▅▄▄██▄▆ data for 2017-05-18–2017-06-05; range: 4.225s–4.807s; 5% slower
09:23 pmurias joined #perl6-dev
10:07 Geth ¦ rakudo/nom: 8a0b7460e5 | (Zoffix Znet)++ | src/core/allomorphs.pm
10:07 Geth ¦ rakudo/nom: Fix allomorph smartmatching
10:07 Geth ¦ rakudo/nom:
10:07 Geth ¦ rakudo/nom: Currently, we always fall to Numeric smartmatch, which gives incorrect
10:07 Geth ¦ rakudo/nom: results when a string is smartmatched against the allomorph.
10:07 Geth ¦ rakudo/nom:
10:07 Geth ¦ rakudo/nom: Fix by implementing own .ACCEPTS in allomorphs that matches as follows:
10:07 Geth ¦ rakudo/nom: Allomorph ~~ Allomorph => check Numeric part only
10:07 Geth ¦ rakudo/nom: Numeric   ~~ Allomorph => check Numeric part only
10:07 Geth ¦ rakudo/nom: Str       ~~ Allomorph => check Str     part only
10:07 Geth ¦ rakudo/nom: Any       ~~ Allomoprh => check both Str and Numeric parts
10:07 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8a0b7460e5
10:08 Geth ¦ roast: 4f48067cc5 | (Zoffix Znet)++ | S02-literals/allomorphic.t
10:08 Geth ¦ roast: Spec allomorphic smartmatching
10:08 Geth ¦ roast:
10:08 Geth ¦ roast: Rakudo impl: https://github.com/rakudo/rakudo/commit/8a0b7460e5
10:08 Geth ¦ roast: review: https://github.com/perl6/roast/commit/4f48067cc5
10:13 Zoffix .oO( wonder if it's illegal to kill pigeons... )
10:13 Zoffix Damn vermin. Harrasing me by landing on my AC and scratching the crap out of it.
10:25 astj joined #perl6-dev
11:20 Geth ¦ roast: 9db1e70a54 | (Elizabeth Mattijsen)++ | S03-operators/union.t
11:20 Geth ¦ roast: Fix comment copypasto
11:20 Geth ¦ roast: review: https://github.com/perl6/roast/commit/9db1e70a54
12:03 Geth ¦ roast: 4e5c102f3c | (Elizabeth Mattijsen)++ | S03-operators/intersection.t
12:03 Geth ¦ roast: Add thorough intersection testing
12:03 Geth ¦ roast:
12:03 Geth ¦ roast: Includes 3 starred out tests that fail at the moment, which need further
12:03 Geth ¦ roast: investigation.
12:03 Geth ¦ roast: review: https://github.com/perl6/roast/commit/4e5c102f3c
12:04 [Coke] .
12:04 Geth ¦ rakudo/nom: 6c6bf36d83 | (Elizabeth Mattijsen)++ | t/spectest.data
12:04 Geth ¦ rakudo/nom: Add thorough set intersection testing
12:04 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6c6bf36d83
12:07 [Coke] there is a place to send security tickets, very similar to the p5 setup, that never got documented. moritz & I have the keys to it, let me get you the email addy.
12:09 lizmat afk until later today&
12:11 moritz [Coke]: oh, I think I tested it once, and it didn't work; I think I have to re-test evventually
12:11 [Coke] I see your test email in the queue.
12:11 [Coke] I gave zoffix the address (I think), we'll see if it shows up.
12:21 AlexDaniel joined #perl6-dev
12:28 * Zoffix seent a test mail
12:30 Zoffix Just to clarify for those who didn't backlog in entirety: there's no security bug. IO::Path.mkdir just takes umask into account, but to me it at first looked like the set chmod was broken, so dir with requested chmod 0o000 might've ended up with 0o777
12:33 [Coke] http://news.perlfoundation.org/2017/06/perl-6-travel-grant-proposal.html
12:34 [Coke] This travel grant falls outside the grants committee process and is being handled directly by the board.
12:43 AlexDaniel Zoffix: so which one is wrong? ‘dd <00>’ for not showing the warning?
12:43 AlexDaniel m: say 0000.5
12:43 camelia rakudo-moar 6c6bf3: OUTPUT: «0.5␤»
12:44 AlexDaniel so dd <00.0+0i> is somewhat consistent, dd <00+0i> warns like it would if you just had a 00 literal somewhere, and <00> doesn't warn but should?
12:45 Zoffix AlexDaniel: IMO kinda none should warn. It's a IntStr. What if I *want* it to have extra zeros?
12:46 AlexDaniel who said that it is an IntStr
12:46 AlexDaniel m: dd WHAT <00.0+0i>
12:46 camelia rakudo-moar 6c6bf3: OUTPUT: «Complex␤»
12:46 AlexDaniel m: dd WHAT <00+0i>
12:46 Zoffix m: dd WHAT <00>
12:46 camelia rakudo-moar 6c6bf3: OUTPUT: «Potential difficulties:␤    Leading 0 has no meaning. If you meant to create an octal number, use '0o' prefix; like, '0o0'. If you meant to create a string, please add quotation marks.␤    at <tmp>:1␤    ------> 3dd WHAT <007⏏5+0i>␤Complex␤»
12:46 camelia rakudo-moar 6c6bf3: OUTPUT: «IntStr␤»
12:46 Zoffix well s/IntStr/allomorph/;
12:46 Zoffix Ah, THAT's why it warns
12:47 Zoffix Ok, never mind then.
12:47 Zoffix m: $ = <00/2>
12:47 camelia rakudo-moar 6c6bf3: OUTPUT: «Potential difficulties:␤    Leading 0 has no meaning. If you meant to create an octal number, use '0o' prefix; like, '0o0'. If you meant to create a string, please add quotation marks.␤    at <tmp>:1␤    ------> 3$ = <007⏏5/2>␤»
12:47 AlexDaniel oh, right
12:47 Zoffix OK. Now it makes sense to me :)
12:47 AlexDaniel it does
12:48 AlexDaniel so if you don't have a WhateverStr, it will warn for integers with leading zeros
12:49 perlpilot joined #perl6-dev
12:53 Zoffix k, $*NL thing is done: https://github.com/zoffixznet/perl6-LN
12:53 Zoffix Toaster time \o/ \o\ |o| /o/
12:54 Zoffix I mean $*LN thing
12:54 Zoffix :)
13:11 timotimo of course perl6 IO is slow if it has to go through the netherlands for everything
13:11 Zoffix .oO( that was a random comment... )
13:28 NeuralAnomaly joined #perl6-dev
14:03 Geth ¦ nqp/master: 12 commits pushed by pmurias++
14:03 Geth ¦ nqp/master: review: https://github.com/perl6/nqp/compare/0ea634c2c3...d5a8007cc2
14:06 Geth ¦ nqp: 39d165dc49 | pmurias++ | src/vm/js/nqp-runtime/core.js
14:06 Geth ¦ nqp: [js] Setup a variable for use by eval
14:06 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/39d165dc49
14:10 Geth ¦ nqp: 329246667c | pmurias++ | .eslintrc.js
14:10 Geth ¦ nqp: [js] Make the .eslintrc really lax, it will be made more strict once we fix known classes of violations
14:10 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/329246667c
14:10 Geth ¦ nqp: c27c99f3da | pmurias++ | src/vm/js/nqp-runtime/io.js
14:10 Geth ¦ nqp: [js] Remove debugging leftover
14:10 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/c27c99f3da
14:10 Geth ¦ nqp: 6d0f60068a | pmurias++ | src/vm/js/nqp-runtime/io.js
14:10 Geth ¦ nqp: [js] Tiny bit of stylistic comformance
14:10 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/6d0f60068a
14:18 Geth ¦ rakudo/nom: ac8b4714fd | (Zoffix Znet)++ (committed using GitHub Web editor) | t/harness6
14:18 Geth ¦ rakudo/nom: Fix shebang
14:18 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ac8b4714fd
14:18 Geth ¦ nqp: 7b7b776391 | pmurias++ | src/vm/js/nqp-runtime/io.js
14:18 Geth ¦ nqp: [js] The exit status of nqp::shell needs to be shifted
14:18 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/7b7b776391
14:18 Geth ¦ nqp: b62e21514d | pmurias++ | t/nqp/086-pipes.t
14:18 Geth ¦ nqp: [js] Test nqp::closefh_i with a exit status we expect
14:18 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/b62e21514d
14:19 Geth ¦ nqp: ccdbeb8649 | pmurias++ | 2 files
14:19 Geth ¦ nqp: [js] Run the now passing pipe test
14:19 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/ccdbeb8649
14:21 Geth ¦ rakudo/js: 2d1fb3fd1e | pmurias++ | tools/build/NQP_REVISION
14:21 Geth ¦ rakudo/js: Bump NQP for js backend fixes
14:21 Geth ¦ rakudo/js: review: https://github.com/rakudo/rakudo/commit/2d1fb3fd1e
14:48 [Coke] moritz, Zoffix: test email/ticket rec'd.
14:49 Zoffix sweet
15:09 Geth ¦ rakudo/proc-using-proc-async: fd3ef6f842 | (Jonathan Worthington)++ | src/core/Proc/Async.pm
15:09 Geth ¦ rakudo/proc-using-proc-async: Give Proc::Async support for merged output.
15:09 Geth ¦ rakudo/proc-using-proc-async:
15:09 Geth ¦ rakudo/proc-using-proc-async: Available through the `.Supply` method, which feels fairly natural.
15:09 Geth ¦ rakudo/proc-using-proc-async: In IO::Socket::Async, .Supply is how you get the data received over
15:09 Geth ¦ rakudo/proc-using-proc-async: the socket; `.Supply` on a process can be seen as "get me the output"
15:09 Geth ¦ rakudo/proc-using-proc-async: without caring which output.
15:09 Geth ¦ rakudo/proc-using-proc-async:
15:09 Geth ¦ rakudo/proc-using-proc-async: Added as part of work towards re-working Proc in terms of Proc::Async,
15:09 Geth ¦ rakudo/proc-using-proc-async: which means giving Proc::Async the features that Proc has. However, as
15:09 Geth ¦ rakudo/proc-using-proc-async: it turns out :merge was busted and disabled for Proc, this counts as a
15:09 Geth ¦ rakudo/proc-using-proc-async: bonus, I guess.
15:09 Geth ¦ rakudo/proc-using-proc-async: review: https://github.com/rakudo/rakudo/commit/fd3ef6f842
15:11 AlexDaniel MasterDuke: should we start changing whateverable according to this ↑ ?
15:11 * jnthn ponders how, in terms of API, binding Proc::Async stdin/stdout/stderr to a file handle should look
15:12 AlexDaniel moritz: I think you were interested in this thing also ↑
15:12 jnthn Overloading stdout and stderr to take IO::Handle feels a tad smelly
15:13 jnthn Since in that case they'd not return a Supply
15:13 jnthn And there'd be a discontinuity with stdin too
15:13 jnthn Maybe should have explicit .bind-stdin, .bind-stdout, and .bind-stderr or so
15:15 jnthn Will leave that here for pondering
15:16 jnthn And maybe implement it tomorrow :)
15:16 jnthn In the meantime, I'll go find my Windows box and see how syncfile changes are there, and perhaps merge that branch if all's good.
15:17 jnthn (That's a Moar branch, fwiw, that gets file handles off using libuv)
15:26 Zoffix Is there any reason why making X::Method::NotFound.message slower would get significant performance loss in stresstest? https://github.com/rakudo/rakudo/commit/2500e50
15:26 Zoffix m: say (121-118)/121
15:26 camelia rakudo-moar ac8b47: OUTPUT: «0.024793␤»
15:26 Zoffix And by significant I mean 2.5% loss
15:27 Zoffix buggable: speed 20
15:27 buggable Zoffix, ▄▃▁▁▂▁▃▅▁▁▄▄↑▃▂▂▆▆▂▄ data for 2017-05-28–2017-06-05; range: 4.344s–4.807s; ~0% difference
15:27 Zoffix buggable: speed 10
15:27 buggable Zoffix, ▄▃▇▂▁▁▅▅▁▃ data for 2017-06-01–2017-06-05; range: 4.400s–4.807s; 1% faster
15:29 jnthn Wow, that's a lot...
15:33 Zoffix prolly should come up with a comprehensive bench to make hunting stuff like this easier. The randomness in stresstest is LTA
15:35 * Zoffix tries using spectest instead
15:36 Geth ¦ rakudo/js: 683728bb58 | pmurias++ | tools/build/NQP_REVISION
15:36 Geth ¦ rakudo/js: Correct NQP bump
15:36 Geth ¦ rakudo/js: review: https://github.com/rakudo/rakudo/commit/683728bb58
15:36 Geth ¦ rakudo/js: 2a64d2f2f2 | pmurias++ | t/spectest.js.data
15:36 Geth ¦ rakudo/js: [js] Correct passing spectest list
15:36 Geth ¦ rakudo/js: review: https://github.com/rakudo/rakudo/commit/2a64d2f2f2
15:51 Geth ¦ nqp: e624c72e92 | jnthn++ | t/nqp/019-file-ops.t
15:51 Geth ¦ nqp: Correct left-behind use of nqp::tellfh.
15:51 Geth ¦ nqp:
15:51 Geth ¦ nqp: Should just be .tell on an NQP file handle object.
15:51 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/e624c72e92
16:00 Zoffix Well, based on my experiment. looks like the ~2.6% stresstest/spectest loss is due to commit https://github.com/rakudo/rakudo/commit/2500e50
16:01 Zoffix Not 100% sure why. Gonna rewrite it in NQP now and see if it'll make that 2.6% go down
16:02 jnthn Odd. It even only does the work if .message is called to actually get the error
16:03 Zoffix Yeah
16:04 Zoffix zoffix@VirtualBox~/CPANPRC/rakudo/t/spec (master)$ grep -FR 'X::Method::NotFound' | wc -l
16:04 Zoffix 58
16:04 Zoffix And barring loops, there isn't that many of them being tested
16:04 jnthn I wonder if something triggers it accidentally somewhere
16:17 Zoffix m: use nqp; my $sug := nqp::hash; nqp::bindkey_i($sug, 'foo', 1);
16:17 camelia rakudo-moar ac8b47: OUTPUT: «MVMHash representation does not support native type storage␤  in block <unit> at <tmp> line 1␤␤»
16:17 Zoffix How do I get one that does?
16:19 Zoffix grepping for bindkey_i suggests only lexpad is that... /me goes with regular nqp::hash
16:23 japhb Zoffix, re: https://gist.github.com/zoffixznet/30ba2fbc0244a9d19395b4c9486bb157 either my eyes are failing me (totally possible), or the old code does 'nqp::istype(b, Numeric)' twice, meaning the only possible values of the add_i are 0 or 2, so the outer nqp::if will always take the true branch.
16:27 Zoffix japhb: a yeah. Indeed it has a mistake. Though, with it fixed, I still get "NEW version is 52.16x faster"
17:19 Zoffix Well, this was a waste of time.
17:30 Zoffix Filed my results: https://rt.perl.org/Ticket/Display.html?id=131515
17:30 Zoffix Giving up with it.
18:13 TimToady do we know if there are going to be any hackathons around the Amsterdam conference?  (starting to think about plane tickies...)
18:21 lizmat TimToady: I think there will be
18:21 lizmat what would be best for you?  Before or after the conference ?
18:42 TimToady the after ones tend to be slightly more productive in my experience, since before people tend to be writing their talks, or attending classes, which tend to be before
18:43 TimToady but either would work for us
18:44 ilmari joined #perl6-dev
18:44 lizmat ok, I'll make sure we discuss this tomorrow at the amsterdam.pm meeting, should have some more info after that  :-)
18:45 TimToady thanks, ticket prices will go up soon...
19:03 Geth ¦ nqp: 0ac9b4f06f | (Elizabeth Mattijsen)++ | tools/build/MOAR_REVISION
19:03 Geth ¦ nqp: Bump Moar to get the latest delibuvied goodies
19:03 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/0ac9b4f06f
19:03 Geth ¦ nqp: version bump brought these changes: https://github.com/MoarVM/MoarVM/compare/2017.05-25-g62bc54e...2017.05-59-gf17adad
19:08 Zoffix m: say 1, 2, 3 …^ 4, 5, 6
19:08 camelia rakudo-moar ac8b47: OUTPUT: «(1 2 3 5 6)␤»
19:08 Zoffix There can be more than one endpoint...
19:10 Zoffix ehhe... I have 2017.05-320-g8ec1810 installed on a box and a module install just failed, telling me my perl6 is too old :)
19:10 Zoffix ZofBot: now that's what I call bleeding edge!
19:10 ZofBot Zoffix, It involves the "serialization" of code and data
19:15 Geth ¦ rakudo/nom: 05f3e9a010 | (Elizabeth Mattijsen)++ | tools/build/NQP_REVISION
19:15 Geth ¦ rakudo/nom: Bump NQP version to get latest MoarVM
19:15 Geth ¦ rakudo/nom:
19:15 Geth ¦ rakudo/nom: Specifically because synchronous file I/O now no longer uses libuv.
19:15 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/05f3e9a010
19:15 Geth ¦ rakudo/nom: version bump brought these changes: https://github.com/perl6/nqp/compare/2017.05-98-g05e5d75...2017.05-145-g0ac9b4f
19:28 * [Coke] wonders if he can host an albany conference just to get people to swing by. :)
19:39 Zoffix ZOFVM: Files=1254, Tests=138733, 120 wallclock secs (22.25 usr  3.47 sys + 2434.76 cusr 167.69 csys = 2628.17 CPU)
19:41 Geth ¦ rakudo/nom: 8efffb1d13 | (Zoffix Znet)++ | src/core/Rakudo/Internals.pm
19:41 Geth ¦ rakudo/nom: Fix IO::Path.exception with Range parts
19:41 Geth ¦ rakudo/nom:
19:41 Geth ¦ rakudo/nom: …when end points are excluded.
19:41 Geth ¦ rakudo/nom:
19:41 Geth ¦ rakudo/nom: Range.minmax already returns points with excluded parts excluded.
19:41 Geth ¦ rakudo/nom: Also, it can't be called on non-int Ranges.
19:41 Geth ¦ rakudo/nom:
19:41 Geth ¦ rakudo/nom: Fix by using .min/.max instead (the excluded end already get subtracted)
19:41 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8efffb1d13
19:41 Geth ¦ roast: a5d933fcef | (Zoffix Znet)++ | S32-io/io-path-extension.t
19:41 Geth ¦ roast: Test IO::Path.extension with excluded Range as parts
19:41 Geth ¦ roast:
19:41 Geth ¦ roast: Rakudo fix: https://github.com/rakudo/rakudo/commit/8efffb1d13
19:41 Geth ¦ roast: review: https://github.com/perl6/roast/commit/a5d933fcef
19:44 lizmat Zoffix: re 8efffb1d1326663d7e81 , wonder how that works on 32bit systems
19:45 Zoffix Wouldn't islt_I take any Int?
19:46 Zoffix Wonder where -2**63 came from actually
19:46 Zoffix m: say -2**64
19:46 camelia rakudo-moar 05f3e9: OUTPUT: «-18446744073709551616␤»
19:46 Zoffix oh "# we have isbig_I, but it tells whether the value is above max int32 value"
19:47 Zoffix Yeah, probably needs fixing. I'll check next time I'm at my 32-bit box
19:47 dogbert17 joined #perl6-dev
19:48 Zoffix I'll go over the whole .extension code. It's noob-Zoffix NQP code :)
19:53 Zoffix BTW that +> ticket was kinda forgotten it seems: https://rt.perl.org/Ticket/Display.html?id=131306#ticket-history
19:57 dogbert17 hmm, is getting SEGV's on two spectest files
19:58 dogbert17 t/spec/S17-procasync/no-runaway-file-limit.t and t/spec/S17-procasync/many-processes-no-close-stdin.t
20:01 dogbert17 Program received signal SIGSEGV, Segmentation fault.
20:01 dogbert17 0xb7c7594c in gc_mark (tc=0x804c458, st=0x8095b78, data=0xa87e438, worklist=0xa87b7e0) at src/6model/reprs/MVMCallCapture.c:54
20:01 dogbert17 54             if (flag_map[flag] & MVM_CALLSITE_ARG_NAMED) {
20:01 Zoffix I noticed more flapping in all the module tests recently.
20:01 Zoffix ZOFFLOP: t/spec/S11-modules/require.t
20:03 dogbert17 I ran my spectest with some debug thingies turned on and with a smaller nursery. Haven't seen these asplode earlier though
20:05 Zoffix buggable: tags
20:05 buggable Zoffix, Total: 1623; BUG: 1064; UNTAGGED: 375; LTA: 168; NYI: 95; JVM: 60; RFC: 57; REGEX: 56; CONC: 55; UNI: 30; SEGV: 28; PERF: 26; TESTNEEDED: 25; IO: 23; REGRESSION: 23; NATIVECALL: 22; POD: 21; @LARRY: 20; TODO: 18; PRECOMP: 13; BUILD: 11; OO: 11; TESTCOMMITTED: 11; STAR: 10; OPTIMIZER: 9; BOOTSTRAP: 5; REPL: 5; GLR: 4; MATH: 4; OSX: 4; WEIRD: 4; SPESH: 3; WINDOWS: 3; MOAR: 2; RT
20:05 Zoffix 1623... ~500 more than when I came to Perl 6.
20:06 dogbert17 and the doc repo had like 70 issues at the start of 2016, now it's like 260
20:06 Zoffix heh
20:07 Zoffix Guess having no tickets is worse, so I shouldn't complain.
20:38 AlexDaniel .help
20:38 yoleaux AlexDaniel: I'm yoleaux. Type .commands to see what I can do, or see http://dpk.io/yoleaux for a quick guide.
20:39 AlexDaniel .in 1.5y What about 1000 open tickets? “<Zoffix> Guess having no tickets is worse, so I shouldn't complain.”
20:39 yoleaux AlexDaniel: I'll remind you on 5 Dec 2018 08:39Z
21:00 ugexe Zoffix: zef#171 seems to be specific to tap::harness6
21:05 Zoffix ugexe: in what way?
21:05 ugexe it doesn't happen if i use --/tap-harness
21:06 ugexe so i'm either passing in the paths wrong or somewhere in tap::harness it executes a proc without including CURFS repos in $*REPOs
21:07 Zoffix yeah, same here. don't happen with --/tap-harness
21:08 ugexe my guess is https://github.com/perl6/tap-harness6/blob/26f7586996c7abc438f53a99972697812c79d6e7/lib/TAP.pm#L943-L945 isn't doing what they want it to
21:10 Geth ¦ tap-harness6: 65d021cfe1 | (Zoffix Znet)++ (committed using GitHub Web editor) | lib/TAP.pm
21:10 Geth ¦ tap-harness6: Use correct way of stringifying IO::Path
21:10 Geth ¦ tap-harness6: review: https://github.com/perl6/tap-harness6/commit/65d021cfe1
21:10 ugexe I'm passing in the includes the same as prove6, so it must be in TAP::Harness
21:11 * lizmat tests whether HARNESS_TYPE=6 now updates TAP.pm6  :-)
21:12 lizmat whee!
21:12 lizmat (it does)
21:12 Zoffix \o/
21:15 Zoffix Ah
21:16 Zoffix ugexe: I think the problem is in my test script actually. I'm running $*EXECTUABLE and passing it -Ilib but it's not getting whatever temp path Digest is unpacked at
21:18 ugexe oh, yeah that will do it
21:20 ugexe m: say "-I" <<~<< $*REPO.repo-chain.map(*.path-spec)
21:20 camelia rakudo-moar 8efffb: OUTPUT: «(-Iinst#/home/camelia/.perl6 -Iinst#/home/camelia/rakudo-m-inst-2/share/perl6/site -Iinst#/home/camelia/rakudo-m-inst-2/share/perl6/vendor -Iinst#/home/camelia/rakudo-m-inst-2/share/perl6 -Iap# -Inqp# -Iperl5#)␤»
21:22 Zoffix wond
21:23 Zoffix wonder how prove makes it work
21:27 Zoffix Yup. The fish thing fixes the issue
21:27 ugexe not just prove, perl6
21:27 Zoffix Fish ==  «~«
21:28 ugexe if you do --/tap-harness --/prove it will do `perl6 -I... t/whatever.t` which also works
21:28 Zoffix ugexe: so should I close the issue or do you think it should be possible to make it work without needing changes in my code?
21:29 ugexe i cant figure out why it worked for --prove
21:29 Zoffix .oO( set PERL6LIB to fish thing in TAP::Harness inc thing )
21:30 ugexe you can close the issue, but i would have expected prove to fail before adding the fish thing.
21:40 lizmat and another Perl 6 Weekly hits the Net: https://p6weekly.wordpress.com/2017/06/05/2017-23-fulfilled-tap/
21:44 lizmat good night, #perl6-dev!
21:45 Zoffix night
21:45 jnthn 'night
21:46 timotimo cool, night lizmat
21:50 Zoffix lizmat++ # good weekly
22:03 AlexDaniel oh no… now we have File::Temp and Temp::path
22:03 AlexDaniel Temp::Path*
22:03 jnthn Isn't there something built in for that too? :)
22:17 Zoffix AlexDaniel: yeah, but File::Temp is shit
22:18 AlexDaniel Zoffix: well, I created an issue for that
22:19 * Zoffix closes it
22:20 * AlexDaniel is confused
22:33 Zoffix AlexDaniel: the module's code is not written for Perl 6 IO. I'm guessing it was written before a billion changes were made for Perl 6 IO. It has awful interface that forces you to close filehandles all the time, without even giving a hint that in Perl 6 it's necessary to do so, and it gives you a Str for a path, instead of proper IO::Path. It takes its arguments as positionals, instead of letting you use
22:33 Zoffix nameds. It doesn't document the template must have at least 4 asterisks. Its code doesn't use proper Perl 6 IO and instead uses $*SPEC. Its code will crash on deleting all the rest of the files if a single dir fails to delete (e.g. when its perms don't program read its contents). And to crown it with a cherry on top. Here's a very simple demonstration of it deleting an imporant customer file (I shortened
22:33 Zoffix the template so I could use a single customer data file, but imagine a directory with some letters, e.g. "jack-bowers".. as that's a possible template name): https://gist.github.com/zoffixznet/b37aea33e78218d01e282dd53e0a7049
22:34 Zoffix And I'm not going to write a paragraph in my docs, slugging off another module. I don't advise using it. If that's not good enough for you, go ahead and use it. I wouldn't care.
22:35 AlexDaniel well, you already say “I advise against using it”. I just think that if you're saying this, why not give some reasons why
22:37 Zoffix I'm just gonna remove that section entirely.
22:43 Zoffix Now I'm sad.
22:46 Zoffix Toaster successfully toasted a first batch of 20 modules \o/
22:46 Zoffix Just need to hook it up to an SQLite DB and bolt a web front end onto it and done.
22:46 * Zoffix calls it day
22:58 BenGoldberg joined #perl6-dev
23:14 timotimo that sounds like good stuff,zoffix
23:14 timotimo also, good rest!

| Channels | #perl6-dev index | Today | | Search | Google Search | Plain-Text | summary