Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-07-12

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

All times shown according to UTC.

Time Nick Message
00:46 tbrowder japhb: it's actually based on PR 112 which lizmat reverted, so I'm not sure how to answer the question.
01:10 dalek roast: 44fc047 | (Zoffix Znet)++ | S02-lexical-conventions/unspace.t:
01:10 dalek roast: [TODO FUDGE] Unspace with sigiless terms
01:10 dalek roast:
01:10 dalek roast: RT#128462
01:10 dalek roast:
01:10 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128462
01:10 dalek roast: Unspace with method calls on sigiless terms must compile and work.
01:10 dalek roast: review: https://github.com/perl6/roast/commit/44fc047be4
01:46 dalek roast: b7d8279 | (Zoffix Znet)++ | S02-literals/quoting.t:
01:46 dalek roast: [TODO FUDGE] fancy quotes work inside qww<> like regular quotes
01:46 dalek roast:
01:46 dalek roast: RT#128304
01:46 dalek roast: review: https://github.com/perl6/roast/commit/b7d827907a
01:46 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128304
03:10 TimToady joined #perl6-dev
03:12 cog__ joined #perl6-dev
05:59 b2gills joined #perl6-dev
06:28 b2gills joined #perl6-dev
06:59 dalek rakudo/nom: 3625369 | (Zoffix Znet)++ | src/core/Any-iterable-methods.pm:
06:59 dalek rakudo/nom: min/max on Failure must fail, not return ±Inf
06:59 dalek rakudo/nom:
06:59 dalek rakudo/nom: Fixes RT#128573
06:59 dalek rakudo/nom:
06:59 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128573
06:59 dalek rakudo/nom: sortiz++
06:59 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/362536920e
06:59 dalek rakudo/nom: d15f626 | lizmat++ | src/core/Any-iterable-methods.pm:
06:59 dalek rakudo/nom: Merge pull request #815 from zoffixznet/fix-Failure-min-max-RT-128573
06:59 dalek rakudo/nom:
06:59 dalek rakudo/nom: min/max on Failure must fail, not return ±Inf
06:59 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d15f6266b3
07:11 dalek rakudo/nom: a6c12a7 | lizmat++ | src/core/Any-iterable-methods.pm:
07:11 dalek rakudo/nom: No point in returning a throw
07:11 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a6c12a7d88
07:55 psch yay, progress.  now the ClassNotFoundException happens in the guts of the objectweb.asm framework instead of inside the URLClassLoader :|
07:59 vendethiel- joined #perl6-dev
08:34 dalek rakudo/nom: 4e5c3e8 | lizmat++ | docs/ChangeLog:
08:34 dalek rakudo/nom: Some ChangeLog entries
08:34 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/4e5c3e88e2
08:43 RabidGravy joined #perl6-dev
09:22 psch alright, i think i've figured out the problem
09:22 psch (1) class loading happens from parent to child
09:22 psch (2) parent class loaders can't load from their children
09:23 psch (3) the codegen framework gets loaded with a parent of the class loader i'm using to load the to-be-adaptor'd class
09:23 psch which, as i understand it, means, that the class loader that exists inside that framework cannot reach into the class loader that actually knows the class we want to build the adaptor for
09:24 psch soo i probably have to preventively start a new class loader that loads the asm framework, and use that same class loader to load classes that are not in CLASSPATH at application start
09:34 [Tux] This is Rakudo version 2016.06-229-g4e5c3e8 built on MoarVM version 2016.06-9-g8fc21d5
09:34 [Tux] test            15.647
09:34 [Tux] test-t           9.014
09:34 [Tux] csv-parser      16.616
09:36 |Tux| joined #perl6-dev
10:43 Zoffix lizmat, so to send a scan of CLA... where would I send it to, would you know? One of the people on http://www.perlfoundation.org/contact_us ?
10:44 lizmat I would send it to Karen, with a CC to Dan
10:44 Zoffix Thanks!
10:57 dalek rakudo/nom: 825d7c8 | lizmat++ | docs/ChangeLog:
10:57 dalek rakudo/nom: Some more ChangeLog entries
10:57 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/825d7c82b6
11:01 dalek rakudo/nom: 72d2d69 | lizmat++ | docs/ChangeLog:
11:01 dalek rakudo/nom: We already mentioned adverbed slices
11:01 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/72d2d696e9
11:14 |Tux| joined #perl6-dev
12:28 * [Coke] will be in Leeds in a few weeks, if anyone is nearby (pretty sure not. :)
12:30 moritz but at least the pound is cheap :-)
12:34 BrokenRobot Emailed my CLA. Second time a charm, I'm hoping :)
12:35 sno joined #perl6-dev
12:36 AlexDaniel joined #perl6-dev
13:02 |Tux| I see some conflict in my logs. Did  my timing post come through today?
13:02 nine |Tux|: it did
13:03 nine 11:34 < [Tux]> test-t           9.014
13:03 |Tux| yep. thnx
13:14 dalek rakudo/nom: c7332c0 | niner++ | src/core/CompUnit/PrecompilationRepository.pm:
13:14 dalek rakudo/nom: Only record a dependency once per precomp file
13:14 dalek rakudo/nom:
13:14 dalek rakudo/nom: If more than one of our dependencies depend on the same dependency, we
13:14 dalek rakudo/nom: recorded this base dependency more than once and also checked it more
13:14 dalek rakudo/nom: than once when loading the precomp file. Filter out those duplicates as
13:14 dalek rakudo/nom: early as possible instead.
13:14 dalek rakudo/nom:
13:14 dalek rakudo/nom: Speeds up loading of complex module hierarchies, e.g. XML::LibXML:
13:14 dalek rakudo/nom: Initial compile: 49s -> 34s
13:14 dalek rakudo/nom: Changing a base module: 39s -> 19s
13:14 dalek rakudo/nom: Loading unmodified: 3s -> 1.9s
13:14 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c7332c0c30
13:14 [Coke] https://github.com/perl6/doc/issues/690 - we should probably encourage these two to talk this out. Do we have anything resembling a community manager?
13:14 [Coke] (last 2 comments)
13:15 [Coke] By my reading, it looks like zoffix overreacted.
13:15 skids joined #perl6-dev
13:22 jnthn nine: Wow. :)
13:22 BrokenRobot "We need a redesign [...] You may find Stylish useful in the interim"... There's no need for "RIGHT NAOW" and "by all means". I opened the Issue on a project, not "Zoffix's stylistic preferences repository"
13:22 BrokenRobot My ignore button works. I've got nothing to talk out.
13:23 tadzik joined #perl6-dev
13:25 perlpilot communication is hard m'kay
13:27 [Coke] BrokenRobot: be that as it may, there's also no need for your response there.
13:33 gfldex BrokenRobot: please don't close valid github tickets just because you didn't like a response
13:34 BrokenRobot gfldex: I closed it because you said we need a redesign and you told me to use Stylish vis-a-vis the Issue.
13:34 AlexDaniel /o\
13:35 BrokenRobot I've read your response as a slightly politer way of telling me to fuck off with my issue. So I closed it and moved on.
13:37 dalek rakudo/nom: 7830ec3 | lizmat++ | src/core/Array.pm:
13:37 dalek rakudo/nom: Make Array.AT-POS 2x faster for unreified elements
13:37 dalek rakudo/nom:
13:37 dalek rakudo/nom: - and about 1% faster for reified elements
13:37 dalek rakudo/nom: - only have a helper method for creating viviable containers
13:37 dalek rakudo/nom: - don't check anything more than once
13:37 dalek rakudo/nom: - allocate when we need a container (if it happens once, it will happen more)
13:37 dalek rakudo/nom: - 2 allocations fewer for each non-reified element seen
13:37 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7830ec3efe
13:37 nine BrokenRobot: did you consider that it was actually not meant the way you read it?
13:38 BrokenRobot nine: yes, that's why my comment was made 45 minutes after I closed the issue. I was considering.
13:39 perlpilot BrokenRobot: was it really just that last sentence that made you think gfldex was telling you to piss off?
13:39 BrokenRobot perlpilot: sure, the rest is valid technical commentary.
13:39 gfldex BrokenRobot: i can assure you that I had no intention to tell you to fuck off
13:39 nine So, it's a simple misunderstanding as they are unfortunately very common in written communications.
13:41 perlpilot BrokenRobot: interesting.  I read it more like "here's something you can do if you want a quick fix"
13:41 gfldex that's why I phrased 'RIGHT NAOW' in lolspeak. It's usually not mean to be understood as literal.
13:41 BrokenRobot OK
13:43 AlexDaniel \o/
13:44 perlpilot gfldex: If your intention was similar to the way I read it, perhaps consider using simple, direct words to that effect next time.  (I don't know if that would help the mitigate communication errors, but it seems like it should)
13:45 BrokenRobot Well, two people other people read it correctly. Obviously the issue was on my end and not how it was written.
13:48 perlpilot BrokenRobot: Any time there's innuendo or "implicit winking" or any other things that would normally be communicated via body language if you were in person, the possibility and probability of misunderstanding goes up if you're only communicating in text.  Moreso if the communication is not real-time-ish (like IRC).  IMHO..
13:56 lizmat m: my @a; for ^1000000 { @a[$_] }; say now - INIT now
13:56 camelia rakudo-moar 7830ec: OUTPUT«1.70064509␤»
13:56 lizmat m: my @a; for ^1000000 -> int $_ { @a[$_] }; say now - INIT now
13:56 camelia rakudo-moar 7830ec: OUTPUT«1.44487401␤»
13:57 lizmat m: say 1.445 / 1.7
13:57 camelia rakudo-moar 7830ec: OUTPUT«0.85␤»
13:57 lizmat ^^^ is there a reason why "for ^10000 {}" isn't equivalent to "for ^10000 -> int $_ {}" ???
13:59 lizmat jnthn timotimo ^^^  ??
14:02 BrokenRobot perlpilot: also https://en.wikipedia.org/wiki/Asperger_syndrome
14:10 jnthn lizmat: Difficult proof.
14:11 jnthn lizmat: Easier to leave it to spesh, I suspet
14:11 lizmat but didn't we special case "for ^1000" already ?
14:11 jnthn Sure
14:11 jnthn But you can't just go transforming stuff based on wishful thinking :P
14:11 lizmat ah, I get you
14:11 jnthn In the case of $_, it's an "is dynamic" iirc
14:12 lizmat its part of the block's signature, duh  :-)
14:12 jnthn Which makes analysis of its usage and later binding a bunch harder
14:12 jnthn Also that, yeah
14:12 jnthn I mean, the analysis for the empty block case is easy, but also not so useful in reality ;)
14:14 gfldex if I use `for ^1000 { .Str }`, will the implicit $_ pop up as part of a signature?
14:14 lizmat m: for ^10 -> int $_ { .Str }
14:14 camelia rakudo-moar 7830ec: ( no output )
14:14 lizmat m: for ^10 -> int $_ { say .Str }
14:14 camelia rakudo-moar 7830ec: OUTPUT«0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤»
14:16 gfldex m: for ^1 { &?BLOCK.signature.perl.say }
14:16 camelia rakudo-moar 7830ec: OUTPUT«:(;; $_? is raw)␤»
14:16 skids joined #perl6-dev
14:27 jnthn You don't even have to use $_ for it to be in the signature
14:28 gfldex sneaky little $_
14:33 lucasb_ joined #perl6-dev
14:37 stmuk
14:41 |Tux| This is Rakudo version 2016.06-233-g7830ec3 built on MoarVM version 2016.06-9-g8fc21d5
14:41 |Tux| test            15.351
14:41 |Tux| test-t           8.637
14:41 |Tux| csv-parser      16.240
14:42 lizmat looking good  :-)
14:43 Ulti yeah whatever has been touched recently has been impacting
14:43 * moritz blames lizmat++ and jnthn++
14:44 moritz and possibly nine++ for precomp speedups
14:44 |Tux| can I be so pushy and annoying to ask for renewed attention for https://rt.perl.org/Ticket/Display.html?id=128184 RT#128184 again please?
14:44 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128184
14:57 gfldex m: grammar G { rule TOP { <r> }; rule r { 'abc' } }; G.^methods>>.say
14:57 camelia rakudo-moar 7830ec: OUTPUT«rule TOP { <r> }␤rule r { 'abc' }␤parse␤parsefile␤subparse␤Method 'say' not found for invocant of class 'NQPRoutine'␤  in block <unit> at <tmp> line 1␤␤»
14:57 gfldex m: grammar G { rule TOP { <r> }; rule r { 'abc' } }; say G.^methods;
14:57 camelia rakudo-moar 7830ec: OUTPUT«Method 'gist' not found for invocant of class 'NQPRoutine'␤  in block <unit> at <tmp> line 1␤␤»
14:57 gfldex m: grammar G { rule TOP { <r> }; rule r { 'abc' } }; say G.^methods>>.perl;
14:57 camelia rakudo-moar 7830ec: OUTPUT«Method 'perl' not found for invocant of class 'NQPRoutine'␤  in block <unit> at <tmp> line 1␤␤»
14:57 gfldex m: grammar G { rule TOP { <r> }; rule r { 'abc' } }; say G.^methods;
14:57 camelia rakudo-moar 7830ec: OUTPUT«Method 'gist' not found for invocant of class 'NQPRoutine'␤  in block <unit> at <tmp> line 1␤␤»
14:58 gfldex m: grammar G { rule TOP { <r> }; rule r { 'abc' } }; G.^methods>>.?say
14:58 camelia rakudo-moar 7830ec: OUTPUT«rule TOP { <r> }␤rule r { 'abc' }␤parse␤parsefile␤subparse␤Method 'dispatch:<.?>' not found for invocant of class 'NQPRoutine'␤  in block <unit> at <tmp> line 1␤␤»
14:58 gfldex m: grammar G { rule TOP { <r> }; rule r { 'abc' } }; G.^methods>>.say
14:58 camelia rakudo-moar 7830ec: OUTPUT«rule TOP { <r> }␤rule r { 'abc' }␤parse␤parsefile␤subparse␤Method 'say' not found for invocant of class 'NQPRoutine'␤  in block <unit> at <tmp> line 1␤␤»
15:02 dalek rakudo/nom: 0941d0f | lizmat++ | src/core/Array.pm:
15:02 dalek rakudo/nom: Array.ASSIGN-POS 4x faster for unreified elements
15:02 dalek rakudo/nom:
15:02 dalek rakudo/nom: - and 10% faster for reified elements
15:02 dalek rakudo/nom: - drop the SLOW-PATH altogether
15:02 dalek rakudo/nom: - don't allocate unless we must
15:02 dalek rakudo/nom: - 7 allocations less per assignment for unreified
15:02 dalek rakudo/nom: - 1 allocation less for assignment to reified
15:02 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0941d0fb9c
15:04 lizmat m: my @a; for ^1000 { @a[$_] = 1 }  # example of assignment to unreified elems
15:04 camelia rakudo-moar 7830ec: ( no output )
15:04 timotimo that's amazing
15:04 timotimo that'll really be worth something for a whole lot of code
15:04 lizmat yeah, thought so too  :-)
15:05 * lizmat hopes someone will make a Rakudo * for the next release
15:06 heatsink joined #perl6-dev
15:24 [Coke] hey, someone had mentioned possibly moving our rakudo html profile page to angular 2 - are you still thinking of that, because I have to learn some Angular 2 for work, and this would be a great learning project. I'm going to start hacking on it, but won't mind if someone beats me to a finished product.
15:24 stmuk lizmat: I was planning to unless someone else gets there 1st!
15:25 [Coke] ++stmuk!
15:25 stmuk also I have a RPI3 and was thinking of releasing binaries for that as well
15:25 BrokenRobot [Coke]: that was me. Go for it. I've not started anything yet
15:26 BrokenRobot And for the thinking part. I kinda want to go through the ticket queue first
15:26 BrokenRobot Only 15% done in the last weekend, so it'll take me a while.
15:28 lizmat ++stmuk
15:28 lizmat afk&
15:28 BrokenRobot And I want to finish the release tool. And hack on some tickets. Plenty of work.
15:29 [Coke] BrokenRobot++ # ticket maintenance.
15:29 [Coke] you can only do so much of that for so long. :)
15:31 timotimo <3
16:02 AlexDaniel hmm how interesting
16:03 AlexDaniel sometimes this link returns 404: https://github.com/perl6/doc/issues/693
16:09 [Coke] Versioning_and_Authership (typo)
16:09 [Coke] not that that impacts the 404
16:10 AlexDaniel [Coke]: I mean, github itself returns 404
16:11 AlexDaniel not anymore! Seems like it had to propagate through their servers or something
16:11 AlexDaniel I wonder if you can observe the same problem on any other newly created issue
16:16 vendethiel joined #perl6-dev
16:26 geekosaur if you can access fast enough, you should be able to reproduce something similar on any web plant backed by a CDN. networks aren't instantaneous over large distances.
16:37 tbrowder I need to talk to someone about pod tables.  My proposed fix will involve extensive changes to rakudo/src/Perl6/Pod.nqp, and I think my upcoming PR should probably become a branch off nom for more inspection and testing by others.
16:50 gfldex tbrowder: the Pod stuff was written many years ago bei Damian, who doesn't frequent this channel. You may have to talk to yourself.
16:50 gfldex tbrowder: I can help you with testing if I got the time to spare when you are ready.
16:52 gfldex tbrowder: that's what I use to look at the structure of $=pod https://gist.github.com/gfldex/64e429b8fa9a518e461b7a2dc38f94df
16:59 tbrowder gfldex: thanks for the info and code.  I'll give it a shot and probably have questions later.  Lizmat has been supportive of my effort, as well as ShimmerFairy and Zoffix, so I hope the branch idea will float okay.
17:01 tbrowder gdflex: I assume the '$' chars at the end of each line are newline markers?  And thanks for the offer of help!
17:05 AlexDaniel tbrowder: if I were you I'd probably talk to ShimmerFairy, as that's someone who has been very interested in pod stuff
17:05 tbrowder roger
17:46 hoelzro tbrowder: I spent a good chunk of time in the POD weeds two years ago, if you have questions
17:47 hoelzro ShimmerFairy was working on very nice (and probably necessary) piece of work that essentially amounted to a new part of the language braid for POD
17:51 tbrowder Okay, pod people (gdflex, AlexDaniel, hoelzro, ShimmerFairy, and lizmat) what's the best way to get my changes out for testing?  Any of you able to start a rakudo branch off nom on the github.com/rakudo repo?
17:51 [Coke] you can fork and work there, also.
17:51 hoelzro tbrowder: do you have a fork of rakudo?
17:51 [Coke] that's standard if you don't have commit bits.
17:51 [Coke] if you can't create a fork, you can't commit to the branch, soo.
17:52 [Coke] er.
17:52 [Coke] if you can't create a branch, you can't commit to it, I meant.
17:55 gfldex tbrowder: if you fork on github I can tell github to watch your fork (I think). Should make it easy for me to spot changes.
17:55 tbrowder Yes, I have my own fork and branch, that's how I have done the PRs for nom in the past; but my changes I believe warrant a branch off nom that I can send PRs for and others can then checkout and test the branch.
17:57 hoelzro tbrowder: if you want to have a branch for this, you can always create it on your fork, and still send PRs in from that branch
17:57 [Coke] .ask pmichaud - I thought the SYN were not the spec, why are you opening tickets citing the syn as the source of truth?
17:57 yoleaux2 [Coke]: I'll pass your message to pmichaud.
17:57 tbrowder I have a fork on github and a local clone of the fork, but my PRs so for have been for changes to nom.
17:58 hoelzro or you're saying you want to merge those changes into a branch on rakudo/rakudo until they're ready for nom?
17:58 [Coke] If they're not ready for merging, why not keep them in a fork?
18:00 tbrowder hoelzro: that's what I'm saying. The changes I believe are too extensive for merging to nom in one PR without others testing it which they can with a nom branch.
18:01 hoelzro tbrowder: if people want to test your changes, they can always pull from your repo
18:01 tbrowder Then we can fiddle with that branch until all are happy.
18:02 hoelzro the disadvantage of having those changes hanging off of a branch on rakudo/rakudo is you would constantly need someone with a commit bit to merge them to that POD staging branch; it may impede your progress
18:02 hoelzro you can still submit PRs against your own repo
18:02 hoelzro and we can all review them there
18:03 tbrowder Well I am at that point with normal PRs. Someone is more likely to accept a PR for a branch than for head I should think.
18:04 gfldex tbrowder: I would need the name of the branch to test against
18:15 timotimo i'm annoyed at github for not showing recently active branches prominently at all on the repository page
18:15 nine It somehow looks like I need to get rid of the "no precomp after changing the repo chain" rule to make custom repos work for real.
18:16 hoelzro timotimo: same
18:20 BrokenRobot They're just a click away: https://github.com/rakudo/rakudo/branches
18:20 tbrowder gdflex: of course, it is in my (and our) best interest to advertise it
18:22 timotimo BrokenRobot: i'd like for the "branches" tab to show if the repo has active branches or if it's a one-branch repo, like a big portion of repos are
18:28 hoelzro timotimo: we could probably easily write a tool to display that information; it would be nice if GH did it themselves =/
18:47 dalek roast: 78ccae4 | (Zoffix Znet)++ | S03-operators/minmax.t:
18:47 dalek roast: Unfudge min/max with 1-arg tests
18:47 dalek roast:
18:47 dalek roast: RT#125334
18:47 dalek roast: RT#128573
18:47 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=125334
18:47 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128573
18:47 dalek roast: review: https://github.com/perl6/roast/commit/78ccae4548
18:49 [Coke] i'm not going to merge a PR from a fork into a branch.
18:49 [Coke] what would the point be?
18:50 [Coke] if you're still doing active developmnt on it, do it in your fork. Until you have collaborators who have commit bits, I see no point in putting into rakudo/rakudo
18:50 [Coke] and even then, you can collaborate in your fork.
19:08 tbrowder Okey doke.
19:08 tbrowder left #perl6-dev
19:13 [Coke] we are historically bad about reviewing submissions, so keep asking if you feel your work isn't getting reviewed in a timely fashion.
19:14 [Coke] I know we've burnt some people out that way.
19:15 [Coke] putting the angular2 stuff here: https://github.com/coke/perl6-profiler-html - my goal is to have this project generate a packed html file that we can then include in NQP.
19:15 [Coke] might even do the packing part first before an angular 2 conversion.
19:18 nine [Coke]: if I were to do a new UI for the profiler, I'd actually make it run on top of a lightweight Perl 6 server. Our current UI cannot cope with large profiles (which values of large that are actually still small)
19:19 nine Performance could be much better if we only give the UI the data that's actually necessary for display and have a simple way for it to request more.
19:20 hoelzro indeed; a few reminders a day wouldn't be a bad idea, considering how many timezones we're stretched across
19:21 perlpilot nine: +1
19:21 [Coke] nine: I'm trying to fix the UI so it -can- run with large profiles
19:22 nine [Coke]: I though browsers were already struggling with decoding the multi megabyte JSON dumps?
19:22 nine [Coke]: anyway if you know a way, go for it! :)
19:23 [Coke] My goal here is to improve performance. regardless of where the data is coming from, we're still going to have some giant HTML page that we have to maintain.
19:24 [Coke] so if I go through this and someone changes the datasource to come from an ajax call to a local server, that's fine. we'll still have a nice maintainable app-in-a-page
19:33 [Coke] regarding the template, it might be nice if we would pass in a flag to set the name of the output file.
19:33 [Coke] ... oh, whoops, digging.
19:37 [Coke] yay, found it. :)
19:46 dalek roast: b777194 | usev6++ | / (6 files):
19:46 dalek roast: Fudge some more failing tests for JVM
19:46 dalek roast: review: https://github.com/perl6/roast/commit/b7771940ba
19:49 dalek roast: c079c6d | usev6++ | S04- (3 files):
19:49 dalek roast: Use 'todo' instead of 'skip' if possible
19:49 dalek roast: review: https://github.com/perl6/roast/commit/c079c6d645
19:54 [Coke] mail to perl6-users asking about a MOAR threading bug.
19:54 [Coke] from mt1957
19:58 dalek rakudo/nom: 2d95de3 | lizmat++ | src/core/Array.pm:
19:58 dalek rakudo/nom: Array.BIND-POS 3.5x faster for non-reified elements
19:58 dalek rakudo/nom:
19:58 dalek rakudo/nom: - and about 5% faster for reified elements
19:58 dalek rakudo/nom: - lose the "ensure allocated"
19:58 dalek rakudo/nom: - allow for Arrays without $!reified set
19:58 dalek rakudo/nom: - actually check for negative indices (didn't before)
19:58 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2d95de34e0
20:04 dalek rakudo/nom: 75971f3 | lizmat++ | docs/ChangeLog:
20:04 dalek rakudo/nom: Mention Array.AT-POS/ASSIGN-POS/BIND-POS improvements
20:04 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/75971f36cf
20:05 lizmat [Coke]  re perl6 profiles: would be nice if we somehow could slice perpendicular to the current setuo
20:06 lizmat aka, have a page that lists one benchmark over a period of time
20:06 lizmat rather than all benchmarks of a given date
20:09 [Coke] lizmat ? that sounds like more of a benchmark than a profile
20:10 [Coke] If you did want actual profiles over time, that sounds like something that has to go into the benchmark, and we can save not just runtimes, but profiles also
20:10 lizmat ah, oops, I was thinking you were working on the pages as seen on http://www.moarvm.org/measurements/perl6-bench/
20:10 [Coke] I might not be getting your ask though.
20:10 lizmat no, mixup entirely on my side  :-)
20:10 [Coke] no, like ./perl6 --profile -e '...'
20:10 lizmat yeah, duh  :-)
20:16 tbrowder joined #perl6-dev
20:36 dalek rakudo/nom: 371b934 | lizmat++ | src/core/Array.pm:
20:36 dalek rakudo/nom: Oops, forgot to add a type smiley
20:36 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/371b934b2d
21:26 gfldex m: sub factory(&before, &after){ return f(){ before(); say 'oi‽'; after(); } }; my %h = one => factory({say 'hello'}, {say 'baby')};
21:26 camelia rakudo-moar 371b93: OUTPUT«===SORRY!===␤Function 'say' needs parens to avoid gobbling block␤at <tmp>:1␤------>  }; my %h = one => factory({say 'hello'}⏏, {say 'baby')};␤Missing block (apparently claimed by 'say')␤at <tmp>:1␤------> ; my %h = one => facto…»
21:27 gfldex m: sub factory(&before, &after){ return f(){ before(); say 'oi‽'; after(); } }; my %h = one => factory({say('hello')}, {say('baby')};
21:27 camelia rakudo-moar 371b93: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Unable to parse expression in argument list; couldn't find final ')' ␤at <tmp>:1␤------> > factory({say('hello')}, {say('baby')};⏏<EOL>␤    expecting any of:␤        argument list␤»
21:27 gfldex m: sub factory(&before, &after){ return f(){ before(); say 'oi‽'; after(); } }; my %h = one => factory({say('hello')}, {say('baby')});
21:27 camelia rakudo-moar 371b93: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Undeclared routine:␤    f used at line 1␤␤»
21:27 gfldex m: sub factory(&before, &after){ return sub { before(); say 'oi‽'; after(); } }; my %h = one => factory({say('hello')}, {say('baby')});
21:27 camelia rakudo-moar 371b93: ( no output )
21:27 gfldex m: sub factory(&before, &after){ return sub { before(); say 'oi‽'; after(); } }; my %h = one => factory({say('hello')}, {say('baby')}); %h<one>();
21:27 camelia rakudo-moar 371b93: OUTPUT«hello␤oi‽␤baby␤»
21:43 Ulti https://gist.github.com/MattOates/c2e19950f46d1a1c241a  this is looking good
21:55 dalek rakudo/nom: ad9103c | (Daniel Green)++ | docs/ChangeLog:
21:55 dalek rakudo/nom: Add IO::[Path,Handle].mode to Changelog
21:55 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ad9103c972
21:55 dalek rakudo/nom: 303e77f | lizmat++ | docs/ChangeLog:
21:55 dalek rakudo/nom: Merge pull request #825 from MasterDuke17/nom
21:55 dalek rakudo/nom:
21:55 dalek rakudo/nom: Add IO::[Path,Handle].mode to Changelog
21:55 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/303e77f610
21:56 gfldex m: my $endings = any <.tar.gz .tar.xz>; say 'file.tar.gz'.split($endings) if 'file.tar.gz'.ends-with(any(<.tar.gz .tar.xz>));
21:56 camelia rakudo-moar 371b93: OUTPUT«any((file ), (file.tar.gz))␤»
22:04 lizmat m: use nqp; my $l := nqp::list(1,2,3,4); nqp::setelems($l,2); dd $l; nqp::bindpos($l,4,42); dd $l  # jnthn, is this expected behaviour of nqp::setelems?
22:04 camelia rakudo-moar 371b93: OUTPUT«(1, 2)␤(1, 2, 3, 4, 42)␤»
22:15 lizmat good night, #perl6-dev
22:15 Zoffix night
22:34 dalek roast: 9236d81 | (Zoffix Znet)++ | S04-statements/repeat.t:
22:34 dalek roast: [TODO FUDGE] repeat inside sub inside a loop executes
22:34 dalek roast:
22:34 dalek roast: ...even when condition is false.
22:34 dalek roast:
22:34 dalek roast: RT#128596
22:34 dalek roast: review: https://github.com/perl6/roast/commit/9236d81b3e
22:34 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128596
22:38 skids joined #perl6-dev
23:14 tbrowder left #perl6-dev
23:53 Zoffix Just noticed: passing TODOs inside of subtests don't fail the run. Don't show up in the "TODOs passed" at the end or anything at all
23:54 Zoffix Hum, but apparently only if there are no todos in the non-subtest areas
23:55 [Coke] ok, https://github.com/coke/perl6-profiler-html now can run a perl6 sample program, extract the rawData for the JSON, "build" a template from source (this is just "copy the old template we already had" right now), insert the new data into the new template, then you can run that. so right now, a really complicated way to just get the existing template back.
23:55 Zoffix m: https://gist.github.com/zoffixznet/d903e0add1e70aea0782edd6253eaf78
23:55 camelia rakudo-moar 303e77: OUTPUT«1..2␤    1..1␤    ok 1 - # TODO foo␤ok 1 - meow␤not ok 2 - # TODO bar␤␤# Failed test at <tmp> line 11␤»
23:55 Zoffix run that with prove -e 'perl6' -vlr   you won't get any TODO PASSED messages, but change the last test to succeed and you'll get a message that 2 TODOs passed
23:56 [Coke] but with this, I can now start splitting up the original HTML source into chunks, or add a build step that minifies the html, etc.
23:56 Zoffix Oh, '2' is the number of the test

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