Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2017-04-23

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

All times shown according to UTC.

Time Nick Message
00:17 astj joined #perl6-dev
01:35 BenGoldberg joined #perl6-dev
01:41 Ben_Goldberg joined #perl6-dev
01:47 Ben_Goldberg joined #perl6-dev
01:49 ilbot3 joined #perl6-dev
01:49 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today
04:14 geekosaur joined #perl6-dev
07:14 stmuk Sounds like having R* 2017.05 off a more stable monthly release is a good idea although I'll probably build a testing one based on 04.x (without official release) to try out things anyway
07:14 yoleaux 22 Apr 2017 10:29Z <Zoffix> stmuk: R* likely delayed you say... I'm thinking of maybe delaying it by 4 weeks and cutting it off 2017.05 or something isn't a bad idea. There are two pretty serious bugs in 2017.04.2 that we'd need to patch or point-release for
07:30 samcv stmuk, let me know how testig one goes
07:31 samcv though idk are there other bugs other than the regex bug i just fixed? i actually don't know what the other point releases were cut for
07:31 * samcv reads changelog on rakudo repo
07:32 samcv oh i see now
07:42 lizmat Files=1191, Tests=56827, 203 wallclock secs (12.15 usr  4.80 sys + 1197.53 cusr 119.23 csys = 1333.71 CPU)
07:42 samcv lizmat, can you tell me about radix?
07:43 lizmat not much more than what I know when looking at the code now
07:43 samcv the nqp op and its return values. see what i just said in #moarvm about plans for the indexic op. since we need to return more data to nqp/rakudo than only the index of the located match
07:43 samcv https://github.com/perl6/nqp/blob/master/docs/ops.markdown#radix
07:44 samcv m: use nqp; my str $thing = '10'; say nqp::radix(10, $thing, 0, 0)
07:44 camelia rakudo-moar ce41e4: OUTPUT: «(10 100 2)␤»
07:45 samcv not sure. it says the that 100 is the radix ** (number of digits converted)
07:45 samcv i would think it'd have converted 2 digits, and it would be at offset 2 (which is the last number)
07:45 samcv so i expected 10, 2,2
07:45 samcv but maybe i'm reading the doc wrong
07:46 Geth ¦ rakudo/nom: 22bd2bbd7a | MasterDuke17++ | src/core/Main.pm
07:46 Geth ¦ rakudo/nom: Escape spaces and quotes in MAIN() usage message
07:46 Geth ¦ rakudo/nom:
07:46 Geth ¦ rakudo/nom: Before, if you had a literal parameter in a MAIN() with a space or
07:46 Geth ¦ rakudo/nom: quote (single or double) the parameter wouldn't get quoted when a
07:46 Geth ¦ rakudo/nom: usage message was printed. This means you couldn't just copy what
07:46 Geth ¦ rakudo/nom: was suggested. Now, qoute if needed.
07:46 Geth ¦ rakudo/nom:
07:46 samcv m: use nqp; my str $thing = ' 10'; say nqp::radix(10, $thing, 1, 0)
07:46 camelia rakudo-moar ce41e4: OUTPUT: «(10 100 3)␤»
07:46 Geth ¦ rakudo/nom: Fixes RT #113954
07:46 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/22bd2bbd7a
07:46 Geth ¦ rakudo/nom: f9ebd24666 | lizmat++ (committed using GitHub Web editor) | src/core/Main.pm
07:46 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=113954
07:46 Geth ¦ rakudo/nom: Merge pull request #1064 from MasterDuke17/escape_spaces_and_quotes_in_usage_message
07:46 Geth ¦ rakudo/nom:
07:46 Geth ¦ rakudo/nom: Escape spaces and quotes in MAIN() usage message
07:46 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f9ebd24666
07:46 samcv m: use nqp; my str $thing = '99'; say nqp::radix(10, $thing, 0, 0)
07:46 camelia rakudo-moar ce41e4: OUTPUT: «(99 100 2)␤»
07:47 samcv but maybe it's just the radix ** 2
07:47 samcv since i see the ** in the docs, but. that's not super clear
07:48 samcv ohhh i get it. ok. it relates to the highest decimal place it parsed
07:48 samcv m: use nqp; my str $thing = '1000.0'; nqp::radix(10, $thing, 0, 0)
07:48 camelia rakudo-moar ce41e4: ( no output )
07:48 samcv m: use nqp; my str $thing = '1000.0'; say nqp::radix(10, $thing, 0, 0)
07:48 camelia rakudo-moar ce41e4: OUTPUT: «(1000 10000 4)␤»
07:48 samcv m: use nqp; my str $thing = '1000'; say nqp::radix(10, $thing, 0, 0)
07:48 camelia rakudo-moar ce41e4: OUTPUT: «(1000 10000 4)␤»
07:48 samcv m: use nqp; my str $thing = '1'; say nqp::radix(10, $thing, 0, 0)
07:48 camelia rakudo-moar ce41e4: OUTPUT: «(1 10 1)␤»
07:53 tadzik joined #perl6-dev
07:53 Geth ¦ nqp: 61b9a2da82 | (Samantha McVey)++ | docs/ops.markdown
07:53 Geth ¦ nqp: [ops.markdown] make radix doc a little more clear
07:53 Geth ¦ nqp:
07:53 Geth ¦ nqp: Was not entirely clear that the words inside stood for a variable
07:53 Geth ¦ nqp: which was the number of digits converted and not human language
07:53 Geth ¦ nqp: grammatical ones. Make it more clear.
07:53 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/61b9a2da82
08:03 timotimo joined #perl6-dev
09:02 [Tux] This is Rakudo version 2017.04.2-28-gf9ebd2466 built on MoarVM version 2017.04-44-gf0db8822
09:02 [Tux] csv-ip5xs        3.221
09:02 [Tux] test            13.123
09:02 [Tux] test-t           5.262 - 5.336
09:02 [Tux] csv-parser      13.096
09:49 Zoffix samcv: whatisit?
09:49 samcv oh. if we're making a point release. the changes are in rakudo
09:50 samcv for the regex thing
09:50 samcv doesn't have to be today, but would be nice if there were a point release. but idk other people think maybe it's not a good idea.
09:51 Zoffix wtf is happening with test-t
09:51 Zoffix buggable: speed 10
09:51 buggable Zoffix, ▁▆▂▄▃▃▃▂▄▃ data for 2017-04-20–2017-04-23; range: 5.102s–5.479s; 3% slower
09:51 Zoffix buggable: speed 20
09:51 buggable Zoffix, ▂▁▃▃▃▃▃▃▄▃▃█▄▆▄▄▄▄▆▅ data for 2017-04-15–2017-04-23; range: 4.994s–5.479s; 4% slower
09:51 Zoffix buggable: speed 30
09:51 buggable Zoffix, ▁▂▄▄▂▃▂▃█▅▂▂▄▃▄▃▃▃▄▃▃█▄▆▅▅▅▄▆▅ data for 2017-04-11–2017-04-23; range: 4.942s–5.479s; 6% slower
09:51 Zoffix buggable: speed 40
09:51 buggable Zoffix, ▂▂▇▁▇▃▇▁▂▁▁▂▄▄▂▄▂▃█▅▂▂▄▃▄▄▃▃▄▃▃█▄▆▅▅▅▄▆▅ data for 2017-04-07–2017-04-23; range: 4.921s–5.479s; 6% slower
09:52 samcv buggable, speed 200
09:52 buggable samcv, Refusing to do more than 100 last entries
09:52 samcv buggable, speed 100
09:52 buggable samcv, ▁▂▁▁▃▂▂▂▂▂▁▁█▂▂▃▁▁▁▁▁▁▁▁▁▁▁▁▁▂▁▁▁▂▂▁▁▁▁▁▁▅▂▁▁▁▁▁▁▁▂▁▃▁▂▁▁▁▁▁▁▁▂▁▂▂▂▁▁▁▁▁▂▂▁▂▁▂▃▂▂▁▂▂▂▂▂▂▂▂▂▃▂▂▂▂▂▂▂▂ data for 2017-03-09–2017-04-23; range: 4.829s–7.664s; 5% slower
09:53 Zoffix samcv: the other bug there weren't a point release for was somethign about <( )> and action methods.
09:53 Zoffix NeuralAnomaly: release
09:53 Zoffix oops
09:53 timotimo haha
09:53 samcv NeuralAnomaly, status
09:53 NeuralAnomaly samcv, [✘] Next release will be in 3 weeks and 6 days. Since last release, there are 7 new still-open tickets (5 unreviewed and 0 blockers) and 45 unreviewed commits. See http://perl6.fail/release/stats for details
09:53 Zoffix Well, even if my VM were up, it's in safemode mode anyway :)
09:54 samcv but my reasoning is regex is something a new user might try out, and if they try it out and it's broken
09:54 samcv it would not be very good of an experience. but the other stuff i guess were problems too. but
09:54 samcv that's just my opinion
09:54 Zoffix samcv: I'd expect the user to use Rakudo Star. Not a tagged rakudo release.
09:54 Zoffix Or rakudobrew, where it's also fixed
09:55 samcv in moar-blead?
09:55 samcv or?
09:55 Zoffix m: say "abc" ~~ m:i/caaaaa/
09:55 camelia rakudo-moar f9ebd2: OUTPUT: «False␤»
09:55 Zoffix Rakudo HEAD
09:55 AlexDaniel joined #perl6-dev
09:55 Zoffix Default `rakudobrew build moar` gets EHAD
09:55 samcv how does that work. also some distros go off our tagged releases i think
09:55 samcv oh
09:55 samcv and moar-blead?
09:55 samcv that's git moar and git rakudo?
09:56 Zoffix You know what. I'm just gonna cut a point release right now.
09:56 samcv k :)
10:00 timotimo oh? things are good?
10:00 Zoffix What things?
10:01 Zoffix Gonna test out the whole "make a branch for release; cherry pick; tag it; then delete it" thing.
10:03 timotimo oh
10:08 Zoffix c: 2017.03 say 'a' ~~ m:i/ab/
10:08 committable6 Zoffix, ¦2017.03: «「a」»
10:08 Zoffix Oh so it's not a new bug?
10:09 Zoffix c: 2017.03 say 'abcdb' ~~ m:i/ab/
10:09 committable6 Zoffix, ¦2017.03: «「ab」»
10:14 Zoffix There's a problem with 'delay release by a month' thing I proposed. Well, kind'f. You can't use that branch to release off roast master. You'd have to check out a commit around the time of last commit to release branch, otherwise you'll have failures.
10:14 Zoffix And if you cherry pick some fixes into it, you'll have even more failures :|
10:20 dogbert17 o/
10:20 dogbert17 stupid question of the day
10:21 dogbert17 does nqp::die write its output to stdout and if so why?
10:21 timotimo it does not
10:21 timotimo it throws an exception
10:21 dogbert17 so where's the exceptions text written then?
10:22 timotimo whatever handler catches it
10:22 dogbert17 looking a bit at RT #130760
10:22 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=130760
10:22 timotimo probably the one in Perl6::Compiler if you don't do anything to catch it yourself
10:24 timotimo could very well be that the standard exception handler isn't even in scope when that exception gets thrown
10:24 Zoffix dogbert17: there's a `say` for usage: https://github.com/rakudo/rakudo/blob/nom/src/Perl6/Compiler.nqp#L81
10:25 * dogbert17 looks
10:26 dogbert17 would it be possible to use 'note' instead?
10:28 Zoffix REMINDER: The proper TAP.pm6 now lives in https://github.com/perl6/tap-harness6
10:28 timotimo yeah, but it should only use note if the usage is printed because of an error
10:28 timotimo i.e. perl6 --help should go to stdout
10:28 dogbert17 timotimo: agreed
10:30 Geth ¦ tap-harness6: 6e6dc5292f | timo++ (committed using GitHub Web editor) | lib/TAP.pm
10:30 Geth ¦ tap-harness6: port fix for file descriptor exhaustion from rakudo
10:30 Geth ¦ tap-harness6:
10:30 Geth ¦ tap-harness6: it used to not close the stdout descriptors for subprocesses, thus
10:30 Geth ¦ tap-harness6: easily reaching multiple thousand open file descriptors
10:30 Geth ¦ tap-harness6: review: https://github.com/perl6/tap-harness6/commit/6e6dc5292f
10:34 Geth ¦ tap-harness6: 4ff82b1d86 | (Zoffix Znet)++ | lib/TAP.pm
10:34 Geth ¦ tap-harness6: Close IO::Hadle/IO::Pipe when done reading
10:34 Geth ¦ tap-harness6: review: https://github.com/perl6/tap-harness6/commit/4ff82b1d86
10:34 Geth ¦ tap-harness6: c19f5cc8d9 | (Zoffix Znet)++ | .gitignore
10:34 Geth ¦ tap-harness6: Ignore precomp files
10:34 Geth ¦ tap-harness6: review: https://github.com/perl6/tap-harness6/commit/c19f5cc8d9
10:35 Geth ¦ rakudo/2017.04.3-release: 5 commits pushed by TimToady++, (Zoffix Znet)++, (Timo Paulssen)++, (Samantha McVey)++
10:35 Geth ¦ rakudo/2017.04.3-release: 2f143f476d | fix )> to work again whenever .MATCH is called
10:35 Geth ¦ rakudo/2017.04.3-release: 6c66c1b88c | Fix REPL failures with Linenoise
10:35 Geth ¦ rakudo/2017.04.3-release: 0150c7b8c5 | bump NQP to get $!to fix for )> construct
10:35 Geth ¦ rakudo/2017.04.3-release: f756b4b54f | bump nqp for eqatic fix
10:35 Geth ¦ rakudo/2017.04.3-release: 25048824c8 | Bump MoarVM/NQP for case insensitive regex fixes
10:35 Geth ¦ rakudo/2017.04.3-release: review: https://github.com/rakudo/rakudo/compare/0000000000...25048824c8
10:39 Zoffix timotimo: "bump nqp for eqatic fix" what was the issue?
10:39 samcv this is the same thing
10:39 Zoffix Ah ok
10:39 samcv yea
10:39 samcv that actually was the indexic fix that he bumped. then i did equatic fix
10:40 samcv but timo wrote the wrong op in the name of the commit
10:40 samcv but regardless. both ops are now rescued
10:41 Geth ¦ rakudo/2017.04.3-release: b4455d875c | (Zoffix Znet)++ | docs/ChangeLog
10:41 Geth ¦ rakudo/2017.04.3-release: Log changes for 2017.04.3 release
10:41 Geth ¦ rakudo/2017.04.3-release: review: https://github.com/rakudo/rakudo/commit/b4455d875c
10:43 timotimo yes, i did it totally wrong
10:43 timotimo i wonder if we should start using git notes for that kind of thing
10:52 samcv what are git notes
10:52 Geth ¦ rakudo/2017.04.3-release: 147ad2ea91 | (Zoffix Znet)++ | docs/announce/2017.04.3.md
10:52 Geth ¦ rakudo/2017.04.3-release: Add 2017.04.3 release announcement
10:52 Geth ¦ rakudo/2017.04.3-release: review: https://github.com/rakudo/rakudo/commit/147ad2ea91
10:53 Geth ¦ rakudo/2017.04.3-release: a4e59cf3b2 | (Zoffix Znet)++ | VERSION
10:53 Geth ¦ rakudo/2017.04.3-release: [release] bump VERSION to 2017.04.3
10:53 Geth ¦ rakudo/2017.04.3-release: review: https://github.com/rakudo/rakudo/commit/a4e59cf3b2
10:53 samcv yay more versions
10:53 timotimo samcv: sort of a thing you can attach to a commit "after the fact"
10:53 samcv oh
10:54 timotimo the ownership is reversed, so that the commit content doesn't have to change
10:54 samcv well i describe it fully in moarvm commit log
10:54 samcv idk. ok i see
10:54 timotimo i could have put a note in saying "oops my commit message was crap"
10:54 timotimo definitely beats force-pushing over master
10:54 * Zoffix just comments on the commit in github in those cases...
10:55 samcv i should try out git notes
10:59 timotimo i have not used them yet
10:59 timotimo and i don't know of anybody who has
10:59 timotimo more or less just a random thought
10:59 samcv i'll try it out sometime on a private repo and see what happens
11:00 samcv i gotta go, may not be back before i go to sleep
11:01 Zoffix m: say 'sta' ~~ m:i/st/
11:01 camelia rakudo-moar f9ebd2: OUTPUT: «「sta」␤»
11:01 Zoffix c: 2017.01,2017.02 say 'sta' ~~ m:i/st/
11:01 committable6 Zoffix, ¦2017.01,2017.02: «False»
11:02 Zoffix c: 2017.03,2017.04 say 'sta' ~~ m:i/st/
11:02 committable6 Zoffix, ¦2017.03,2017.04: «「sta」»
11:02 Zoffix So that's another regression :/
11:02 Zoffix Ah no
11:02 Zoffix just works differently but still brokenly
11:03 samcv yeah that's going to be an issue until i retrofit indexic/eqatic ops
11:03 dogbert17 timotimo: using note works, the question then is what's the best way to figure out wheather 'note' or 'say'. Looking at the string parameter sent to the 'usage' method feels a bit brittle. Any suggestions?
11:03 samcv to return information about the length of the expansion
11:04 samcv since atm it has no clue those cp's expanded and thinks the needle of 2 graphemes will hae matched a 2 grapheme section of string
11:04 samcv that is next on the list
11:04 dogbert17 s/wheather 'note'/wheather to use 'note'/
11:05 timotimo dogbert17: where does it decide that an error should be displayed?
11:05 samcv well i have eqatic somewhat working. but. need changes in regex compiler to have it do things
11:05 timotimo like "unknown flag: --foobar"
11:05 samcv night o/
11:05 samcv happy point release tidings to you and your kin
11:06 Zoffix night
11:06 timotimo night samcv
11:07 dogbert17 timotimo: not sure but I belive it might be in the class HLL::CommandLine::Parser found in nqp/src/HLL/CommandLine.nqp
11:08 dogbert17 that's where I found the string 'No such option'
11:08 timotimo we can introduce some API for that
11:08 timotimo does that function call the usage function?
11:09 dogbert17 that's the problem (for me) I can't find that place, if it does it is not obvious to me :(
11:10 timotimo let me see
11:10 timotimo ok, it just dies
11:13 timotimo aha
11:14 timotimo nqp/src/HLL/Compiler.nqp
11:14 dogbert17 what have you found
11:14 timotimo line 333
11:14 timotimo it has a try with .parse in it, when it CATCHes, it calls usage
11:14 dogbert17 nice find
11:15 dogbert17 is there a way to figure out if usage has been called from that code or is that the wrong approach?
11:17 timotimo we could set a dynamic variable
11:18 dogbert17 something like $*CALLED-FROM-EXCEPTION ?
11:23 timotimo i wouldn't call it that :)
11:23 dogbert17 :)
11:23 timotimo $*USAGE-FOR-FAILURE?
11:23 dogbert17 naming is one of the hardest things ...
11:25 Geth ¦ rakudo: zoffixznet++ created pull request #1065: 2017.04.3 release
11:25 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/pull/1065
11:26 Zoffix Well, release is done, but do I merge the branch or what?
11:27 Zoffix c: 2017.04.3 say "wut"
11:27 committable6 Zoffix, ¦2017.04.3: «No build for this commit»
11:30 Zoffix git describe
11:30 Zoffix 2017.04.2-28-gf9ebd24
11:30 Zoffix That's off a fresh clone. Gonna merge and see, I guess
11:31 timotimo what did i get put into the contributors list for? for the bump commit? :)
11:32 Zoffix I guess
11:33 dogbert17 or your TAP.pm6 fix
11:33 Zoffix It's not in release
11:33 dogbert17 ah
11:33 timotimo i don't really understand what "whenever MATCH is called" means in the fix comment
11:34 Geth ¦ rakudo/nom: 9 commits pushed by TimToady++, (Zoffix Znet)++, (Timo Paulssen)++, (Samantha McVey)++
11:34 Geth ¦ rakudo/nom: 2f143f476d | fix )> to work again whenever .MATCH is called
11:34 Geth ¦ rakudo/nom: 6c66c1b88c | Fix REPL failures with Linenoise
11:34 Geth ¦ rakudo/nom: 0150c7b8c5 | bump NQP to get $!to fix for )> construct
11:34 Geth ¦ rakudo/nom: f756b4b54f | bump nqp for eqatic fix
11:34 Geth ¦ rakudo/nom: 25048824c8 | Bump MoarVM/NQP for case insensitive regex fixes
11:34 Geth ¦ rakudo/nom: b4455d875c | Log changes for 2017.04.3 release
11:34 Geth ¦ rakudo/nom: 147ad2ea91 | Add 2017.04.3 release announcement
11:34 Geth ¦ rakudo/nom: a4e59cf3b2 | [release] bump VERSION to 2017.04.3
11:34 Geth ¦ rakudo/nom: c0c4084fd5 | Merge pull request #1065 from rakudo/2017.04.3-release
11:34 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/compare/f9ebd24666...c0c4084fd5
11:35 Geth ¦ rakudo/nom: version bump brought these changes: https://github.com/perl6/nqp/compare/2017.04...2017.04-12-ga5337ff
11:35 timotimo huh
11:35 timotimo 2017.04 was 12:13 am, 2017.04.1 was 11:59 am?
11:36 timotimo how did that happen
11:37 Zoffix By releasing within 6 hours?
11:37 Zoffix 12
11:38 Zoffix 12:13am is midnight; 11:59am is noon
11:39 Zoffix c: $*PERL.compiler.version
11:39 committable6 Zoffix, I cannot recognize this command. See wiki for some examples: https://github.com/perl6/whateverable/wiki/Committable
11:40 Zoffix c: HEAD $*PERL.compiler.version.say
11:40 committable6 Zoffix, ¦HEAD(f9ebd24): «v2017.04.2.28.gf.9.ebd.2466»
11:40 Zoffix .tell stmuk well, 2017.04.3 did happen after all. Should be OK to base R* off of, if needed
11:40 yoleaux Zoffix: I'll pass your message to stmuk.
11:41 timotimo m(
11:42 timotimo this time format is simply ridiculous
11:42 timotimo how can anybody ever think it's a good idea to keep this garbage
11:42 * Zoffix is used to it
11:51 Zoffix ZOFFLOP: t/spec/integration/advent2012-day15.t # 6.c-errata version
11:54 Zoffix m: say 19/365
11:54 camelia rakudo-moar c0c408: OUTPUT: «0.052055␤»
11:54 Zoffix In the past year, I took a break for only 5% of days, according to my github contributions chart :|
11:57 timotimo yikes
11:57 timotimo get yourself a little bit of vacation some time :)
12:01 Zoffix Today's as good a time as any, I guess
12:01 * Zoffix & for rest of the day :)
12:02 timotimo \o/
12:02 timotimo have a good one!
12:02 dogbert17 Zoffix: play some games
12:02 timotimo don't you dare come back before you're relaxed ;) ;)
12:03 dogbert17 timotimo: this dynamic variable we spoke about, doesn't it have to be declared in two places?
12:03 timotimo nah, we'd just check if it exists (with a try, probably) when we output the usage
12:03 timotimo oh hold on
12:04 timotimo this is a method we're talkin gabout
12:04 timotimo why not just pass a named parameter
12:04 timotimo that way other users of HLL::Compiler don't have to change for this
12:04 dogbert17 that sounds too easy :)
12:04 dogbert17 but clever
12:07 Zoffix ZOFVM: Files=1241, Tests=133802, 121 wallclock secs (23.86 usr  3.43 sys + 2531.13 cusr 130.78 csys = 2689.20 CPU)
12:09 dogbert17 could you send an appropriate output handle as a named param and then use that when outputting or should we settle for a boolean?
12:11 timotimo i think boolean is enough
12:12 dogbert17 ok, can try to cobble something together when I get back from a short errand &
12:12 dogbert17 unless you want to do it of course
12:13 timotimo i'm working on something else right now :)
12:13 dogbert17 cool, will try to fix it in a bit then :)
13:14 Ven joined #perl6-dev
13:20 Geth ¦ star: d45c907fb8 | (Steve Mynott)++ | tools/star/Makefile
13:20 Geth ¦ star: Increase Rakudo a point version to .3
13:20 Geth ¦ star: review: https://github.com/rakudo/star/commit/d45c907fb8
13:23 stmuk Zoffix: is there going to be a nqp-2017.04.1.tar.gz release based on 2017.04-24-g87501f7b?
13:23 yoleaux 11:40Z <Zoffix> stmuk: well, 2017.04.3 did happen after all. Should be OK to base R* off of, if needed
13:24 Zoffix uuuhhhhh
13:24 Zoffix shit >_<
13:30 stmuk I think it's only needed by star anyway and a usual rakudo build uses a git checkout of nqp so maybe its just a case of sticking that tarball up and not changing rakudo
13:31 Zoffix stmuk: what about packagers and stuff? You now can't build 2017.04.3 rakudo without fetching stuff over net
13:31 stmuk I don't think you can do that anyway AFAIK it uses git .. its only R* that doesn't
13:32 stmuk not 100% sure since I've not looked for a while
13:32 Zoffix stmuk: you can usually make an NQP package off NQP release and make Rakudo package off a rakudo release but 2017.04.3 uses an unreleased NQP version
13:33 stmuk yes this is true
13:33 Zoffix omg...
13:34 Zoffix 2017.04.4....
13:34 timotimo we're learning so much this month!
13:36 dogbert17 such action
13:39 Zoffix The largest point release before was .2
13:40 stmuk I sort of wish things were a bit simplier
13:40 Zoffix Normally they are.
13:40 Zoffix You switch the bot off safemode, tell it to release, and ~40m later release is done.
13:41 stmuk if you shipped a 2017.04-24-g87501f7b.tar.gz you could probably avoid bumping rakudo
13:42 stmuk a bit ugly looking
13:43 Zoffix + that'll also need tweaking download page to recognize such releases + we'll have a release without a tag
13:43 Zoffix Re-releasing is simpler
13:43 Zoffix ummm
13:43 Zoffix samcv: timotimo but... the regex fix. It was in moarvm wasn't it?
13:44 Zoffix So it's kinda pointless to re-release anything without a point moarvm release.
13:44 stmuk it doesn't break the https download page :) but yes re-releasing is probably safer
13:45 Zoffix NeuralAnomaly: status
13:45 NeuralAnomaly Zoffix, [✘] Next release will be in 3 weeks and 6 days. Since last release, there are 7 new still-open tickets (5 unreviewed and 0 blockers) and 54 unreviewed commits. See http://perl6.fail/release/stats for details
13:46 Zoffix At least things will auto-resolve in 3 weeks and 6 days :}
13:51 Zoffix stmuk: does Rakudo Star need a released nqp?
13:53 Zoffix Looking at El_Che's packages, looks like they include moarvm + nqp.
13:53 stmuk the process which automatically builds the final tarball has to ftp the 3 tar balls but I could probably stick in a different tar ball quite easily
13:55 Zoffix Then maybe best do that. Instead of re-releasing moarvm + nqp + yet another point release of rakudo :o
13:56 Zoffix or wait until 2017.05
13:56 * Zoffix headdesks
13:56 stmuk how important is the moarvm fix?
13:58 Zoffix c: 2017.04 say 'abcdefgz' ~~ m:i/az/
13:58 committable6 Zoffix, ¦2017.04: «False»
13:58 Zoffix c: 2017.04 say 'abcdefz' ~~ m:i/za/
13:58 committable6 Zoffix, ¦2017.04: «「z」»
13:58 Zoffix stmuk: ^ it fixes that
13:59 Zoffix c: 2017.04 say 'abcdefz' ~~ m:i/zafsdfsdfsdfsdfsdfsdfdsfsdfsdfsdfdsfsdfsdfdsfsdf/
13:59 committable6 Zoffix, ¦2017.04: «False»
13:59 Zoffix c: 2017.04 say 'abcdefz' ~~ m:i/zafsdfsdfsdfsdfsdfsdfdsfsdfsdfsdfdsfsdfsdfdsfsdfa/
13:59 committable6 Zoffix, ¦2017.04: «False»
14:00 stmuk I'll experiment with this shortly
14:00 stmuk which moarvm hash should I use?
14:01 Zoffix stmuk: current, HEAD, I believe. f0db882
14:03 stmuk is El_Che nxadm I forget>
14:03 Zoffix probably
14:07 dogbert17 so how the heck do you build nqp in a proper manner 'perl Configure.pl --backends=moar  --prefix=/home/dogbert/repos/rakudo/install/' followed by 'make', 'make test' and 'make install' doesnt do the trick. What am I missing?
14:08 Zoffix --gen-moar
14:08 dogbert17 I'm using the same prefix as when I'm configuring moar
14:09 dogbert17 Zoffix: will try 'perl Configure.pl --backends=moar --gen-moar --prefix=/home/dogbert/repos/rakudo/install/'  ?
14:12 dogbert17 nope, 'Unhandled exception: Missing or wrong version of dependency '/home/dogbert/repos/rakudo/install/bin/../share/nqp/lib/MAST/Nodes.nqp' (from 'src/Perl6/Pod.nqp')'
14:22 Zoffix dogbert17: you don't need --gen-more if you're building it yourself and that error sounds like you're running `nqp` instead of `./nqp`
14:23 AlexDaniel “The largest point release before was .2” what was it?
14:23 AlexDaniel not after Christmas
14:23 AlexDaniel ah I see, 2015.07.2
14:29 dogbert17 Zoffix: I tried without now. Do you have to rebuild rakudo afterwards?
14:30 Zoffix dogbert17: yes
14:31 dogbert17 thx, that explains it
14:43 dogbert17 timotimo: the change we talked about seems to be working now :)
15:32 ggoebel joined #perl6-dev
15:45 stmuk I guess 2015.07.2 was GLR related?
16:16 nine stmuk: no, GLR was merged in September 2015
16:18 pmurias joined #perl6-dev
16:18 Geth ¦ nqp/master: 9 commits pushed by (Pawel Murias)++
16:18 Geth ¦ nqp/master: 1cb412411f | [js] Implement VMHash repr.
16:18 Geth ¦ nqp/master: 248dd51215 | Test VMHash repr
16:18 Geth ¦ nqp/master: 8855369a56 | Test serialization/deserialization of VMHash repr
16:18 Geth ¦ nqp/master: e2e2e3067f | [js] Support composing P6int with integer info but without bits
16:18 Geth ¦ nqp/master: 6519d68e47 | [js] Make serialization context boolify as true
16:18 Geth ¦ nqp/master: c49e4a3fe6 | Test nqp::istrue on a serialization context
16:18 Geth ¦ nqp/master: e07b6471cd | [js] Fix 'static' variables when we don't deserialize an sc
16:18 Geth ¦ nqp/master: d2692f47ae | Revert "[js] Clean up wval setup"
16:18 Geth ¦ nqp/master: d8dddb77ce | [js] Fix bug with nqp::clone
16:18 Geth ¦ nqp/master: review: https://github.com/perl6/nqp/compare/61b9a2da82...d8dddb77ce
16:23 stmuk ah
16:38 BenGoldberg joined #perl6-dev
16:43 Zoffix New blog post: "The Failure Point of a Release": https://perl6.party/post/The-Failure-Point-of-a-Release
16:45 Geth ¦ rakudo/nom: 1bce2e5472 | (Zoffix Znet)++ (committed using GitHub Web editor) | docs/release_guide.pod
16:45 Geth ¦ rakudo/nom: Log all 2017.04 point releases
16:45 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1bce2e5472
16:45 nine jnthn: is "muchly" really a word? :)
16:46 timotimo it's acceptably a neologism
16:47 TimToady it's neologismly
16:49 pmurias m: use nqp;say(nqp::getobjsc(nqp::hash()));say('lived')
16:49 camelia rakudo-moar c0c408: OUTPUT: «(signal SEGV)»
16:50 pmurias m: use nqp;say(nqp::isnull(nqp::getobjsc(nqp::hash())));say('lived')
16:50 camelia rakudo-moar c0c408: OUTPUT: «1␤lived␤»
16:52 pmurias lizmat: is the BEGIN nqp::hash a big performance gain?
17:05 pmurias lizmat: the problem with that is that nqp::hashes are usually special cased during serialization but a BEGIN nqp::hash serializes them like a regular object
17:15 Geth ¦ star: a0b4754ec7 | (Steve Mynott)++ | tools/star/Makefile
17:15 Geth ¦ star: versions of nqp/MoarVM for 2017.04.3
17:15 Geth ¦ star: review: https://github.com/rakudo/star/commit/a0b4754ec7
17:20 Geth ¦ star: ef0ccebbda | (Steve Mynott)++ | 3 files
17:20 Geth ¦ star: test-mock now needs oo-monitors
17:20 Geth ¦ star: review: https://github.com/rakudo/star/commit/ef0ccebbda
17:33 lizmat pmurias: well, you probably know more about that than I
17:34 lizmat when I last checked, the nqp::hash was *not* HLLized
17:34 lizmat if that's what you're referring to
17:36 lizmat afk again&
17:54 Geth ¦ star: a7e7b6b92b | (Steve Mynott)++ | 16 files
17:54 Geth ¦ star: Bump versions of submodules which at least install fine on Linux
17:54 Geth ¦ star: review: https://github.com/rakudo/star/commit/a7e7b6b92b
18:02 samcv Zoffix, good per6.party post
18:05 timotimo Zoffix: a typo in your post: you wrote "create directories / and /bar." but it should read / and /foo instead
18:05 timotimo but the rest of the post is good
18:05 Zoffix Fixed. Thanks.
18:06 samcv ok so we're not having another point release? or we are? or we had one for rakudo but not nqp and moar? so
18:06 timotimo we can put non-release moarvm and nqp into the rakudo star
18:13 Geth ¦ star: cdbe6e45b3 | (Steve Mynott)++ | docs/announce/2017.04.md
18:13 Geth ¦ star: import announce
18:13 Geth ¦ star: review: https://github.com/rakudo/star/commit/cdbe6e45b3
18:21 pmurias Zoffix: re Perl6 VIP, it seems a good idea.
18:42 stmuk http://pl6anet.org/drop/rakudo-star-2017.04-RC0.tar.gz
18:47 Ven joined #perl6-dev
18:55 MasterDuke_ aren't there some utility functions somewhere to test the output of stderr?
18:57 moritz yes
18:57 moritz t/spec/packages/Test/Util.pm
18:57 MasterDuke_ thanks
19:00 Zoffix MasterDuke_: is_run in particular
19:01 MasterDuke_ also thanks
19:02 timotimo hm
19:03 timotimo long ago i think we used to upload release tarballs before announcing them and only posted the announcement and uploaded to the "official place" when people have given a thumbs-up
19:03 Zoffix what "people"?
19:05 MasterDuke_ does anyone have an opinion about whether tests for RT #113954 should go in roast or rakudo's tests?
19:05 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=113954
19:05 * Zoffix would put in rakudo
19:07 MasterDuke_ was leaning there myself. but do those tests have access to is_run?
19:08 Zoffix nope
19:09 MasterDuke_ any reason not to copy it over?
19:09 Zoffix nope
19:14 timotimo Zoffix: well, more than just one person
19:15 timotimo no matter who in particular. just get reports from a few people that things work out fine
19:16 MasterDuke_ isn't there some thought of lib/Test.pm6 going away? if not, should i just copy t/spec/packages/Test/Util.pm to lib/Test/Util.pm6? or just copy the functionality i need directly into the .t file?
19:16 Zoffix What do you do with the tag then tho? You can't move it even if people find some issues
19:16 Zoffix MasterDuke_: no such thought
19:17 timotimo you can forcepush over a tag
19:17 Zoffix MasterDuke_: and don't add it to lib add it somewhere in t/
19:18 timotimo hm. does the tag need to exist in order for a release tarball to be attempted?
19:18 timotimo maybe i'm thinking of star the whole time!?
19:19 MasterDuke_ t/lib? t/packages? anyone care?
19:56 geekosaur joined #perl6-dev
20:03 MasterDuke_ got err: "Usage:\n  getout-2074-60417.code 'foo bar' \n"
20:03 MasterDuke_ expected err: "Usage:\n  -e '...' 'foo bar'"
20:07 timotimo that's fun.
20:08 Zoffix Yeah, it creates files it runs then
20:08 Zoffix Just use `like` and read from `run` instead of using is_run
20:08 Zoffix is run(:err, $*EXECUTABLE, <some arts>).err.slurp(:close), 'somethinng or other'
20:09 Zoffix *args
20:09 MasterDuke_ is_run is broken?
20:09 Zoffix .
20:10 MasterDuke_ `is_run('multi MAIN(q|foo bar|) {}', 'foo bar', { err => "Usage:\n  -e '...' 'foo bar' \n" }, 'description'` is what i ran
20:11 Zoffix MasterDuke_: what's `'foo bar'`?
20:12 MasterDuke_ a string with a space in it
20:12 Zoffix ...
20:13 MasterDuke_ i'm testing that foo bar doesn't work and suggests 'foo bar'
20:14 Zoffix You're sending it to STDIN tho
20:15 MasterDuke_ yeah, and then it doesn't hit the multi and the usage goes to stderr
20:16 Zoffix Then why send it at all?
20:16 Zoffix It'll still have the getout* part because is_run uses a file
20:17 MasterDuke_ to match the use case in the RT
20:17 MasterDuke_ but, true, it's not necessary
20:17 Zoffix That's not the usecase in the ticket. In the ticket you're passing as args, not STDIN
20:18 Zoffix There are docs for is_run: https://github.com/perl6/roast/blob/master/packages/Test/Util.pm#L305-L351
20:19 MasterDuke_ oh, right, STDIN
20:23 MasterDuke_ `is_run('multi MAIN(q|foo bar|) {}', :args['foo', 'bar'], { err => "Usage:\n  -e '...' 'foo bar' \n" }, 'description'` still has the getout... in the 'got err'
20:52 MasterDuke_ so i'm ending up with `is run(:err, $*EXECUTABLE, ['-e', 'multi MAIN(q|foo bar|) {}', 'foo', 'bar']).err.slurp, ...`
20:52 MasterDuke_ fwiw, slurp(:close) would cause `The spawned command '<...>' exited unsuccessfully (exit code: 2)`
21:11 Geth ¦ rakudo: MasterDuke17++ created pull request #1066: Add tests for RT #113954
21:11 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/pull/1066
21:11 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=113954
21:12 BenGoldberg joined #perl6-dev
21:14 Geth ¦ rakudo: MasterDuke17++ created pull request #1067: Add missing '.' to end of expected text
21:14 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/pull/1067
21:48 MasterDuke_ m: role R { method r {...} }; class C is R { method r { say "hi" } } # RT #118263
21:48 camelia rakudo-moar 1bce2e: OUTPUT: «5===SORRY!5=== Error while compiling <tmp>␤Method 'r' must be implemented by R because it is required by roles: R.␤at <tmp>:1␤»
21:48 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=118263
21:50 timotimo you want "does" not "is" there
21:50 MasterDuke_ you can "is" roles though
21:50 timotimo yeah
21:50 timotimo it'll stub the role so the class can derive from it
21:50 MasterDuke_ m: role R { method r { say "hi" } }; class C is R { }; C.new.r
21:50 camelia rakudo-moar 1bce2e: OUTPUT: «hi␤»
21:50 timotimo but it'll give you a stranger error message, still
21:52 MasterDuke_ i feel like the error message would end up a paragraph
21:54 timotimo haha
21:54 MasterDuke_ Method 'r' must be implemented by R because it is required by roles: R. Classes can't implement methods stubbed in roles when they "is" the role, did you mean to "does" the role?
21:54 timotimo "can't pun a role with required methods"
21:58 MasterDuke_ timotimo: that's pretty good
21:59 timotimo if it could also figure out that the pun was made because of a "is" instead of "does", that'd be better
22:00 MasterDuke_ the backtrace starts in `sub trait_mod:<is>`, so hopefully should be possible
22:04 MasterDuke_ is there a way to tell if the $parent in `sub trait_mod:<is>(Mu:U $child, Mu:U $parent)` has required methods?
22:04 samcv can people here weigh in on these non-standardized tags in META https://gist.githubusercontent.com/samcv/b7785ded237a92a254d4298881a1eed8/raw/f089f8f751492b472d7c031fcec3e441f60af8fe/-
22:05 samcv some are wrong, and some maybe should be added to the spec. but what do you guys think
22:19 timotimo i didn't even see that i had already answered that bug!
22:19 timotimo that bug is *old*
22:20 timotimo oh
22:20 timotimo i was also the one who created that bug
22:21 MasterDuke_ heh, yeah. was just looking through old bugs to see if there was anything easy to pick off
22:28 MasterDuke_ P6opaque: must compose Whatever before allocating
22:28 MasterDuke_ don't think i've seen that before
22:28 timotimo it's when you try do put a * very early into the core setting
22:29 MasterDuke_ ah, thanks. should be easy to work around
22:38 MasterDuke_ m: role R { method a {...}; method b { say "bee" }; method c {...} }; class C is R { }; C.new
22:38 camelia rakudo-moar 1bce2e: OUTPUT: «5===SORRY!5=== Error while compiling <tmp>␤Method 'a' must be implemented by R because it is required by roles: R.␤at <tmp>:1␤»
22:38 MasterDuke_ locally that says `Can't pun role R with required methods: a, c`
22:39 MasterDuke_ timotimo: think that ^^^ looks ok?
22:39 timotimo that makes it sound like you're "punning a role with methods"
22:39 timotimo like, you're using methods to pun something
22:40 timotimo i'd phrase it more like "role R because it has required methods: ..."
22:42 MasterDuke_ yeah, that's better
22:43 timotimo i think i'll be going to bed soon
22:46 MasterDuke_ "Class C can't pun role R because it has required methods: a, c"
22:46 MasterDuke_ how's that?
22:47 timotimo do you actually find out that it's Class C that's doing it?
22:47 MasterDuke_ well, i find out the C part
22:47 timotimo yeah, that sounds good
22:47 MasterDuke_ can anything else "is" things?
22:47 timotimo can we find a phrasint that'd somehow hint at "maybe you wanted to use 'does' instead?"
22:48 timotimo everything can is things :P
22:48 timotimo trait_mod:<is> has many candidates
22:48 timotimo m: say &trait_mod:<is>.^candidates
22:48 camelia rakudo-moar 1bce2e: OUTPUT: «No such method 'candidates' for invocant of type 'Perl6::Metamodel::ClassHOW'␤  in block <unit> at <tmp> line 1␤␤»
22:48 timotimo m: say &trait_mod:<is>.candidates
22:48 camelia rakudo-moar 1bce2e: OUTPUT: «(sub trait_mod:<is> (Mu:U $child, Mu:U $parent) { #`(Sub|31809120) ... } sub trait_mod:<is> (Mu:U $child, :$DEPRECATED!) { #`(Sub|31816264) ... } sub trait_mod:<is> (Mu:U $type, :$rw!) { #`(Sub|31816416) ... } sub trait_mod:<is> (Mu:U $type, :$nativesize…»
22:49 timotimo c: HEAD .say for &trait_mod:<is>.candidates
22:49 committable6 timotimo, https://gist.github.com/180c435a12693e68cce480df29d97237
22:50 MasterDuke_ ah, but on the one does `^add_parent`
22:50 MasterDuke_ *only
22:51 timotimo yeah, only the Mu:U, Mu:U one, i bet
22:51 timotimo set a dynamic variable $*PUNSTER :D :D :D
22:52 MasterDuke_ dinner &. will PR after
22:53 timotimo cool. i'm going to bed soon
22:54 MasterDuke_ later...

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