Camelia, the Perl 6 bug

IRC log for #perl6, 2015-08-31

Perl 6 | Reference Documentation | Rakudo | Niecza | Specs

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

All times shown according to UTC.

Time Nick Message
00:26 laouji joined #perl6
00:40 laouji joined #perl6
00:44 llfourn joined #perl6
00:49 laouji joined #perl6
00:52 japhb o/
01:13 BenGoldberg joined #perl6
01:20 vendethiel joined #perl6
01:36 ab5tract_ joined #perl6
01:37 zacts joined #perl6
01:38 ggoebel2 joined #perl6
01:40 Fleurety joined #perl6
01:41 cgfbee joined #perl6
02:13 thou ugexe: FYI: this works like I want it to. https://gist.github.com/sof​tmoth/1da0f8e722e98c2363bf
02:13 noganex_ joined #perl6
02:14 thou ugexe: regarding tap/act, I still contend that you've got a race there. See commented-out .act() calls in this code ^; the note() calls for stderr and stdout will overwrite each other.
02:14 thou I have to use another channel to properly merge the lines. I believe your process code has the same issue.
02:16 thou erm, sorry, nevermind! I do think it's still a potential problem, but this one is due to my running multiple threads a the same time
02:17 thou I'd need to do some random sleeping in the act() closures to test my hunch
02:17 thou jobs
02:45 llfourn joined #perl6
03:01 ugexe if i want the lines properly merged, i access stdout and stderr directly. i would not use $merge
03:01 ugexe but thats assuming im using one persons definition of proper
03:02 kaare_ joined #perl6
03:02 llfourn joined #perl6
03:05 ugexe your problem is because you are calling say/note whatever inside your tap/act
03:07 ugexe if multiple threads are printing out $*OUT, you will get what appears to be incorrectly merged lines
03:11 ugexe if you are really trying to merge the output to the terminal, that is a different thing. a blog post that shows the general idea http://ugexe.com/create-a-p​erl6-terminal-progress-bar/
03:13 peteretep joined #perl6
03:13 peteretep homebrew wants to know if I want to install rakudo-star with pcre
03:13 ugexe but you have to make choices. if you receive a line of text without a line break at the end, and then another process fires its own tap to print, what do you do?
03:13 peteretep Is there a quick overview of the implications of that choice?
03:15 ugexe zef::cli::statusbar puts a lock inside $*OUT's print method
03:30 rurban joined #perl6
03:44 laouji joined #perl6
03:46 _itz joined #perl6
03:46 prevost joined #perl6
03:51 _itz joined #perl6
03:52 thou joined #perl6
03:53 llfourn joined #perl6
03:55 TEttinger joined #perl6
04:02 _itz_ joined #perl6
05:00 TimToady checking out breakfast at the Abba, then off to the Alhambra
05:01 TimToady we have one free ticket if anyone happens to be here and awake enough to leave 8:00 or so
05:19 ShimmerFairy TimToady: how many more stops until you're home (or if not that, your next destination)? A quick look for Alhambra places at least two in Spain and California, so it's not like I could guess :)
05:36 emilbarton joined #perl6
05:36 emilbarton hi, why does $str.eval returns "method not found for invocant of class string"?
05:39 ShimmerFairy it's spelled EVAL in Perl 6 now. so it'd be EVAL($str) or $str.EVAL :)
05:39 emilbarton ok thanks
05:52 cognominal joined #perl6
05:55 TimToady ShimmerFairy: "The" Alhambra is in Granada, Spain
05:55 ShimmerFairy ah, I somehow missed the "the" there :)
06:05 ShimmerFairy By the way, I think I'll be ready to push my val() and allomorphic types as a branch tonight :) . The only main issue is that my implementation of IntStr dies on bigints, and I'm not sure how to fix that.
06:06 ShimmerFairy (The only thing I can think of is a fallback Int attribute, and then reimplementing all of Int's methods in IntStr, but ick)
06:09 lizmat joined #perl6
06:09 ShimmerFairy m: say "".Numeric  # this is surprising, and feels wrong to me
06:09 camelia rakudo-moar 5ba44f: OUTPUT«0␤»
06:09 GLRelia rakudo-moar 36ea47: OUTPUT«0␤»
06:12 nine ShimmerFairy: that's probably Perl 5 heritage
06:13 ShimmerFairy nine: replacing .Numeric with a val() call in my code caught that in a test, and I'm quite OK with saying "but I don't see a number in that string"  :)
06:13 ShimmerFairy (replacing the implementation of Str.Numeric, that is)
06:15 dalek rakudo/glr: cb9b821 | FROGGS++ | t/01-sanity/19-say.t:
06:15 dalek rakudo/glr: remove test of say() that does not work with .gist, laben++
06:15 dalek rakudo/glr: review: https://github.com/rakudo/rakudo/commit/cb9b821b08
06:15 dalek rakudo/glr: cf3b24c | timotimo++ | t/01-sanity/19-say.t:
06:15 dalek rakudo/glr: Revert "remove test of say() that does not work with .gist, laben++"
06:15 dalek rakudo/glr:
06:15 dalek rakudo/glr: Was supposed to be on the glr branch
06:15 dalek rakudo/glr:
06:15 dalek rakudo/glr: This reverts commit cb9b821b08095c6d4c1157085e7fc2084f2dd72d.
06:15 dalek rakudo/glr: review: https://github.com/rakudo/rakudo/commit/cf3b24c6c8
06:15 dalek rakudo/glr: 5ba44fc | hoelzro++ | tools/build/NQP_REVISION:
06:15 dalek rakudo/glr: Bump NQP_REVISION
06:15 dalek rakudo/glr:
06:15 dalek rakudo/glr: For close_stdin fix in MoarVM
06:15 dalek rakudo/glr: review: https://github.com/rakudo/rakudo/commit/5ba44fc3dd
06:15 dalek rakudo/glr: 78b8f29 | (Stefan Seifert)++ | tools/build/NQP_REVISION:
06:15 dalek rakudo/glr: Merge branch 'nom' into glr
06:15 dalek rakudo/glr: review: https://github.com/rakudo/rakudo/commit/78b8f293d1
06:15 ShimmerFairy nine: I could see it might've been some sort of guarantee for  ?$string === ?$string.Numeric.Stringy  , back when "0" was False :)
06:16 nine Darn...really have to leave for the office now
06:28 FROGGS joined #perl6
06:32 [Sno] joined #perl6
06:37 [TuxCM] joined #perl6
06:49 aborazmeh joined #perl6
06:49 aborazmeh joined #perl6
06:50 xinming joined #perl6
06:52 rurban joined #perl6
06:58 ChoHag joined #perl6
07:05 masak morning, #perl6
07:07 ShimmerFairy o/
07:07 itz joined #perl6
07:10 laben joined #perl6
07:13 laben hello #perl6
07:13 abraxxa joined #perl6
07:19 ShimmerFairy o/ laben
07:27 xfix joined #perl6
07:31 Ven joined #perl6
07:35 RabidGravy joined #perl6
07:38 darutoko joined #perl6
07:42 Ven joined #perl6
07:47 Ven joined #perl6
07:56 laben as i feared, Template::Mustache doesn't work because the grammar or the actions fail to create the correct result. now, where do i sign up for a crash course in grammars and actions?
07:56 rurban joined #perl6
07:57 nine laben: http://doc.perl6.org/language/grammars is quite good
07:58 laben nine: do you know if Grammars and/or Actions internally use some sort of push that would need to be adjusted for glr?
07:58 DrForr laben: I've got a pretty extensive grammar/actions combo in the ANTLR module in the ecosystem.
08:05 FROGGS laben: do you get error messages you could gist?
08:05 yoleaux 30 Aug 2015 21:00Z <laben> FROGGS: the LAST phaser doesnt seem to fire when the last execution is finished with next, see "for 1 {LAST {say "last"};next}"
08:05 FROGGS ohh, the LAST phaser thing might be some very easily fixable...
08:05 laben just a quick question, the TOP action is called last, after every other action has been called, right?
08:06 rarara joined #perl6
08:08 nine laben: my Apache-To-Nginx seems to parse just fine and that's a largish grammar. From what I can see, there's just a couple of GLR fixes necessary in the converter code itself.
08:09 nine FROGGS: do you by chance still have my short micro USB cable?
08:11 JimmyZ isn't the TOP first?
08:12 laben somehow an action doesnt get called, which means the correct ast is never made. let's see why...
08:15 espadrine joined #perl6
08:16 [Tux] joined #perl6
08:18 FROGGS nine: damn, it was yours?
08:18 FROGGS ó.ò
08:18 FROGGS I convinced tadzik to take it -.-
08:19 nine This...may even be an advantage. You aren't coming to YAPC are you?
08:19 FROGGS JimmyZ: the TOP token is called first, so the TOP action method is called last
08:19 FROGGS nine: I'm not, correct
08:19 colomon joined #perl6
08:19 nine tadzik may be, so I may even have a chance of getting it back ;)
08:19 FROGGS :S
08:19 FROGGS nine: Sorry
08:19 nine No worry.
08:19 tadzik nine: I'm not, sorry :(
08:20 tadzik nine: but I'm still in Zurich :)
08:20 tadzik if you want to catch me
08:20 nine I'm already back in Linz
08:20 tadzik ah :/
08:21 nine I should have charged my phone in Switzerland at least once. Then I'd have noticed the missing cable.
08:21 nine Curse you large battery!
08:21 rindolf joined #perl6
08:21 colomon joined #perl6
08:22 tadzik :)
08:23 DrForr DAMN YOU FOR BEING SO EFFICIENT
08:24 nine Maybe I should just give in and register a Google account to suck my battery empty in a day ;)
08:24 laben gosh, the bottom action doesn't get called even once on glr, that's why it does not create any ast. at the same time, why the heck is nom calling the same action 2 times, with the same argument, when there's evidently only one occurrence?
08:25 nine laben: sounds like we overcompensated?
08:26 lizmat_ joined #perl6
08:28 domidumont joined #perl6
08:30 laben seems like it's called twice because it's used in an assertion. but the assertion totally doesn't get called in glr. what's going on?
08:42 laben need to go now, have a good day #perl6
08:51 lizmat_ good *, #perl6
08:51 lizmat messages?
08:51 yoleaux 30 Aug 2015 10:39Z <nine> lizmat: I figured out what's wrong with zip-latest.t! It's just that every way we tried to fix the issue hit an underlying bug. Array.clone re-uses the same $!reified buffer for the clone and List.from-iterator creates a new $!reified but uses the same scalar containers for the values. So they are still linked.
08:52 lizmat .tell nine Good to know what caused the issue...  :-)
08:52 yoleaux lizmat: I'll pass your message to nine.
08:52 nine lizmat: It's also fixed by now :)
08:52 yoleaux 08:52Z <lizmat> nine: Good to know what caused the issue...  :-)
08:52 lizmat cool!
08:53 nine There's now only 2 GLR related spec test failures: Sequences do not yet throw the correct exceptions on invalid input and one of the advent tests is not properly GLRed yet.
08:54 telex joined #perl6
08:54 lizmat I saw 5: 1 of them is a known issue on OS X with IO::Socket::INET
08:54 lizmat the other two I'm not sure about
08:55 nine t/spec/S02-literals/numeric.t fails also on nom
08:55 nine ShimmerFairy++ fixed numeric.t to actually test what it's claiming to test and that's why it now shows failures. They should just be todo'd
08:56 cdc .tell [Coke] about http://irclog.perlgeek.de/p​erl6/2015-08-29#i_11138242, I use for each arguments of a Bash command-line: sub escape-arg(Str $arg is copy) { if $arg ~~ s:g/([ ' ' | \' | \\ | <:Control> ])/{ $0.ord.fmt('\\u%04x') }/ { [~] <$'>, $arg, <'> } else { $arg } }
08:56 yoleaux cdc: I'll pass your message to [Coke].
08:56 nine many-processes-no-close-stdin.t was also a nom issue that has since been fixed by hoelzro++ (NQP bump was missing)
08:58 araujo joined #perl6
08:58 araujo joined #perl6
09:00 ShimmerFairy sorry for not TODOing those tests, by the way; I think my idea was "oh, I messed around in that part of the grammar recently, I could fix that soon!" and then didn't :P
09:03 espadrine_ joined #perl6
09:13 ShimmerFairy m: say :10<-42>  # this is how I guessed you negate adverbial numbers, but apparently I'm wrong :)  (though to be fair the spec says nothing about a negative number of this form)
09:13 camelia rakudo-moar 5ba44f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/jP8rFHCpOz␤Malformed radix number␤at /tmp/jP8rFHCpOz:1␤------> 3say :10<7⏏5-42>  # this is how I guessed you negate␤    expecting any of:␤        number in radix notation␤»
09:13 GLRelia rakudo-moar 78b8f2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/zkxkvTeU3r␤Malformed radix number␤at /tmp/zkxkvTeU3r:1␤------> 3say :10<7⏏5-42>  # this is how I guessed you negate␤    expecting any of:␤        number in radix notation␤»
09:16 RabidGravy If I need to reference a type 'B' in a role 'F' that may be consumed in term by 'F'   is doing something like "self ~~ ::('B')" in the role valid way of not having to load the class (and get blown out for circular dependency)?
09:17 ShimmerFairy do you want ::?CLASS, per chance?
09:17 RabidGravy no I don't
09:18 RabidGravy the role may be consumed by a variety of things, it needs to check if something is one of them
09:18 RabidGravy (actually self was a bad example there as the actual code doesn't do that)
09:20 ShimmerFairy you can always stub the class(es) in question, and hopefully be able to use them well enough in the role.
09:23 RabidGravy the actual code is something more like:
09:23 RabidGravy m: role F { method b(Mu:D $f) { say $f ~~ ::("B"); } }; class B does F { }; B.new.b(B.new)
09:23 camelia rakudo-moar 5ba44f: OUTPUT«True␤»
09:23 GLRelia rakudo-moar 78b8f2: OUTPUT«True␤»
09:24 RabidGravy where obviously the circumstances of the creation of the second B are somewhat more complicated
09:26 Ven joined #perl6
09:27 RabidGravy I'm going to go with it and see what happens
09:38 mprelude joined #perl6
09:45 nine ShimmerFairy: could you please todo them now? I'm not sure about the protocol there. For example, do we require RT tickets for todos?
09:48 ShimmerFairy nah, I'll get to it. Just a case of focusing on other stuff, but it'd only take a second. :)
09:48 nine That's even better :)
09:49 nine Btw. sorry if I've been a bit grumpy about this. Past few weeks have been exhausting.
09:51 ShimmerFairy That's alright :)
09:56 dalek roast: 6b38ee7 | ShimmerFairy++ | S02-literals/numeric.t:
09:56 dalek roast: TODO a couple tests.
09:56 dalek roast:
09:56 dalek roast: nine++ for catching that they needed it.
09:56 dalek roast: review: https://github.com/perl6/roast/commit/6b38ee747d
10:04 dalek roast/glr: 8fe513c | ShimmerFairy++ | S02-literals/allomorphic.t:
10:04 dalek roast/glr: Updating allomorphic tests.
10:04 dalek roast/glr:
10:04 dalek roast/glr: In particular coercing the allomorphic types for 'is' tests, since
10:04 dalek roast/glr: allomorphic types probably shouldn't implement === for an allomorphic
10:04 dalek roast/glr: type and one of its component types.
10:04 dalek roast/glr: review: https://github.com/perl6/roast/commit/8fe513c789
10:04 dalek roast/glr: d47db92 | ShimmerFairy++ | S02-literals/listquote.t:
10:04 dalek roast/glr: Fix accidental pure string expectation
10:04 dalek roast/glr:
10:04 dalek roast/glr: Since the test in question isn't meant to test allomorphic types, I
10:04 dalek roast/glr: changed the tested key/value pair to not make an allomorphic type.
10:04 dalek roast/glr: review: https://github.com/perl6/roast/commit/d47db92d59
10:04 dalek roast/glr: 6b38ee7 | ShimmerFairy++ | S02-literals/numeric.t:
10:04 dalek roast/glr: TODO a couple tests.
10:04 dalek roast/glr:
10:04 dalek roast/glr: nine++ for catching that they needed it.
10:04 dalek roast/glr: review: https://github.com/perl6/roast/commit/6b38ee747d
10:04 dalek roast/glr: ba3393d | (Stefan Seifert)++ | S02-literals/ (3 files):
10:04 dalek roast/glr: Merge branch 'master' into glr
10:04 dalek roast/glr: review: https://github.com/perl6/roast/commit/ba3393d285
10:07 Ven joined #perl6
10:21 rurban joined #perl6
10:27 Loren_ joined #perl6
10:45 espadrine_ joined #perl6

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

Perl 6 | Reference Documentation | Rakudo | Niecza | Specs