»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
00:02 RabidGravy left 00:11 ThisGeekChick left
llfourn .tell nine I've done some more investigation on RT #128156. Results on the ticket :) 00:15
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128156
yoleaux llfourn: I'll pass your message to nine.
dalek pan style="color: #395be5">perl6-examples: e3cea7e | raiph++ | categories/best-of-rosettacode/24-game.pl:
s/casting/coercion and removed old info
00:21
00:27 adu joined
dalek pan style="color: #395be5">perl6-examples: eaab00f | raiph++ | categories/best-of-rosettacode/ackermann-function.pl:
Update doc links
00:28
00:38 pierre__ joined 00:41 mr-foobar left 00:59 kurahaupo left 01:06 stmuk joined 01:07 stmuk_ left 01:13 astj joined 01:16 cdg joined 01:28 Actualeyes joined 01:29 BenGoldberg joined 01:37 ssotka joined 01:41 molaf left 01:48 pierre__ left 01:49 finanalyst joined 01:50 pierre__ joined 01:54 molaf joined 02:03 dunkel joined 02:04 cpage_ left 02:05 cdg left 02:07 pierre__ left 02:08 kid51 left 02:14 AlexDaniel left 02:15 pierre__ joined 02:17 BenGoldberg left, pierre__ left 02:19 BenGoldberg joined 02:27 finanalyst left 02:40 TreyHarr1 joined 02:51 pierre__ joined 02:52 gregf_ left 02:53 gregf_ joined 02:54 noganex_ joined 02:57 noganex left 03:01 |2701 left, pierre__ left 03:05 skids joined 03:16 jack_rabbit joined 03:20 BenGoldberg left 03:22 BenGoldberg joined, gregf_ left
MasterDuke .tell AlexDaniel PR 775 addresses the first part of RT #128214 by failing when a path part doesn't exist, instead of continuing, but not resolving. the second part wasn't really a problem with resolve, you were just trying to '.dir' on a non-existent path 03:28
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128214
yoleaux MasterDuke: I'll pass your message to AlexDaniel.
03:32 gregf_ joined 03:35 pierre_ joined 03:38 Xliff joined
Xliff \o 03:38
Anyone around?
Xliff is now playing: Frederic Robinson - Walk The Distance feat. Vicky Harrison (Original Mix)
03:41 BenGoldberg left 03:46 BenGoldberg joined 04:02 BenGoldberg left 04:03 gregf_ left 04:04 gregf_ joined 04:13 khw left, skids left 04:20 lizmat joined 04:32 huggable joined 04:36 jcallen joined
Xliff Methinks IO::Handle.slurp-rest on an 90M file might not be a good idea. 04:39
timotimo if you slurp it with :bin, it'll be less problematic
04:39 cpage_ joined, Cabanossi left 04:42 Cabanossi joined
Xliff Even though it is text? 04:42
timotimo well, there's multiple aspects to this 04:43
one, we normalize to NFG, which means combining characters will turn into sinlge graphemes
Xliff And the read never finished before I re-ran using :bin
timotimo when you spurt it back out to utf8, you might end up with different output
Xliff Oh, and YAMLish doesn't work on YAML output from a game I am playing. :( 04:44
timotimo the other thing is, we internally store all codepoints as 32bit integers, so if your text is almost exclusively what fits into ascii, it goes from 8 bits per character to 32bits per character
which is a big inflation
Xliff I may need to see why it is broken and submit a patch.
Ah. Yes, that is.
timotimo i didn't know YAMLish survives its own tests
Xliff But this is a lot of data. I really wish they would have continued to use SQLite exports. 04:45
LOL! I don't think it's configured.
timotimo as far as i know yamlish is b0rked. or perhaps i'm thinking of an older module and somebody else built a YAML parser/emitter in the mean time?
Xliff I may just need to get the MySQL dumps and make my own mapping files. *sigh* 04:46
timotimo in any case, you can still use a yaml module from perl or python with the Inline:: modules
Xliff No. That's the one. Honestly, I was optimistic and just installed it and ran tests.
OK. How do you use an installed Python module using Inline::Python? The docs don't say.
I looked at that, first. 04:47
The only thing I see is Inline::Python::run($script) .... which is bad on many fronts.
timotimo yeah, hmm.
github.com/niner/Inline-Python/blo...call.t#L73 - there's also this kind of thing 04:48
Xliff Yeah. Parsing YAML or large JSON isn't going to work when all I need are ID to Name mappings.
Oh, I guess I could use import and run it that way. 04:49
Hmmm.... thanks!
lizmat good night, #perl6! 04:50
04:50 lizmat left 04:54 pierre_ left 05:00 sno left 05:08 dunkel left 05:10 tokuhirom left 05:15 xiaomiao joined
Xliff timotimo: If I were making a module to connect to a game's API service, what would be a good namespace for it? 05:16
The game in question is Eve Online (www.eveonline.com/)
So Games::EveOnline::API? Just EveOnline::Api looks kinda naked. 05:17
llfourn EveOnline has an API? 05:18
05:26 finanalyst joined 05:39 pierre_ joined 05:40 pierre_ left 05:50 domidumont joined 05:54 CIAvash joined, domidumont left 05:55 domidumont joined 05:59 pierre_ joined 06:02 domidumont left 06:10 sno joined 06:15 domidumont joined 06:24 CIAvash left 06:28 firstdayonthejob joined
Xliff llfourn, yup! 06:29
wiki.eveuniversity.org/EVE_API_Guide
06:30 CIAvash joined
llfourn hmmm it doesn't look like it lets you control your character? 06:30
this is what I'm interested in wrt MMO APIs :( 06:31
06:38 wamba joined 06:39 ssotka left
Xliff No. That would be scripting, which they don't allow. 06:41
Weird. I can run a query via MySQL console and get lots of results, but when I run it through DBIish, I don't get anything.
06:42 gregf_ left
Xliff DUH! Forgot to .execute() 06:43
06:43 pierre_ left 06:46 rurban joined, pierre_ joined 06:48 Actualeyes left 06:49 kaos01 joined 06:50 gregf_ joined 06:53 jjido joined 06:58 rurban left, jjido left
CIAvash Xliff: I think web services should use "WebService", it probably should be WebService::EveOnline 06:59
Xliff CIAvash, good point. Thanks! I had actually started with Net::
But I like WebService better. 07:00
07:01 Actualeyes joined 07:05 mr-foobar joined 07:08 wamba left 07:09 mr-foobar left
Xliff Now I wonder how well perl6 will handle a 3M file that's just a Hash definition. 07:09
07:09 wamba joined 07:10 _mg_ joined 07:11 cpage_ left
CIAvash Xliff: kind of relevant: pause.perl.org/pause/query?ACTION=...odules#Net 07:12
Xliff Ahh... Thanks. 07:13
07:13 brrt joined
Xliff m: $s = "Marvelou's"; $s ~~ S:g/\'/\\'/; 07:14
camelia rakudo-moar 2c4506: OUTPUT«5===SORRY!5=== Error while compiling /tmp/rOmCyEMiEX␤Variable '$s' is not declared␤at /tmp/rOmCyEMiEX:1␤------> 3<BOL>7⏏5$s = "Marvelou's"; $s ~~ S:g/\'/\\'/;␤»
Xliff m: my $s = "Marvelou's"; $s ~~ S:g/\'/\\'/;
camelia rakudo-moar 2c4506: OUTPUT«Potential difficulties:␤ Smartmatch with S/// can never succeed because the string it returns will fail to match. You can use given instead of ~~.␤ at /tmp/ORHXyTL8Ph:1␤ ------> 3my $s = "Marvelou's"; $s ~~ 7⏏5S:g/\'/\\'/;␤»
Xliff ???
moritz good morning 07:15
Xliff \o moritz 07:16
m: my $s = "Marvelou's"; $d = $s ~~ S:g/\'/\\'/;
camelia rakudo-moar 2c4506: OUTPUT«5===SORRY!5=== Error while compiling /tmp/b2RIeyGJHU␤Variable '$d' is not declared␤at /tmp/b2RIeyGJHU:1␤------> 3my $s = "Marvelou's"; 7⏏5$d = $s ~~ S:g/\'/\\'/;␤»
Xliff m: my $s = "Marvelou's"; my $d = $s ~~ S:g/\'/\\'/;
camelia rakudo-moar 2c4506: OUTPUT«Potential difficulties:␤ Smartmatch with S/// can never succeed because the string it returns will fail to match. You can use given instead of ~~.␤ at /tmp/taf8uqoIq8:1␤ ------> 3my $s = "Marvelou's"; my $d = $s ~~ 7⏏5S:g/\'/\\'/;␤»
brrt good * 07:17
Xliff I thought S/// was substitute and return.
CIAvash m: say S:g/\'/\\'/ given "Marvelou's" 07:18
camelia rakudo-moar 2c4506: OUTPUT«Marvelou\'s␤»
moritz Xliff: it is. And ~~ is smart-match
07:18 wamba left
moritz Xliff: so, you're substituting, and the return value is then used for smart matching 07:19
Xliff Right. And I foolishly did that after reading Zoffix's excellent article on the subject. *sigh*
This is what I get for coding after 3am
moritz nearly all of my coding is done after 3am :-) 07:20
kaos01 jesus
Xliff Yeah, but I bet you do coffee
moritz I don't 07:21
but I do sleep between 3am and coding :-)
07:24 rindolf joined
Xliff Yeah. I haven't yet. 07:26
OK. Why am I getting "Method 'match' not found for invocant of class 'Any'" when the field should be VARCHAR. 07:27
moritz sounds like you're getting a NULL
Xliff Nope. It's "#System"
And I'm doing "if $field.defined" before I perform the match. 07:28
07:28 RabidGravy joined
moritz weird 07:28
please show the code
Xliff One sec. 07:29
shadowpaste "xliff" at 217.168.150.38 pasted "WAT??" (10 lines) at fpaste.scsys.co.uk/517506 07:31
Xliff The first say is "#System" the second say is "Str" 07:32
07:33 pierre_ left
raydiak isn't ($d<typeName> given S:g/\"/\\"/) backwards? 07:34
07:34 abraxxa joined
moritz it is 07:36
07:36 masak joined
masak morning, #perl6 07:36
moritz \o masak
raydiak morning masak
masak something... rebooted since yesterday? 07:37
diakopter ur face 07:38
moritz yes, hack, I think
07:38 wamba joined
masak I experienced problems ssh-ing into hack yesterday 07:38
wanted to report it, but was too lazy to use webchat 07:39
07:40 abraxxa left 07:41 abraxxa joined 07:45 pierre_ joined
masak I wrote a thing in Perl 6. it runs in 2m31.850s. then I ported it to Perl 5. it runs in 0m5.990s. 07:49
I wonder if I should submit this thing to some benchmarking suite.
m: my ($minutes, $seconds) = 2, 31.850; my $time = 60 * $minutes + $seconds; say $time / 5.990; 07:50
camelia rakudo-moar 2c4506: OUTPUT«25.350584␤»
moritz m: say ( 2 * 60 + 31.85) / 5.99
camelia rakudo-moar 2c4506: OUTPUT«25.350584␤»
masak "great minds..." :P
srsly, that thing's freaky sometimes :P 07:51
moritz "... don't always type at the same speed"
masak I'll note that I was faster *and* more self-documenting :P
07:51 bjz joined
moritz masak: especially given how little common face time we had 07:51
07:52 adu left
masak .oO( IRC soulmates, forever ) 07:52
next up: gonna re-write the same script in Python 3, to see how it comes out. my wild guess: within 1x-2x of the Perl 5 runtime. 07:54
Xliff Oh. LOL!
nine masak: it's a number, it should be optimized, it should go into the benchmarking suite. Especially since it's probably something real-life related and not an artificial benchmark.
yoleaux 00:15Z <llfourn> nine: I've done some more investigation on RT #128156. Results on the ticket :)
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128156
masak nine: thanks; that's all the encouragement I needed. <3 07:55
07:55 pierre_ left 08:02 araujo joined 08:03 sergot joined 08:11 avalenn joined 08:12 bjz left, darutoko joined
masak japhb: if I want to add a benchmarks which takes 2m31s to run in Rakudo, does it go in minibenchmarks.pl ? 08:17
08:20 ocbtec joined 08:23 bjz joined
Xliff Forcing perl6 to compile a 3M .pm6 08:25
Just to see if something breaks.
Wow! Still compiling. 08:28
gregf_ Sw33tAppl3!23 08:29
yoleaux 20 May 2016 18:08Z <raiph> gregf_: [Nil] becomes [Any]; see irclog.perlgeek.de/perl6/2016-05-20#i_12517566
masak gregf_: if that was a password... maybe time to change your password now...? :) 08:30
Xliff gregf_, wrong window? 08:31
masak: LOL!
===SORRY!=== 08:32
Frame 2 local access out of range
^^ That's what I got. The short code never ran.
08:32 huggable left
Xliff Trying to get Rakudo to compile a 3M hash definition. 08:32
masak Xliff: are you able to find the exact cutoff where it doesn't work? 08:33
Xliff masak: No. That's all I got.
masak I mean, can you play around with the hash size, essentially bisecting it?
Xliff masak: Yeah. I'll play around with that, tomorrow.
masak I've gotten similar bailouts when I've put ridiculously many lexical variables in the same scope.
Xliff Remind me. I'm too tired to do that right now and I need a nap. I need to be up in another 3 hours. 08:34
I'm timing it now. I will paste results before I snooze.
gregf_ ah - man 08:35
din see that ..... *hides*
monitors were off - so i retyped 08:36
Xliff gregf_, no worries if you change that, now.
gregf_ :) 08:37
Xliff Otherwise I might be forced to script log into everything on your subnet using that…
(nice password BTW)
Now I wonder if I can do unicode passwords with Linux. Would make it impossible to log into systems without XCompose or something compatible. 08:38
Or I would be forced to build password via cut&paste which would defeat the purpose.
(BTW - I was just kidding about the script thing...)
masak: 08:39
===SORRY!===
Frame 2 local access out of range
real 3m53.154s
user 3m51.848s
sys 0m1.264s
I will get you an actual size, after I wake up.
masak it's not a nice password, for the record. far too short for 2016. and xkcd.com/936/ applies. 08:47
moritz really depends on the use case 08:48
if it's something that can only be attacked through the network, 13 chars is plenty
also, the question is, what standards do you apply
masak all valid points.
moritz if you compare it to most of leaked passwords out there, it's quite good in comparison 08:49
08:49 dakkar joined
masak "there are worse passwords", sure ;) 08:49
this one is essentially word + word + 123
which is eminently dictionary-able
08:50 CIAvash left 08:53 skingsbu left
masak Rakudo build fails for me: 08:55
./perl6-m tools/build/install-core-dist.pl /home/masak/ours/rakudo/install/share/perl6
Failed to open file /home/masak/ours/rakudo/install/share/perl6/short/6D917C2DF32BB3F8FB6B5F8E529BB779C3DCA6F7/3FD70CCCD6914FAEC84AFAE6F97AF461A3EE1588: not a directory
in any at ././CORE.setting.moarvm line 1
:(
08:55 CIAvash joined
masak oh, it seems to fail during the `make install` part. 08:56
moritz tries
08:56 pierre_ joined
moritz yes 08:56
Xliff masak: Depends on the dictionary attack algorithm.
masak do we no longer support local installs?
Xliff: of course it does.
moritz masak: that would greatly surprise me
masak "this password is vulnerable." -- "depends how you attack it."
tadzik try "iddqd" :P 08:57
masak moritz: this was after a realclean + Configure (moar) + make + make install on Rakudo 2c45068.
09:00 pierre_ left
Xliff masak: I'm still scratching my head over how Randal came up with the number of bits. 09:00
Then again... I need sleep, too! 09:01
OK. Recompiling Rakudo 1fa1e36 to see if I can reproduce error, and then I'm calling it a night. 09:02
I'm doing "rakudobrew build moar"
moritz masak: fwiw installation worked fine here 09:04
Xliff Correction, I was at ce5dc00 and am now at the same commit as masak.
It's been a while since I've freshend up my rakudo install.
09:04 gregf_ left, AlexDaniel joined
Xliff Fails with "main::build_impl("moar", undef, "") called at /home/cbwood/.rakudobrew/bin/rakudobrew line 120" 09:05
Stage parse...
If someone wants me to do more for a bug report, speak now.... 09:06
moritz Xliff: maybe try updating your rakudobrew first?
Xliff: or get some sleep first :-)
#perl6 can wait
masak is currently trying to build an older Rakudo 09:07
keep getting "failed to load library 'dynext/libperl6_ops_moar.so'" even on older checkouts
09:07 labster left
masak I don't remember *building* rakudo being this fragile before -- anyone else having similar troubles lately? 09:08
guess this is turning into a #perl6-dev discussion
moritz masak: maybe try a "git clean -xdf" in your rakudo git repo (make sure there are no interesting, untracked files in there)
no, I don't remember such trouble, though I haven't built rakudo all that often myself
Xliff moritz: All good suggestions! I've updated rakudobrew! =) 09:09
masak oh, good idea
09:09 pierre_ joined
RabidGravy m: use NativeCall; class Foo is repr("CStruct") { has CArray[Foo] $.children; } 09:09
camelia rakudo-moar 2c4506: OUTPUT«5===SORRY!5=== Error while compiling /tmp/MDMmgvD9m9␤An exception occurred while parameterizing CArray␤at /tmp/MDMmgvD9m9:1␤Exception details:␤ Cannot call infix:<===>(Foo, Str); none of these signatures match:␤ ($?)␤ (\a, \b)␤…»
moritz m: use NativeCall; class Foo is repr("CStruct") { has $.children } 09:10
camelia rakudo-moar 2c4506: OUTPUT«5===SORRY!5=== Error while compiling /tmp/aAm4PYMaLx␤CStruct representation only handles int, num, CArray, CPointer, CStruct, CPPStruct and CUnion␤at /tmp/aAm4PYMaLx:1␤»
RabidGravy leaving aside the LTAness of what happens, someone came up with a work around for that the other day and I can't remember what it was
moritz m: use NativeCall; class Foo is repr("CStruct") { has CArray[int] $.children }
camelia ( no output )
moritz RabidGravy: inheriting from Any as well?
RabidGravy yep that was the one :)
moritz m: use NativeCall; class Foo is repr("CStruct") is Any { has CArray[Foo] $.children; }
camelia ( no output )
RabidGravy I knew it was simple but I just couldn't remember 09:11
Xliff irclog.perlgeek.de/perl6/2016-05-16#i_12487930 09:12
Dammit, moritz! Stop being so fast!
=)
09:12 gregf_ joined
masak no, please keep it up 09:13
Xliff Heh!
masak moritz: good news: `git clean -xdf` fixed the problem in HEAD, too
moritz: am I wrong in thinking it's... unsatisfactory that a target called *realclean* doesn't really clean things enough, but `git clean -xdf` does? 09:14
09:14 TEttinger left
moritz masak: yes 09:15
Xliff Maybe "realclean" should test for .git and add that command?
moritz masak: so the next time this happens, please do a "make realclean" first, then a "git clean -xdf", and note which files weren't deleted 09:16
Xliff: noooo
09:16 bjz left
moritz Xliff: people occasionally write small test scripts etc. in the root directory, and "make realclean" shouldn't delete them 09:16
Xliff Oooh.
moritz Xliff: it should only delete files generated during the build
masak moritz: will do :)
Xliff Well, that's my new thing learned for the day. 09:17
Of course, something tells me that after I've done "git clean -xdf" I might need to reinstall all of my modules.
masak Xliff: besides which, it doesn't sound healthy that the build toolchain have a dependency on Git, at least not for that target.
Xliff masak: You have a point.
masak for doing releases, it makes sense, I guess. 09:18
moritz masak: I guess the biggest difference is that "make realclean" doesn't delete install/
Xliff Yup. Works. So I wonder what "git clean -xdf" did to fix the problem?
masak moritz: TIL.
moritz: any good reason for that? 09:19
Xliff And yes, it's a clean rakudo installation.
moritz masak: that's not how things are done :-)
masak: if rakudo is installed into /, should "make realclean" delete / as well?
Xliff Shit.
Shell::Command failed.
masak moritz: n... no. :)
moritz "install" is just a default :-) 09:20
shadowpaste "xliff" at 217.168.150.38 pasted "Rakudobrew build fail" (63 lines) at fpaste.scsys.co.uk/517602
masak moritz: I'm glad I asked. ;)
moritz :-) 09:22
Xliff: looks like that happens during the panda bootstrap... so rakudo already installed
Xliff No. The binaries did not get installed 09:23
moritz huh
no wonder the panda bootstrap failed.
Xliff Wonder why the regression, there. 09:24
error: pathspec 'nom' did not match any file(s) known to git. 09:25
panda for nom not found
moritz Xliff: what command did you launch?
Xliff "rakudobrew build moar"
Which has always worked up until now.
moritz sounds legit 09:26
RabidGravy it's been doing that for a while, but carries on to work anyway 09:28
Xliff And now "rakudobrew build panda" doesn't work. 09:29
(cd ~/.rakudobrew/bin; for a in ../moar-nom/install/bin/*; do ln -s $a; done) <-- fixed rakudo but not panda 09:30
09:32 _mg_ left
Xliff Well, now that the links are in place, let's see if "rakudobrew build moar" fixes the panda install. 09:33
09:33 _mg_ joined
Xliff Yup. That fixed it. 09:35
09:36 skingsbu joined 09:38 labster joined 09:43 pierre_ left 09:46 gnull left 09:49 iH2O joined 09:50 brrt left 09:55 pierre_ joined 09:59 cpage_ joined 10:11 pierre_ left
Xliff masak: So far I have determined that a hash of 15k items works, and a hash of 20k items doesn't. Trying to narrow it down further. 10:16
However, I am wondering if it is more size of the represented data, than the number of key/value pairs.
17500 items fails. 10:17
16000 items succeeds. 10:21
10:21 pierre_ joined
RabidGravy is reminded why he set out to write a new internet radio station software in Perl 6 in the first place 10:25
Xliff RabidGravy: And what was the reminder? 10:27
RabidGravy Airtime giving me one of those periodic WTF moments
iH2O :( 10:28
i say perl6 is better than LISP for artificial intelligence
RabidGravy it uses ID3 tags in the audio files for its own purposes, including those created by the show recorder, except if you download one of the recorded shows and then try to re-upload it for for some reason it craps out because it can't parse its own tags :( 10:29
sjn Xliff: how about 16383 and 16384 (2^14) 10:31
Xliff sjn: Good idea. I'll cue those up. 10:34
Trying 16200 now.
But that seems to be lining up with what I am seeing. 10:35
Takes rakudo 3 minutes to precompile the module. That adds up.
10:36 ocbtec left
Xliff sjn: Why would static hash definitions be limited, though? 10:36
jnthn The magic number will be around the 16-bit unsigned limit. 10:40
And that's because passing pairs to a hash constructor is argument passing.
Xliff Yeah. 16384
jnthn And there's a limit to how many arguments you may pass.
(I don't expect this to change any time soon.)
Xliff Means I will have to break this up into pieces, then. 10:41
jnthn aye
lunch &
Xliff 16383 fails 10:46
So I'll try one more and if that works, I'll use that as a hard limit. 10:47
16340 works. 10:49
10:51 finanalyst left
masak Xliff: 16384 is the cutoff. 10:52
er. wait. no. 10:53
masak sucks at backloggin'
10:53 aindilis joined
masak and at reading in general. 10:53
10:54 iH2O left
Xliff Nope. So far it's 16340. 16382 failed, so if it's close to a power of 2, it's not that close. 10:56
I'm reading an e-Book while I do this, so it's mindless testing.
masak "close to a power of 2" is still significant 10:59
if there's a header to the data structure which takes up some constant space
10:59 brrt joined
RabidGravy I just had a moment of struggle switching to my inner "python mode" 11:01
masak ...in order to speak parseltongue with one of your evil minions?
hahainternet is interviewing for a python consultant position today i think 11:02
11:02 cpage_ left, kaare_ joined 11:03 tbrowder joined
tbrowder good morning p6 people! 11:03
11:03 cpage_ joined
RabidGravy it's the "indent has to be the same kind of whitespace" thing that gets me every time 11:04
hahainternet my favourite python thing is "do not align assignments"
masak tbrowder: top of the localtime to you!
hahainternet why? well just because ok, you don't think the people who wrote PEP8 are 20x smarter than you? :D
then i used go and gofmt aligns things horizontally for you, i fell in love
masak hahainternet: if you feel strongly about it, write a Python formatting tool that is configured the way you want :) 11:05
tbrowder continuing my problems with p6 pod: I have scrubbed a bad pod file down to I think a minimum, and used the suggested method to test it
masak can't be that hard, Python doesn't have much syntax :P
hahainternet masak: :grin:
RabidGravy I dunno, this code here the assignments are all aligned neatly
hahainternet masak: but it's more the fact it goes against the 'official' formatting guidelines
masak tbrowder: sounds like great news. do you have a gist for us?
hahainternet RabidGravy: non pep8 compliant code!!!!! (moer exclamations required!)
masak hahainternet: who cares? honestly. if you and your team are happy with your own guidelines. 11:06
(and feel you have good reasons to uphold them)
hahainternet masak: i absolutely agree, i'm just parodying the python community's approach :)
tbrowder I'll put it in a gist shortly, but besides the bad pod, the worst part is perl6's response to it leaves no cluse as to what the problem is
hahainternet i'll stop noising up #perl6
RabidGravy well to be fair this particular application (Airtime) is an evil mix of python and PHP, I'm just amazed it works at all half the time ;-)
masak hahainternet: I don't necessarily want to defend the "only one way" mentality... but I can see how its opposite "anything goes" worldview can sometimes lead to trouble, too 11:07
hahainternet masak: it's why i fell in love with gofmt
no more arguing
Xliff masak: 16360 passes. I think I'll stop there.
tbrowder for example, executing 'perl6 bad.pod' yields exactly "===SORRY!=== Cannot iterate object with P6opaque representation" 11:08
hahainternet an entirely unrelated question: why is posix_memalign the only libc malloc class function without an alternate symbol name :/
11:13 ocbtec joined 11:15 skingsbu left
masak tbrowder: yes, that's an unacceptable error message. I agree. 11:15
tbrowder the gist with the short good and bad Perl 6 pod examples is here: gist.github.com/tbrowder/23c5ef38c...a2ca61659d 11:16
masak nice. short. 11:17
problem reproduced.
masak submits rakudobug
11:23 finanalyst joined
AlexDaniel “It seems like you forgot to attach a file. You wrote "is attached" in your message, but there are no files attached. Send anyway?” – gmail 11:24
yoleaux 03:28Z <MasterDuke> AlexDaniel: PR 775 addresses the first part of RT #128214 by failing when a path part doesn't exist, instead of continuing, but not resolving. the second part wasn't really a problem with resolve, you were just trying to '.dir' on a non-existent path
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128214
11:24 Amnez777 left
Xliff Damn. Splitting the data into 2 hashes still hits that "Frame 2 local access out of range" error. 11:25
AlexDaniel MasterDuke: in other words, ‘../foo/..’ is not going to resolve?
11:26 labster left
masak tbrowder: rt.perl.org/Ticket/Display.html?id=128221 11:29
tbrowder masak: thanks! can you please take a look at this gist for an htmlify.p6 problem: gist.github.com/tbrowder/b83da3fa5...c75747af54 11:33
masak takes a look 11:34
11:34 kid51 joined
masak tbrowder: a variable called $source-name on line 171 of the file htmlify.p6.latest is undefined 11:35
tbrowder: I could delve deeper and find out why, but I'm... kinda at $work
11:36 wamba left 11:38 adu joined
tbrowder masak: sorry--didn't mean to interrupt--I have tried to find the variable but no luck so far--no biggie--seems not to cause a problem, though 11:38
it would be nice to get a better error message to help pin-point it like using carp or cluck 11:39
11:44 kid51 left
Xliff masak: Attempting to split the hash into 2 modules. 11:44
That worked. 11:45
This is a piss poor way to do things, though. Really I may end up using SQLite to store this data. 11:46
11:46 cpage_ left
RabidGravy just keys and plain values? 11:47
I think there is some working hash file storage thingy in the ecosystem 11:48
Xliff One is keys and values, another is keys and records
Xliff checks Ecosystem. 11:49
11:49 telex left
RabidGravy my work on gdbm is still in a state of indefinitely postponed 11:49
11:50 telex joined 11:51 cpage_ joined
Xliff Yeah. That's the next step. SQLite databases should be easier to distribute and use rather than this Frankenstein thing I'm currently dealing with. 11:53
stmuk another thing that looked horrid to nativecall was Tk :) 11:54
Xliff Can DBIish create SQLite files/databases? 11:55
Hrrmn. "touch file.db" seems to be the easiest thing to do. 11:57
masak Xliff: pretty sure I've been creating a SQLite database using DBIish. 11:58
at some point
tbrowder: no worries. just explaining why I can't be of more assistance right now.
tbrowder: what is it about the error message that you don't consider clear? it gives you the name of the variable, the cause of the warning, the line and the file. 11:59
tbrowder: granted, there'll be more things causing the variable being undefined, but you'd basically need to start debugging and/or tracing the data flow backwards in the program in order to find that out. which is Hard and not something I'd expect the compiler to do for me. 12:00
tbrowder masak: yes, that's the starting point, and I intend to track it down, but haven't looked outside the doc repo yet--I suspect it's in rakudo
12:01 pierre_ left 12:02 |2701 joined
masak tbrowder: disconfirmed. `find -name htmlify\*` turns up nothing in the rakudo repo. 12:02
dalek pan style="color: #395be5">perl6-examples: 452df0d | (Steve Mynott)++ | categories/ (4 files):
link to doc.perl.org not spec/design
12:04
tbrowder masak: I think the key is finding variable "$source-name"...I'll search now that I'm back on a real computer 12:09
masak it should be on that line in that file :) 12:10
tbrowder but it's not--I remember I tracked it down to an apparent problem with CompUnit and sure enough it's there in rakudo--I'll keep checking--thanks for waking me up 12:12
masak tbrowder: a `find` on perl6-all-modules digs up this file: github.com/perl6/doc/blob/master/htmlify.p6 12:16
moritz++
but... plot thickens... no variable `$source-name` in that file
however, as you imply, this variable occurs in both src/core/CompUnit/PrecompilationRepository.pm and src/core/CompUnit/Repository/Installation.pm 12:17
in the former, even on line 171
tbrowder masak: exactly--it's over in rakudo in ./src/core/CompUnit/PrecompilationRepository.pm and ./src/core/CompUnit/Repository/Installation.pm
masak yes. 12:18
tbrowder masak: I'll do some more debugging on that later today... 12:23
12:27 adu left 12:32 huggable joined 12:35 Amnez777 joined
dalek k-simple: d1fcc6e | azawawi++ | examples/0 (16 files):
Replace underscores with dashes and .pl6 as an extension
12:49
12:53 _4d47 joined 12:58 skids joined 13:01 pierre_ joined 13:02 cdg joined 13:03 cdg left 13:04 cdg joined 13:09 pierre_ left
Xliff What does this warning mean: "Asked to remove 20 spaces, but the shortest indent is 16 spaces in any trim_heredoc at gen/moar/m-Perl6-Actions.nqp line 451" 13:12
And how can I get rid of it?
timotimo it means you have a heredoc where the terminator is further to the right than the shortest line is 13:13
hm. it really could do with reporting a line number if it can
Xliff Yeah, and I really didn't have any heredocs where the terminator was further right. 13:14
But I unindented the terminator and the error went away.
Thanks, ("timo" xx 2) 13:15
timotimo did you have tabs in there? 13:17
by default we interpret tabs to be 8 spaces, unless you set $*TABWIDTH or what it is
Xliff Ah... that might be it. 13:18
Now... to see if DBIish will want to load the entire .sqlite3 file before it will run queries. 13:19
Which would be bad.
timotimo it doesn't do anything to the file, it just uses the sqlite library
moritz right 13:20
and I guess sqlite3 doens't do that
timotimo it probably mmaps, which is much cheaper
13:20 Sgeo left
Xliff Excellent! 13:20
However inserting is slow.
RabidGravy yeah it does
Xliff Might be coz index.
moritz also, sqlite isn't designed for heavy, concurrent writes 13:21
that simply requires a server model
Xliff This isn't a concurrent write use-case.
masak haven't thought about it that way, but sqlite is kind of interesting *because* the tradeoffs it makes translates to fairly common use cases 13:24
RabidGravy you can turn journalling off with a pragma if you don't think you need it
masak for example, it's excellent on mobile for local storage 13:25
RabidGravy also in a non concurrent bulk write kind of situation setting the locking mode to exclusive may speed it up 13:26
Xliff masak: That's why I thought of it for this project.
RabidGravy as it won't bother with row locks
13:26 ocbtec left
Xliff And once the database is written, it will be read-only, so turning off journaling after the data is loaded would be god. 13:27
s/god/good/
masak right
Xliff Also, are primary keys auto-indexed for SQLite files?
masak there's probably a fancy term for "the low end of use cases that don't require the full ACID hoopla"
hoelzro Xliff: yes 13:28
RabidGravy many and varied pragma for tuning www.sqlite.org/pragma.html
hoelzro the primary key is the clustering key for SQLite, iirc 13:29
13:30 cpage_ left, cdg left
Xliff Cool. Now I think I have this thing working. 13:32
Thanks, everyone.
13:33 ocbtec joined
Xliff Now the next thing to test is if multiple tables in a single file will degrade performance. 13:34
RabidGravy I've run quite large applications with tens of tables on sqlite for testing purposes, it's not great but it not entirely un-usable 13:37
Xliff Cool.
13:37 wamba joined
Xliff I have tons of data that I basically need name -> ID and ID -> Name mappings. 13:38
And this needs to be in a distribution so I don't have the luxury of an RDBMs 13:39
13:41 domidumont left 13:47 _4d47 left 13:50 adu joined 13:56 khw joined
RabidGravy is there something in the ecosystem that already defines the time.h stuff from POSIX? 14:00
annoying USB library uses a timeval
moritz it seems the videos from GPW2016 are being uploaded to youtube: www.youtube.com/watch?v=D1YxXfqP1K...e=youtu.be 14:05
tadzik sweet 14:07
moritz ... and the introductions are cut off :(
El_Che what time slots are available for talks at yapc::eu? (No info without login in)
tadzik seeing masak's name asciified to Maesak, I can't wait to see what mine will be :o if the lighning talks get uploaded, at least
15, 30 and 60 I think? 14:08
moritz yes
and lightning, I hope :-)
El_Che thx
tadzik El_Che: Maesak
erm
imgur.com/N53P4UK
that
[Coke] mst: did you write down the desired redirect channel so the next time someone asks a perl5 question, we can safely redirect them?
tadzik so 90 too
14:08 sufrostico joined 14:09 lizmat joined
Xliff What outputs to STDOUT without automatically adding a newline? 14:17
moritz print
[Coke] is github slow today or just me?
Xliff Thanks.
mst [Coke]: eh? there's already a perl5 channel on this network - #perl - which normally people send them to. you're the first person to ignore the obvious answer and do something inexplicable instead, so I dunno where I should write down "don't be [Coke]" :D 14:23
yoleaux 22 May 2016 20:05Z <moritz> mst: to get that blog post out
[Coke] mst: I am not the first person to do that. there's no need to be rude.
14:23 itaipu joined
mst [Coke]: first person I've seen in here do that. I don't know what's rude about the truth. 14:24
[Coke] "don't be [Coke]". 14:25
gfldex mst: so the truth is »"don't be [Coke]"« 14:26
lizmat
.oO( don't be mst )
[Coke] OK. We'll keep not writing it down. I'll try to remember the correct answer next time.
masak print 14:27
mst [Coke]: if you think there *is* a reason to send somebody to another IRC network, irc.perl.org's website has a channel list on it www.irc.perl.org/channels.html
masak heh, stale backlog :P
mst lizmat: that's also very good advice.
masak otherwise I'd've beat moritz to the punch, I'm sure :P
moritz punches to the beat 14:28
14:28 mojca1 joined
mojca1 [Coke]: I tried to update MoarVM, nqp and rakudo in MP, but all tests from t/01-sanity/53-transpose.t on fail 14:28
mst [Coke]: ok, look, I have no idea why, on a network with a very long established #perl channel, you decided to send people to an entire other network instead, and therefore I'm not sure what to write down or where to try and avoid somebody else doing that. if you want to sulk instead of trying to solve the problem, your choice, but that was a genuine statement, not an insult 14:29
[Coke] mst: Historically, that was the channel that we sent people to. It's just inertia on my part, that's all.
masak I don't mind my last name Mäsak being ASCIIfied to Maesak. but on IRC I greatly prefer just "masak".
mst [Coke]: wut
mojca1 I get "Dubious, test returned 1 (wstat 256, 0x100)", I don't know what that means
mst [Coke]: #perl6 has always sent people to #perl on here and it's fine
[Coke] mst: "always" is very strong and not true.
mst [Coke]: you, OTOH, tried to send them to #perl on irc.perl.org which is completely different, not something I've seen in here before, and wrong
[Coke] I'm probably the only dinosaur left. That's fine, I'll change my stripes. 14:30
mst ok, well, I don't recall 'irc.perl.org #perl' receiving anybody from here
err, up until a few years ago that would've been even more wrong than it is now. I had to play a flamethrower over quite a few of the old regulars to make irc.perl.org #perl habitable ;)
moritz that's why we sent them to #perl-help or something 14:31
or used to, at least
mst #perl on here or #perl-help over there are both fine. it was [Coke] trying to send them to #perl over there that I went "eep" at, because I hadn't seen that before and it isn't a good idea
masak mst: fwiw, I've been known to send people to #perl over on irc.perl.org
14:32 JimmyZ joined
[Coke] mst: you went quite more than eep. 14:32
mst masak: well, you can [Coke] between you can figure out where to write down "don't do that" then
mojca1 [Coke], are you willing to look into test errors on Mac and report them to a suitable place?
[Coke] mojca1: for what now?
mst [Coke]: yeah, well, I don't like throwing newbies into shark tanks and I expect regulars to prefer not shitting on newbies too
masak mst: I don't know much about either channel, but my impression is that the #perl over at irc.perl.org is higher-traffic and higher-quality in some sense
mst masak: the irc.perl.org #perl is a social channel and not somewhere to send people with tech questions
mojca1 [Coke], I'm not sure, I get a zillion "Dubious, test returned 1 (wstat 256, 0x100)" when running "sudo port -v test rakudo" for 2016.05 14:33
[Coke] masak: I'm happy to defer to mst about which perl5 stuff goes where.
mst irc.perl.org itself documents #perl-help but not #perl for a reason
but obviously nobody reads the website for the network
so I'm still not sure where to write this down
masak mst: got it. that does sound familiar.
I also remember something about #metallica :)
mst [Coke]: I'm not refusing to write this down, like you keep pretending
I just don't know where, since evidently you didn't read the existing docs 14:34
masak is it just me, or is it a little tense in here?
mst the question is, where *would* you look, and how do we patch that?
moritz masak: it is
mojca1 [Coke]: but I hope you will manage to figure out the details; I'm not a perl6 user myself (hardly even a perl5 user)
mst if somebody tells me that, I'm happy to help. otherwise, not a lot I can do, sorry
masak looks around for people who might need hugs 14:35
dalek c: 6db1d03 | titsuki++ | doc/Type/IO/Special.pod:
Fix a typo in the IO::Special page
c: 97ef172 | titsuki++ | doc/Type/IO/Special.pod:
Merge pull request #538 from titsuki/fix-typo

Fix a typo in the IO::Special page
14:39 ptolemarch joined
lizmat afk again& 14:41
14:41 lizmat left
perlpilot masak: who doesn't need a hug? :) 14:45
mst perlpilot: hugs are not necessary to human existence and you can perfectly well survive without them.
they are, however, *awesome*, and I am always sad when somebody turns one down 14:46
masak TIL :)
the whole tension thing was worth it, just for me learning that :P
gfldex i do agree on "survive" but must reject "perfectly well".
moritz wonders if a newborn can actually grow up without any hugs 14:47
perlpilot mst: There's "survival" and then there's "living" ... I tend to favor the latter
masak moritz: I think so, but the results are not so good.
perlpilot moritz: I'm willing to bet that someone somewhere did such an experiment.
masak moritz: I remember some science on monkeys and a barbed-wire "mother"
damn, now I'm sad too :(
tadzik damn 14:48
moritz masak: well, there was a famous experiment where children were raised, and nobody was allowed to speak to them, to see what language was the "original" language of mankind
they all died
masak stahp
perlpilot masak: so ... now we're back to "who doesn't need a hug?" :-)
masak times five hundred 14:49
masak .oO( our science conclusively shows the original language to be... DEATH )
masak .oO( yay! science! )
nemo en.wikipedia.org/wiki/Language_dep...In_history 14:50
moritz right, the fource paragraph is what I was refering to
*fourth
third, actually. Counting is hard 14:51
moritz hugs masak
masak ahhh.
perlpilot methinks that hug may have gone on a little too long ;> 14:52
masak perlpilot: clearly you need more high-quality hugs :)
masak hugs perlpilot
moritz hugs [Coke] and mst, for balance 14:53
perlpilot Just so you know ... I'm smiling right now IRL :-)
[Coke] moritz: Ow.
masak perlpilot: haven't let go of you yet :P
14:53 s_kilk joined
moritz hugs [Coke] 14:53
masak hugs [Coke]
perlpilot hugs [Coke] 14:54
mst: your turn.
masak hug REQUIREMENT!
there's a lady in our neighborhood who basically goes around hugging people she meets.
[Coke] attempts to not actually visualize the hugs scrunching up his wounded rotator cuff.
masak as far as I can tell, that's her dayjob.
perlpilot [Coke]: these are gentle hugs
mst HUGPILE 14:55
mst glomps everybody
masak "he's not from here..." :P
perlpilot I'm not normally ebullient, but it's always nice to keep one of those people near by.
masak .oO( they come to our channel, take our hugs ) 14:56
moritz
.oO( they love us for your hugdom )
erm, hate
whatever :-)
masak moritz: LoVe, HAte :)
moritz hate is love! war is is peace! and all of that
14:57 brrt left
masak perlpilot: I can imagine some people are bothered by the hugs the Hug Lady hands out, or don't want to play sometimes. 14:57
me, I think it's super cool. I bet her hugs with me are the longest :)
she's like a walking Free Hugs service
[Coke] I'm of two minds, one of which is "please don't touch me." 14:58
masak well. the first time I was... surprised. 14:59
mst oh, you're talking about somebody local to you, not woolfy 15:00
ilmari masak: does she impose the hugs on people who don't want them?
masak mst: haha
masak: I've sometimes seen her request hugs from some of the children in our yard, and get a "no". she's a bit insistent, but eventually she backs off. 15:01
er, ilmari*
15:02 luiz_lha joined
masak ilmari: my wife and I have also decided to each hug her, but to discreetly steer our 17mo son around and behind her when we're out together. that tends to work. 15:02
15:02 luiz_lha is now known as Guest41960
ilmari fair enough 15:02
15:03 Guest41960 is now known as luiz_lha 15:08 domidumont joined 15:10 mojca1 left
perlpilot masak: hugs are how she brainwashes you before the rest of her species arrive on planet for the invasion. 15:11
the upside is that you probably live longer and are healthier, so it's a win-win. 15:12
masak no argument there
I for one welcome our new Hug Lady overlords 15:13
15:14 s_kilk left 15:21 |2701 left 15:23 raiph left, domidumont1 joined 15:25 Actualeyes left 15:27 domidumont left 15:36 xue_ joined 15:38 domidumont1 left, Actualeyes joined, finanalyst left, domidumont joined 15:39 ssotka joined 15:45 Zoffix joined
Zoffix stmuk, I think your blog aggregator config is missing a hyphen here: github.com/stmuk/pl6anet.org/blob/...anetrc#L71 In other news, I've sent you a PR to have my blogs point to Perl6.Party and not blogs.perl.org: github.com/stmuk/pl6anet.org/pull/4 15:48
stmuk: ^ (just recalled you have a crappy client that doesn't highlight on commas 😜)
15:49 _mg_ left
stmuk actually I changed to weechat now! 15:54
15:55 xue_ left
stmuk hmm I seem to be having GH problems with sessions 15:57
I'll merge when I can actually get it to work!
16:02 kurahaupo joined 16:12 itaipu left 16:14 itaipu joined
kaare_ How come that the Anguish blog post is always on top at perl6.org/ ? 16:15
Even though there are newer posts, even from Zoffix himself 16:16
mst because programming is all about angst, anguish, existentialism and hubrisism 16:17
kaare_ and beer
Zoffix :/ is China DoSing GitHub again? 16:19
diakopter or Mars 16:20
kaare_: maybe it goes by modifica date
Zoffix Mars? 16:21
diakopter the planet
Zoffix Ahhh... diakopter++
Probably. I fixed a typo recently and it was after my lastest post.
diakopter kind of like google's date-filtered search results trust the reported date instead of when it actually first appeared 16:23
(I've seen search results claiming to be from the 1970s)
when Google first indexed it is usually more relevant to me 16:24
kurahaupo believe it or not, computers existed before the internet. And history goes back even further 16:27
diakopter I mean, the dates were obviously wrong, and from the 1970s 16:29
perlpilot or ... someone in google invented a time machine 16:30
kurahaupo doesn't see a problem with a document that was created in 1840 having that as its mtime when it's transcribed onto a computer, and the webserver serves that date
pochi I need to pop up a window where the user can draw strokes with the mouse. Is there a module that can help me with that (I tried the module for SDL, but it seems to be broken) 16:31
kurahaupo (except that 32bit epoch seconds dont go back that far)
kaare_ Nope, time began August 4, 1997, when Cyberdyne switched on Skynet
16:32 ptolemarch left
kurahaupo diakopter: there are unix systems with real files whose timestamps are in the 1970s; just not very many. 16:32
16:32 huggable left 16:33 sno left
kurahaupo everyone knows that Cyberdyne is a division of Alphabet Inc. 16:33
diakopter kurahaupo: I appreciate your help, but I already know all of that, and yet I still said what I said. I'll clarify again. The documents were created after the 70s. 16:34
kurahaupo Just it's taking Skynet a bit longer than planned to become selfaware
diakopter: how is Google to know which are the true and bogus timestamps? 16:35
diakopter I wasn't blaming google, FFS
let alone complaining 16:36
merely commenting
jfc
kurahaupo ok
16:37 dakkar left 16:40 Util joined 16:42 ggoebel114 joined
dalek c: af876ce | titsuki++ | doc/Type/Routine.pod:
Add missing semicolons
16:47
c: 4d200a8 | titsuki++ | doc/Type/Routine.pod:
Merge pull request #539 from titsuki/fix-yada

Add missing semicolons
16:53 acrussell joined 16:55 abraxxa left 16:58 wamba left, wamba1 joined 17:07 spider-mario joined 17:10 ptolemarch joined, _mg_ joined 17:11 adu left, araujo_ joined 17:15 araujo left 17:18 dha joined
dogbert17 wonders if dalek has kicked the bucket 17:23
17:24 Amnez777 left
dha Daleks *always* come back. 17:25
17:27 araujo_ left
jdv79 are they here? 17:27
dha: is there a meeting anytime soon? jim thought not. 17:28
dha Oh hell. I meant to call one for wednesday. OSCON has messed my mind up.
Think it's too late to do it?
jdv79 is wed the normal day? why not thurs? 17:29
17:29 Amnez777 joined
jdv79 it might be a bit late given the holiday weekend... 17:30
dha yeah, that's why we were talking about wednesday.
Maybe I should make it for wednesday or thursday next week. 17:34
Zoffix It's a long weekend in US too? 17:37
jdv79 sounds better
is memorial day on monday
dalek c: d0e337d | (Jan-Olof Hendig)++ | doc/Type/Parameter.pod:
Fixed typo and added a missing C<>
jdv79 *its 17:38
Zoffix Oh. I thought it was about today....
Victoria Day here in Canukistan.
dha Ah. 17:40
RabidGravy I've just blinded myself with a raspberry pi 17:41
Zoffix I hope not permanently :)
dha We're short on Queens here, so we have to use regular words for our holidays. :-)
RabidGravy - congratulations? 17:42
Also, alert Thomas Dolby!
RabidGravy well in doing so I've proved that it's software not hardware that is being weird
shop.pimoroni.com/collections/rasp...cts/piglow to be specific, really, really bright 17:44
17:50 acrussell left 18:01 labster joined 18:04 Actualeyes left, sno joined
nine llfourn: "This precompilation is complicated stuff!" Hey, I arrived at the same conclusion :) 18:07
18:10 jevin joined 18:14 spider-mario left 18:17 _mg_ left 18:30 adu joined 18:32 dwarring joined 18:37 spider-mario joined 18:39 mojca joined 18:40 mojca left, spider-mario left 18:42 AlexDani` joined 18:44 AlexDaniel left 18:46 AlexDani` left 18:53 mojca joined
dalek c: de581cb | (Jan-Olof Hendig)++ | doc/Language/5to6-perlfunc.pod:
Fixed three broken links
18:55
18:55 cpage_ joined 18:57 cpage_ left 18:58 cpage_ joined 18:59 cpage_ left 19:00 jjido joined, domidumont1 joined 19:01 astj left 19:02 domidumont left 19:06 jjido left 19:09 wamba1 left 19:13 spider-mario joined 19:16 wamba joined 19:18 domidumont1 left 19:30 mojca left 19:39 CIAvash left, rindolf left 19:44 rindolf joined
dogbert17 hello :), does anyone know if the descripton of the method 'isa' in S32 is correct? design.perl6.org/S32/Basics.html#Any 19:47
19:47 hahnara joined
timotimo hm, i think isa also takes strings for types 19:47
dogbert17 timitimo: and it seems to have moved to Mu as well
timotimo also, it looks like it got damaged in editing or something
the design docs aren't really kept up to date any more :S 19:48
dogbert17 I was thinking of 'stealing' the text and update Mu.pod but I don't want to add nonsense :)
dha timotimo - that's... unfortunate.
19:48 hahnara left
dha Especially since I still hear people saying things like "I couldn't find that in the user docs, so I went to the design docs". 19:49
timotimo we're a bit low on people-power
dogbert17 timotimo: should I add the text it to Mu.pod? 19:50
s/it//
dha Ok. Is there a way to fix that? I would certainly be willing to try to put in some time if I had a clear idea of what was needed.
(Note: I ask that with the knowledge that the answer to my question may well be "no") 19:51
timotimo sorry, i'm not confident in deciding whether the design docs should follow what rakudo does or whether rakudo should change to do what the design docs say or whether it's okay to have them diverge
dha That is certainly a reasonable question. So I guess I move on to the question "is there a way we can resolve that?" (again, realizing there may not be a good answer to that either) 19:52
RabidGravy surely, if the design docs say the thing that is not true as regards implemented and tested behaviour then they're wrong
mst perhaps the answer in that case is to prepare a patch to the design docs, make sure it's hunked so each difference with rakudo is separate, then get jnthn/larry/whoever to review? 19:53
Zoffix Wasn't there a decision that Roast is the spec and the docs are the human version of that spec. The `design` is outdated and there's no real point in trying to keep it up to date. Hence why we call the specs "speculations" and not "specifications" now.
RabidGravy but, the actual docs should be the tested behaviour
[Coke] the design docs are more historical at this point. 19:54
stmuk the tests are Golden Source of Truth
[Coke] Add a note to this effect at the top of each one, point people at docs.perl6.org, and update things there instead.
RabidGravy yeah
mst yes, hence my thought that you prepare updates to the relevant docs to bring them in line with the tests, but then get somebody to double check
seems like generally the updates will be correct, but belt and braces 19:55
dogbert17 so you're ok with me adding 'isa' to Mu.pod?
RabidGravy I am
dogbert17 then it will be done :) 19:56
perlpilot dogbert17: I'm in favor of motion over stagnation :)
dogbert17 perlpilot: agreed :)
RabidGravy though maybe noting that a smart match against a type object is probably "better"
dha I understand the idea the the design docs can be considered historical at this point, but, from a user perspective, since the user docs are often lacking (not a criticism, just how thing are, due to personpower, etc.), that's where people go looking for info. 19:57
Until that is not the case, the out-of-date-ness seems problematic. to a new user, in any case.
dogbert17 I'll whip something up and get back to you all in a bit
dha I am, of course, wholly in favor of the user docs being exhaustive to avoid this problem. ;-) 19:58
RabidGravy well fixing the design docs to fix it seems silly when it's a similar effort to write proper documents
mst there is always the option of "both" 19:59
[Coke] dha: of course it's problematic.
I agree with you completely there.
mst: we don't have time to do it once, let alone twice. 20:00
so it's not really an option, IMO.
mspo there should be a way to annotate sources to satisfy the "spec", generating docs?
but keeping two things in sync won't ever work 20:01
[Coke] awwaiid has some plans in that direction.
mst [Coke]: I'm more thinking "it may be that in some cases tweaking the spec now is doable and writing the docs isn't yet, so doing that first and filling the docs out later may be the least worst optioN"
dha Ok, so we agree on there being a problem. Should we have a conversation about how to solve the problem? Is there some specific group of people who should have that conversation? (honestly not being snarky here. Just trying to see if there's some way to fix this.)
RabidGravy feel free etc
Zoffix Then you have source mixed with docs and the potential range of contributors much smaller.
mst I leave decisions as to that up to the people actually doing the work, just making sure we look at the permutations
dha And, yes, I agree that if we have to choose between fixing docs and fixing spec, docs should take precedence. But, if docs can't be fixed, maybe a minor tweak to the spec would be helpful. 20:02
Zoffix Personally, I don't agree that there's a problem. We have a large language and few people to document it. It's a work in progress. Just because someone realized the docs aren't complete and went to read some sort of a technical document which is out of date doesn't mean there's a problem with anything.
RabidGravy the only thing preventing the docs being improved is people time 20:03
dha Zoffix - Then maybe there should be an explicit statement that those documents are not to be fully trusted?
20:04 cpage_ joined
Zoffix All the specs, as Coke suggested above. 20:04
[Coke] I think adding a warning to the design docs is in order, esp. if it points people at where to open a github issue saying what they were looking for and couldn't find. In terms of getting docs written, I'd encourage volunteers to write for docs.perl6.org before fixing up the design docs. I don't know if there is more of a plan to be had there.
Zoffix There's actually already text of that kind on design.perl6.org 20:05
Not too obvious, but at least it's there :)
dha Ok, maybe it should be more obvious. :-) 20:06
mst making sure every page on the design docs has it in big letters at the top would help
I've often gone there via direct links
[Coke] added github.com/perl6/specs/issues/110 20:07
dha And having people submit things they were looking for and couldn't find, and have those submissions in an obvious, central place would be tremendously useful
Zoffix Pffftt.... users and their crazy ways of using websites! *shakes head at mst*
[Coke] also added github.com/perl6/doc/issues/541 20:08
dha [Coke]++ 20:09
mst Zoffix: yeah, well, I know I'm unusual for clicking the links people give me when I ask questions and actually reading them, but there's the occasional other person with the same superpower 20:10
20:10 darutoko left
Zoffix Crazy! :) 20:10
20:11 kaare_ left
dalek ecs: 4791822 | (Zoffix Znet)++ | html/style.css:
Add note that specs may be outdated on each page (Closes #110)
20:13
dha Zoffix++
20:14 |2701 joined
dogbert17 ok, here's a draft for method 'isa', how does it look? gist.github.com/dogbert17/6a325518...17a3a0823e 20:14
20:15 mojca joined
[Coke] dogbert17: is the uppercase \SELF used elsewhere? 20:15
dalek ecs: 337e775 | (Zoffix Znet)++ | html/perl.css:
Add outdatedness note to second stylesheet used by the site as well
20:16
dha Silly question. Is there any way to find what's in the test suite that's not documented. Other than actually going through it manually, of course.
dogbert17 [Coke]: in Mu.pm yes
can't say it's a syntax that I'm familiar with :( 20:17
RabidGravy I'd leave the name of the invocant out 20:18
[Coke] ok. looks overly verbose to me, but if there's prior art, then no worries in you using it.
dogbert17 Perhaps Mu:D: instead?
RabidGravy just Mu: 20:19
dogbert17 RabidGravy: thx, will fix
dha dogbert17 - the second part of the signature uses "$name" but you don't mention it in the explanation. Should it also be "$type"? 20:20
RabidGravy Zoffix, you appear to have added a big red banner on the docs as well as the specs
Zoffix :(
dha RabidGravy - is that a bad thing?
dogbert17 dha: I think it should be $type, thx 20:21
RabidGravy well yes
dha What do you have against big red banners?
[Coke] yup, please revert.
dalek ecs: d9fab81 | (Zoffix Znet)++ | html/perl.css:
Revert "Add outdatedness note to second stylesheet used by the site as well"

This reverts commit 337e775a8511ad6cef00924cb9a71a57cc6b7785.
The sheet is used on other sites too.
20:22
mst dha: DANGER: COMMUNISM
masak DANGER DANGER COMRADE WILL ROBINSON
dha You're just afraid of STICKING IT TO THE MAN.
dalek c: b9efb74 | (Tom Browder)++ | doc/Language/variables.pod:
make a working example
20:23
c: fcd329e | (Tom Browder)++ | doc/Language/variables.pod:
Merge pull request #540 from tbrowder/grammar-example-2nd-try

make a working example
RabidGravy dogbert17, actually none of the Mu methods have the invocant documented
dogbert17 RabidGravy: so what should I write then? 20:25
RabidGravy just leave the invocant out
e.g. "method isa($type)"
dogbert17 RabidGravy: fixed, please reaload the page 20:26
s/reaload/reload/
RabidGravy cool 20:27
dogbert17 thx guys, I'll commit it 20:28
RabidGravy do it! 20:29
dogbert17++ # jfdi
20:29 _dolmen_ joined, Ven joined
dalek ecs: 04b3cc6 | (Zoffix Znet)++ | html/perl-with-historical-message.css:
Work around the issue of multiple sites using perl.css
20:31
RabidGravy Zoffix++
20:32 huggable joined
dalek : 72af323 | (Zoffix Znet)++ | util/update-design.perl6.org.sh:
Use the CSS with the outdatedness message included during generation
20:33
c: 1fb27c0 | (Jan-Olof Hendig)++ | doc/Type/Mu.pod:
Added docs for method 'isa' in Mu. #perl6++
dogbert17 if this works we might perhaps be able to close github.com/perl6/doc/issues/274 20:35
dha I lost track. Is the big red banner on doc.perl6.org pages supposed to be gone by now? I only ask because I know it was mentioned somewhat upthread, but it's still there. 20:38
dalek c: ff833e0 | (Steve Mynott)++ | doc/ (3 files):
fix minor typos
dogbert17 dha: impossible to miss :) 20:40
dha Oh, you'd be surprised what I can miss. :-)
Zoffix dha, it'll update in ~30 minutes on the next run of cronjob 20:42
20:47 jjido_ joined 20:48 itaipu left 20:49 dha left 20:51 dha joined
dalek c: 47a5910 | (Jan-Olof Hendig)++ | doc/Language/objects.pod:
Another broken link fixed
20:53
RabidGravy red thing gone :)
Zoffix Looks like someone needs to connect to hack and update this script: github.com/perl6/mu/blob/master/ut...rl6.org.sh
20:54 pmurias joined
Zoffix Note to self: always have a jug of water for emergencies... I just brewed a cup of coffee using water from my steamer :( 20:54
moritz Zoffix: why not just update it in the repo?
oh, it doesn't pull itself? 20:55
pull'd
Zoffix moritz++
dha Yes. no more red thing. Yay. 20:56
20:57 AlexDani` joined, AlexDani` is now known as AlexDaniel
dha Incidentally, looking at notifications, I see there's a mention that qqw is searchable, but qqww isn't. I think the fact that it's not actually in the docs might explain that. What the heck is qqww supposed to do, anyway? 20:57
Ah, synchronicity. 20:58
Zoffix What's that?
dha What's what? qqww or the synchronicity?
20:58 skids left
Zoffix Both! :) 20:59
dha Well, I have no idea what qqww is (see github.com/perl6/doc/issues/533 though) and the synchronicity is that AlexDaniel is the one who submitted that. :-) 21:00
AlexDaniel qqww is «» 21:01
dogbert17 Isn't Synchronicity a Police album?
Zoffix AlexDaniel, that explains nothing to me :)
AlexDaniel design.perl6.org/S02.html#Adverbs_on_quotes 21:02
m: .say for <'hello world' test>
camelia rakudo-moar b348f2: OUTPUT«'hello␤world'␤test␤»
dha Well, the docs seem to say that 􏿽xAB􏿽xBB is qqw. I have no idea what qqww would actually do.
21:02 synopsebot6 left
AlexDaniel m: .say for «'hello world' test» 21:03
camelia rakudo-moar b348f2: OUTPUT«hello world␤test␤»
21:03 synopsebot6 joined
AlexDaniel that's the difference 21:03
Zoffix Ah, AlexDaniel++ Thanks
AlexDaniel dha: rt.perl.org/Ticket/Display.html?id=128177
dha ah. didn't see the rt version. just the github. 21:04
And I don't know if you've seen it, but I commented on the github issue. 21:05
specifically that qqww isn't documented, so searching for it is probably a lost cause.
AlexDaniel dha: I was addressing your “the docs seems to say that X is qqw” (github.com/perl6/doc/issues/532)
dha: but yes, I've already answered and changed the title :) 21:06
dha Ah. I'm old. it's hard for me to keep up with you kids. :-)
AlexDaniel nah, it's just that I am over energized right now. Had a dead line 5 minutes ago 21:07
dha Ah.
Zoffix - I'm amused that your comment on that issue refers to the design docs, given the red banners and such. :-) 21:08
dalek c: a4576d6 | (Zoffix Znet)++ | doc/Language/objects.pod:
Make term mixin searchable (Closes #542)
21:09
AlexDaniel dha: speculations have a lot of useful information
Zoffix dha, 'outdated' ne 'useless'
AlexDaniel dha: most of it is outdated, but there are some things that we can still steal for the docs :)
dha Oh, of course. I said amused, not horrified. :-)
AlexDaniel Zoffix: “Note: these documents may be out of date.” uh oh!
Zoffix: isn't it a little bit too forgiving? 21:10
dha AlexDaniel - Indeed. The trick is knowing what fits that description.
AlexDaniel these docs are out of date, that's it
21:11 pecastro joined
Zoffix AlexDaniel, bikeshed.com/ 😜 21:11
dha Hm... looks like there should be some discussion of w vs ww in the quoting docs. 21:12
21:13 cdg joined
AlexDaniel dha: yes. That'd be great. Can you write it? 21:13
21:13 Ven left
RabidGravy dha, github.com/perl6/doc/issues/376 - keep forgetting to do it 21:14
dha Looking into that now. Also, unless I'm confused, that design document indicates that w and ww are switched wrt what they're doing now.
unless, I don't understand what "quote protection" means.
AlexDaniel oh, there's one more issue! Great… 21:15
dha Am I confused? Or are they actually reversed now?
AlexDaniel m: .say for «'I am protected'» 21:16
camelia rakudo-moar b348f2: OUTPUT«I am protected␤»
AlexDaniel m: .say for <'And I am not'>
camelia rakudo-moar b348f2: OUTPUT«'And␤I␤am␤not'␤»
RabidGravy the problem with the design docs and quoting, is that half of what's speculated isn't implemented or is completely different
21:16 jjido_ left
AlexDaniel m: .say for qqw􏿽xE2􏿽x8C􏿽xA9'I am protected'􏿽xE2􏿽x8C􏿽xAA 21:16
camelia rakudo-moar b348f2: OUTPUT«'I␤am␤protected'␤»
AlexDaniel m: .say for qqww􏿽xE2􏿽x8C􏿽xA9'I am protected'􏿽xE2􏿽x8C􏿽xAA # whoops 21:17
camelia rakudo-moar b348f2: OUTPUT«I am protected␤»
Zoffix :/
AlexDaniel yeah I messed it up
Zoffix Ah. Those show up as null bytes in my IRC client :/
AlexDaniel Zoffix: unicode support. 2016 style. 21:18
Zoffix I see them in the browser :) HexChat on Windows just sucks :)
pmurias Updating the synopsis seems helpfull
AlexDaniel Zoffix: Can you see these ⟨⟩? 21:19
pmurias A lot of the stuff in docs.perl6.org seems tutorialish
which is horrible when I just want to get all the info about something
AlexDaniel pmurias: why don't you read the source code then
okay, just kidding, you've got a point 21:20
Zoffix AlexDaniel, I just get codes: i.imgur.com/RhKELUC.png
AlexDaniel Zoffix: oh come on… :(
pmurias AlexDaniel: I had to read the tests for native call stuff as docs.perl6.org was wrong
RabidGravy well maybe that needs to be fixed 21:21
AlexDaniel m: say join ‘’, (0..0x1FFFF ==> grep { .uniname ~~ /:i angle/ } ==> map {.chr}) 21:22
camelia rakudo-moar b348f2: OUTPUT««»͉̚‹›⃤∟∠∡∢⊾⊿⌧〈〉⍼⎊⏃⏄⏅⏩⏪⏫⏬⏭⏮⏯⏴⏵⏶⏷▬▭▮▯▲△▴▵▶▷▸▹▼▽▾▿◀◁◂◃◢◣◤◥◬◭◮◸◹◺◿⛛❬❭❮❯❰❱➝➞➟➠⟀⟁⟨⟩⟪⟫⦑⦒⦛⦜⦝⦞…»
RabidGravy I think Natve/call needs two docs
AlexDaniel m: say join ‘’, (0..0x1FFFF ==> grep { .uniname ~~ /:i angle/ & /:i bracket/ } ==> map {.chr})
camelia rakudo-moar b348f2: OUTPUT«〈〉❬❭❰❱⟨⟩⟪⟫⦑⦒⧼⧽〈〉《》︽︾︿﹀␤» 21:23
dha Hm. currently, we have docs for qw and qqw. Should I look to document the difference between w and ww separately, or just add a ww explanation in both qw and qqw?
AlexDaniel Zoffix: ↑ can you see anything?
dha: /o\ 21:24
dha Did I score a goal?
AlexDaniel dha: I think that qqww should exist anyway 21:25
dha: just like qqw exists: doc.perl6.org/syntax/qqw.html
dha: so maybe you can start from that, and then just link both from each other? Dunno
21:26 _dolmen_ left
Zoffix AlexDaniel, some of it. IRC: i.imgur.com/5DRXqcw.png Browser: i.imgur.com/0Ie3NSh.png 21:26
21:27 adu left
dha Well, I'm on the fence there. The docs have different entries for interpolation vs. non-interpolation. :w is discussed only under qw. So what you're suggesting *should* lead to an entry for qww as well as qqww. 21:27
labster Thanks for the docs Zoffix. And thanks for all of the HN karma off of your Anguish post.
21:27 TEttinger joined
dha I'm not sure that's worthwhile. 21:27
AlexDaniel dha: it is. People searching for qqww should get a qqww page
Zoffix labster, and thanks for putting my Anguish post to #1 spot for a brief period :) 21:28
dha ok. so I'll do a qww as well, then.
AlexDaniel dha: honestly, it does not matter so much as long as /something/ is there
labster I had to change the title to make it a bit more clickbaity. dang changed it back later, which is why it dropped from #1 so fast :(
dha well, yes. but I'd like to do it right if I can. :-)
AlexDaniel .u 〈〉 21:29
yoleaux U+3008 LEFT ANGLE BRACKET [Ps] (〈)
U+3009 RIGHT ANGLE BRACKET [Pe] (〉)
21:29 spider-mario left
AlexDaniel m: ‘〈〉’.uninames.say 21:30
camelia rakudo-moar b348f2: OUTPUT«(LEFT ANGLE BRACKET RIGHT ANGLE BRACKET)␤»
AlexDaniel it does not say that these are full-width, but they are
.u〈〉
m: ‘〈〉’.uninames.say
camelia rakudo-moar b348f2: OUTPUT«(LEFT ANGLE BRACKET RIGHT ANGLE BRACKET)␤»
AlexDaniel m: ‘〈〉’.ords.say 21:31
camelia rakudo-moar b348f2: OUTPUT«(12296 12297)␤»
AlexDaniel m: ‘〈〉’.ords.say
camelia rakudo-moar b348f2: OUTPUT«(12296 12297)␤»
dha Oh. Also :w and :ww don't seem to be documented. So, do those need entries as well?
21:31 spider-mario joined
dha This leads us to entries on w, ww, q, qq, qw, qww, qqw and qqww. 21:31
Bleah.
AlexDaniel it does not look like a great idea anymore :D 21:32
dha Oh, here's a question. You say people should be able to search for qqww. But wouldn't they be more likely to search for 􏿽xAB􏿽xBB?
21:32 synopsebot6 left, synopsebot6 joined
AlexDaniel dha: search for what? 21:32
dha: what is \253\273 ? :/ 21:33
dha Oh, did that not render?
AlexDaniel nope
Zoffix nope
dha double angles
<<>>
pmurias AlexDaniel: there is a huge difference between docs oriented towards new towards something and ones for someone who wants all the minute details
AlexDaniel pmurias: no, you are completely right. But I'm not sure what would be the solution to this 21:34
pmurias: ideally, somebody has to work on the specs too, but there's not enough manpower
dha So, what I'm thinking is that we can have the qw docs mention the qww version and similarly for qqw. So, qww and qqww wouldn't be searchable, but <> and <<>> would be.
Zoffix Make Perl 6 super popular to get more people working on the docs!
:)
dha Not necessarily as a permanent solution, but so that we have the actual documentation there for now.
AlexDaniel dha: «» is already searchable 21:35
dha and deal with searchability at some later point.
right.
pmurias AlexDaniel: docs.perl6.org is intended to fullfill both the reference and introduction roles?
AlexDaniel it just points to doc.perl6.org/language/quoting
dha qww and qqww aren't.
pmurias maybe we should merge the synopsis with docs.perl6.org? 21:36
Zoffix merge how? 21:37
AlexDaniel gradually, perhaps
Zoffix: move stuff from specs to docs, one by one
dha I'm thinking, since it seems to me that people would be *more* likely to search for the operators, rather than the qw versions, we put the w/ww distinction in the qw and qqw sections for now, and worry about searches for qww and qqww at some later point, but put the distinctions in the existing entries so the documentation is at least written and available if not wholly searchable.
masak pmurias: the synopses and docs.perl6.org have different target demographics. 21:38
21:38 jjido joined
masak users and implementors, respectively. 21:38
Zoffix AlexDaniel, wouldn't that make docs.perl6.org overly complex?
dha Also, the synopses are not up to date. Wait, didn't we have this discussion earlier? :-)
pmurias masak: how do those demographics differ?
Zoffix Kinda like Perl 5's perldocs. You get a giant wall of text for everything :(
masak pmurias: are you kidding me? :)
pmurias masak: it's a serious question 21:39
masak pmurias: as an implementor of a Perl 6 implementation, I need to know what a lexical closure is. what a continuation is. what bounded serialization is.
AlexDaniel Zoffix: not if we assign you to do that. Look at your blog posts, they're fun to read, they give basic introduction, and yet they talk about some details
masak as a user of the language, I don't, necessarily.
dha masak - it's not as ridiculous a question as it sounds. user docs are incomplete, so people often wind up looking at the design docs to find answers not covered in the user docs.
pmurias masak: what's bounded serialization? 21:40
dha Again, I think there's a distant echo in here. :-)
masak pmurias: it's something Rakudo does to communicate between BEGIN time and runtime.
21:40 spider-mario left
masak dha: maybe my distinct lack of backlogging is at fault here. if so, apologies. 21:41
pmurias masak: that's what serialization is ;)
AlexDaniel masak: although I partially agree with you, I'm not sure why these things can't be on docs.perl6.org. Perhaps not amidst tutorial-ish stuff, but still
dha masak - it was a while back. no worries.
masak pmurias: no arguments there. in Rakudo it's called "bounded serialization" for reasons I either never learned or can't remember. 21:42
dha WRT the differentiation between reference and tutorial, I note that there are several p5 perldocs that are explicitly tutorials. Is there a reason that extensive tutorial material can't be in separate docs?
masak pmurias: the modifier "bounded" is however not vital to my point, which is: as a non-implementor user, I usually don't have to worry about serialization on the compiler/runtime level. I wouldn't expect to find (much of) it in the user docs. 21:43
21:43 jjido left, ocbtec left
AlexDaniel dha: actually, we already have some tutorials here: doc.perl6.org/language.html 21:43
dha So, I guess the answer is "no, there is no reason that can't happen". :-)
pmurias masak: IMHO that depends on the type of a user 21:44
perlpilot masak: it's bounded because we only serialize things in the current compilation unit IIRC. Even if they reference items in other compilation units. (or something like that) 21:45
pmurias masak: it seems to me there is a huge difference between learning perl 6 and looking up what happens to your objects at runtime
perlpilot masak: But, I'm sure if you really care about why it's called "bounded serialization", you can ask jnthn :) 21:46
RabidGravy isn't that a new document then
21:46 mr-foobar joined
Zoffix stmuk, any idea why my blog posts disappeared from pl6anet.org ? I'm guessing my RSS XML is missing some keys.... any idea which? 21:46
pmurias masak: when I want to deeply understand something I have similiar needs to the implementor 21:48
timotimo MadcapJake: i just finished your iterators/generators post. nice :) 21:50
pmurias Zoffix: I love wall of text style perldocs ;)
AlexDaniel
.oO( it would be interesting to have a “GIMME MORE INFO” button. Each time you press it, more stuff appears )
dha Out of curiosity, why are there quoting adverbs? Are they ever useful? I mean in their ':q', ':w', etc. forms? 21:51
timotimo .o( method pull-a-fast-one ... )
dha I mean, I can't see a reason to ever use, say, Q :q :w// and I'm not sure how else you would use them. 21:52
timotimo dha: sure there are, if you want to write a string literal that's supposed to contain { }, you can :!c in front and not have to escape the curlies
and of course for heredocs you always use the :to adverb 21:53
Zoffix stmuk, never mind. Reading the source tells me I'm missing the modified date metacpan.org/source/DAVECROSS/Perl...erlanet.pm
dha And you would do that instead of not using curlies as your delimiters... why?
21:53 _dolmen_ joined
timotimo not for delimiters 21:53
dha oh, right.
timotimo m: say q:!c"hey, dha, {how are you?}"
camelia rakudo-moar b348f2: OUTPUT«hey, dha, {how are you?}␤»
timotimo m: say q"hey, dha, {how are you?}"
camelia rakudo-moar b348f2: OUTPUT«hey, dha, {how are you?}␤»
timotimo er, with q that's dumb, of course 21:54
m: say qq"hey, dha, {how are you?}"
camelia rakudo-moar b348f2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ojeDENk83f␤Bogus postfix␤at /tmp/ojeDENk83f:1␤------> 3say qq"hey, dha, {how are you7⏏5?}"␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end…»
timotimo that's what i meant.
m: say qq:!c"hey, dha, {how are you?}"
camelia rakudo-moar b348f2: OUTPUT«hey, dha, {how are you?}␤»
dha ok. so the useless ones are there for consistency? :-)
timotimo they are also how we implement the different quoting constructs 21:55
dha ok. So they should actually be documented somewhere. I'm not tackling that at the moment...
timotimo but yeah, i suppose, consistency
dha huh. interesting. Qc works, but Q!c does not. the latter requires the colon while the former does not. 21:57
timotimo yeah, probably because we have a shortcut for that for convenience 21:58
just like qw
dha That's... annoying. But quite possibly unavoidable, I suppose.
masak pmurias: I hear what you're saying -- you're the kind of user who often asks questions similar to what an implementor would.
pmurias: the synopses have often been described as "dry". the reason for this has been that they don't need to be verbose and descriptive to users, only enough for implementors. 22:00
to me, "merging" doesn't sound like "combining the strengths of both" in this case. it sounds like "trying to serve two wildly different groups with a single thing" 22:01
dha oop. S02 says if you want to abbreviate quoting constructs, "just define a macro". I feel odd saying "just do this" with something that spits at you with EXPERIMENTAL FEATURE!
timotimo hehehe 22:02
22:06 Sgeo joined
dha Ugh. I think we're going to have to define "quote preservation" before I document the difference between :w and :ww. In particular, what counts as a quote in this context. 22:09
masak 'night, #perl6 22:11
22:11 pmurias left
Zoffix night 22:11
AlexDaniel m: .say for «aaa /hello world/ » 22:12
camelia rakudo-moar b348f2: OUTPUT«aaa␤/hello␤world/␤»
AlexDaniel m: .say for «aaa ‘hello world’ »
camelia rakudo-moar b348f2: OUTPUT«aaa␤hello world␤»
AlexDaniel m: .say for «aaa (hello world) »
camelia rakudo-moar b348f2: OUTPUT«aaa␤(hello␤world)␤»
AlexDaniel m: .say for «aaa «hello world» »
camelia rakudo-moar b348f2: OUTPUT«aaa␤«␤hello␤world␤»␤»
AlexDaniel m: .say for «aaa 「hello world」 »
camelia rakudo-moar b348f2: OUTPUT«aaa␤「hello␤world」␤»
AlexDaniel I have no idea :/ 22:13
dha Yes, there are a whole bunch. That's the problem. :-)
I'm going to have to dig into source code, aren't I? :-(
AlexDaniel m: .say for «aaa ’hello world’ »
camelia rakudo-moar b348f2: OUTPUT«aaa␤’hello␤world’␤»
AlexDaniel m: say ’hello world’ 22:14
camelia rakudo-moar b348f2: OUTPUT«hello world␤»
22:14 kid51 joined
dha Also, how the heck do you type a japanese quotation mark on os x? 22:17
MadcapJake timotimo: thanks! Glad you liked it! 22:18
AlexDaniel dha: copy… paste… 22:19
dha Heh.
Well, that does work... :-) 22:20
RabidGravy right that's me done for the day
22:23 Zoffix left 22:25 RabidGravy left 22:28 adu joined
dha Ok, this probably needs some tweaking, particularly with respect to the <> and <<>> operators, but how horrible is this as a first pass? 22:33
gist.github.com/dha/409794ce0679da...2602d39374
22:35 adu left 22:38 slobo joined 22:41 rindolf left 22:44 perlpilot_ joined, john51 joined
dha Anyway, time to go. I'll look for more input anon. 22:44
22:44 dha left 22:47 dustinm`_ joined, pierrot_ joined, dwarring left, obfusk_ joined 22:48 sjn_ joined, bonsaikitten joined, jdv79_ joined, diakopte1 joined, El_Che_ joined, mls_ joined 22:49 petercom1and joined 22:50 kid51 left, boegel|quassel joined 22:51 kid51 joined, lnr joined, andrewalker_ joined 22:52 sQuEE joined, sftp_ joined, eviltwin_b joined, romrador joined, Celelibi_ joined 22:53 ptolemarch left, ssotka left, khw left, xiaomiao left, TreyHarr1 left, pierrot left, andrewalker left, obfusk left, eviltwin_b is now known as geekosaur, sftp_ is now known as sftp, lnr is now known as lnrdo 22:57 xfix joined 22:59 wamba left 23:01 _dolmen_ left 23:04 bjz joined 23:05 firstdayonthejob left 23:14 adu joined 23:16 TreyHarr1 joined 23:20 bjz left
AlexDaniel .tell dha now I understand your point about quote protection 23:22
yoleaux AlexDaniel: I'll pass your message to dha.
AlexDaniel .tell dha I may be wrong, but it looks like you missed the point that under :ww 'hello world' is a single element
yoleaux AlexDaniel: I'll pass your message to dha.
AlexDaniel .tell dha anyway, I've left a comment on your gist
yoleaux AlexDaniel: I'll pass your message to dha.
23:25 avar left 23:27 cdg left 23:30 Bucciarati joined 23:31 literal joined 23:32 awwaiid left, avar joined, awwaiid joined 23:33 TeamBlast joined 23:38 ssotka joined, ptolemarch joined, khw joined, inokenty joined, adu left, flaviusb joined 23:39 pnu_ joined 23:43 eam joined
AlexDaniel m: say ‘../foo/..’.IO.resolve 23:43
camelia rakudo-moar b348f2: OUTPUT«"/home/foo/..".IO␤»
23:43 drvvv joined
AlexDaniel m: say ‘../foo/..’.IO.resolve.dir 23:44
camelia rakudo-moar b348f2: OUTPUT«Failed to get the directory contents of '/home/foo/..': chdir failed: Unknown system error 2␤ in any at /home/camelia/rakudo-m-inst-1/share/perl6/runtime/CORE.setting.moarvm line 1␤ in block <unit> at /tmp/qUE6YGvVTs line 1␤␤Actually thrown at:…»
23:45 musca joined 23:47 BrassLantern joined, BrassLantern left, BrassLantern joined 23:54 cpage__ joined
MasterDuke AlexDaniel: .resolve currently stops "resolving" and just returns what's left of what it was given when it hits a part of the given path that doesn't exist 23:54
AlexDaniel MasterDuke: I see, that's fine
I don't think that it's a bug anymore, I was wrong 23:55
MasterDuke i'd say it's LTA though
AlexDaniel exactly. That's what I wrote on #128214
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128214
AlexDaniel MasterDuke: so your commit makes it less LTA, right? 23:56
MasterDuke my PR has it fail if it hits something that doesn't exist
which doesn't contradict the docs or the spectest, i think it's a little better 23:57
AlexDaniel write a test for it and that's -1 ticket
m: chdir ‘wat’ 23:58
camelia rakudo-moar b348f2: OUTPUT«Failed to change the working directory to '/home/camelia/wat': does not exist␤ in block <unit> at /tmp/bHCBsomLWp line 1␤␤Actually thrown at:␤ in any at gen/moar/m-Metamodel.nqp line 3055␤ in block <unit> at /tmp/bHCBsomLWp line 1␤␤»
AlexDaniel MasterDuke: ah, I see 23:59
so you say that .resolve itself should fail?
now, I'm not sure :/