»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 Zoffix on 25 May 2018.
LeboyX where is the $var.list method documented? 00:04
tobs LeboyX: it's a method of Any 00:06
LeboyX Ah. I see that now
Thanks!
Kaiepi mkdir /share: Permission denied at /usr/libdata/perl5/ExtUtils/Command.pm line 299. 00:48
?
benjikun Kaiepi: what's that from? 01:19
trying to use the `mkdir` routine?
Juerd It looks like Any.List just calls Any.list... Why do both exist? 02:06
TimToady m: say [1..10].list.WHAT 02:10
camelia (Array)
TimToady m: say [1..10].List.WHAT
camelia (List)
Juerd Thanks. What's the use case for .list? 02:11
TimToady it's really short for "make something Positional" 02:12
but that's kinda long
Juerd And not much less confusing :) 02:13
MasterDuke m: my $a = ^4; .say for $a; .say for $a.list 02:14
camelia ^4
0
1
2
3
TimToady whereas something capitalized like .List pretty much implies you want to end up with that exact type 02:15
Juerd I'll try to get some sleep, hoping I'll understand in the morning :) 02:18
TimToady Just because I'm explaining doesn't mean I understand. :)
Zoffix .tell AlexDaniel recall you had issue with using «» in stuff like `run`? It may or may not interest you that the error for `m: «` referrs to them as "shell-quotes" 02:24
yoleaux Zoffix: I'll pass your message to AlexDaniel.
rindolf hi all 02:25
sup?
AlexDaniel .tell Zoffix yeah, because «» do shell style quoting 02:28
yoleaux 02:24Z <Zoffix> AlexDaniel: recall you had issue with using «» in stuff like `run`? It may or may not interest you that the error for `m: «` referrs to them as "shell-quotes"
AlexDaniel: I'll pass your message to Zoffix.
rindolf TimToady: I created github.com/shlomif/rakudo-git-master-install 02:29
rindolf TimToady: btw, is the word "creator" in English reserved to divine creation (like Hebrew "boré") or can it be a more mundane one ("yotzer")? 02:33
geekosaur english has only the one
rindolf geekosaur: yes 02:35
warriors how do we type those (not on my keyboard) characters, shown in many Perl6 examples, the lazy evaluation example on Perl6.org have a few :)
geekosaur to the extent thatthere's a difference, "create" implies something new, "make" means following someone else's pattern. but not everyone uses them that way; it's a much looser distinction than in Hebrew 02:36
rindolf geekosaur: well, there is www.gnu.org/philosophy/words-to-av...ml#Creator 02:37
geekosaur um. I assume it is known that the p6 doc site has a bunch of bad links?
yes, it varies. there's religious groups that can get snippy at times. they don't run the show
rindolf geekosaur: in hebrew we have `assah for "make" 02:38
geekosaur warriors, docs.perl6.org/language/019-unicode_entry
warriors :) thanks 02:39
warriors google returns a bad link, i just created an issue on github for it 02:40
geekosaur and on windows I use github.com/samhocevar/wincompose
oh, they just changed all the links and I had to reload the page I had open
rindolf geekosaur: thanks! Seems like the FSF are being paranoid in this regard 02:44
geekosaur some groups are. "creator" was a perfectly good word until *recently* certain pseudo"Christian" groups decided to throw their weight around 02:45
rindolf geekosaur: ah 02:46
warriors abbrev mode in emacs seems to do the trick 02:53
rindolf geekosaur: thanks 02:54
warriors emacs is good 02:57
emacs + Perl6 is bettar
TimToady I dunno, why do you need two operating systems? :-) 03:02
rindolf TimToady: two are better than one!
TimToady: I collect OSes
TimToady: github.com/shlomif/shlomif-email-s...s.txt#L701 03:04
benjikun Perl6 operating system :thinking: 03:05
warriors :)
(define-abbrev-table 'global-abbrev-table '(("Perl6" "♥"))
rindolf benjikun: will it have a good text editor?
benjikun rindolf: I'm a vim guy :( 03:06
rindolf benjikun: can it send mail?
warriors i become an emacs fan after i successfully configured org mode
rindolf benjikun: so am i.
warriors org mode is soo good
benjikun A perl6 system shell :thinking: 03:07
rindolf benjikun: neovim seems promising but it doesnt have sth as good as gvim yet
benjikun great for general scripting
rindolf: meh I just used customized vim
rindolf benjikun: my gvim is customised too 03:08
warriors i did notice that most Perlers prefer vim .. but well
benjikun rindolf: Why use gvim over normal vim?
warriors again org mode is just too good
rindolf warriors: you dont have to use vim for perl
benjikun I suppose if you don't use cli as much for some stuff
warriors i know
rindolf text is text
benjikun: i use cli 03:09
warriors all we really need is repl as good ad utop for ocaml
rindolf benjikun: gvim has better colours and menus
warriors and i will be very happy
benjikun warriors: I love ocaml, I certainly prefer perl6 for general system-wide scripting though
rindolf i use alf+f+s sometimes
or alt+f+x
warriors i was just mentioning it for the utop thing :) 03:10
benjikun I've been exploring nim lately
warriors utop is a really nice repl, i like how it does autocomplete 03:11
benjikun warriors: github.com/bduggan/p6-jupyter-kernel 03:11
if you don't mind doing it in a webbrowser :3 03:12
warriors interesting
benjikun You can also get history in the p6 cli repl and arrow key usage with Linenoise 03:13
rindolf benjikun: i sometimes use vim and sometimes kate 03:13
benjikun rindolf: I use vim or joe normally 03:14
rindolf benjikun: ah, joe
rindolf benjikun: joe was my first unix console editor 03:15
benjikun Mostly same for me, other than nano
rindolf benjikun: but i stopped using it
or maybe i used pico a little before that 03:16
warriors the emacs have syntax highlight, if we get utop for perl6, i wont mind opening a repl for the auto complete,not the most efficient, but will be satisfacotry
the emacs perl6 mode i meant 03:17
rindolf benjikun: I dreamt in gvim a few times
rindolf benjikun: I got used to gvim on windows nt/2000 03:18
benjikun: it was installed on the technion computer lab
benjikun: for a while i used :source /mswin.vim 03:19
but i stopped using it too
vim has a syntax file for emacs lisp and emacs has a mode for vimscript ;) 03:25
warriors I strongly , strongly , strongly recommend, that the official docs, always include examples with the ascii equivalent in addition to that with the unicode char 04:06
Zoffix warriors: github.com/perl6/doc/issues/new 04:10
Zoffix Statements on IRC, especially on a Friday night usually get lost :) 04:10
warriors good idea :) i will 04:11
Zoffix ZofBot: rock it with me, bud: www.youtube.com/watch?v=01g5uQiSdWM 04:12
rindolf Zoffix: hey 04:14
Zoffix \o
benjikun how are you doing Zoffix 04:33
Zoffix 👍 04:42
rindolf Emojjix 04:48
[Coke] tbrowder_: what's the status on the broken links? those need to be changed back asap - this is a breaking change to the website. 05:13
if this means reverting the whole shebang, that's preferable to breaking all those links. 05:14
[Coke] Looks like they're still broken. 05:14
Geth doc: coke assigned to tbrowder Issue URGENT: work to order entries on docs.perl6.org/language.html has broken long standing links. github.com/perl6/doc/issues/2202
c1f1bb5954 | (JJ Merelo)++ | doc/Language/syntax.pod6

Addressing this comment in the IRC channel: colabti.org/irclogger/irclogger_log...07-21#l229
05:17
warriors subscribe to this one github.com/perl6/doc/issues/2190 05:20
i think this is where they will solve it
jmerelo status 06:21
releasable6: status
releasable6 jmerelo, Next release in ≈12 hours. 4 blockers. 0 out of 233 commits logged (⚠ 78 warnings)
jmerelo, Details: gist.github.com/8ea7e88e965bde18b0...88eeb456c0
jmerelo Someone to help this SO poster? stackoverflow.com/questions/514327...r-goes-out 06:22
Kaypie getting a netbsd vm set up for testing, fingers crossed the installer doesn't fuck up the disk again
oh websockets i could give it a go
jmerelo Kaypie: thanks! 06:24
rindolf Kaypie: i know vbox has poor support for fbsd 07:12
Kaypie worse, i'm using qemu on openbsd since vmd doesn't recognize the netbsd iso as bootable 07:13
Kaypie i can get it to install, but not boot 07:13
araraloren How can I add a pragma before enter the Perl 6 REPL ? 07:15
jmerelo araraloren: You can use it _when_ you enter. Same effect, I guess. 07:22
geekosaur perl6 -MMONKEY (or whatever)
araraloren for example `no strict` 07:23
I found no way to add this working for everyline
I mean like -M in command line 07:24
xinming m: my $take-me:<home> = 'When the glory has no end'; 07:50
camelia ( no output )
xinming In this case, what does m: mean? or it's just copy&paste junk? 07:51
docs.perl6.org/language/072-syntax <--- It's in this url
stmuk looks like copy&paste junk to me 08:13
Xliff \o 08:15
What's the best way to write a regexp that will allow for an expression like "a.b.c" 08:16
But will fail "a.b."
jmerelo xinming: correct. Copy/paste junk. Will fix. 08:19
Xliff: /\w\.\w\.\w/? 08:21
m: say "a.b.c" ~~ /\w\.\w\.\w/
camelia 「a.b.c」
jmerelo m: say "a.b." ~~ /\w\.\w\.\w/
camelia Nil
Xliff jmerelo: But what about aa.bb.cc.dd. 08:22
?
synopsebot Link: doc.perl6.org/language/syntax
Xliff So more like / \w+ % '.' /, but that allows for a following . 08:23
jmerelo m: say "aa.bb.cc" ~~ /\w+\.\w+\.\w+/ 08:24
camelia 「aa.bb.cc」
Xliff 'You are still using a set number. I can't use that. 08:25
Because it can be aa.bb
Or just aa
Just can't have the trailing .
I am thinking I have to settle on just matching and using an action class to throw an exception if that following . is detected.
jmerelo Xliff: use negative lookahead assertion docs.perl6.org/language/067-regexe...Assertions 08:26
Geth doc: 915eb4951a | (JJ Merelo)++ | doc/Language/syntax.pod6
Reflow here and there
08:32
doc: 58983f4232 | (JJ Merelo)++ | .gitignore
:see_no_evil:
synopsebot Link: doc.perl6.org/language/syntax
jmerelo OK, I got the problem with the page build 08:48
There was a leftover directory which hadn't been deleted when updating source. 08:49
I really don't know why
jmerelo I wonder if we should add this stackoverflow.com/questions/149915...irectories to housekeeping in the perl6 docs site 08:50
basically add git clean -fd to make html 08:51
CIAvash[m] .tell Xliff '%' doesn't allow a trailing '.', '%%' does. 08:54
yoleaux CIAvash[m]: What kind of a name is "Xliff"?!
stmuk jmerelo: I was wondering that myself ... maybe go for it! 08:55
CIAvash[m] .tell Xliff '%' doesn't allow a trailing '.', '%%' does.
yoleaux CIAvash[m]: I'll pass your message to Xliff.
jmerelo stmuk: still some stuff to clean up first... builds in travis are also failing. I'm fixing the stuff there, if that works I'll get to that. 08:56
stmuk: I'll create an issue so that I don't forget.
stmuk jmerelo: given how urgent fixing the site is it could be worth trying that one line fix before anything else? 09:01
or even just "rm -rf build" in "make html" 09:02
jmerelo And we're back to normal URLs docs.perl6.org/language/5to6-nutshell 09:03
stmuk: it's not enough.
stmuk: the problem is that when you delete a whole directory with files in it it's not deleted from the repo. Even if I kept removing it from here and there, it kept issuing the error. 09:04
stmuk: I had to look up the new script to see what it was doing, and it was copying everything under doc/Language And there was this untracked directory, 0-html-whatever, which was causing the error
stmuk: so I removed it by hand, rebuilt, sync... 09:05
Geth ¦ doc: JJ self-assigned Clean directories when building the site github.com/perl6/doc/issues/2203 09:06
stmuk so 0-html-whatever isn't under "build"? 09:07
jmerelo stmuk: it was, but because it was copied over and over again from under Language/
stmuk this sounds quite messy 09:08
stmuk oh you mean "repo checkout" not "repo" above 09:11
stmuk jmerelo: I was refering to two problems in #2202 -- the one you just opened as #2203 and a further one with the "html-nohighlight" target 09:14
jmerelo stmuk: let me see 09:31
tbrowder_ wellll, looks like jmerelo and stmuk saved my bacon! 09:38
jmerel: thanks
jmerelo: thanks
stmuk: thanks
sarna o/ 09:39
tbrowder_ sarna:
sarna: \o
fingers not awake yet
jmerelo: i just looked at the build log stmuk showed, it still doesn't show the build dir clean and gen-pod6-source BEFORE running htmlify. 09:49
what is running that script? docker?
stmuk that's an old build log .. it seems to have deployed ok since then due to jmerelo's change 09:50
or maybe something else done on the server .. not sure 09:51
tbrowder_ ok, any more problems with links? 09:52
stmuk I don't think "git clean" in makefile targets is sensible see comments on github.com/perl6/doc/issues/2203#i...-406784734 09:55
tbrowder_ i'm starting another PR to address the file "ddd-" prefix but want to know if there are any current problems with links? google search?
tbrowder_ i see searching for "perl 6 concurrency" gives the right page--great! pressing on... 09:56
jmerelo tbrowder_: sure! 10:01
tbrowder_: the docker file is just used for testing
tbrowder_ yes, but the trick is the build dir must be current first with the current tooling or the unexpected may happen 10:02
jmerelo stmuk: I cleaned up stuff on the server... That seemed to work.
tbrowder_ are you still in japan
does anyone have any problems with the current dir layout? 10:03
jmerelo tbrowder_: no, just back
At home now since yesterday at 1AM
A bit jetlagged, I guess 10:04
tbrowder_: I think that if we do git clean -fd before updating dirs it shouldn't be a problem
tbrowder_ i would, in a separate move later, like to move all fixed or original sources, except doc, under one dir
El_Che happy to see you guys working on the docs! 10:05
appreciated!
tbrowder_ i don't see target "clean-fd"
El_Che: \o
jmerelo and stmuk got us back online again, whew! 10:06
jmerelo tbrowder_: no, I would add it to make html 10:07
tbrowder_ also we need to figure out how to handle the chinese translation so generously provided
jmerelo tbrowder_: I have already said it on the issue. I would really rather not.
stmuk jmerelo: github.com/perl6/doc/issues/2203#i...-406784734
tbrowder_ jmerelo: i understand the concern about bitrot and non-sync info. 10:08
jmerelo tbrowder_: plus the bigger problem is that pro translations are not just done directly. You need a lot of tooling, including terminology, whole sentences... There's a program called Omega (whatever) that handles that 10:09
tbrowder_ can the perl6.org give a separate site for each language, and there warn the user the canonical info is the the main site? 10:10
jmerelo tbrowder_: I used it in a mini-translation hackathon during YAPC in Granada, and it really helps speed up translations and keep it coherent.
stmuk git clean will destroy any developer's work in progress which isn't checked in. Neither should we assume people are running from a git checkout 10:11
jmerelo tbrowder_: that would be ideal; to have forks with the whole tooling and stuff published to, let's say, zh.docs.perl6.org
stmuk: OK, I see what you mean. Dangerous thing to do locally. I can add git stash there, just in case. 10:12
tbrowder_ AlexDaniel agrees that we also need something like "test.docs.perl6.org" to avoid my disastrous live push!!
jmerelo tbrowder_: it's not been the first, it's not gonna be the last. 10:13
sarna tyil: looks really nice! personally I'd add more stuff to Env, but that could be left as an exercise for the reader :)
stmuk the whole thing is quite fragile
tbrowder_ but a test site would be cool, if it's not too much work
jmerelo tbrowder_: so no big deal. A huge amount of work has a non-zero likelihood of problems when going to production. Still very grateful for the work done. 10:14
sarna tyil: also, your bracing style is really interesting, I've never seen stuff like `) {` on its own line :D
tyil :o 10:14
stmuk jmerelo: my other point is clean up shouldn't be specific to one version control system. .git directories wouldn't exist in a CPAN release tarball or in R* for example
tyil now you have :3
jmerelo tbrowder_: for starters, I think we would have to set up secondary DNSs, which I did for the last time in the 90s. But I guess it could be done.
tyil I might redo some parts of it again to make it easier to follow along in a tutorial 10:15
tbrowder_ part of my motivation is to help make the site more robust and untangle data and presentation as much as possible.
jmerelo stmuk: but the objective of this is precisely to clean up directories that are no longer under source control. So this is kinda specific. I don't know if it affects stuff it's never been under source control 10:15
tbrowder_ jmerelo: i know you are pushing templates--i agree!
jmerelo tbrowder_: that's the way to go and paves the path for a future dynamic generation...
stmuk jmerelo: the clean up should be explicit and targeted. We should know what files/dirs are being created 10:16
jmerelo stmuk: hum. We need the git log for that...
stmuk its nothing to do with version control. Someone could be running from a tarball 10:17
jmerelo stmuk: another option is to just not do anything at all. So far, this is the first time I've seen this problem. Let's add some notice to the infrastructure document to tell how it was fixed and that's that
sarna tyil: though, the original lispy was just one file, yours feels quite big (even if it isn't)
tyil it is bigger, yes 10:18
I can golf it down and put it in one file
jmerelo tbrowder_: finally going to Glasgow?
tyil but I'd rather have something more generic that can be easily expanded upon
stmuk tbrowder_: I suggest any files and directories created be under one non-revision control dir (maybe we have that now with "build"?) and that directory be removed before the build?
jmerelo Maybe do a BoF about doc tooling and roadmap while we're there?
tbrowder_ hm, afraid not
tyil I'm not sure how it'll turn out in a couple iterations
I'm learning to deal with grammars better myself everytime I change it up 10:19
jmerelo stmuk: yep, that's the thing... html is both used for some files and for files generated by htmlify.p6
tbrowder_ i'm starting another doc wiki page, we can use it like a blackboard 10:19
stmuk I still think the system should be able to delete the data it creates rather than using git to do it! 10:20
jmerelo stmuk: still, this particular problem was created by copying files that were no longer under source control. When you do a git pull and all files in a directory have been eliminated by git rm, git does not remove the directory or the files.
sarna tyil: no no, don't golf it, lispy's point was being a simple interpreter you could understand fairly quickly
tyil tbrowder_: if you want to have a page to keep notes with multiple people, you can use etherpad 10:21
I host one on my own site that everyone is allowed to use pad.tyil.nl/
jmerelo stmuk: maybe I can just change how stuff is done in docs.perl6.org, no need to touch the rest of the stuff.
tbrowder_ the wiki can be edited by anyone
stmuk jmerelo: I understand the problem but I don't think you understand why your solution is wrong 10:22
jmerelo stmuk: I have proposed several solutions, but I'm all ears... I'm not committed to either.
tyil sarna: I'll try to keep it readable then, but I'll rewrite it a couple times regardless to come to a good, clean solution that I can incorporate into an article :>
I'm not too unhappy with the current state, but I think I can do better 10:23
sarna tyil: alright, good luck :)
tyil thanks ^_^
stmuk jmerelo: I've tried to explain on the ticket and IRC already
tyil I'll poke you when I have more updates I'd like you to look at, if you don't mind
sarna tyil: please do 10:24
jmerelo stmuk: I'm looking now into changing the update-and-sync script. That's only on the server side, it wouldn't hurt devs or delete things that are under control. 10:24
tbrowder_ i started the problem because i was intentionally adding dirs because i saw travis failing and not finding a dir 10:25
jmerelo tbrowder_: water under the bridge. Let's see now how to avoid this kind of thing happening in the future (other will surely happen) 10:26
tbrowder_ ultimately it was because i couldn't see what was happening is part of the travis process, nor how the site is actually generated
jmerelo tbrowder_: that's a thing that only now I'm beginning to understand. Basically, update-and-sync is run every 5 minutes on the server, all stuff is generated, and then rsync'ed somewhere else. 10:27
tbrowder_ ergh!
jmerelo stmuk: do you see a problem with running git clean -fd right after git checkout, line 11 of update-and-sync?
tbrowder_: don't look at me. It was already there when I arrived... 10:28
tbrowder_: there's a hack in update-and-sync, and that's because html has some files under source control. Maybe we could follow stmuk's advice and have those files somewhere else, maybe move them to html-source or something like that 10:29
tbrowder_ ah! i'll just add a couple of lines then...
stmuk jmerelo: Yes that looks fine to me. Most people working on the docs wouldn't run that and it's a system script explicitly for deploy 10:30
jmerelo stmuk: OK. Let's try that then. :-) Thanks! 10:31
tbrowder_ i agree, that's what i'll show in the wiki before i do anything "dangerous" (been reading Vince Flynn's 'Kill Shot')
jmerelo tbrowder_: didn't know the author. Worth the while? I'm into Harlan Coben, Connelly, Hiaasen, American noir in general. 10:33
stmuk jmerelo: I'd perhaps backup the existing files on the server into a tarball before running that "git clean -fd" just in case there are files manually copied in we need
jmerelo tbrowder_: well, world noir in general. I love Icelandic noir, for instance.
tbrowder_ never mind, the "make html" does the right thing
jmerelo stmuk: but were those copied by hand? Or were they part of the script?
stmuk jmerelo: as a devop type I'm a paraniod wondering whether there are some essential copied by hand stuff 10:34
jmerelo stmuk: I see force-update, build-log and some other stuff
tbrowder_ icelandic noir: i've read a couple novels, man named ove, and some police procedural
stmuk jmerelo: also you could try a manual dry run with -n to see what gets deleted
jmerelo stmuk: yep, that's right. Very few systems are build from scratch, and most are jury-rigged one way or the other. 10:35
stmuk: OK, I'll try that.
stmuk jmerelo: it will probably delete the existing server logs and its directory if "build-log" is there
which is Bad
jmerelo tbrowder_: They are _very_ good. Arnaldur Indridasson does the procedural. 10:36
stmuk: right. Thing is, it shouldn't be there.
stmuk: that's defined in a variable, LOGDIR 10:37
stmuk: it's probably there so that it's copied to the right place, but we can copy it from wherever it is using other command
stmuk its probably just there because its under the webroot 10:38
jmerelo stmuk: right, but files get copied from there to the production site. We can just change the origin directory, it's in sync-build-log
stmuk: there's also an update.log which is empty. 10:39
stmuk: it's not a good idea to keep logs in the same place as source controled files
stmuk jmerelo: true but we shouldn't be doing dev on a live system either 10:40
maybe wait until the dev/staging instance before radical deletion/move type changes 10:41
tbrowder_ i started the doc wiki page
afk&
jmerelo stmuk: we're doing that... kinda. Anything pushed to master goes, every five minutes, to the "staging", which then gets rsynced to production 10:44
stmuk: but master should be ready to go into production, anyway.
stmuk: the way I see it, it's been a combination of things that shouldn't happen too often: test system broke down, there was an error with a no-longer-source-controlled directory, and the veterans running the place were not available (never enter production Friday afternoon)... 10:45
stmuk: ... and I was in a flight back from Japan. So I don't foresee that combination of things happening any time in the future. 10:46
stmuk: we don't have so many resources as to run a staging server, monitor, check... and then put it into production. We have tests doing the hard work for us. It passes tests, it should go into production. 10:47
stmuk upload.wikimedia.org/wikipedia/en/.../Jenga.JPG
:)
jmerelo stmuk: we're not testing major infrastructure changes, that's the main problem. So maybe the solution would be to boost our tests.
stmuk you're assuming tests don't have bugs .. they are just software too 10:48
timotimo we might want to grep our logs for urls that have been referred to from the outside and make sure, in the tests, that there's no 404s there (unless we want them to)
stmuk anyway off for a bit 10:49
jmerelo timotimo: well... that would need to be run on the server... Can't do it on travis
stmuk: have fun
timotimo we can install an apache on travis :)
jmerelo timotimo: and we would have to generate every single link, test it... Hum.
timotimo: the whole point of testing is to have it fast. If we are back to 40 minute build it defeats the whole purpose of testing. 10:50
timotimo that's just a "curl" with an input file of urls
jmerelo timotimo: problem is not only that. It's that htmlify.p6 is a 1K-line file with no tests. We don't know how many index entries there should be. Sometimes whole swaths of entries disappear at a time, and we don't even notice until the issues start to crop up. 10:51
timotimo so ... maybe we want to have at least a diff of available pages for each commit? 10:52
jmerelo timotimo: we need to run htmlify.p6 for that... It's run as a test, but just as a smoke test. If it breaks down, something is wrong. If there's some difference in the number or name of generated files, it does not complain. 10:53
timotimo: I mean, look at this: github.com/perl6/doc/issues/561+
timotimo: I think that by now most 404 links should have been fixed, but still there are tons of broken anchor links that are there. 10:54
timotimo: once again, we should have a BoF in Glasgow to talk about all this... 10:55
timotimo: should I propose it to someone?
timotimo: I see it's just a wiki. OK
timotimo what is? 10:59
jmerelo The BoF thing. Just created a page: act.perlconference.org/tpc-2018-gla...tion%20BoF 11:02
But I'm not sure how to actually do it. Do we have to find out the place and time that's more suitable for everyone?
timotimo ah 11:03
sorry, i won't be in glasgow
jmerelo timotimo: too bad... 11:04
timotimo way cool, my "get host and port from listen socket" patch is working 11:07
timotimo makes stresstest 11:11
timotimo yay it's clean 11:24
xinming say @letters[lazy 3..7]; # OUTPUT: «(d e f)␤» 11:27
say @letters[ 3..*]; # OUTPUT: «(d e f)␤»
What are the differences between these 2 examples?
The behavior when indexing a collection via lazy subscripts is different than when indexing with their eager counterparts. When accessing via a lazy subscript, the resulting slice will be truncated.
docs.perl6.org/language/subscripts <-- from this url.
It seems they are the same. *_*
xinming and BTW, perl6 WILL be the language for next centuries... :-) 11:28
AlexDaniel jmerelo: so docs are not fixed still?
ah… they are… but
docs.perl6.org/build-log/ ? 11:29
jmerelo AlexDaniel: I did it by hand 11:41
jmerelo AlexDaniel: it should be updated next time it's pushed 11:42
AlexDaniel I see
AlexDaniel jmerelo++ 11:42
jmerelo: fwiw you didn't sound very inclusive in #2201 :)
jmerelo xinming: right. Can you please create an issue? 11:43
AlexDaniel jmerelo: in a sense that if I was OP, I'd be pissed to read your answer :)
jmerelo AlexDaniel: that's bad. Sorry.
AlexDaniel: sorry about that. What could I do to improve my answer? Or is your reply enough? 11:44
AlexDaniel jmerelo: I'd say my reply is a bit more correct historically
that particular policy is not there because we love unicode sooo much, but because it will consistify things and we'll be able to satisfy both ways of doing things 11:45
s/not there/there not/ 11:46
jmerelo AlexDaniel: I am not aware of the history. I pretended to be inclusive: if you really don't like those examples, please point at some... As a general rule, the policy is the opposite of what he(or she) is asking.
jmerelo AlexDaniel: I still gave him or her the change to support that policy. 11:47
AlexDaniel: but please, no JavaScript thing.
AlexDaniel jmerelo: how else can we implement it then?
the website is not dynamic, so are we calling it blocked for now? 11:48
jmerelo AlexDaniel: just point out to some examples whose readability could be improved by _not_ using Unicode and provide hand-crafted alternatives for them, or expand the explanation, or add a comment to the code.
tbrowder_ I just more or less finished my new strawman proposal for a doc reorg on the wiki—comments or edits welcome 12:03
tbrowder_ several issues relating to bad links, etc., i think can be closed now (by someone other thsn moi who started the problems) 12:08
sarna hey, I'm reading one of jonathan's presentations and he mentions "storing the result of map and grep in a Scalar" 12:36
what happens if I store a collection in a Scalar?
pmurias jmerelo: there is some value in showing people who prefer not to use unicode in their code how to avoid using it 12:37
timotimo sarna: it'll keep the Seq around, instead of iterating over it and making it an array, like when you assign to @foo
pmurias my @powers-of-eleven = lazy 1, 11, 121 … 10¹⁰⁰ # EVEN BETTER 12:39
sarna timotimo: so.. it's lazy?
pmurias shouldn't it be 11**100 as the last element?
timotimo i'd call it that, i guess 12:40
if you do that, you'll have to be careful about only iterating intermediate Seq objects once
m: my $foo = ^100 .grep(*.is-prime); say $foo.tail; say $foo.tail 12:41
camelia 97
The iterator of this Seq is already in use/consumed by another Seq
(you might solve this by adding .cache on usages of the Seq, or
by assigning the Seq into an array)
in block <unit> at <tmp> line 1
timotimo that's the error you get when you try to take values from a Seq twice, i.e. it won't keep the values around automatically
you have to use .cache or store into an array if you need that
as the error message points out
sarna timotimo: hmm, so I can use this to .map() several times only iterating over the collection once? 12:44
timotimo i'm not sure i understand, but i don't think so 12:46
pmurias jmerelo: the hard part I think is how to show the ascii version
timotimo when i spectest the jvm backend, how do i tell if my changes are bad :| 12:47
wow 170 seconds for stage parse of the core setting 12:50
jmerelo pmurias: of course. But the consensus reached the opposite conclusion, so everything would have to be changed. It's much better to try and provide alternatives or additional explanation. 12:55
El_Che (count me on the the party of people that don't use unicode chars in code when those are not on my keyboard) 12:56
timotimo Exception in thread "Thread-8" org.perl6.nqp.runtime.SaveStackException
at org.perl6.nqp.runtime.Ops.continuationcontrol(Ops.java:7339)
:<
pmurias jmerelo: what I mean is that how we display the alternatives is the hard to figure out part
timotimo j: react { my $tap = do whenever IO::Socket::Async.listen("localhost", 0) { say "hello there!" } } 12:57
camelia ( no output )
timotimo i don't know why i get that error with my port acquisition patch 12:59
ah, sure, now it works :| 13:01
SmokeMachine What do you guys think should be this module's name? github.com/FCO/ChainableSeq/blob/m...t/02-all.t 13:11
moritz SeqSlicer 13:38
SmokeMachine moritz: SeqSlicer is a good name!!! Thanks! 13:40
moritz my pleasure :) 13:42
timotimo well done, moritz
Juerd m: my $a = (1, 2, 3); for $a { .say }; for $a<> { .say }; for $a.list { .say }; for $a.List { .say } 14:08
camelia (1 2 3)
1
2
3
1
2
3
1
2
3
Juerd Is any of these better?
That is, out of .<>, .list, and .List, which one should be used in which circumstance? 14:09
moritz there is also for @$a { } 14:10
which I believe is the same as .list
El_Che I would say .list
moritz .List is for when you want to have an object of type List, not just anything list-like
El_Che .List is a coercion, iirc
moritz I tend to go with .list
Juerd moritz: Does .list ever return something that isn't a List?
El_Che .<> is unreadable
my 2c
timotimo Juerd: could return an Array, though Array Is List iirc 14:11
moritz m: my @a; say @a.list.^name
camelia Array
moritz m: my @a; say @a.List.^name
camelia List
moritz m: say Array ~~ List 14:12
camelia True
El_Che m: say List ~~ Array
camelia False
tbrowder_ El_Che: i’m with you on unicode. i wish i had a keyboard with a good number of them, even one of the addons but i’m not sure i could use them with all the editors i use. 14:20
does anyone have an addon that works under linux with at least xterm, vi, and emacs? 14:21
tbrowder_ working with comma would be nice, too 14:22
tobs m: my $a = (1, 2, 3); for |$a { .say } # Juerd: another one for your .list 14:29
camelia 1
2
3
El_Che tbrowder_: even if a out on a Limb and configure my Linux machine and remember all the magic shortcuts, I will hava a hard time using the macbook I use on the train. And we're not even discussing colleagues that will adapt the code on Windows machines... 14:31
tbrowder_ yeah, it’s a bucket of worms, i know. but for use when appropriate it would be nice to have. 14:32
El_Che tbrowder_: indeed, it's nice to read 14:33
being the 100-year language we still have 97 years for someone to come with a keyboard replacement :) 14:34
tbrowder_ i just don’t know anyone personally who has tried such a thing. i guess i will just buy one and see if i can get it to work. 14:35
El_Che I tried to add shortcuts in macos and it's terrible 14:36
it's easier on Linux, but I forget the shortcuts :)
tbrowder_ very astute observation. i have seen msny koreans use a special keyboard for their language and wish someone made a similar one based on english + unicode. 14:37
El_Che tbrowder_: on the other hand I have seen Israelis with suck setup say "fuck it" and write mails in English to their parents 14:38
tbrowder_ and i understand that. 14:39
maybe i need a keyboard like so many long-time p6 seem to have so i can use all the funny quotes 14:42
long-time p6 people
back to business—any comments on my doc wiki about reorg? 14:43
afk&
jmerelo releasable6: status 15:45
releasable6 jmerelo, Next release in ≈3 hours. 4 blockers. 0 out of 234 commits logged (⚠ 79 warnings)
jmerelo, Details: gist.github.com/a6a988bf79e99160b7...8f32ffce05
xinming say "11" ~~ /(\d) {} :my $c = $0; ($c)/; 16:18
evalable6 「11」
0 => 「1」
1 => 「1」
xinming :m say "11" ~~ /(\d) {} :my $c = $0; ($c)/;
m: say "11" ~~ /(\d) {} :my $c = $0; ($c)/;
camelia 「11」
0 => 「1」
1 => 「1」
xinming hmm, What does the : before my mean?
AlexDaniel xinming: it's just a way of declaring variables in a regex docs.perl6.org/language/regexes#in...ntry-%3Amy 16:19
jmerelo xinming: As AlexDaniel says, declares the scope of a variable. 16:20
xinming THanks
That page is where I'm reading. :-)
Later, It has explanation. 16:21
Zoffix Any good ideas for ~2hr worth of a Perl 6 talk? 16:37
moritz target audience? 16:38
Zoffix Local Perl Mongers meeting 16:39
moritz personally, I'd go deep into regexes, grammars and parsing 16:40
Zoffix Also, I have like ~2 days to prep for it :P 16:41
moritz there's a certain book you could use as template... :)
Kaiepi nativecall?
Zoffix heh 16:42
moritz but in that case, more like a tour through all of the Perl 6 features you like, and are familiar with
"Perl 6, Zoffix's Pick"
Zoffix Hm.. Maybe I should write that "Your™" compiler in Perl 6 rather than nqp and use that as a parsing example. Two birds with one code
moritz "Hand-Select Perl 6"
Zoffix I actually did my "Wow, Perl 6" talk in 2016 to basically the same group already: tpm2016.zoffix.com/
Zoffix But yeah, grammars are a good suggestion, based on the survey results of favourite features. moritz++ 16:43
jnthn Yeah, it's a good idea ;) 16:45
*:)
That or concurrency but that might need more prep
Zoffix m: /:P5 / 17:21
camelia 5===SORRY!5=== Error while compiling <tmp>
Unrecognized regex modifier :P5
at <tmp>:1
------> 3/:7⏏5P5 /
expecting any of:
term
Zoffix There's no way to use P5 regexes in grammars is there? 17:22
(no non-hackish way, I mean)
pmurias Zoffix: what will Your be? 17:47
Zoffix man, I need to name it something else. "Your" is impossible to search for 17:50
pmurias: I already answered that the last time you asked. It's somewhere in the logs
timotimo PureL6 17:51
pmurias Zoffix: I found the answer 17:54
jnthn On the upside, if it's too fast, somebody can file a bug report "Your too slow" and it'll be correct grammar :P 17:58
hah, if it's too slow :)
Zoffix :) 17:59
pmurias m: say qq{#!/usr/bin/perl\nprint "1..1\\nok 1\\n"}
camelia #!/usr/bin/perl
print "1..1\nok 1\n"
pmurias isn't that a spec compliant Perl 6 to Perl 5 compiler? 18:00
timotimo hm, the harness doesn't check if the plan matches, does it?
timotimo i mean whether the plan matches the one in the file 18:01
pmurias timotimo: I don't think so
timotimo in that case, yeah, that seems spec compliant ...
we might want to have some verbiage that makes that clear 18:02
pmurias timotimo: what do you want the verbiage to make clear? 18:03
timotimo "cheating is only okay if you don't get caught" 18:04
pmurias I don't think of a good way to avoid the "problem" of hardcoded implementations when defining a language just by a test suit and not a specification 18:06
a strong enough test suit couldn't be hardcoded but ours is more a list of examples 18:07
timotimo suit not strong enough? put more starch on it 18:10
pmurias timotimo: not strong as in accepts things that aren 18:17
't really Perl 6 implementations 18:18
timotimo: I don't think there is really a practical way to avoid that problem
pmurias timotimo: like for a simple language you could compare random programs to a reference implementation 18:19
geekosaur the flip side of this is that if we ever reach a point where playing the spec matters, arguably we've arrived 18:30
timotimo maybe zoffix should call his implementation "haterlang" to mirror "fanlang" 19:06
El_Che trollang 19:11
pmurias strawmanperl 19:22
El_Che pmurias++ 19:23
great one
timotimo swyne6 19:28
x[LGWs4x4i]uG2N0 m: my %h=(a => [[1],[2]]); for %h<a> { say $_ } 20:25
camelia [[1] [2]]
x[LGWs4x4i]uG2N0 m: my %h=(a => [[1],[2]]); for |%h<a> { say $_ }
camelia [1]
[2]
x[LGWs4x4i]uG2N0 i don't understand why i have to flatten, if i want to iterate over all the elements 20:26
El_Che becasue the elements are the containers 20:28
x[LGWs4x4i]uG2N0 i guess i have to read more about containers 20:29
Zoffix x[LGWs4x4i]uG2N0: you don't have to flatten, you only have to decont: 20:33
m: my %h =(a => [[1],[2]]); for %h<a><> { say $_ }
camelia [1]
[2]
x[LGWs4x4i]uG2N0 ah 20:34
thanks
Zoffix x[LGWs4x4i]uG2N0: and you only have to do that because you've storing stuff in an object that automatically sticks stuff into containers. Often you can avoid that in the first place:
m: my %h is Map =(a => [[1],[2]]); for %h<a> { say $_ }
camelia [1]
[2]
Zoffix x[LGWs4x4i]uG2N0: there's this article on containers: perl6advent.wordpress.com/2017/12/02/
x[LGWs4x4i]uG2N0 tyvm 20:34
b2gills .tell Xliff Was this more of what you were looking for: "aa.bb.cc" ~~ /^ [\w+]+ % '.' $/ 22:24
yoleaux b2gills: I'll pass your message to Xliff.
xinming in docs.perl6.org/language.html There is a page describe what my $var:xy<abc> thing does, Anyone here tells me which page it is? 22:43
I just go through all the pages, I can't find which page it is. I do remember I read that description today
xinming Got it, It's in syntax 22:56
xinming docs.perl6.org/language/grammars <- Here rule calc-op:sym<sub> { <num> '-' <num> } <--- is here the "sym" special? or I can have other name? 23:03
I mean something like rule calc-op:op-name<sub> 23:04