»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.
pilne am i wrong in my feeling that perl6 is probably the easiest way to code if you want/must glue together other languages due to grammars? 00:22
like... if i don't care right now about performance, but for some reason i want to/need to mix say... c, ruby, python, and perl5 code.... 00:23
perl6 would probably be the most natural way to do that (as compared to anything else, except maybe doing it "bottom up" with c?)
like if perl5 is glue, perl6 is superglue :D
raschipi I don't see the relation to Grammars, though... 00:24
pilne aren't grammars what allow the interop to be kept perl6 based? 00:25
in terms of syntax? or is my brain finally just melting down a little?
raschipi No, it's the foreign function interface 00:26
i.e. NativeCall 00:27
Rakudo uses NativeCall to call into other languages.
raschipi But Perl6 was built from the start to be able to do these things, so it's not just NativeCall. 00:28
pilne fair enough
Geth doc: f7eefbd890 | (Zoffix Znet)++ | 3 files
Document behaviour with degenerate cue intevals

  - Could use less stilted language
Impl: github.com/rakudo/rakudo/commit/df01ad97e5 Spec: github.com/perl6/roast/commit/baa8ccbb30 Extra tests: github.com/rakudo/rakudo/commit/031f8cf77c
01:19
raschipi u: — 01:23
unicodable6 raschipi, U+2014 EM DASH [Pd] (—)
Herby_ o/ 01:44
HoboWithAShotgun m: my %h = gather { take :p(1) }; %h.say 01:55
camelia Unexpected named argument 'p' passed
in block <unit> at <tmp> line 1
HoboWithAShotgun m: my %h = gather { take p => 1) }; %h.say 01:56
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing block
at <tmp>:1
------> 3my %h = gather { take p => 17⏏5) }; %h.say
expecting any of:
postfix
statement end
statement modifier
statement …
HoboWithAShotgun m: my %h = gather { take p => 1 }; %h.say
camelia Unexpected named argument 'p' passed
in block <unit> at <tmp> line 1
HoboWithAShotgun bummer 01:56
Zoffix HoboWithAShotgun: you're passing it as named arg to &take. Pass it as a positional instead 01:58
m: my %h = gather { take "p" => 1 }; %h.say
camelia {p => 1}
Zoffix m: my %h = gather { take "p" => 1; take :42z.Pair; :100y.take }; %h.say 01:59
camelia {p => 1, y => 100, z => 42}
Zoffix I think it's described in detail here: docs.perl6.org/language/traps#Named_Parameters
Zoffix .put-into: $bed 02:00
HoboWithAShotgun m: say (y => 1) == ('y' => 1) 02:02
camelia Cannot resolve caller Numeric(Pair: ); none of these signatures match:
(Mu:U \v: *%_)
in block <unit> at <tmp> line 1
HoboWithAShotgun good night 02:04
me too
ugexe m: my %h = gather { take $(:p(1)) }; %h.say 02:09
camelia {p => 1}
Herby_ question 02:31
disclaimer: I'm new to Linux
in the terminal, if I type "zef", i get the options for zef
if I type "sudo zef", it says command not found
what gives 02:32
raschipi do you have sudo installed? 02:36
Herby_ yep
geekosaur root often has a restricted $PATH
Herby_ Should I need sudo to: zef install HTTP::UserAgent ? 02:37
raschipi Yep, sudo will clean up the environment before running commands.
geekosaur sudo is usually configured to clean its environment and install a sanitized $PATH by default; see /etc/sudoers and the doucmenttion thereto
raschipi And it's better to leave it that way. Instead use the complete path to the zef executable. 02:38
Herby_ hmmm
geekosaur and you should only need to sudo if you have rakudo installed in a system directory. be cautious in that case if it got there by a package manager 02:39
Herby_ i'm clueless linux rookie... i used checkinstall to install rakudo
raschipi can you give us the result of `which perl6`? 02:40
Herby_ '/home/chrx/rakudo/rakudo-star-2017.07/install/bin//perl6 02:41
raschipi hum, doesn't seem like you'd need to use sudo to install modules 02:42
Herby_ when I try: zef install HTTP::UserAgent
Herby_ I get a wordy error, but the gist is "permission denied" 02:42
Failed to open file /home/chrx/.zef/store/MANIFEST.zef: Permission denied 02:43
geekosaur sounds like you used sudo a previous time and now you have root-owned files under $HOME 02:44
Herby_ definitely possible...
raschipi yeah, I would do `chown -R chrx:chrx /home/chrx/.zef`? 02:45
geekosaur (this might even break zef, I think it assumes something installed under $HOME is guaranteed writable and it can re-precomp or etc.)
or maybe CURLI makes the assumption
the chown needs to be done via sudo though
raschipi right, you'll need sudo 02:46
Herby_ so I should: sudo chown -R chrx:chrx /home/chrx/.zef
geekosaur yes
raschipi have a look at this: explainshell.com/explain?cmd=sudo+...hrx%2F.zef 02:47
Herby_ progress. got further in the install then got this: CompUnit::Repository install target is not writeable/installable: /home/chrx/auto 02:50
Need a valid installation target to continue
geekosaur same issue, sounds like. but why is zef creating a dir there?
raschipi you have the same problem on that path probably
Anyway, it should be safe to execute the same command in that path: `sudo chown -R chrx:chrx /home/chrx/auto` 02:54
Herby_ chown: cannot access '/home/chrx/auto': No such file or directory 02:55
raschipi try `mkdir ~/auto`
geekosaur now I think something else is wrong 02:56
like, I didn't think it should be writing there to begin with. wrong path in zef's CUR somehow?
or maybe it's a bug in the package 02:57
I would be rather annoyed if installing a random package started putting stuff in random dirs outside the CUR playground
much as I would be if apt-get dropped something in $HOME (a debian no-no)
Herby_ i think my install is all jacked 03:02
I was able to install http::useragent
but when I run a simple script 03:03
Could not find HTTP::UserAgent at line 2 in:
/home/chrx/.perl6
/home/chrx/rakudo/rakudo-star-2017.07/install/share/perl6/site
/home/chrx/rakudo/rakudo-star-2017.07/install/share/perl6/vendor
/home/chrx/rakudo/rakudo-star-2017.07/install/share/perl6
CompUnit::Repository::AbsolutePath<77908064>
CompUnit::Repository::NQP<59934952>
CompUnit::Repository::Perl5<59934992>
ugexe its all configurable. github.com/ugexe/zef/blob/master/r...onfig.json 03:13
HoboWithAShotgun Color::Named - Because colors are people too. Do you think that could be somehow politically incorrect? i'm super bad at judging such things 03:15
raschipi One thing that comes to my mind is that there are people that don't have names. 03:16
HoboWithAShotgun i mean it as a reference to "aliens are people too" 03:17
meh, better not before somebody calls me racist or something
geekosaur its possible. 'colored' was at one point a racial epithet 03:27
in the US at least
AlexDaniel` .seen azawawi 04:31
yoleaux I saw azawawi 12 Oct 2017 03:47Z in #perl6: <azawawi> github.com/Microsoft/language-serv...rotocol.md # Seems like the direction Atom is taking with future language integrations (autocomplete, find definitions, warning/error diagnostics, ..etc)
AlexDaniel` so Imlib2 is too bitrotten, MagickWand is too incomplete, and Image::PNG::Portable is too slow 06:07
AlexDaniel` maybe Image::Magick:from<Perl5> ? Hmm 06:10
jonas1 test.rg.avisita.com/admin/agenda/ 07:10
wander Can we add a precedence info of proto-regexes? 07:41
moritz wander: are you talking about grammars in general? or the Perl 6 grammar in Rakudo?
wander_ rule expression:sym<+> (precedence => 1) { <expression>+ % <sym>}; rule expression:sym<*> (precedence => 0) { <expression>+ % <sym>}; 07:42
something like these
moritz no 07:43
wander_ in Rakudo
moritz there' no operator precedence parser built into the proto mechanism
wander_ since I have several operators to handle, I try to find some method to deal with them in similar form 07:45
moritz a good OPP supports quite some features (precedence, different operator types (infix, prefix, postfix, circumfix), different associativies, including non-associative operators)
and we haven't found a good way yet to structure the API for one
moritz this is something that could be well experimented with in a module 07:46
wander_ sadly 07:47
robertle in perl6 (as oppsosed to perl5), can I create circular structures and they get garbage collected? or do I need to do some sort of reference weakening? 09:05
jnthn Yeah, circular structures are no problem 09:06
GC is based on reachability
robertle cool, thanks! 09:07
Geth ecosystem: 16a8a777be | (Alexey Melezhik)++ (committed using GitHub Web editor) | META.list
Move Sparrowdo::RemoteFile to CPAN
09:21
buggable New CPAN upload: Sparrowdo-RemoteFile-0.0.1.tar.gz by MELEZHIK cpan.metacpan.org/authors/id/M/ME/...0.1.tar.gz 09:31
melezhik buggable 09:32
sorry, last line has been printed occassionally , never mind it :)) 09:33
steeznson what would be a good beginners project for someone interested in learning Perl 6's unique functionality? I've written applications in Java and C# before 09:47
ufobat doing something with grammars? 09:59
steeznson i 10:01
ufobat and i am allways looking for ppl to contribute to Bailador ;) 10:02
steeznson *i'm interested in using unicode co-extensively with functions and exploring rationals
sena_kun parsing, concurrency, scripting as a good points to start... Basically, I'd better go with a software you want/need, not the one Perl 6 makes(does it?) you to write, I guess.
steeznson thanks, i'll mull it over. I'm taking more of an academic interest in perl for a grad school project. 10:03
vijayanat say hello 10:07
sena_kun m: say "Hello"; 10:08
camelia Hello
sena_kun oh, new hyper/race is merged. \0/ jnthn++ 10:10
sproctor So I've found a YAML parser that doesn't explode when I give it a swagger file. Next up cro include and I've got all I need (I think) to make a CRO swagger plugin. :D 12:08
(Sorry OpenAPI) 12:09
jnthn Nice 12:09
sproctor: By "cro include" do you mean the include feature in the router?
sproctor That's the bunny yes.
jnthn Nice timing, I implemneted it last weekend :) 12:10
sproctor Cool. I've been thinking about it for a while. The idea being you should be able to decorate the OpenAPi with OperationId's that match controller methods and the just do an include $plugin->parse_file( "path/to/file" ) 12:11
It'll generate all the routes and also validation sub types based on the spec. So you just need to create the controllers... That's my thinking anyway. 12:12
Similar to the Mojo plugin.
I prefer that way of working to the auto generation thing in most cases. As it means you can update your spec file later and not have to regenerate all your system. Anyway I'll keep an eye on things and play about with this over the weekend. 12:13
(Note to self. Finish slides for LPW) 12:14
DrForr Ack, I have to finish my abstract :/ 12:15
yoleaux 16 Oct 2017 09:39Z <HoboWithAShotgun> DrForr: On the site for Perl6Tidy (github.com/drforr/perl6-Perl6-Tidy) you claim it could be installed by zef, but it isn't in the eco system (modules.perl6.org/search/?q=tidy).
DrForr .tell HoboWithAShotgun I'll probably add it over the weekend. I was thinking it should do something less trivial than bracing, but braces are good enough for now, I suppose. 12:18
yoleaux DrForr: I'll pass your message to HoboWithAShotgun.
[Coke]_ (JSON::Fast) did you look at any of the other JSON modules, btw? 13:34
timotimo is there something wrong with JSON::Fast? 13:35
i'll probably merge the "allow sorting hashes by key" thing soon
[Coke] again finds the irc logs needing a verrrry wide display to be read. 13:47
jdv79 the web version? sometimes it doesn't fit. looks like an html fail. 13:48
timotimo it's pretty bad, yeah
raschipi What about the raw text version in the colabti.org logs? 13:50
buggable New CPAN upload: Sparrowdo-RemoteFile-0.0.2.tar.gz by MELEZHIK cpan.metacpan.org/authors/id/M/ME/...0.2.tar.gz 14:11
Geth ecosystem: c2fda9a172 | (Zoffix Znet)++ (committed using GitHub Web editor) | META.list
Add Reminders.pm6 to ecosystem

  "Reminders: Class for managing reminders about task and events"
See: github.com/zoffixznet/perl6-Reminders
14:15
[Coke] zoffix, testing out a modified update-rakudo that reinstalls all my modules. 14:48
(might be too much, may end up with just adding a bunch of installs of the stuff I know I use daily so that experimental stuff isn't resurrected.)
timotimo you're only reinstalling stuff to get the tests to run again, right? 14:49
or to get newer versions i guess 14:50
[Coke] reinstalling because my normal upgrade process wipes out all the installed modules. 14:51
timotimo ah ok
[Coke] is HyperIterable on nom yet, or still in a branch? 15:04
timotimo the hyper work has landed in nom 15:14
not sure if HyperIterable is an actaul thing 15:15
ilmari no, that got _deleted_ 15:16
it's HyperSeq now 15:17
ilmari (and RaceSeq) 15:18
jnthn HyperIterable was part of the old internals of the now-discarded hyper/race implementation 15:23
jnthn Well, the original idea was to make those "internals" API 15:23
But (a) the design didn't work so well, and (b) the new one seems to work better, but I'd like freedom to change it in the future 15:24
[Coke] HyperIterable is in the docs, failing an example build.
line 150, doc/Type/Any.pod6
timotimo ah, because there's a map method on HyperIterable 15:26
jnthn Huh, that sounds confused 15:28
[Coke] docs.perl6.org/type/Any#method_map 15:30
jnthn oh, right 15:33
No, it ain't going to work that way any more
The map method is on HyperSeq and RaceSeq now
haxmeister paste.pound-python.org/show/5qzOIR...0Kj0qtaEg/ 16:42
timotimo haxmeister: can you debug PowerShell? 16:49
in this case powershell syntax 16:50
haxmeister not much of a power user on windows 16:51
anything specific I might look at?
timotimo the thing that's exploding here is the sha256 implementation that uses the powershell so it doesn't have to pull in any dlls or slow perl6 code 16:52
timotimo all of this is just for downloading and verifying the dlls we have on the gtk-dlls.p6c.org site 16:52
you can manually git clone the gtk simple source repository, put the DLLs into the resources/ folder (and check the sha256 signatures manually) and then just throw out the Build.pm file 16:53
then "zef install ." inside the code directory
haxmeister kk
ugexe zef install . --/build 17:04
timotimo oh i did not know about that
nice.
ilmari --/?
timotimo maybe i should reach out to the twittersphere and find somebody to fix the sha256 thing 17:05
ugexe false
--/ = Bool::False, -- = Bool::True (when applied to a boolean type anyway)
ilmari what happened to the traditional --foo/--no-foo? 17:06
ugexe personally i don't like that convention, which means writing both params in your &MAIN
sub MAIN(Bool:D :$foo, Bool:D, :$no-foo) { } 17:07
ilmari why couldn't --no-foo set Bool:D :$foo to False?
ugexe because then MAIN is parsing its params fairly different than other subs 17:08
what happens if i have an actual :$no-foo for instance
b2gills --no-foo could be added as an option that you can turn on 17:14
ugexe well sure, but any GetOpt thing could technically be added as an option. 17:15
haxmeister timotimo: ok I cloned gtk-simple.. now what DLLs do I need to move, and where is this resources folder? 17:22
timotimo you'll find all the URLs in Build.pm
timotimo github.com/perl6/gtk-simple/blob/m...r/Build.pm 17:22
it turns out gtk-simple's Build.pm will try to load libgtk-3-0.dll and if it succeeds won't try to download anything or verify hashes 17:23
haxmeister man that's kind of a nuisance for windows users 17:35
nine Yep, it'll be good when Build.pm is finally gone 17:39
timotimo weird. it looks like all you have to do to get a hash is Get-FileHash /foo/bar -Algorithm SHA256 | select Hash | Format-List 17:41
why is our code so much longer?
function get-sha256 { param($file);[system.bitconverter]::tostring([System.Security.Cryptography.sha256]::create().computehash([system.io.file]::openread((resolve-path $file)))) -replace \"-\",\"\" }
for comparison
"this should work all the way back to powershell v1" that's why, of course 17:42
the ms docs for powershell only have it since 5.1 17:43
haxmeister v1 is an awfully long time ago 17:47
timotimo cool, you can PowerShell.exe -Version 2
timotimo i have a powershell version 4.0 here but it has Get-FileHash for some reason 17:49
oh
what the docs page means with "the closest version" is apparently "two versions further away from the closest version" 17:50
haxmeister /cygdrive/c/Users/jday/AppData/Local/Temp/libgmodule-2.0-0.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libgobject-2.0-0.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libgtk-3-0.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libiconv-2.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libpango-1.0-0.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libpangocairo-1.0-0.dll 17:51
/cygdrive/c/Users/jday/AppData/Local/Temp/libpangoft2-1.0-0.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libpangowin32-1.0-0.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libpixman-1-0.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libpng15-15.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libxml2-2.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libatk-1.0-0.dll
/cygdrive/c/Users/jday/AppData/Local/Temp/libcairo-gobject-2.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libffi-6.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libfontconfig-1.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libfreetype-6.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libgdk_pixbuf-2.0-0.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libgdk-3-0.dll
/cygdrive/c/Users/jday/AppData/Local/Temp/libgio-2.0-0.dll /cygdrive/c/Users/jday/AppData/Local/Temp/libglib-2.0-0.dll
oh sorry
have no idea why it did that
windows :/
liblzma-5.dll is missing?
patrickz Hey! I'm currently trying to get buttons on the raspberry pi working using interrupts. 17:54
Using NativeCall + the wiringPi library doesn't work, because wiringPi spawns a thead and thus causes a MVM panic when the interrupt hits. 17:55
Has somebody already done interrupt stuff with perl6 and can point into a direction? 17:56
timotimo oh, interrupts? like, posix signals?
patrickz think so
patrickz if I looked it up correctly wiringPi uses ioctl + poll on a filehandle internally 17:57
but there is no IO::Handle.poll in p6...
timotimo an strace would perhaps be enlightening
mst is it not possible to create a callback that can be passed via NCI and transfers back to the MVM thread first? 17:59
timotimo we do have something for cross-thread callbacks; at least there's a test that has some of those words in its filename 18:00
what does this mvm panic look like, patrickz? 18:01
haxmeister timotimo: thanks for the help.. zef completed with no errors.. will try some code here shortly
timotimo maybe all that's needed is to remove the \ in front of the " signs in the hash function? 18:03
patrickz timotimo: MoarVM panic: native callback ran on thread unknown to MoarVM 18:16
timotimo yes, that makes sense. you might have to use some other ipc mechanism 18:17
patrickz I thought about replicating the ioctl + poll thing in p6 land. 18:19
patrickz No idea if those functions are already available in p6. A quick grep on the source only yielded stuff in libuv :-/ 18:21
patrickz stuff in libuv/src/unix/poll.c is not called in moarvm. I guess p6 has no native poll support. 18:23
mst patrickz: poll() is just one way to determine if a filehandle is readable 18:24
patrickz: libuv has multiple backend implementations to do that, of which poll is one.
raschipi Probably because they develop to the lcd of linux and windows 18:25
mst patrickz: it would make no sense for moarvm to call that directly
timotimo shouldn't it be enough to have a thread that read()s in a loop?
steeznson Sorry to go a little off topic: Could someone briefly elaborate a little on grammars here? Am I correct in thinking they just function as classes of regexes?
mst raschipi: poll() is often not the best backend on linux anyway
raschipi you prefer real time signals? 18:27
timotimo grammars are actually just a kind of class, and regexes/tokens/rules in them are just a kind of method 18:27
steeznson @timotimo thanks that clears things up a little 18:28
timotimo being able to derive from and mix into grammars is super useful 18:29
patrickz timotimo: just looked at the strace of my little test script failing with wiringPi. It looks like the C poll() just does read() in a loop :-P 18:30
timotimo well, that's at least easy to replicate in pure perl6 code 18:31
steeznson great, looking forward to playing with grammars while exploring more of perl's functionality 18:31
raschipi At least now Linux hackers are working on making aio_read() actually non-blocking. 18:32
wander how to write rules to get the precedence expected? 18:40
rule mul { | <id> '*' <mul> | <id> '/' <mul> | <id> '%' <mul> | <id> }
I write down this and find multiplicative-expression assoc<right>
while expected is left 18:41
wander oh, it has nothing to do with precedence, but associativity 18:42
tyil travis-ci.org/scriptkitties/perl6-.../289455352 does anyone know why zef throws these errors? 18:48
ugexe probably because you are using stuff in your depends you should not be using 18:55
tyil I took it from the META6 spec, but thats out of date then 18:56
is there an up to date page with the full META6 spec 18:57
tyil the docs refer to design.perl6.org/S22.html#META6.json 18:58
which has the depends that I based that META6.json on
so I guess according to the docs, zef doesnt support stuff it should be supporting 18:59
ugexe those arent "the docs" 19:00
tyil ttps://docs.perl6.org/language/modules#index-entry-META6.json-META6.json refers to it
if docs.perl6.org isnt "the docs", what is?
ugexe what does that url, which is different than the previous url, point at that zef does not support? 19:02
tyil "See the full META specification for more details as well as further options available for use in META6.json files."
that "full META specification" is the S22.html
so the docs quite clearly state that S22.html contain the full spec, and the docs.perl6.org page is just a sample 19:03
ugexe well S22.html at the top clearly states its not the spec
tyil "clearly" in a bit overblown
ugexe its a bright red header 19:04
tyil since it says on S22.html it "may" be out of date
and the docs.perl6.org points to the "full META spec", to S22.html
ugexe "for specs, see the official test suite"
tyil which is not actually linked it 19:05
to
ugexe ok, anyways, you know why it doesnt work now.
tyil regardless, zef and the docs are out of sync, and the error message could certainly get some improvement
ugexe go for it
tyil well, I can guess why it doesnt work now, but I still dont have a clear link to the actual spec that zef supposedly supports 19:06
ugexe its supposedly in the code and CUR/rakudo
tyil if you have a real pointer I'd gladly check it out later today, but I'm getting a strong vibe that you're annoyed that I want better docs and zef to adhere to it 19:07
wander seems Perl 6 grammar use LL() grammar, how to build rules with left-associativity 19:08
tyil so I'll just leave it at "zef and the docs arent in sync", and let it rest for a bit
moritz or do something about it 19:09
tyil moritz: hence I asked for a clear link to the actual spec ;) 19:09
then I can update the docs with the correct link, write it out if I have more time
"its supposedly here and there" doesn't really help me 19:10
or saying "the docs are wrong", that doesn't really help anyone either
if he doesn't want to help me find out where the correct info is, I can't help to correct the docs
ugexe part of your responses are in part to how you framed your questions 19:11
tyil and because you seem pretty bent on not providing me with the simple answer to my simple question I wanted to let off some steam on both sides 19:12
because all I get from this is that you're annoyed that I found a possible issue with zef
and I'm getting annoyed at you evading trying to actually help me
tyil and this is helping nobody so far 19:12
raschipi if the answer was easy, it would have been provided to you already 19:13
tyil my question right now is, where can I find the META6.json spec directly 19:14
raschipi docs.perl6.org/language/modules#in...META6.json
moritz tyil: there's stuff on docs.perl6.org and in the synopsis 19:15
tyil which refers to the "full spec" which is either incorrect or not cerroctly implemented
moritz they are out of sync, and out of sync with the implementation
hence, the need for improvement which you mentioned
tyil yes, but I cant improve it if nobody wants to link me the full spec
ugexe again with the framing... maybe not-yet-implemented would have been a better choice
raschipi There's a note in the spec that says it's known to be incomplete.
samcv so i think the answer is just to not make sub items inside of depends
and then zef should be able to handle it?
raschipi I linked it above for you, that's it, all there is. 19:16
moritz tyil: sure you can, if you want to dig in and find out what's actually going on
ugexe yes
moritz this is not waterfall, where a Complete Spec[tm] is written, and everything is derived from it
perlpilot tyil: what samcv said.
tyil perlpilot: I know thats the solution, we're past that point for quite a while now
perlpilot tyil: ah, then update the docs to reflect reality :-) 19:17
tyil "zef does not support the full META6 spec", I can just add that yes 19:18
raschipi tyil: the docs were written as forward-compatible with the new and improved specification you're planning to write.
this way you'll won't have to rewrite the docs after writing the specification.
ugexe we dont know if it supports the full meta spec or not. what is in the SPECULATION documents is not neccesarily the spec 19:19
tyil it clearly says "specification" in the link on docs.perl6, not speculation 19:20
idk why you're making such a big deal out of a simple question
ugexe uh ok 19:21
tyil I'm willing to update the docs on this to avoid confusion, and if there currently is no actual spec to adhere to outside of the sample on docs.perl6.org, thats fine too
raschipi the word 'specification' on those documentsz is there because history. 19:22
perl6 actual specifications are written in the form of tests 19:23
perlpilot tyil: that section where it says "META specification" also talks about the "META.info file" too, so it's nicely out of date.
tyil perlpilot: yes, and it refers back to perl6.org which is also out of date, which just causes confusion
which is why I want to update it
but on this particular issue I can't seem to get a clear and straightforward answer 19:24
masak greetings #perl6 19:25
raschipi heya masak
tyil: there's no specification.
masak I just learned from a Guy Steele talk that the `?` in regexes was probably invented by Alfred Aho back in 1975
moritz \o masak, long time no see 19:27
masak yeah. I haven't defected or anything, I'm just being "put to good use" at $work :)
moritz btw, while you were away, I've written two books :-) 19:28
Geth doc: Tyil++ created pull request #1614:
Remove link to outdated META6 spec
masak I noticed! congratulations!
moritz++
moritz thanks :)
No. 2 isn't quite published yet 19:29
but the manuscript is submitted, and I keep running into topics that I could have covered
masak I've *started* on one of my two books, but there's ways to go still 19:30
moritz which one?
masak the cat theory one 19:31
raschipi about the physical state of cat matter? 19:32
are cats solid or liquid?
moritz and if it's too much at once, we'll replace it with more theory or less theory :-)
callyalater m: my $s = "p" ~ 0x304.chr; 20:08
camelia ( no output )
callyalater m: my $s = "p" ~ 0x304.chr; say $s; say $s.samemark("a"); say $s.ords; .say for $s.uninames; say $s.codes; say $s.chars; 20:10
camelia
p
(112 772)
LATIN SMALL LETTER P
COMBINING MACRON
2
1
callyalater m: my $s = "p" ~ 0x304.chr; say $s.uninames; 20:11
camelia (LATIN SMALL LETTER P COMBINING MACRON)
Zoffix tyil: the clear and straightforward answer is there's no "official" and complete spec for META contents at the moment. So far we're "winging" it with what zef does and what the slightly outdated speculation you saw documents. Making a concrete spec is listed as "would be good" on 6.d agenda github.com/perl6/6.d-prep/blob/mas...6json-spec , but I haven't
seen anyone willing to champion that. There's a separate #perl6-toolchain channel for this stuff. I don't sit in it, but it's possible some of the spec was discussed (logs irclog.perlgeek.de/perl6-toolchain/2017-10-18 )
tyil: so for your PR to docs, perhaps it better to clarify it's a work in progress and keep the link rather than removing it 20:12
geekosaur META format still changes every couple weeks, in fact
I'm not sure we're anywhere near the point of having a spec as such
Zoffix FWIW, it might be best to spec (as roast spec) just the basics, leaving extended stuff as a separate spec
Zoffix I'm not surprised. In logs I saw a person interested in the spec getting annoyed at vague responses 20:14
raschipi Well, it would help if someone in the chgannel knew the answer. 20:15
Zoffix :) 20:16
Zoffix goes back to the dev cave 20:18
pmurias if we are going to have a same name multiple authors for a given module which one will be installed when I type 'zef install Foo'? 20:33
tyil ZofBot: thats fine too, but as it stands its only causing confusion, so I want to change that at least 20:35
tyil samcv suggested to change the wording to say that there is a bigger target to work towards, but that it currently is not supported by the existing package managers, so one should stick to the sample on the docs.perl6 page 20:36
.tell zoffix thats fine too, but as it stands its only causing confusion, so I want to change that at least (I highlighted it to your bot by accident) 20:37
yoleaux tyil: I'll pass your message to zoffix.
ugexe pmurias: whatever the first repo/mirror in the zef config that reports the highest version 20:56
if you want a specific auth explicity, then be explicit. if an ecosystem allows itself to make the implicit invocation problematic then user/company policy would likely remove them from the config as a source 20:59
similar to when you say `use Foo` vs `use Foo:auth<bar>` 21:00
pmurias ugexe: so the plan is that the ecosystem will enforce name ownership? 21:13
ugexe that is up to whatevery ecosystems policy
its not enforced in the current p6c ecosystem, but its been able to maintain sanity 21:14
pmurias all the default zef ones must enforce that eventually 21:15
otherwise people who type zef install DBIish will install DBIsh:auth<evilbitcoinminer> 21:17
ugexe it just mirrors in general 21:20
you can change the default to true/false in the config for what does exists, or with --/p6c --/cpan etc (commands also pulled from config file). so the tooling to do that exists 21:23
github.com/ugexe/zef/blob/master/r...on#L32-L56
HoboWithAShotgun .tell skaji I hacked on mi6, check out github.com/skaji/mi6/pull/28 22:33
yoleaux 12:18Z <DrForr> HoboWithAShotgun: I'll probably add it over the weekend. I was thinking it should do something less trivial than bracing, but braces are good enough for now, I suppose.
HoboWithAShotgun: I'll pass your message to skaji.
Herby_ o/ 23:40
tyil \s
\o
*
Herby_ \o/
Anyone know if there is a plan to put a CSV module in with Rakudo Star? 23:41
tyil I dont know of any such plans, but there might be more knowledgable people around
Herby_ a module for basic CSV reading/writing would definitely be used I think 23:42