»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend! | YAPC::NA talk submissions due end of Month. Have you submitted your talk?
Set by toddr_ on 18 March 2013.
00:03 xenoterracide left 00:04 xenoterracide joined 00:06 colomon left, Chillance left 00:13 frd|afk left, frdmn joined 00:29 gcole left 00:31 census joined 00:34 census_ joined 00:37 census left 00:38 colomon joined 00:40 census_ left 00:45 adu_ joined 00:46 frd|afk joined 00:50 frdmn left 00:52 census joined
census anybody familiar with running perl out of debian and virtual VM ? 00:54
00:55 anuby joined 01:15 census left 01:18 araujo left 01:27 kurahaupo left 01:28 adu_ left 01:31 adu_ joined 01:33 araujo joined 01:39 kurahaupo joined 01:52 adu_ left 02:02 census joined 02:10 adu_ joined 02:15 preflex left 02:17 preflex joined 02:22 census left 02:26 adu_ left 02:33 lustlife joined 02:41 japhb_ left 02:57 orafu left 02:58 orafu joined 03:01 preflex_ joined, preflex left, preflex_ is now known as preflex 03:03 preflex left 03:04 preflex joined 03:11 adu_ joined, leprevost_ left 03:30 colomon left 03:34 lichtkind_ joined 03:37 lichtkind left 04:07 adu_ left 04:08 adu_ joined 04:22 anuby_ joined 04:23 anuby left 04:24 anuby joined 04:25 anuby left, anuby joined 04:26 anuby_ left, anuby_ joined 04:30 anuby left 04:37 thou left 04:49 dayangkun left 04:52 adu_ left 04:55 thou joined 05:02 thou_ joined, thou left 05:05 Tene left 05:23 thou_ left, thou joined 05:26 thou_ joined, thou left 05:33 kurahaupo left 05:44 dmol1 joined 05:47 thou_ left, thou joined 05:52 odoacre_ is now known as odoacre 05:53 SamuraiJack joined 06:00 adu_ joined 06:03 anuby_ left 06:06 rindolf joined, drbean left 06:07 kaleem joined, thou left, thou_ joined 06:13 drbean joined 06:28 thou_ left, thou joined 06:33 adu_ left 06:35 dmol1 left 06:37 FROGGS joined, SamuraiJack_ joined 06:38 xalbo left 06:40 SamuraiJack left 06:45 baest_ is now known as baest 06:46 SamuraiJack_ left, SamuraiJack__ joined 06:48 thou_ joined, thou left 06:58 thou_ left 06:59 thou joined 07:04 domidumont joined, thou left 07:06 daxim joined 07:08 domidumont left 07:09 domidumont joined
moritz sorear: fwiw I've stopped people to come here for ilbot help, there's now #ilbot 07:18
jnthn morning, moritz o/ 07:19
moritz \o jnthn 07:20
timotimo good day perl6ers 07:21
i feel a little like my fascination or at least excitement for perl6 is running a bit dry. anyone feel like wowing me back into full perl6 fanboy mode? 07:22
diakopter .give timotimo hype hype hype 07:25
masak morning, #perl6
timotimo or maybe hype hype hype just leaves me burnt out again more quickly and i should keep kind of a warm flame going? 07:27
moritz yes, maybe
and don't forget to take pauses and vacations etc. 07:28
\o masak, timotimo, diakopter, *
diakopter .give timotimo perpetual energy machine
sorear o/ moritz, jnthn, timotimo, masak 07:30
arnsholt timotimo: I think the warm flame thing is best in the long run, yeah
masak timotimo: what moritz said. pauses matter.
jnthn Indeed. :)
masak timotimo: also, Perl 6 is not a means to an end. never was.
timotimo right. it's not like all of a sudden i'll carry perl6 to 100% production readyness over a month :P 07:31
masak timotimo: all we want to do is make programming better.
jnthn is always more motivated/productive after a break.
timotimo i suppose my frustration with not being able to handle python on a big project lead me here, so making programming better is something i like
arnsholt timotimo: I've been thinking about this iperl6 thing BTW. D'you think you could give me a quick sketch of how ipython segments things? 07:33
timotimo uh, i guess. segment on what level?
arnsholt On the one hand, having everything in one process would be really convenient, but on the other you sort of want to separate the interpreter doing the evaling from the parts doing IPC 07:34
timotimo well, in ipython itself, there is one process that does the ipc and the evaling
arnsholt Especially you don't really want the REPL code to be able to trample the IPC talking to the frontend
Right. So with sufficient knowledge of the internals, you can craft code that does weird things to the IPC bits? 07:35
timotimo well, the kernel is doing some "i'll give you my own globals and locals" thing
but yeah, that's not a safe sandbox 07:36
on the other hand, you can *hardly* write code that tramples on the IPC by accident
arnsholt Right
timotimo and nobody would ever trust the kernel thing to safely separate user supplied code from the system
arnsholt True that
timotimo so that's not something ipython tries to do. it's not designed to be safe from manipulation, it's more like a super advanced repl that you can do awesome stuff with 07:37
arnsholt Not unreasonable
07:38 SamuraiJack__ left, woolfy1 joined 07:39 SamuraiJack__ joined 07:52 rindolf left 07:53 rindolf joined 07:55 fhelmberger joined
lizmat rn: my %bar; say %bar.end 07:55
p6eval niecza v24-35-g5c06e28: OUTPUT«Unhandled exception: Unable to resolve method end in type Hash␤ at /tmp/AT1vB_nu9S line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4299 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4300 (module-CORE @ 583) ␤ at /home/p… 07:56
..rakudo ffe441: OUTPUT«-1␤»
lizmat I'm surprised to see .end work on a hash
masak too
arnsholt jnthn: How hard would it be to make something vaguely similar to rakudo-debugger also be implemented using Perl 6 code?
lizmat my %bar= :foo<bar>; say %bar.end 07:57
rn: my %bar= :foo<bar>; say %bar.end
p6eval niecza v24-35-g5c06e28: OUTPUT«Unhandled exception: Unable to resolve method end in type Hash␤ at /tmp/fE5JAw9x7R line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4299 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4300 (module-CORE @ 583) ␤ at /home/p…
..rakudo ffe441: OUTPUT«0␤»
FROGGS arnsholt: pretty hard, right? because you have no access to the underlying guts
moritz .end is just tied to the listy view that a hash can also have
jnthn lizmat: It probably is dong .list.end
*dong
... 07:58
07:58 rindolf left
jnthn *doing 07:58
masak :P
moritz :-)
FROGGS .oO( ring a ding dong? )
lizmat :-)
good morning all :-)
moritz or just .elems - 1
FROGGS morning \o/
jnthn
.oO( need to re-learn typing... )
07:58 rindolf joined
moritz more like "morning /o\ " here 07:58
jnthn arnsholt: The Big Issue at the moment is that handling NQP-ish objects from Rakudo is a pain. 07:59
arnsholt: I'm gonna deal with that Very Soon, though.
timotimo jnthn: i discovered that when i tried to make ctxsave work properly for my own eval in perl6
FROGGS jnthn: I have trouble installing the v5 module :/
lizmat the only place I see .end mentioned, is in S02, and then only one line
jnthn arnsholt: Your VMArray branch needs it (and you've already done some of the hard work) and Rakudo on JVM needs it.
lizmat so at least it is underspecced :-)
arnsholt jnthn: Yeah, I've realized that branch is wildly misleadingly named as well 08:00
lizmat rn: my %bar= :foo<bar>; say %bar.list.end
p6eval rakudo ffe441, niecza v24-35-g5c06e28: OUTPUT«0␤»
jnthn arnsholt: Well, and it may want to be multiple branches.
lizmat rn: my %bar; say %bar.elems 08:01
jnthn arnsholt: That is, one for HLL stuff, one for the VMArray and nqp::list switch.
p6eval rakudo ffe441, niecza v24-35-g5c06e28: OUTPUT«0␤»
jnthn arnsholt: Are the two fairly split up commit wise already? If so, we just rebase -i our way there... :)
lizmat fitness&
arnsholt Yeah, they're more or less separate I think 08:04
But there's a pile of other commits as well that I think want to be reverted
I've hacked away at the code to get NQP actually running again, but I suspect many of those changes are better handled differently
08:05 rindolf left, rindolf joined
arnsholt And yeah, I've been considering making a new branch and rebasing the old stuff into better shape =) 08:05
08:06 arlinius left
masak funny alternate timeline: www.dadhacker.com/blog/?p=1976 08:07
08:10 SamuraiJack joined
labster hi all 08:10
08:11 rindolf left
labster is that just good general advice to avoid Stallman? 08:11
masak think so. 08:12
08:12 rindolf joined
labster files a Perl 5 bug for File::Spec while working on the P6 version. 08:13
08:13 SamuraiJack__ left
labster some of that code for OS2/VMS/Epoc hasn't been looked at in years. 08:14
Is Perl 6 seriously going to target any ecosystem for Perl 6 that isn't POSIX, Win32, or Cygwin? 08:16
I don't want to put in a lot of effort for something that will never be used. 08:17
masak labster: I'm not aware of any OS2 Perl 6 users. 08:19
08:20 rindolf left
masak invokes YAGNI 08:20
08:20 rindolf joined 08:21 nbezzala joined
nbezzala irc://irc.perl.org 08:22
08:24 dakkar joined
masak nbezzala: welcome. that's not how it works. :) 08:24
nbezzala thanks. I know now :)
masak nbezzala++ 08:25
08:25 nbezzala left 08:29 nitish joined 08:30 nitish is now known as nbezzala 08:31 rindolf left, rindolf joined 08:39 rindolf left, sqirrel joined, rindolf joined 08:48 rindolf left 08:49 rindolf joined 08:55 dakkar left 08:56 dakkar joined
FROGGS labster: abandon OS2 and others... 09:01
nwc10 VMS is very active in Perl 5 land. I would have run NQP-JVM on VMS, except that the system I have access to only has Java (1.)6 09:02
I think the key things are
1) maintain concepts of volume, directories and filename
2) catfile and catdir need to be different operations, even if on most platforms they are the same 09:03
and be aware that on Win32 at least, it's a current directory per volume.
FROGGS nr: our @a = 7 09:05
p6eval niecza v24-35-g5c06e28: ( no output )
..rakudo ffe441: OUTPUT«No such method 'STORE' for invocant of type 'Any'␤ in block at /tmp/0wduKWCGi5:1␤␤»
FROGGS jnthn: can you give me a hint what I need to do there?
jnthn: because copying some code from the scope-'my' branch of sub Perl::Actions::declare_variable to 'our' doesn't help 09:07
moritz FROGGS: do you understand the underlying problem? 09:08
jnthn Wildly copy-pasting code rarely fixes deeper problems :) 09:09
FROGGS moritz: no
that's why I ask
jnthn Well, the key probably lies in the semantics of Stash.
moritz FROGGS: the problem is that symbol tables (Stash) are just glorified hashes 09:10
FROGGS: and support no type constraints
FROGGS: and no defaults
jnthn Which at present don't pay any attention to the names of the things being put in them
moritz so my @a is actually closer to my Positional @a := Array.new()
(ok, Positional @a is wrong, but you get the idea) 09:11
masak why does everyone keep making that mistake? :)
moritz there's a type constraint of Positional, and a default of Array.new
and while lexical scopes support that concept, Stash does not
so 'our @a' initializes @a to Any 09:12
and list assignment calls @a.STORE
jnthn I think the eventual solution is that we want Stash to have a smarter at_pos that knows how to vivify things besides scalar containers, going on sigil or so
moritz jnthn: at_key instead?
jnthn uh, yes.
But today, Array and Hash containers lack any container-y thing we can attach such whence-y behavior to. 09:13
Which is one of the things pmichaud and I have been pondering may change.
FROGGS well, sounds like a plan 09:20
I come around that problem because this is a common Perl 5 idiom: BEGIN { unshift @INC, '.' }, and declaring+assigning to our @vars 09:21
for @INC I can dispatch that to @*INC though 09:22
<jnthn> Wildly copy-pasting code rarely fixes deeper problems :) <--- but if you do that you get a feeling what you can do with the current ast or not, because it can't handle instantiate_generic for exampel 09:24
jnthn FROGGS: Yeah. We have some interesting design questions here, but largely they boil down to, "how smart should stashes be"
Oh! We could say that writing "our @a" actually installs something into the appropriate Stash at compile time 09:25
We could even bind a Scalar container into it to support the type constraints
Then we always know it's already there by the time we hit compile time...
labster FROGGS: Actually, from looking at the P5 OS/2 module, I decided that Win32 could probably handle it better than OS2.pm can. Some code seems to handle UNC paths, others ignore it. (Not to mention the random no-op substitutions). 09:30
nwc10: I am aware of the per-volume cwd thing on windows, though I don't have a working Cwd module, so I haven't thought *that* much about it. 09:31
09:34 lichtkind_ is now known as lichtkind
labster VMS is up next, though I suspect it might give me some trouble. The question I wanted to deal with is whether having IO::Path split into IO::File and IO::Dir subclasses (as currently specced) makes more sense, or to deal with just IO::Path objects. 09:38
09:40 am0c joined
FROGGS would be nice if you give the user an IO::File if it actually is a file 09:42
so a user could smartmatch against it and doesn't has to check on its own
nwc10 is "File" vs "Dir" the correct abstraction? After all, on this file system I can also have at least symlinks, named pipes and character special devices 09:43
(to list things that non-root programs might want to interact with) 09:44
labster ~~ IO::File vs. ~~ :f?
FROGGS I'd say a symlink to a file is a file, pipes and block devices are more like a file than a dir
hmmm 09:45
labster: you have a point there
nwc10 yes, with that I'd agree. directories hold other files, and have a different set of operations. The others all have a similar set of operations
*until*
you start having systems that can do fun things such as map archive files as directories
(stuff can behave as the union of file-like and directory like)
FROGGS ya, but there is some glue code between that 09:46
09:46 seq99 joined
labster knowing that it's a file could save a stat call (well, 2 calls, given our current implementation). But on most systems (i.e. not VMS), you don't know if a path is a file before calling stat anyway 09:49
09:53 orafu left 10:06 Gruu_ joined 10:21 stevan_ left, stevan_ joined 10:22 stevan_ left, kresike joined
kresike hello all you happy perl6 people 10:22
FROGGS hi kresike 10:25
kresike hello FROGGS o/
10:25 kurahaupo joined 10:27 alino2 joined 10:28 kurahaupo left 10:29 lizmat left, woolfy1 left 10:59 xinming left 11:03 xinming joined 11:18 alino2 left 11:20 colomon joined 11:21 Bzek joined 11:23 census joined 11:41 shlomif joined, rindolf left 11:49 shlomif left, rindolf joined 11:51 rindolf left 12:04 nbezzala left
dalek : bb0fa8e | (Tobias Leich)++ | lib/Perl6/P5 (2 files):
dispatch @INC to @*INC
 12:16
12:24 kaleem left
[Coke] I am dealing with java's "magically map zip files as filesystems" stuff in 1.7 now. Had to workaround their desire to slurp the file into memory before writing it out to the zip file. :P 12:33
(so don't do that particular mistake)
masak leaky abstractions strike again. 12:34
12:38 Chandra joined
Chandra hi 12:38
hy
[Coke] hi
Welcome to #perl6
Chandra how i can create IRC Channel?
thanks
how i can create IRC Channel?
[Coke] try to /join a channel that doesn't exist. 12:39
tadzik did you wish to create a Perl 6 channel? There already is one :)
Chandra no
i want to create IRC Channel for samp server
masak Chandra: welcome. :) it is as they say, just do "/join samp-server" or so 12:40
Chandra not create?
masak entering a nonexistent room creates it. 12:41
Chandra how?
masak that's an interesting question, though to a first approximation, that's just how IRC does it.
Chandra ,,,
... 12:42
masak whether or not you understand the mechanics of it, that's how it is done. :)
FROGGS Chandra: klick on #samp-server
(maybe right click)
masak oh, a webchat user.
Chandra i am join samp-server now
masak \o/
FROGGS :o)
masak FROGGS++ 12:43
Chandra: do you live in Russia?
12:45 Chandra left
pmichaud good morning, #perl6 12:49
moritz good am, pm
masak pmichaud! \o/ 12:52
FROGGS hi pmichaud 12:54
12:56 fhelmberger left 12:57 stevan_ joined
[Coke] how does #perl6 end up being the first place people who don't know how to use IRC come to? 12:57
weird.
FROGGS [Coke]: no idea 12:58
moritz [Coke]: maybe some weird google karma sending people the freenode web chat when they google for IRC + something
masak maybe all channels get that to some extent. 13:00
or all popular (>150 users) channels, or something. 13:01
13:01 sqirrel left
masak speaking of which, why are there always 200 nicks in here? is there an upper limit of 200? 13:02
pmichaud google "irc newcomer perl" and see what you get :-)
13:02 gtodd joined 13:03 stevan_ left
gtodd is there a single place/page where one can find "testimonials" re: perl 6 ? 13:03
pmichaud afk, kid to school 13:05
gtodd people just use perl5 and let it speak for itself but "new" languages seemingly have marketing departments ... fsharp.org/testimonials/
masak gtodd: hi, I'm a satisfied Perl 6 user. what would you like to hear? :) 13:06
gtodd heheh please describe your latest "enterprise solution" or theoretical physics breakthrough haha 13:07
masak :)
13:08 fhelmberger joined
masak gtodd: I am runnning my blog on Perl 6. have been since September 2010. it is very stable and reliable. I have full control and can extend it however I want. 13:08
gtodd: strangelyconsistent.org/blog/dog-fo...rl6-flavor
gotta catch a train. I bet there are others in here who can give you more in terms of testimonials, however. 13:09
gtodd I have good news and bad news ... parrot and perl6 allowed me to see several quasars ... unfortunately they are headed this way and will arrive within the next 50000 years ... so there may not be time to develop perl7 ... badumdump! 13:10
masak: thanks for that
13:11 arlinius joined
tadzik gtodd: maybe not your usual testimonial, but I used to make my living on Perl 6 and I've built my entire professional career on that :) 13:14
so at least the educational value is high
13:16 bluescreen10 joined 13:18 kaare_ joined
gtodd tadzik: :) good ... I notice Microsoft technologies (and other languages too but esp. Microsoft) seem to target "molecular analysis" "genetics data" etc. etc. and areas where perl is strong and sometimes wonder if it is a concerted strategy. Personally I just have tiny scripts that have been fun and fairly easy to convert to perl6 code so I have no career at stake :) 13:18
13:20 PacoAir joined
timotimo well, many (at least a few, didn't check in a long time) of the problems from rosalind have solutions in perl6 13:21
13:42 mtk left 13:46 rindolf joined 13:57 stevan_ joined, stevan_ left 14:01 Gruu_ left 14:04 prammer__ is now known as prammer 14:06 fhelmberger left 14:10 seq99 left, seq99` joined 14:11 census_ joined, xilo left 14:12 census left, census_ left, census joined 14:22 xalbo joined 14:23 Psyche^ joined, fhelmberger joined 14:26 Patterner left, Psyche^ is now known as Patterner 14:33 daxim left, daxim joined
[Coke] Hey, nwc10, you gonna help us add EBCDIC to rakudo? 14:35
[Coke] runs.
14:39 xilo joined 14:42 mtk joined 14:50 mtk left
japhb [Coke], as well you should run! 14:52
nwc10 [Coke]: I have no access to any EBCDIC platforms. Even if I did, I wouldn't help. It's a counter-productive diversion of limited resources. 14:53
14:53 mtk joined
masak ah, the old "serious reply to tongue-in-cheek question" comeback. :) 14:54
daxim I imagine he's grinning like he's wont to 14:57
nwc10 it's the polite version for a publicly logged channel 14:58
now, please keep distracting me whilst I type my root password into the wrong window...
[Coke] nwc10: there is still ebcdic support in p5, aye? 14:59
nwc10 s/aye/sigh/ 15:00
[Coke] ISTR you were nosing around ripping it out at one point.
nwc10 it was severly bitrotted
daxim user: chocolate pw: teapot
nwc10 someone from IBM, at the 11th hour and then some
has started actually trying to get it back to working
but this also consumes committer time 15:01
and it's not clear, at the end of this, whether anyone will set up a smoke server to let us know that it's still working
meanwhile, as long as the UTF-EBCDIC code remains (not just the EBDIC vs ASCII code), the UTF-8 handling code has to be aware of EBCDIC 15:02
which is a lot of complexity cost, for the benefit of about 0 users. "About", because it's just recently moved from 0 active to 1 active. 15:04
after maybe 5 years of 0 contributing users.
[Coke] the web page for utf-ebcdic seems to eschew it in favor of utf-16 15:05
kresike bye folks
15:05 kresike left
nwc10 I wasn't aware of that. But as far as the Perl 5 internals go, they make a massive set of assumptions that the internal character set they are dealing with is some variant where characters in the ASCII set (whether encoded as ASCII or EBCDIC) are represented as 1 octet 15:07
so using UTF-16 on an EBCDIC platform is about as useful as using UTF-16 on a "UTF-8" platform
anyway, the flippant answer is "no, I won't help you add it. But I'll help you remove it"
15:08 stevan_ joined
[Coke] nwc10: good answer. I only brought it up because some folks on my alum chat server were talking about ebcdic in perl earlier this morning. 15:10
15:12 salv0 left, stevan_ left 15:15 domidumont left 15:22 yx left 15:23 awwaiid left 15:26 dmol joined 15:27 kurahaupo joined 15:28 kurahaupo left, salv0 joined, yx joined 15:29 kurahaupo1 joined 15:31 xenoterracide left, xenoterracide joined 15:33 renormalist joined, kaleem joined 15:34 kurahaupo joined 15:35 kurahaupo1 left 15:36 xenoterracide left, xenoterracide joined, mmcleric left 15:38 thou joined 15:40 xalbo is now known as xalbo_ 15:41 xalbo_ left 15:45 am0c left 15:58 fhelmberger left 15:59 dmol left, kaleem left 16:04 kurahaupo left 16:06 dmol joined 16:11 FROGGS left
arnsholt jnthn: A question about hllconfig stuff. On JVM things are stored in the HllConfig class, which lives in a couple of different things AFAICT. On Parrot, does it make sense to just have a HLL config hash living in the global context and fetching stuff from it with ops? 16:16
Or did you have something different in mind?
16:22 kurahaupo joined 16:27 awwaiid joined 16:49 lizmat joined 16:50 Chillance joined
[Coke] takes a nap while rakudo does a build and test run. 16:54
16:58 PacoAir left, PacoAir joined 17:02 daxim left 17:03 lizmat left, lizmat joined 17:04 FROGGS joined 17:05 dakkar left 17:15 kurahaupo left 17:18 lizmat left
jnthn arnsholt: Well, lookups of hll stuff can be hotpath-y. And doing everything with virtual calls and hash lookups is kinda a performance anti-pattern. 17:19
arnsholt That's definietly true 17:21
nwc10 but think of the poor CPU manufacturers. And their shareholders
17:21 domidumont joined
arnsholt What things in NQP/Parrot correspond to the JVM thingies we hang the configs off, though? 17:22
jnthn arnsholt: That's the bit I need to ponder a little. 17:23
Parrot already has some notion of HLL stuff.
arnsholt True, true 17:24
jnthn And we already convey current HLL that way.
arnsholt Indeed 17:25
jnthn We maybe shouldn't worry too hard on the performance angle of this, though. I mean, I suspect even if it's hot-path-ish, it'll still be swamped by stuff like Parrot's invocation overhead. 17:27
arnsholt That's a possibility as well, yeah
In that case I'll implement it that way for now, and we can worry about refactoring once we have something measurable 17:28
jnthn *nod*
arnsholt I'm timesharing my NQP tuits with the annoying callback segfault as well 17:29
I've no idea what's going on =(
jnthn To me, much of this work is really about getting Rakudo/NQP on Parrot sufficiently aligned with how it'll be on every other backend so we don't have to fight bugs/breakages or maintain different versions of stuff 'cus of discrepancies. 17:30
arnsholt Definitely
jnthn On the callback thing - it isn't a missing write barrier, I guess? 17:31
arnsholt That's another reason I want to set up something emmentalery for NQP and Rakudo
Oh!
That might be it
masak .oO( Emmentalery, my dear Watson )
17:32 thou left
geekosaur pretty cheesy if you ask me 17:32
arnsholt (Re: Emmentalery) I suspect that with multiple backends, we really can't rely on developers testing manually to keep tabs on the state of things, which is more or less what we have ATM 17:33
dalek rl6-roast-data: 145a7c1 | coke++ | / (3 files):
today (automated commit)
jnthn arnsholt: Yes, agree it will become very important.
17:33 SamuraiJack left, thou joined 17:36 SamuraiJack joined 17:38 lizmat joined 17:43 lizmat left 17:44 djanatyn joined 17:55 atta_ left 17:57 atta_ joined 18:10 SamuraiJack left
jnthn dinner & 18:10
18:12 lizmat joined 18:17 sqirrel joined 18:18 spider-mario joined 18:19 SamuraiJack joined 18:29 SamuraiJack left 18:48 Tene joined
arnsholt Further continuing my testing thoughts, what I want, ideally, is something I can set up to be run as a cronjob every night on my work machine for example 18:48
masak I wonder how self-contained it could be made. 18:50
something that you set up once, and then it builds and tests nqp and Rakudo.
regularly.
[Coke] masak: kind of like perl6-roast-data? :P 18:52
arnsholt masak: Exactly! Since I'm completely Unix-brainwashed, that means cronjob =)
[Coke] (I'll definitely add rakudo-jvm once that can run the test suite.)
arnsholt And with some further hackery (and a webservice on feather), it can automatically submit the reports afterwards and we can look at them in the aggregate 18:53
I want it to be as self-contained as at all possible, so that it's essentially fire-and-forget 18:54
Both for one-off runs and repeated ones
masak yes, that's what I mean. essentially a package that you can download and deploy easily. 18:58
arnsholt Yup 19:01
I think that's the only way we can expect to get non-core people to contribute sporadically 19:02
colomon I've been meaning to get a cron smoke test running myself, if someone else does the hard work I'd certainly be happy to have my Linux box run it. 19:03
19:06 census left
raiph .oO ( p6 might be a good fit for Ubuntu Kylin ) 19:10
diakopter "Another assumption has been that if we don't talk about something in these Synopses, it's the same as it is in Perl 5. Soon we plan to fill in the gaps with the Perl 5 details though." - S01 - I wonder how big this set of features is... very big at all? 19:11
[I'm sure at the time the last sentence was written, the language was thought to be more similar to Perl 5, and the Synopses were far shorter.] 19:12
masak diakopter: it almost feels like that first sentence is more for guiding language design than for explaining gaps in the synopses. 19:13
but yes, maybe that's an after-the-fact thought.
lizmat gong through S02, I found some places that referred to perl 5 19:15
maybe they should be improved to not mention perl 5 ?
masak mentioning Perl 5 per se isn't so bad. no-one is denying the intellectual heritage between the two languages. :) 19:17
flussence p6 has the unenviable task of specifying unspecced behaviour that people have been relying on for 2 decades, sort of like the WHATWG's work... (except nobody complained that they took over 10 years)
PerlJam masak: aye, but sentences like "Non-container types define truthiness much as Perl 5 does." could use a little less P5 IMHO 19:18
masak oh, agreed.
diakopter well "Soon" in S01 definitely needs fixed, unless the comedic effect is now intended - www.nntp.perl.org/group/perl.perl6....26170.html 19:19
PerlJam "Soon we plan", not "soon we will fill the gaps" Maybe the plan has already been sooned :) 19:21
19:21 domidumont left
masak heh - "needs fixed". :) 19:22
19:26 census joined
labster good afternoon, #perl6 19:29
19:29 rindolf left
labster I've decided that I need to improve S05 by defining the backslash sequences somewhere, instead of just hitting man perlreref. 19:32
PerlJam labster: you mean for the character class shortcuts? like \s and \w and such? 19:37
labster Yeah. It also needs to mention what was taken out in S02, like \l and \u. 19:38
PerlJam I *think* backslash sequences are _only_ character class shortcuts now.
\b is now <|w>
(for instance)
and yes, \l \L \u and \U are gone 19:39
FROGGS \Q and \E are gone too AFAIK 19:40
labster yeah.
"The \L...\E, \U...\E, and \Q...\E sequences are gone. In the rare cases that need them you can use <{ lc $regex }> etc. The \G sequence is gone. Use :p instead"
PerlJam If that's true, then a nice section header would be "Character class shortcuts" :-) 19:41
labster It's tells you what's gone... just not what's present!
census labster ++ 19:42
[Coke] karma bot? 19:47
timotimo oh, you can just call lc $regex? that sounds quite awesome 19:49
hm, i wonder now what kinds of manipulations one can apply to a regex in perl6 19:50
moritz you can call it
you can embed it into another regex
which is roughly the same that you can do with any other callable 19:51
timotimo is there any sense to wrapping a regex?
flussence
.oO( regexes - boil 'em, mash 'em, stick 'em in a stew... )
 19:52
timotimo :) 19:54
.o( wut's rexes eh? )
geekosaur rex regis
moritz timotimo: as in / ^ \w+ $other_regex / 19:55
tadzik rebellis! \o/ 19:56
timotimo moritz: i kind of thought about actually using the wrap method on the regex object
19:57 orafu joined
census <[Coke]> :( 19:58
lizmat rn: my Set %set= <foo bar baz>; say %set 19:59
p6eval niecza v24-35-g5c06e28: OUTPUT«Unhandled exception: Unmatched key in Hash.LISTSTORE␤ at /tmp/r0GKCNucO2 line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4299 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4300 (module-CORE @ 583) ␤ at /home/p6eval/niec…
..rakudo ffe441: OUTPUT«Nominal type check failed for parameter '$x'; expected Set but got Str instead␤ in method STORE_AT_KEY at src/gen/CORE.setting:6796␤ in method STORE at src/gen/CORE.setting:6730␤ in block at /tmp/moV6hgBUG5:1␤␤»
lizmat seems I miss some adverb to turn the quoted word list into a set 20:00
I know we have map, but wonder whether you can do without in Perl 6
timotimo rn: my Set $s = <foo bar baz>; say $set;
p6eval niecza v24-35-g5c06e28: OUTPUT«===SORRY!===␤␤Variable $set is not predeclared at /tmp/T5HcagPruI line 1:␤------> my Set $s = <foo bar baz>; say ⏏$set;␤␤Potential difficulties:␤ $s is declared but not used at /tmp/T5HcagPruI line 1:␤------> my S…
..rakudo ffe441: OUTPUT«===SORRY!===␤Variable '$set' is not declared␤at /tmp/Uvw2rPniq_:1␤------> my Set $s = <foo bar baz>; say $set⏏;␤ expecting any of:␤ postfix␤»
timotimo er, oops
rn: my Set $s = <foo bar baz>; say $s; 20:01
p6eval rakudo ffe441: OUTPUT«Type check failed in assignment to '$s'; expected 'Set' but got 'Parcel'␤ in block at /tmp/yKo2E3Uixh:1␤␤»
..niecza v24-35-g5c06e28: OUTPUT«Unhandled exception: Nominal type check failed for scalar store; got Parcel, needed Set or subtype␤ at /tmp/OWTHpEM4iV line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4299 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 43…
timotimo hm, well ...
lizmat indeed, sorry for the variable mismatch :-)
timotimo no, i made the variable mismatch
er ... or something
PerlJam r: my $s = set <foo bar baz>; say $s;
p6eval rakudo ffe441: OUTPUT«set(foo, bar, baz)␤»
lizmat aha! 20:02
20:02 leont joined
colomon apologize for not reading at the moment someone was asking Set questions.... 20:04
PerlJam colomon: that's an odd thing to apologize for :)
colomon PerlJam: there was a lot of experimentation there doing stuff I could have answered in a moment. :)
timotimo the solution was surprisingly obvious and should have been clear to me the moment i saw what the error message was :| 20:09
PerlJam timotimo: well, if there's one thing I've learned over the years on #perl, is that even "obvious" things sometimes aren't. that's why the Postel Principle is so important. 20:11
(well, one of the reasons it's important :) 20:12
lizmat rn: my $s= <foo bar baz> :set; say $s
p6eval rakudo ffe441: OUTPUT«===SORRY!===␤You can't adverb that␤at /tmp/NkRQ2L9Gzn:1␤------> my $s= <foo bar baz> :set⏏; say $s␤ expecting any of:␤ pair value␤»
..niecza v24-35-g5c06e28: OUTPUT«===SORRY!===␤␤You can't adverb that at /tmp/DLpGhW9nWY line 1:␤------> my $s= <foo bar baz> :set⏏; say $s␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1443 (die @ 5) ␤ at /hom…
timotimo ah, mhm.
lizmat rn: my $s= <foo bar baz>.set; say $s
p6eval rakudo ffe441, niecza v24-35-g5c06e28: OUTPUT«set(foo, bar, baz)␤»
moritz consistent error messages \o/
lizmat hehe 20:13
PerlJam (using "adverb" as a verb)++ :-)
[Coke] you mean "the Evil Postel Principle". :P 20:14
lizmat was never any good with natural languages 20:15
timotimo moritz: the number of newlines don't match up! :o
lizmat anyway, what I wanted was a quick way to check if all elements in an array where the same 20:16
rn: my @a= <foo foo foo>; say @a.set.elems == 1
p6eval rakudo ffe441, niecza v24-35-g5c06e28: OUTPUT«True␤»
lizmat is what I winded up with in the end
timotimo r: my @a = <foo foo foo>; say all(@a) == all(@a); my @b = <foo foo foo bar>; say all(@b) == all(@b); 20:18
p6eval rakudo ffe441: OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '⏏foo' (indicated by ⏏)␤ in method Numeric at src/gen/CORE.setting:10620␤ in sub infix:<==> at src/gen/CORE.setting:2919␤ in sub infix:<==> at src/gen/CORE.setting:2917␤ in s…
timotimo er, wrong comparison
r: my @a = <foo foo foo>; say all(@a) eq all(@a); my @b = <foo foo foo bar>; say all(@b) eq all(@b);
p6eval rakudo ffe441: OUTPUT«all(all(True, True, True), all(True, True, True), all(True, True, True))␤all(all(True, True, True, False), all(True, True, True, False), all(True, True, True, False), all(False, False, False, True))␤»
timotimo r: my @a = <foo foo foo>; say so all(@a) eq all(@a); my @b = <foo foo foo bar>; say so all(@b) eq all(@b);
p6eval rakudo ffe441: OUTPUT«True␤False␤»
timotimo :D
this solution is probably only 100000x slower
(for sufficiently small input lists) 20:19
PerlJam github.com/perlpilot/benchmark/ time them :-)
timotimo good idea 20:20
the algorithmic complexity class for one is far below the other, though
so not only is there probably a huge constant factor due to how slow junctions are (have to be constructed at each run), but also the efficiency of using a hash will cause some major differences 20:21
10000 may have been a bit much for such slow code :D 20:27
20:28 sorear left 20:30 kurahaupo joined, sorear joined
lizmat rn: my @a= <foo> xx 10000; say @a.uniq.elems == 1 20:34
p6eval rakudo ffe441, niecza v24-35-g5c06e28: OUTPUT«True␤»
lizmat just learned of using .uniq rather then set 20:35
[Coke] you could also use a KeySet. 20:38
er... sorry, a Bag.
FROGGS a BagSet? 20:39
masak .oO( $bilbo = Bag.end ) 20:40
[Coke] (the bag lets you know how many foos you had, which you may not care about.) 20:41
20:44 lizmat left 20:50 bluescreen10 left 20:51 kaare_ left
timotimo wow, what? the difference in time for junction vs set isn't that huge. i should add more items. 20:52
gtodd in chromatic's perl6 critical agonistes (www.modernperlbooks.com/mt/2012/12/...rl-6.html) the LuaVM is mentioned ... is there a serious effort to make perl6 run on the LuaVM ?
20:53 bbkr left
rurban p2 is a serious effort to run p6 on something based on the luavm (perl11.org/p2), and fperrad uses LuaJIT also as p6 backend 20:53
gtodd cool thanks ... since I'm not a language implementor the advantages/disadvantages of parrot vs luavm vs whatever are beyond me .. I'm only interested in there being good ways to run perl6 :) 20:55
FROGGS hmmm, good way to run perl6 for are like: sudo apt-get install perl6 20:56
I don't really care what vm it is
20:56 am0c joined
[Coke] I don't think I've ever seen someone talk about running perl6 on lua in this channel. I wonder how recent fperrad's efforts are. 20:57
gtodd (although that article seems fairly disheartening about parrot ... so much so I thought it was some kind of meta pun/joke regarding the monty python parrot sketch)
rurban I couldn't get a decent answer from chromatic, but I guess he meant pcc and m0 20:58
[Coke] gtodd: chromatic had a falling out with the perl6/parrot efforts some time back.
gtodd " Parrot's current use case. "What's it for?" he asked. "What does it do well? What is it supposed to do?" " :-|
ah ok
20:58 grondilu joined
diakopter rurban: you've abandoned potion? 20:59
rurban He wanted to have less C, because he timed the C call overhead as major culprit
gtodd [Coke]: well hopefully everything will be all fine and modern in the future
rurban diakopter: The perl part of potion is called p2
[Coke] rurban: the context switchign between C and parrot, aye?
diakopter rurban: how does lua fit in?
rurban potion is the vm , but I changed a lot so far. 21:00
gtodd Modern Parrot ... that is
21:00 dmol left
rurban diakopter: potion is based on the lua vm (with soime mixos from IO and other vms) 21:00
diakopter I see, ok; thx
rurban The internal bytecode is 90% lua
Coke: Yes the context switching he was talking about 21:01
Which would have been fixed with lorito
[Coke] rurban: how close are you to running, e.g. the spec test suite using this setup? (if not yet, let me know when so I can add you to the perl6-roast-data info. 21:02
PerlJam FROGGS: I used to care about the VM for Perl 6, but lately I've been more like you. I don't care what VM is used as long as it can run Perl 6 and at least have promise of speed.
gtodd potion!?!? wasn't that a mysterious why the lucky stiff thing?
PerlJam gtodd: mysterious? It's just another little language/VM
diakopter mysterious because he and therefore all his stuff is mysterious, I surmise 21:03
rurban gtodd: yes, unfortunately it has an dead upstream author
I tried to contact home when I was in Provo, but apparently the ruby guys tortured him too much
21:03 dmol joined 21:04 dmol left
rurban => contact him 21:04
gtodd mysterious as in "why the lucky stiff" vanished || became a professional deep powder snowboarder || invented bitcoin or ....
_why had many projects
rurban syck sucks, but potion was good 21:05
21:06 dmol joined
masak .oO( I quaffed it and it was good ) 21:06
21:06 bbkr joined
FROGGS .seen pmurias 21:07
yoleaux I saw pmurias 1 Apr 2013 11:40Z in #perl6: <pmurias> pastie.org/7267265
timotimo masak: don't just quaff anything and everything you find on the dungeon floor!
21:08 japhb_ joined
gtodd is it somehow more possible to write/rewrite/modify/invent vm "backends" because of nqp ? or is nqp only a layer for parrot<->perl6 ? 21:08
sorry for naïve questions ... 21:09
[Coke] gtodd: there is work to port nqp to the jvm.
right now it is is functional on parrot, and actively being targetted to the jvm. (hopefully future backend targets will be easier due to the initial porting work by jnthn++ to the jvm) 21:10
rurban nqp is kinda stage1 bootstrap compiler for perl6, for parrot(pir), jvm, ...
timotimo yeah, and if you have a nqp that outputs bytecode on your new vm, you can crosscompile an nqp to live on that vm and then you can try to make rakudo work on top of your vm with not terribly much work
PerlJam gtodd: you could say that nqp allows for a layer of abstraction such that other VMs are possible.
japhb_ gtodd, NQP is a smaller, somewhat lower level subset of Perl 6, so easier to port it to other backends than a direct Rakudo-on-Parrot would be, and easier to write complex features in than a pure Rakudo-on-VM (any VM) would be, because the implementor can think at a higher level than the raw VM.
PerlJam gtodd: what japhb_ said :) 21:11
masak timotimo: but my inventory is full...
gtodd ok grokked .... since perl6/parrot seems "fast enough" for everything I do I just expected it to evolve to be fast enough for everyone ... I only have bad experience with java (but this is from more than 8 years go) and never associate per land java in the same mental space :-)
PerlJam masak: you've been eaten by a grue.
masak oh shucks. 21:12
PerlJam masak: luckily, he swallowed you whole and so you've been able to pass through his gastroinestinal tract with very little damage (save the scars on your soul), and now you're slowing exiting the grue's sphincter. 21:14
timotimo someone mentioned there's a parrot-based opengl library. is that thing still somewhere to be found and could it work? i'd like to see if i can get any decent framerate at all when displaying a few triangles or something :) 21:15
japhb_
.oO( ... you might want to find the shower on the next level )
timotimo, it worked when I wrote it. :-)
But I haven't touched that code in years, and it was ripped out of Parrot's sixparrot branch a month-ish ago.
rurban but sixparrot stripped opengl 21:16
it still works on current parrot and perl6 though
japhb_ It was a sad moment, but one I understand.
timotimo japhb_: do you have an intuition how hard it would be to write a binding for perl6 that parses header files or something to generate a zavolaj binding?
PerlJam speaking of sixparrot ... does anyone know off-hand how that's progressing? Coke? 21:17
rurban There was a working cpp and c parrot language once.
japhb_ timotimo, I fake-parsed the OpenGL headers (regexes looking for easy patterns), and aside from a few things poorly supported in Parrot, it went pretty well.
The biggest single problem was callback handling, which was really suboptimal in Parrot back then. 21:18
timotimo but it's better in zavolaj except there's a bug with the gc :D
japhb_ Also, ISTR someone is writing a C grammar, so you could probably real-parse the headers with that.
[Coke] PerlJam: I worked on it for a few weeks, and then $DAYJOB.
timotimo hm. but i don't really see where callbacks happen at all in opengl
japhb_ GC bugs: hate 'em.
[Coke] same problem with any parrot development. Just not enough tuits.
masak PerlJam: my gratitude knows no bounds. 21:19
japhb_ timotimo, the windowing library: GLUT, EGL, etc.
timotimo ah, yeah, that makes sense
japhb_ s/windowing library/windowing system interface/abstraction library/
timotimo TBH, i'd try to get SDL to work for that
japhb_ Back when I was working on that, SDL's OpenGL support was there, but LTA. Hopefully it has improved. 21:20
timotimo when i try to do opengl in some way, i've - until now - always used SDL, Qt or (once) sfml2
if you mean SDL itself, then yeah, that support is there. if you mean the binding: all that's really needed is the OPENGL flag for creating the primary surface
FROGGS japhb_: I'm afraid but SDL's opengl stuff has not improves... 21:22
*improved
timotimo enlighten me, what's missing?
FROGGS timotimo: well, you can pass the flags :o)
flussence I wonder if sdl2's any better...
timotimo flussence: i hear it has lots more stuff working on opengl
also, a promise for multi-mouse support, which i find pretty neat (but the last time i looked it was completely missing) 21:23
FROGGS flussence: had no chance to look at it, but will due the perl6 bindings next year or so
flussence sweet, maybe that'll motivate me to write graphics code :)
timotimo wow, "next year"? you're planning long-term! 21:24
diakopter giggles at the use of "working" above
FROGGS timotimo: well, this year is v5 year :o) 21:25
timotimo oke
diakopter "it didn't format my hard disk!"
masak 'night, #perl6 21:29
timotimo night masak!
japhb_ o/ masak 21:30
FROGGS night masak
21:31 kurahaupo left
census goodn ight masak! 21:31
timotimo the benchmark is still running :| 21:32
i may have added too much data
lichtkind hej good advertising for perl 6 my p6 talk in berlin got great ratings
timotimo lichtkind: is there a recording?
FROGGS timotimo: not yet available
lichtkind it will be online soon
timotimo OK 21:33
i'd like to get pinged at that time if that's okay
FROGGS timotimo: will do 21:34
21:35 spider-mario left 21:36 leont left
timotimo i hope the audio quality is sensible 21:36
Tene timotimo: years ago, I put together EFL bindings for parrot. You could consider resurrecting EFL bindings for Perl 6. 21:37
timotimo that's the stuff from enlightenment, yes? 21:38
21:38 xilo left 21:39 rindolf joined 21:40 Guest61242 joined
timotimo Tene: sell me, what's the cool stuff it allows? 21:40
census hey! may i ask for some help with installing a p5 package into cygwin or is that 100% improper? 21:42
japhb_ improper
PerlJam census: this is still #perl6
diakopter that's like the 20th time
PerlJam census: #perl _6_
Tene timotimo: stateful 2d canvas is nice; edje is a high-level library for intelligent gui objects that react to signals from the application. You can take a look at the half-assed game I started at github.com/tene/shittygame
FROGGS census: go to #win32 on irc.perl.org
sorear census: I would not object
japhb_ tene, that URL pretty much defines underselling the product .... 21:43
21:43 bbkr left
Tene japhb_: that url oversells its contents. 21:43
japhb_ Wow
Tene It's just half-assed screwing around. 21:44
diakopter census: I'm curious what your question is, because I'm curious how difficult it is to find the answer on google
Tene you fly this around on a field of squares: raw.github.com/tene/shittygame/master/ship.png
If I ever get anything fun together, I'll rename it. 21:45
FROGGS what I always miss in these repos is a screenshot of how the game looks like (even if there is not that much to see)
Tene FROGGS: install the EFL, get it running, and submit a pull request. ;) 21:46
timotimo looks like quite a bit of boilerplate, or am i just reading it wrong? 21:47
anocelot diakopter: lmgtfy?
FROGGS Tene++ # hehe, yeah 21:48
diakopter anocelot: yep, once I construct the right query
sorear I don't remember #perl6 being this *hostile* 21:49
diakopter I should be more clear. I would like to be helpful to census by imparting search engine skills 21:50
Tene timotimo: I'd think you're reading it wrong, but I'm curious how it differs from what you'd expect. Make an object, position it, size it, display it.
timotimo: I'd be interested if you could paste a snippet that looks boilerplatey to you and what you'd like an ideal API to have provided.
at least some of the difference will be due to C, I suspect. 21:51
FROGGS census: lmgtfy.com/?q=how+to+install+a+perl...+on+cygwin
21:51 xenoterracide left
timotimo right, C probably does a lot for that 21:51
FROGGS census: the first answer (the one with the big 5 on the left) is good
diakopter: see ----------^
timotimo i was surprised to see you build your own vector rotate function and then have a function from EFL do some rotation as well
21:51 xenoterracide joined, xenoterracide left
timotimo but i see why that's the way it is now 21:51
diakopter FROGGS: I suspect census was asking about a particular module, but maybe not 21:52
timotimo yeah you're right, it kind of does look near optimal for C code
FROGGS diakopter: ahh, mabbe, dunno :o) 21:53
grondilu guys, any idea on how to make this faster: gist.github.com/grondilu/5296557 ? 21:54
is $a < $x < $b slower than $a < $x and $x < $b for instance?
sorear in niecza, $a < $x < $b will be infinitesimally slower when $x is a variable, and much faster if $x isn't 21:55
diakopter perhaps infinitesimally slower to compile, I'd imagine, but not otherwise
timotimo or $x ~~ $a ^...^ $b
sorear because it stores a copy of the middle value to avoid evaluating it twice 21:56
timotimo sorear: doesn't it have to do that to comply to specs?
census FROGGS: yes it was a very specific module. i have installed about 20 today no problem. cannot get this last one.
grondilu then I wonder why this code is so slow.
sorear $x ~~ $a ^...^ $b is a syntax errror. $x ~~ $a ^..^ $b is going to be rather slower because it creates a Range object on each evaluation
grondilu has an idea 21:57
21:57 xilo joined
timotimo oke 21:57
hm, so infix:<..> isn't marked as pure. i wonder why that is. 21:58
diakopter census: does it [or one of its dependencies] fail tests?
census yes
21:59 xenoterracide joined
census so then i tried to install the dependency separately 21:59
diakopter have you looked at the platform compatibility matrix on search.cpan.org for that module/cygwin?
(because it's quite possible the maintainer just hasn't made it supported on cygwin, for any number of a thousand reasons, including that it's impossible) 22:00
census i have installed it before on active perl without difficulty.
i did not look for compatability matrix.
i did not know there were compatability matrices
22:01 dmol left
diakopter cygwin's entries are kinda sparse, though 22:01
22:02 bbkr joined
grondilu tried to spare one loop: gist.github.com/grondilu/5296557 22:02
diakopter activeperl falls under mswin32, along with strawberry, iiuc
census diakopter: may i ask how to find a comptability matrix?
diakopter sure; click on the link at the top of a code listing to get to the module's page
census i googled that term and did not find it for the module of interest
i'm at the module's page
diakopter I didn't use the right words
timotimo oh, interesting
diakopter search for matrix
22:03 lustlife left
timotimo rn: sub bark { say "arf"; 10 }; say so 5 < bark < 10; # syntax error 22:03
p6eval rakudo ffe441: OUTPUT«===SORRY!===␤Unable to parse expression in quote words; couldn't find final '>'␤at /tmp/s3ajonnQbE:1␤------> }; say so 5 < bark < 10; # syntax error⏏<EOL>␤ expecting any of:␤ argument list␤ prefix or term␤ …
..niecza v24-35-g5c06e28: OUTPUT«===SORRY!===␤␤Preceding context expects a term, but found infix < instead at /tmp/uLlBQd9jYF line 1:␤------> bark { say "arf"; 10 }; say so 5 < bark ⏏< 10; # syntax error␤␤Parse failed␤␤»
census look for the word matrix on the page ?
diakopter it's on the CPAN Testers row
timotimo probably not too surprising
diakopter yes
timotimo rn: sub bark { say "arf"; 10 }; say so 5 < bark() < 10; # syntax error
p6eval rakudo ffe441, niecza v24-35-g5c06e28: OUTPUT«arf␤False␤»
census i don't see it on the page
diakopter at what url are you?
census the url of the module ... 22:04
do you want the exact url? 22:05
diakopter yes :)
FROGGS hehe
22:05 PacoAir left
census oh ok ... y'all won't get met? 22:05
mad ?
FROGGS I won't, promise
diakopter pasting 1 line is almost always ok unless it's super long
census thanks FROGGS. i'm more worried about diakopter and geekosaur 22:06
search.cpan.org/~mikem/Net-SSLeay-1...SSLeay.pod
diakopter click the link at the top Net-SSLeay-1.54
FROGGS census: well, they'll tell you when it is time to stop I guess
census source? thart link? 22:07
FROGGS census: Net::SSLeay is always a problem, even under strawberry sometimes... can't you use the cygwin package installer to install the prebuilt package?
census i got it working in activeperl somehow 22:08
and even debian
dalek pan style="color: #395be5">perl6-examples: 03a5e4f | grondilu++ | rosalind/cstr-grondilu.pl:
Create cstr-grondilu.pl

  (rosalind) CSTR solution.  This was almost too slow, I had to kill the program just before five minutes, hoping that enough lines had been processed.
census i keep getting some warnings pertaining to C and gcc not being installed ... althoug i think i've installed them
diakopter I'd venture to suggest debian and activeperl have far more heavy perl users than cygwin, so they're more supported generally
timotimo grondilu: how many lines are there? 22:09
grondilu at most 100
timotimo mhm, ok
grondilu though what really matters is the number of letters in each line, which is at most 300 22:10
timotimo tries to understand the code properly 22:11
grondilu in general, it really seems to me that loops in perl6 are slow, no matter what they actually do.
timotimo oh, it looks like you run .comb every time you want to access one letter
is that one comb of the line per letter?
22:11 sqirrel left
census diakopter: so your final answer is that this is not meant to be? 22:11
timotimo oh, no it isn't
diakopter no
grondilu timotimo: good point, I should have factorized .comb indeed
diakopter census: you haven't mentioned whether you found the matrix for that distribution 22:12
timotimo this way you .comb every line once per @c. i suppose that's a bit of a loss, but shouldn't be terribly terrible
diakopter census: looks ok for cygwin, at least with the few perl versions tested
census no, i did not find the matrix. i did not know which link you were referring to when you said click the link ..
diakopter census: there are only two links on that page with that text in them 22:13
census googling around i found these posts: cygwin.com/ml/cygwin/2003-12/msg00082.html cygwin.com/ml/cygwin/2003-12/msg00082.html
diakopter one of them is a tar.gz download link
22:13 rking left
census yes .. i asked you the link by clickikng on "source" ? 22:14
i'm sorry if that is a bad q :(
diakopter it's not a question
census which "link" are you asking me to click on?
you said "top of the page" 22:15
i'm looking at the top and i saw "source"
diakopter same thing I said before: click the link at the top [whose text is] Net-SSLeay-1.54
FROGGS right above "source"
census search.cpan.org/~mikem/Net-SSLeay-1.54/ ?
FROGGS right 22:16
then search for a link with "matrix" in its text
22:16 rking joined
census i don't see the word matrix on that page. 22:17
i'm sorry. please don't be mad
FROGGS you can press ctrl+f within your browser to search for a text btw
diakopter which browser do you use? Chrome is case-insensitive when finding on a page
dalek pan style="color: #395be5">perl6-examples: ad9ae17 | grondilu++ | rosalind/cstr-grondilu.pl:
Update cstr-grondilu.pl

much faster version		 22:18
grondilu timotimo++
timotimo grondilu: did you measure the speed improvement?
grondilu I did not measure it because it was obviously much faster
like: seconds against minutes
timotimo wow!
that's actually amazing.
census Froggs: yes. i don't see the word matrix in my brower useing the find search.cpan.org/~mikem/Net-SSLeay-1.54/ 22:19
i'm using mozilla
diakopter search for Matrix
census oh! it was indeed case sensitive
i'm sorry
matrix.cpantesters.org/?dist=Net-SSLeay+1.54 !
so that is the matrix, eh? 22:20
22:20 kurahaupo joined
diakopter yes, you took the blue pill 22:20
er. red.
census oh ok :) en.wikipedia.org/wiki/Red_pill_and_blue_pill 22:21
22:21 stevan_ joined
census so based on that matrix, it should work in theory ? 22:21
FROGGS yes, if the dependences are met 22:22
diakopter do you have one of the versions of perl with green beside it?
FROGGS so if you see error messages about not having an compiler (gcc), that usually means something
census yes!
that is the warning i'm getting
about gcc
but i thought i instaleld it!
diakopter did you install it using the cygwin package manager? 22:23
census yes
FROGGS beware that you might need to install gcc3 and not just gcc4
census but maybe the wrong one?
22:24 xenoterracide left
census i beleive i have downloaded: gcc, gcc-core, gcc-g++, gcc-mingw-ada, gcc-minwg-core, libgcc1 22:24
22:24 xenoterracide joined
dalek pan style="color: #395be5">perl6-examples: 7168248 | grondilu++ | rosalind/cstr-grondilu.pl:
Update cstr-grondilu.pl

slightly simpler code		 22:27
census FROGGS; are you saying to do a search in the packages for gcc4 ? 22:29
timotimo that does look a lot simpler, yes
FROGGS census: no, since I think you dont need gcc4, but gcc3 if that is available
22:29 tgt left 22:30 grondilu left
census FROGGS: do you mean then that i should be looking to install gcc3? sorry i'm not following 22:30
FROGGS census: and I would search for net-ssleay perl module within that installer
census: right, search for gcc3 and/or net-ssleay module
census FROGGS: nothing comes up for gcc3 22:31
FROGGS k, search for ssleay 22:32
census and nothing comes up for net-ssleay or ssleay.. did i do something incorrrect? i tried searching for those terms in the Cygwin Setup Search
nope, nothing comes up for ssleay
is this relevant? stackoverflow.com/questions/5667364...c-compiler
FROGGS census: well, you can test that: 22:33
type: gcc -V
type: cc -V
type: perl -MConfig -E'say $Config{cc}' 22:34
and tell me if it can find gcc, if it can find cc, and what the last line gives you
census could i copy and paste the result? or should i nopaste? 22:35
labster back on topic, which behavior is correct?
rn: say "foo foo" ~~ /\bfoo/; say "\b" ~~ /\b/;
p6eval rakudo ffe441: OUTPUT«#<failed match>␤｢｣␤␤»
..niecza v24-35-g5c06e28: OUTPUT«===SORRY!===␤␤Unsupported use of \b as word boundary; in Perl 6 please use <?wb> (or either of « or ») at /tmp/n3d6VJFVbo line 1:␤------> say "foo foo" ~~ /\b⏏foo/; say "\b" ~~ /\b/;␤␤Parse failed␤␤»
timotimo census: always nopaste if it's more than one line
census it is 1 very long line ... VERY LONG
FROGGS census: come to #help-census 22:36
timotimo the irc server will cut off a big part of the line in that case
diakopter maybe. the web client might split it neatly
22:37 foofofsds joined, foofofsds left
labster would it be better to warn on the use of \b in regex (niezca), or match backspace (rakudo)? 22:38
census thanks diakopter!
diakopter yw
census and i'm so glad we got along :) 22:42
sorear std: /\b/ 22:49
p6eval std 86b102f: OUTPUT«===SORRY!===␤Unsupported use of \b as word boundary; in Perl 6 please use <?wb> (or either of « or ») at /tmp/QprXgzMxUr line 1:␤------> /\b⏏/␤Parse failed␤FAILED 00:00 41m␤»
labster std: "\b" 22:50
p6eval std 86b102f: OUTPUT«ok 00:00 41m␤»
colomon matching \b to backspace is just plain wrong, IMO.
colomon had no idea word boundary wasn't b any more, and silently making it *something completely different* would lead to some really hard to figure out bugs... 22:51
labster like \E does now? 22:52
timotimo colomon: where's the confusion? rakudo should obviously get the error from std, or am i misunderstanding something? 22:53
labster It's just different than string semantics, that's all. probably rakudobug. 22:54
colomon timotimo: look at labster's question about ten lines back. 22:55
timotimo oh, yes.
colomon rn: say "this is a test" ~~ /\bis\b/
p6eval niecza v24-35-g5c06e28: OUTPUT«===SORRY!===␤␤Unsupported use of \b as word boundary; in Perl 6 please use <?wb> (or either of « or ») at /tmp/xfCxExpPSw line 1:␤------> say "this is a test" ~~ /\b⏏is\b/␤␤Parse failed␤␤»
..rakudo ffe441: OUTPUT«#<failed match>␤»
colomon rn: say "th\bis\b is a test" ~~ /\bis\b/ 22:56
p6eval rakudo ffe441: OUTPUT«｢is｣␤␤»
..niecza v24-35-g5c06e28: OUTPUT«===SORRY!===␤␤Unsupported use of \b as word boundary; in Perl 6 please use <?wb> (or either of « or ») at /tmp/7AN48RO4Dh line 1:␤------> say "th\bis\b is a test" ~~ /\b⏏is\b/␤␤Parse failed␤␤»
22:56 rindolf left
FROGGS LHF I'd say 22:56
labster LHF? 22:58
colomon low-hanging fruit
something that's easy to fix, maybe even good for a beginner.
to dip their toes into Rakudo's insides. 22:59
regexes are probably on the complicated end of LHF, though. or at least they've always been mysterious to me...
FROGGS colomon: well, you have just to many pieces that interact with each other 23:01
and it is hard to find th place where you have to fiddle with the code
23:11 lizmat joined 23:13 tgt joined, stevan_ left 23:19 stevan_ joined 23:27 lizmat_ joined, lizmat left, lizmat_ is now known as lizmat
labster mails a CLA. 23:30
You've finally convinced me.
23:59 tgt left

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!