Parrot 5.3.0 "W00tstock Parrot" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 17 April 2013.
Alxandr I do not understand the Grammar.pm file. I've done several context-free grammars in bison and other parser-generators before, however I still have no idea what some of the stuff in the Grammar.pm file means, is there some reasource I could read, somebody who could help me with that? 00:25
tangentstorm which Grammar.pm, Alxandr? can you link to it? 00:37
Alxandr Well, it's the autogenerated one. Pasted it here: monobin.com/1565/ 00:39
For instance; what is <?> 00:40
tangentstorm I was wondering about that one myself. the synopsis for rules should explain it 00:42
Alxandr And **
tangentstorm perlcabal.org/syn/S05.html
<?> always returns true.. hmm 00:43
perhaps { <statement> | <?> } is somehow more efficient than { <statement>? } 00:44
** is supposed to mean quantification... 00:46
i'm not sure how that applies to an argument of ';'
Alxandr Or { [ <statement> | <?> ] ** ';' ] } is somehow different to { [ <statement>? ** ';' }
tangentstorm it seems to indicate repetition separated by the ';' character.
perlcabal.org/syn/S05.html#Bracket_...nalization yeah. 00:49
Alxandr / x ** $y / # same as / x [$y x]* / 00:50
tangentstorm ** apparently does all kinds of things
hrm.. no i misread 00:52
oh wait.. yeah. 00:53
Alxandr Also, wth does the magic <EXPR> match? 00:54
tangentstorm so x ** 2 matches xx and x ** y matches xyx and xyxyx
HLL::Grammar::EXPR
you said this Grammar.pm was generated... What generated it? (still catching up on the compiler tools) 00:56
Alxandr Well, to jumstart a new language I was told to run $> perl $PARROTDIR/tools/dev/mk_language_shell.pl <name_of_lang> 00:59
also relevant: docs.parrot.org/parrot/devel/html/e...1.pod.html
Oh, and tangentstorm; look at this: docs.parrot.org/parrot/devel/html/e...tor_tables 01:00
That entire section just provided me with so much needed knowledge
tangentstorm remembers reading this a year or two ago 01:01
this does look helpful. thanks for reminding me of it. 01:02
Alxandr Yeah, but that little part about how EXPR works! 01:03
And also the fact that it automagically looks up term-values
Explains SO much
tangentstorm so i was completely wrong about EXPR. cool :) 01:06
Alxandr Now, tangentstorm, do you have any idea how to add support for C-like comments (both single-line and block)?
Is the <ws> also a magic entity?
01:07 saikiran_a joined
tangentstorm i'm not convinced it's magic yet 01:08
or that i was completely wrong
i think i was wrong about being wrong, and it really is calling HLL::Grammar::EXPR
and i think you can just override it.
Alxandr What's calling EXPR? 01:10
It seems <ws> is used implicitly (= magic) 01:11
tangentstorm EXPR is defined here: github.com/perl6/nqp/blob/master/s...ar.pm#L403 01:13
<EXPR> is what calls it
Alxandr I'm not convinsed reading that will help my understanding in any way -.- 01:16
Also, does line-breaks (\\n) in token/rules mean anything? 01:18
01:23 Psyche^ joined
Alxandr tangentstorm: Do you know what <ww> is? 01:24
(cause I'm guessing <!ww> means not that)
tangentstorm line breaks are just whitespace inside patterns 01:29
to match the end of the line i think you use $$
end of /a/ line i mean.
pretty sure i saw ww defined in that same file that EXPR was in 01:30
Alxandr I used { \\N* \\n? | \\s+ } to match "till end of line"
I didn't find ww
think it's builtin 01:31
tangentstorm i don't think so 01:32
if it's builtin it would be \\ something not <ww>
i know i just saw ww though
it's not where i thought it was
Alxandr Hmm
<ws> seems to be white-space anyhow (ignored code). 01:33
I just changed my grammar to this: (notice the <comment>) monobin.com/1566/ and I got C-style comments
tangentstorm nice 01:37
totem is your language?
Alxandr Yeah 01:38
Or rather, my language's named Totem. If you know of some other totem, it's not mine 01:39
I also do IronTotem (totem on .NET) 01:40
tangentstorm looking at your github repos, but i don't see an example of the language yet... 01:49
ah github.com/Alxandr/IronTotem-3.0/b...tion.totem 01:50
Alxandr Oh
Well, I don't probably have one xD
I haven't been writing much tests
tangentstorm you have three versions of this compiler and are apparently working on a fourth, but you don't have any code in the language? :D 01:51
Alxandr tangentstorm: Here are some tests: github.com/Alxandr/IronTotem/tree/...otem.Tests
Yeah, you see, whenever I get close to something that's usable (the prev revison, the one I just linked too) I start seeing flaws/things I want to do different in the language design 01:52
Like scoping of variables was a problem in the prev itteration
So I apply changes to the design of the language
And start over xD 01:53
The prev itteration had a basic web-server working iirc. Though the networking/processing of the requests were done in C#, the processed request was handed to totem, which then returned a response 01:54
tangentstorm you were right about <ww> ... it's predefined. "Matches between two word characters (zero-width match)." perlcabal.org/syn/S05.html#Predefined_Subrules 01:55
cool :)
yeah, i've been through quite a few language drafts myself.
i guess i've done the same thing, come to think of it.
i have a little language called scarlet that was meant to transpile to c#, java, and haxe/actionscript 01:57
i don't think i hand-wrote any scarlet code... i just translated actionscript to scarlet :)
Alxandr Ah 01:58
Well, I plan on using totem as a event-based scripting language in some apps
So if you have an app that you want to let the user be able to script things in, it'll be a simple language made for just that 01:59
Most of the languages design is just javascript with added event-handling. Though the event-handling is pretty flexible 02:00
Just like functions are first-class citicens in javascript, so are events in totem
So you can do stuff with them xD
(other than just listening)
tangentstorm so like generalized message passing? 02:02
Alxandr No, more like (imagine a IRC client with scripting capabilities). The normal thing to listen for would be { client.message }, but you would also be able to listen for (for instance) { client.message.from('tangenstorm') }. And this from-function on the message-object is just a user-based function that returns a new EventObject 02:05
tangentstorm interesting. 02:10
Alxandr: i was going to say that since that one i've tried to use a little example-driven testing protocol for compilers that actually run the code and test the expected results 02:21
it seems to work well
github.com/nickelsworth/noct/tree/...test/given
Alxandr Yeah. I must say that I really like the testing-system built into the Parrot Compiler Tools
tangentstorm for example... when the tests run, output is generated to ../expected
Alxandr The .t-files
tangentstorm oh! this is new to me
i know perl has a standard testing system... looking 02:22
02:22 larks joined
Alxandr Is there a list of pirops anywhere? 02:45
tangentstorm yes... but where? :D 02:48
docs.parrot.org/parrot/devel/html/ops.html 02:49
unless you mean operators instead of opcodes, in which case: docs.parrot.org/parrot/devel/html/d...e.pod.html 02:50
02:53 quias joined 02:54 saikiran_a joined
Alxandr I'm basically trying to figure out what is viable to put in { :pirop<neg> } 02:56
And how to make a prefix that does nothing -.-
02:56 preflex_ joined
Alxandr That should not be so hard 02:57
tangentstorm: Do you have any idea how to enable a + prefix? 03:04
tangentstorm Alxandr: github.com/perl6/nqp/blob/master/s...mar.pm#L19 03:07
so maybe token prefix:sym<+> 03:08
Alxandr Yeah, yeah, I know that; but I have this: token prefix:sym<+> { <sym> <O('%unary-sign')> }
problem is that it complains about &prefix<+> not being a method
So I tried adding it to the Actions.pm
But I don't know what it takes as parameters (what the parameters are) 03:09
I need it to simply ignore the '+' 03:10
What can I do with the $/ object-thingy? 03:18
tangentstorm method prefix:sym<+>($/) { }
$/ is the match object
Alxandr I don't want to return nothing, do I?
tangentstorm you can do like $[0] or $<whatever>
those are syntacic sugar for dealing with $/
Alxandr $[0] gave me a compile-error
tangentstorm return ""
$/ has to be in scope 03:19
also i believe $[0] is the first group, so it only works if your rule has parens
Alxandr Right. But this is the prefix-rule 03:20
It uses <O(something)>
tangentstorm oh i see 03:21
you need to look at that Grammar.pm file and see what it's doing
O is implemented in there too
Alxandr It's a has-storage if I understand it correctly
tangentstorm github.com/perl6/nqp/blob/master/s...ar.pm#L132
not pretty 03:22
Alxandr: i believe :pirop<add> is perl6 for %hash[pirop]='add' on the hash that's being built. 03:26
so you can put anything you want in there.
Alxandr But it's beeing used 03:27
Can I dump $/ somehow?
to console
tangentstorm Alxandr: and i think you want $<O><pirop> 03:28
perl5 has Data::Dumper iirc 03:30
not sure about the perl6 equivalent
oh!
say $/.perl
dunno if that works in nqp though
Alxandr I get compile errors 03:31
Probably can't use it there.
Totem::Grammar.O
method prefix:sym<+>($/) { print $/.perl; }
(tired with say too) 03:32
tangentstorm yeah ... it's probably not implemented in nqp yet :/ 03:37
Alxandr Hmm. I think I figured out what's going on
Coke any function calls in nqp require parens.
print(3) works, print 3 will die.
Alxandr Coke: Why does "make $/.ast" work then? 03:38
Coke make is special.
also, if you're using the nqp that came with parrot, it's not the same as the nqp that rakudo is using. 03:39
also bedtime.
-> zzz
Alxandr I see.
Can a method in Actions.pm return an empty subtree (remove the node on reduce?)
tangentstorm make "" 03:40
make null maybe? i forget the perl word for it. 03:41
Alxandr make nil compiled
but didn't do anything
Gah. Fuckit *table-flipping*. I'm going to bed. 03:44
04:56 kurahaupo left
arnsholt Alxandr: You probably don't want a noop operator to return nil, but rather the AST for the operand of the op 05:48
dalek rrot: 2468b97 | dukeleto++ | ChangeLog:
[ci skip][doc] Add a news note about Google Summer Of Code 2013
06:24 Mike-PerlRecruiter_ joined 07:14 preflex joined
dukeleto ~~ 08:06
arnsholt o/ 08:08
10:24 mtk joined
dalek p/qrpa_experiment: b136109 | (Timo Paulssen)++ | src/pmc/qrpa.pmc:
give qrpa the methods instead.
 10:55
p/qrpa_experiment: 91353c0 | (Timo Paulssen)++ | src/pmc/qrpa.pmc:
steal the quicksort implementation from parrot
p/futile_qrpa: af36e92 | (Timo Paulssen)++ | src/ (12 files):
search&replace nqp::list -> nqp::qlist; change .push to nqp::push etc
p/futile_qrpa: 491eec9 | (Timo Paulssen)++ | src/ (6 files):
so many search and replace. probably too many.
p/futile_qrpa: ccfaf88 | (Timo Paulssen)++ | src/ (7 files):
lots and lots of nqp::un/shift/push/pop.
p/futile_qrpa: be526ca | (Timo Paulssen)++ | src/NQP/World.pm:
final fixups
p/futile_qrpa: 6c090d2 | (Timo Paulssen)++ | src/pmc/qrpa.pmc:
give qrpa the methods instead.
10:56 woosley1 joined 13:25 bluescreen joined
dalek p/jvm-support: 613ccd6 | jnthn++ | src/vm/jvm/stage0/ (10 files):
Update stage0 to get sep-comp fixes.
 13:33
p/jvm-support: a550ae8 | jnthn++ | / (8 files):
Bring in the various sep-comp fixes.
p/jvm-support: 08c8cb5 | jnthn++ | / (2 files):
Final tweaks to get a working stage1.
p/jvm-support: 04782d5 | jnthn++ | tools/build/Makefile-JVM.in:
Add stage2 build to Makefile.

Works.		 13:45
p/jvm-support: c321800 | jnthn++ | tools/build/Makefile-JVM.in:
Various Makefile additions.
 14:09
p/jvm-support: 459c817 | jnthn++ | .gitignore:
Update .gitignore.
14:20 PacoAir joined
dalek p/jvm-support: 6918dfd | jnthn++ | / (5 files):
Add convenience .bat/.sh to run NQP on JVM.
 14:23
14:35 PacoAir joined
alester ack2 is out! Whee! 14:39
PerlJam alester++ congrats. 14:41
alester Thanks, it only took 18 months
dalek p/jvm-support: c422bb3 | jnthn++ | src/vm/jvm/QAST/JASTNodes.nqp:
Add forgotten JAST node construction opt.
 14:54
p/jvm-support: 0397e83 | jnthn++ | / (2 files):
Add testing targets.

Will need some triage.
p/jvm-support: 635e421 | jnthn++ | / (20 files):
Add bootstrap-files target; clean up.

The file names used for modules in the stage0 from nqp-jvm-prep were not the same as those produced by the build; this gets things to be consistent. Updates the bootstrap as part of doing so.		 15:04
15:09 darbelo joined
Alxandr arnsholt: But I don't want any op. I have a tree like: <operation><term></operation>, and I want to reduce that to simply <term> 15:10
15:39 vinookuri joined
vinookuri Hey, i'm interested in contributing to Parrot starting with GSOC this year, can anyone guide me as to how i should begin? 15:43
Coke vinookuri: welcome aboard. 15:44
A really good first step would be to join the mailing lists, and see if you can build the latest version of parrot from git.
arnsholt Alxandr: Yeah, that's more or less what I mean 15:51
The key thing here is that the operator doesn't generate too much of the AST. Most of the work is done in the operator precedence parser (the <EXPR> subrule) 15:52
Also relevant is whether you're using NQP from github/perl6/nqp or NQP-rx from the Parrot repo 15:53
vinookuri Coke: I shall do that now, will get back to you if i get stuck up anywhere. 16:13
Coke Excellent. let us know if you want to talk about potential projects, etc. 17:10
17:15 davidfetter joined
dalek rrot: 581541d | util++ | t/library/osutils.t:
[GH #958] Add tests for chomp from osutils.
 17:54
rrot: 8ee2f4c | util++ | runtime/parrot/library/osutils.pir:
[GH #958] Fix chomp to only remove a newline when it ends the string.
 17:55
rrot: 014f0fd | dukeleto++ | examples/threads/tasks.pir:
[doc] Explain our Task PMC example
 18:00
rrot: 1ff60e5 | util++ | / (3 files):
Improve release process by changing crow.pir to auto-insert message digests.
 18:10
18:22 Mike-PerlRecruiter_ joined 18:23 darbelo joined 18:28 Psyche^ joined 18:43 contingencyplan joined 20:02 timo joined
timo good day 20:03
i'm looking into IO::Socket::INET::get in perl6 and it seems like the io_readline_encoded_string function in parrot stops reading before a newline even though there is more data to be had.
what can i do? implementing my own get/readline is hard, because apparently the get method strips the newline wether it exists or not, so i can't simply get until i get an actual newline. and reading byte-per-byte in perl6 is too slow; my connection gets killed before my program can respond to the server :( 20:06
tangentstorm r: say $IN.^methods 20:08
p6eval rakudo f3c3fe: OUTPUT«[31m===[0mSORRY![31m===[0m�Variable '$IN' is not declared�at /tmp/6ALMvI6Uqv:1�------> [32msay $IN.^methods[33m�[31m<EOL>[0m� expecting any of:� method arguments�»
tangentstorm r: say $*IN.^methods
p6eval rakudo f3c3fe: OUTPUT«open close eof get getc lines read seek tell write opened t print slurp copy chmod Str flush d e f l r s w x z modified accessed changed say <anon> <anon> <anon>â¤Â»
tangentstorm i could have sworn there was something in there that dealt with newlines. 20:09
r: say $*IN.^attributes 20:10
p6eval rakudo f3c3fe: OUTPUT«$!PIO $!ins $!chomp $!pathâ¤Â»
tangentstorm r: say $*IN.^$!chomp
p6eval rakudo f3c3fe: OUTPUT«[31m===[0mSORRY![31m===[0m�Variable $!chomp used where no 'self' is available�at /tmp/L3cslsZSDi:1�------> [32msay $*IN.^$!chomp[33m�[31m<EOL>[0m�»
tangentstorm r: say $*IN.^!chomp
p6eval rakudo f3c3fe: OUTPUT«No such method 'dispatch:<.^!>' for invocant of type 'IO'⤠in block at /tmp/oLFzeNi3DU:1â¤â¤Â»
tangentstorm r: say $*IN!chomp
p6eval rakudo f3c3fe: OUTPUT«[31m===[0mSORRY![31m===[0m�Private method call to chomp must be fully qualified with the package containing the method�at /tmp/BcyUe52iNd:1�------> [32msay $*IN!chomp[33m�[31m<EOL>[0m� expecting any of:� method arguments�»
tangentstorm beats me, timo ... but i think maybe you can set this $!chomp somehow 20:11
timo i'll have a look 20:12
IO::Socket::INET doesn't have anything like chomp
tangentstorm :/ 20:19
where is the io_readline_encoded method you mentioned? 20:21
tangentstorm breaks out grep
dalek kudo/nom: 203e3aa | masak++ | docs/release_guide.pod:
sign up Coke and me
 20:22
timo now i have a not very nice get implementation that does lots and lots of buffering over the already existing buffering parrot doe and the buffering that perl6 sockets apparently also do
tangentstorm are you passing in that PIO_READ_SIZE_ANY ? 20:27
dalek rrot: efd2ef2 | dukeleto++ | examples/threads/alloc_test.pir:
[doc] Add some POD to a thread example
tangentstorm +If you can explain what this example does in plain English, be my 20:30
+guest.
really? :D
Util Would a prior Release Manager please grant me admin authority to the web frontend on www.parrot.org? My ID is 'Util'. 20:44
I am trying to do `Administer -> Site building -> URL Redirects` as instructed in release_manager_guide.pod, but I see no "Administer" option.
cotto, dukeleto, rurban, allison, kid51, coke: ^^^ and TIA
allison Util: sure, just a sec... 20:46
rurban they moved it somewhere...
allison Util: you should be set now with admin access 20:50
dalek p/qrpa_experiment: 7d9d1ec | (Tobias Leich)++ | src/ (10 files):
FROGGS++ did all of the remaining things, like stage0
dukeleto tangentstorm: yes. Be my guest :) 21:03
tangentstorm dukeleto: it's not really a good example of anything. it has no output as far as i can tell, and there's no automated test condition. 21:04
what's the point of keeping this?
probably better to just delete it
dukeleto tangentstorm: it is an example usage of the Task PMC api
Alxandr arnsholt: I'm using the one with the parrot source
dukeleto tangentstorm: why delete it when you can improve it? I think it is better than nothing
tangentstorm but the tasks don't do anything interesting
improve works too 21:05
dukeleto tangentstorm: i agree, there are various "examples" starting from "as simple as possible" to "zomg this is complicated"
tangentstorm okay, so what would i use a task for?
dukeleto tangentstorm: what would you use parrallel algorithms for?
tangentstorm: what would you use those 3 extra cores for? 21:06
tangentstorm does not know what a task is yet
this example didn't help :)
dukeleto tangentstorm: ah! mea culpa
tangentstorm so a task is more or less a thread?
dukeleto tangentstorm: I agree. it is badly documented and that is why I am trying to improve the situation :)
tangentstorm: you get to decide how many tasks per core, or cores per task you want. A task is like the atom of building parrallel algorithms in Parrot 21:07
tangentstorm: this is lovingly called an "M:N hybrid thread", because you get to decide M threads per N cores 21:08
tangentstorm: you can do M=1,N=1 for normal non-threaded behavior
tangentstorm dukeleto: can i pass messages between the tasks?
dukeleto tangentstorm: they have data proxies
tangentstorm: some are read-only, some are read-write. I don't fully understand it, but there is a masters thesis about it if you want the link :)
tangentstorm sure 21:09
tangentstorm considers porting superpascal to parrot...
dukeleto tangentstorm: niner.name/Hybrid_Threads_for_the_Parrot_VM.pdf 21:10
tangentstorm: www.parrot.org/content/hybrid-threads
Util: the stupid Drupal UI keeps changing. I highly recommend we convert over to Github pages about 3 months ago :) 21:11
tangentstorm dukeleto: it might be nice to have a little demo that just showed how many cores there were, and let you schedule tasks to run on each core... like maybe just have it do some O(huge) task. 21:15
dalek rrot: f38fdeb | rurban++ | examples/threads/alloc_test.pir:
[doc] #936 Fix doc for examples/threads/alloc_test.pir
tangentstorm that's much better :D 21:17
i like this review link bot thing.
dukeleto tangentstorm: dalek tells us useful things :) 21:18
tangentstorm: you see, "be my guest" was an invitation, and rurban++ could not help but to take it :) 21:19
rurban: what is the "sleep 10s
tangentstorm i'm still not sure it's a useful example though.
there's no output or any way to see what it's doing 21:20
dukeleto tangentstorm: ah, you must "see the ball roll" as my old advisor used to tell me :)
tangentstorm ? 21:21
dukeleto tangentstorm: visualization
tangentstorm: i enjoy it is as well :)
tangentstorm but it's not easy to visualize something that you don't understand to begin with 21:22
dalek rrot: 7143ea2 | dukeleto++ | examples/threads/chameneos.pir:
[doc] Add POD about Chameneos, a cooperative p2p multi-core algorithm
tangentstorm So this example is only useful to people who don't need it. :D 21:23
only understandable to peolpe who don't need an example i mean.
also it makes no sense. it's called sayer but it doesn't say anything.
dalek rrot: 6864a66 | dukeleto++ | / (2 files):
Remove a useless threads example
 21:25
tangentstorm will try to think of a better example to show off this thing. 21:26
21:29 bluescreen joined
dukeleto tangentstorm: there is an example about multiplying matrices 21:30
tangentstorm: but still, that isn't very visual
tangentstorm: we need some kind of GUI visualization of the algorithm happening
tangentstorm: but that might slow it down considerably :)
dukeleto has to run, late for an important date!
dalek rrot: 4b1b60c | rurban++ | examples/threads/chameneos.pir:
[doc] #936 Add doc for examples/threads/chameneos.pir
 21:31
dukeleto rurban: you committed partially duplicate POD below my POD 21:36
Util allison: Works great. Thanks! 21:39
dalek rrot: ab56b10 | rurban++ | examples/threads/chameneos.pir:
[doc] merge both #936 chameneos.pir docs
 21:43
allison Util: great :) 21:52
21:52 kurahaupo joined
Util FYI, I added a redirect from release/devel to (the ftp dir) to match the 'devel' name used elsewhere. 21:53
I changed the release/developer redirect to point to release/devel. 21:54
I tested the double redirection; it works correctly.
rurban Now that parrot will bitrot (nqp being bootstrapped on jvm, and soon in master) shouldn't we reschedule our release dates. every quarteryear maybe? 22:16
tangentstorm does not think parrot ought to bitrot :/ 22:20
allison rurban: that's a rather gloom-and-doom way of putting it :( 22:21
but, I do think quarterly releases would make sense 22:22
the monthly release schedule was started for reasons that haven't been relevant for years now
rurban: I don't know if nqp plans to continue dual-hosted 22:23
rurban: or move exclusively to jvm
rurban: either way, it doesn't really change anything for Parrot
rurban they want both, but I fear parrrot will not get much test coverage 22:24
tangentstorm sure it does.
allison rurban: what do you mean by test coverage? They run the Parrot tests as part of the Rakudo build 22:25
tangentstorm without languages to run on this thing, it probably will bitrot.
rurban for now they do it, but how long?
tangentstorm Parrot is like a consulting company that's about to lose its biggest client.
allison it just means Parrot has to compete on equal footing 22:26
which is generally a healthy thing
tangentstorm We need to Jerry Maguire this thing... Find us a Cuba Gooding Jr language. :D
allison :)
How about Python? :) 22:27
tangentstorm How about Oberon?
rurban anyway, quarterly would make more sense, as there are not many changes
allison nice
tangentstorm Quarterly does make sense. 22:28
rurban and whiteknight, fperrad and me are gone...
allison one thing I think it does bring into question is whether a "Rakudo focused" Parrot is a good idea
or, whether we ditch that branch
I do think trimming Parrot down is the right next step
rurban sixparrot is something they wanted, so keep it.
esp. 6model integration 22:29
and fix 6model with threads

Powered by Cro and the Raku® Programming Language.

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

Thank you!