00:16 benabik left 00:17 census left
jnthn hhm, dalek didn't report the commit I just did... :( 00:26
Anyway, NQP running on the JVM now passes all the tests that the cross-compiler passed. 00:27
lue \o/ jnthn++
japhb_ w00t! 00:28
00:28 benabik joined
japhb_ Awesome, jnthn. 00:28
If I was there, I'd buy you a beer just for that milestone. :-)
lue recalls some comment about socket reading having a limit of 0xFF chars or somesuch a long while ago 00:30
dalek p-jvm-prep: f7fc166 | jnthn++ | docs/ROADMAP:
A ROADMAP update.
00:33
jnthn Enough time at the keyboard for one day. :) 00:34
'night o/
japhb_ o/ 00:35
00:36 census joined
japhb_ phenny, tell jnthn You might want to tag this nqp-jvm milestone, because I would think "can compile itself, but doesn't have the other stuff needed by rakudo" would be an interesting marker for people wishing to port to other platforms. 00:37
phenny japhb_: I'll pass that on when jnthn is around.
00:41 lustlife joined, benabik left 00:49 spider-mario left, Em_ joined 00:51 census left
Em_ r: say( "Hi I'm Emily i"m 8 years old"); 00:52
p6eval rakudo 357e60: OUTPUT«===SORRY!===␤Unable to parse expression in argument list; couldn't find final ')'␤at /tmp/78SpRtwFAJ:1␤------> say( "Hi I'm Emily i"⏏m 8 years old");␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infi…
Em_ r: say( "Hi I'm Emily i'm 8 years old"); 00:53
p6eval rakudo 357e60: OUTPUT«Hi I'm Emily i'm 8 years old␤»
raiph r: say 5 + 10 00:54
p6eval rakudo 357e60: OUTPUT«15␤»
raiph (showed Emily code.org and here we are... :) 00:55
japhb_ raiph++ 00:58
Welcome, Em_
Em_ r: Hey
p6eval rakudo 357e60: OUTPUT«===SORRY!===␤Undeclared name:␤ Hey used at line 1␤␤»
Em_ Hello 01:00
japhb_ Hello there. :-)
Em_ I'm confuesd and raiph is siting next to me hehe 01:03
japhb_ Confusion is completely normally when you're getting started.
Actually, it's pretty normal no matter how long you've been coding. :-) 01:04
Em_ LOL
vut you are SOOO right
geekosaur try switching between perl* and haskell, if you want confused :) 01:05
Em_ s/v/b/
japhb_ Woah, already using s///
Em_ well raiph is sitting next to me waching and telling 01:06
japhb_ understands
01:07 anuby joined
Em_ ya 01:08
raiph i now you dont have a cule what is going on butttt Me but now i know 01:10
hehe
Em_
sorear I started at about 8 too but I didn't have IRC back then :D 01:11
japhb_ Oooh, you just managed to find a bug in my chat client.
sorear 01:12
japhb_ sorear, yeah, 7-9 seems about ideal
geekosaur heh, also found a bug in my unansifier script 01:13
Em_ what?????
no clue 01:14
japhb_ Em_, no worries, we're just amused that a few minutes after coming to visit, you already (accidentally) found bugs in our code. :-)
Em_ ohhh thank you 01:16
lue *That's* what those error message on my screen are :P (I though someone pasted them)
Em_ i have to go be right back have to have disert
japhb_ One surely cannot miss dessert. ;-) 01:17
Em_ ya
geekosaur the most amusing thing is I'm pretty sure japhb_ and I did ours independently, but apparently have the same or very similar bugs :)
swarley Although currently not possible, a future implementation of the PIR language will allow you to use multiple heredocs within a single statement or directive:
japhb_ geekosaur, I can't even test the fix from within my chat client, because it won't let me do that. :-)
swarley ;) And with nesting of heredocs, the lexer is complete 01:18
japhb_ Multiple heredocs in *PIR*?
diakopter swarley: hi! 01:19
geekosaur huuuh. this may be an xchat bug
diakopter swarley: you can pretty much ignore that pdd
lue something like Perl6's q:to/END1/, q:to/END2/; Hi END1 There END2 (forget if that comma should be a semicolon)
swarley gist.github.com/swarley/5061726 01:20
raiph japhb: not sure jnthn has gotten to the point you're thinking. he's got nqp to cross-compile but broken in several ways (failing several tests). now he's gotten nqp to self-host to the same point (still broken in the same ways). still awesome of course, and might be what you already knew
swarley diakopter; why ignore it? :( 01:21
diakopter swarley: wat.
swarley <diakopter> swarley: you can pretty much ignore that pdd
Em_ Hi I'm Back it was pudding
diakopter 'cuz on #parrot recently they bandied about the notion of removing PIR
japhb_ raiph, he says in the ROADMAP that the remaining failing tests are probably for features of NQP that aren't used by NQP itself, but rather by Rakudo. 01:22
01:22 hypolin joined
swarley well, good thing I saved the documentation offline. Because I'm a fan of the intermediate representation, which is why I'm making it available for use outside of parrot 01:22
japhb_ Em_, :-D
Em_ hehe 01:23
:-D
01:24 fgomez left
Em_ Hi all of you i'm Emily 01:24
japhb_ Oooh, target audience for Camelia! 01:26
TimToady speaking of finding bugs...
raiph japhb: ah. thanks. then i think u're right that this is worth bringing to folk's attention.
lue suddenly wonders what task infix:<ö> is worthy of, just so he can use the non-dwimmy hyper form :) 01:27
japhb_ The Röck Döts operator? 01:28
Em_ raiph: Do you want some dessert???????? :-D
japhb_ It makes your code ROCK, of course.
01:28 woosley joined 01:29 Em_ left
lue icuc, me testing user-selected brackets from the Ps/Pe/Pi/Pf categories: gist.github.com/lue/5061801 01:39
skids r: my @a = 1,2,3,4,5; say @a[* «-« (1,2)]; say @a[* X- (1,2)]; 01:40
p6eval rakudo 357e60: OUTPUT«5 4␤Cannot call 'Numeric'; none of these signatures match:␤:(Mu:U \v: Mu *%_)␤␤ in method Numeric at src/gen/CORE.setting:865␤ in sub infix:<-> at src/gen/CORE.setting:2842␤ in block at src/gen/CORE.setting:12899␤ in sub coro at src/gen/CORE.setting:5740␤ in …
01:48 fgomez joined 01:53 gcole left 02:00 rking joined 02:27 benabik joined 02:31 rindolf joined 02:41 japhb_ left 02:59 gcole joined 03:03 gcole left 03:21 orafu joined
[Coke] runs a rakudo program through --profile and sees that 50% of the program is in sink. 03:32
03:33 ikrs joined 03:41 uvtc joined
uvtc What is the reason for being able to do this: `my @a = <foo bar baz>; my $b = @a;` ? I can index into both @a and $b the same (`@a[1], $b[1]`)... 03:48
Both @a.WHAT and $b.WHAT --> Array()
colomon $b can be anything, and an Array is anything. 03:49
uvtc One difference I see now, is:
r: my @a = <foo bar baz>; my $b = @a; say @a.perl; say $b.perl
p6eval rakudo 357e60: OUTPUT«Array.new("foo", "bar", "baz")␤["foo", "bar", "baz"]␤»
uvtc So, $b is non-flattening...
geekosaur you could think of it as $b is getting a reference to the array, while @a is getting the array itself 03:50
(it's more complex than that but supposed to look roughly like that)
TimToady I prefer to think of it as a singular/plural distinction
is a class one thing or a group of things? 03:51
geekosaur if I look at that output with perl5 eyes I see the first as a list and the second as an arrayref
TimToady is a car one object or a group of objects?
uvtc r: my @a = <foo bar baz>; my $b = @a; my @d = (1, $b, 2); say @d.perl; # I see $b doesn't flatten. 03:52
p6eval rakudo 357e60: OUTPUT«Array.new(1, ["foo", "bar", "baz"], 2)␤»
uvtc Ok. 03:53
TimToady: Am I correct in saying that the names which go into lexpads are refererences (though, they're generally auto-deref'd for you everywhere, so you tend not to notice)? 03:54
$x, @a, %h <--- names like those
(moritz explained to me yesterday that lexpads are something like runtime scratchpads where lexicals are kept track of.) 03:55
TimToady well, they're names associated with references by the symbol table
uvtc Oh. Interesting. So the names are mapped to references, and the references may refer to containers, or maybe directly to objects (such as an Int). Is that correct? 03:56
skids If you call both the link from name -> container and from container -> value "references" then it gets confusing.
[Coke] wonders how to get an infinite list of 10s. 03:57
uvtc skids: Ok. TimToady just used the word "associated": "The name is associated with the reference". So, that language sounds fine. :) 03:58
colomon r: say 10, 10, 10 ... * 03:59
p6eval rakudo 357e60: OUTPUT«10 10 10 10 ...␤»
colomon [Coke]: ^
[Coke] heh. I just got there, danke. (only need 2 tens.)
uvtc Is "symbol table" == lexpad?
colomon r: say 10, +* ... * 04:00
p6eval rakudo 357e60: OUTPUT«10 10 10 10 ...␤»
skids uvtc: well since we manipulate such associations with := "bind" works too.
sorear uvtc: yes and no
geekosaur roughly. "symbol table" ususally means something global and woth some permanence, a lexpad is like a scratchpad symbol table that exists only in the current lexical scope
sorear also, lexpads are created anew each time you invoke a sub 04:01
[Coke] colomon: I was hoping for something like 10,x5; ah well.
sorear "lexpad" is a technical Perl 6 term
colomon r: say 10 xx * 04:02
sorear "symbol table" is not a specific technical term, it's a general term of compiler art
p6eval rakudo 357e60: OUTPUT«10 10 10 10 ...␤»
[Coke] ah, xx. Danke.
TimToady symbols are just recipes, that, taken together with some structure, allow you to find the first pointer :)
[Coke] it's been 10 years, hopefully I'll starting picking this language up soon. :|
uvtc skids: Ok. So the terminology then, is: names are made of identifiers. Names are associated with references. References are bound to either containers or directly to other objects.
colomon [Coke]++ 04:03
skids r: say 10 xx Inf # should this maybe work?
p6eval rakudo 357e60: OUTPUT«Cannot coerce Inf to an Int␤ in method Numeric at src/gen/CORE.setting:10515␤ in sub infix:<==> at src/gen/CORE.setting:2886␤ in sub infix:<xx> at src/gen/CORE.setting:6180␤ in block at /tmp/o5IkRw8Qps:1␤␤»
TimToady in the case of lexicals, the structure in question is the current activation record or frame, that contains the actual data for this invocation
[Coke] r: say Inf.Int
p6eval rakudo 357e60: OUTPUT«Cannot coerce Inf to an Int␤ in method gist at src/gen/CORE.setting:10517␤ in sub say at src/gen/CORE.setting:7602␤ in block at /tmp/3nBpfBabJc:1␤␤»
uvtc TimToady: uh, oh. You just added another word that my brain needs to find another place for re. Perl 6: symbols.
TimToady well, here I just mean it semiotically, not in any CS sense 04:04
so, just a 'name'
uvtc Phew. :)
(In, for example, Clojure, symbols are something concrete.)
TimToady it's just something you use to stand for something else that you would otherwise have difficulty talking about :) 04:05
yes, FP languages have it as a technical term
I think it's a bad technical term, but there...
uvtc Are the major/common containers: Scalar, Array, and Hash?
(As in, names like, "$x, @a, and %h are references to containers".) 04:06
TimToady and Routine :)
uvtc Nice. Ok.
TimToady Routines are actually mutable, or you couldn't call .wrap on one
uvtc I don't know what calling .wrap on anything would mean. 04:07
TimToady it's sort of an AOP thing, don't sweat it
skids uvtc: maybe I'm not up to speed but: a name is bound to either reference(s) or value(s). In the case of $ names that's singular. References reference values. Values can, however, be anonymous references (e.g. a Parcel).
uvtc Is a Parcel a container? 04:08
TimToady not in the sense of being mutable
uvtc Ok.
TimToady it's kinda like a Tuple in Python
but more directly mapped to the structure of the program, in general 04:09
if your expression, has ($a, @b, &c, %d), you have a Parcel with four items
you have no idea how those will be flattened; it depends on what kind of signature you bind it to 04:10
so it's where we have a lazy value that does not yet know its item vs list context
if bound to ($,$,$,$) in a signature, there are just 4 arguments 04:11
if bound to ($, *@), there's one positional, and the rest come in variadic (slurpy)
or you could bind to (*@) and they'll all come in flattened
(the way everything comes in in Perl 5, as it happens, at least without prototypes) 04:12
uvtc TimToady: thanks. 04:14
TimToady and we call it a Parcel because you don't know what's in it till it's unwrapped :)
uvtc Hehehe. Good mnemonic. :) 04:15
TimToady well, and "parenthesis cell", but it's really the commas that do it
uvtc Oh, it's the commas that make the cell, not the parens. Ok. 04:16
TimToady r: my \p = 1,2,3; say p.WHAT 04:18
p6eval rakudo 357e60: OUTPUT«(Parcel)␤»
TimToady r: my $p = 1,2,3; # should warn 04:19
uvtc r: my \p; say \p.WHAT;
p6eval rakudo 357e60: ( no output )
rakudo 357e60: OUTPUT«===SORRY!===␤Method 'ast' not found for invocant of class 'NQPMu'␤»
uvtc r: my \p; say \p.WHAT;
p6eval rakudo 357e60: OUTPUT«===SORRY!===␤Method 'ast' not found for invocant of class 'NQPMu'␤»
TimToady r: my $p = 1,2,3; 42; # should warn
p6eval rakudo 357e60: OUTPUT«WARNINGS:␤Useless use of "," in expression "my $p = 1,2,3" in sink context (line 1)␤»
TimToady there we go
uvtc: \ means something else on the RHS 04:20
uvtc What does it mean on the LHS? "Sigilless" variable?
(that always looks like "silliness" to my eyes. :) )
TimToady it means bind the raw parcel without commiting to item vs list context
uvtc Oh. Don't open the parcel. 04:21
TimToady yes, just put another sticker on it so you can forward it :)
uvtc :) Or, maybe: don't apply and context to it.
...just yet.
TimToady right, keep being lazy about deciding the context 04:22
p6 is all about the control of laziness
yesterday I decided that try couldn't be lazy anymore (in the list sense)
try should encapsulate a computation better than that 04:23
you can still fake out a try, but you'd have to use "gather try ... take" for that
uvtc Thanks. Have not yet gotten to `try`, `gather`, `take`, etc. 04:28
TimToady someone should write a book about this...why is everyone looking at me? 04:29
diakopter •.• 04:30
TimToady there's a sense in which Rats are just lazy division too 04:32
uvtc I see a quotable quote coming: "Perl 6 is all about controlling laziness, $something impatience, and $something-else hubris." 04:33
geekosaur $something-else is surely "encouraging" :) 04:34
TimToady well, controlled laziness, controlled impatience, and controlled hubris 04:35
geekosaur $something may be "reducing"
TimToady it depends on whether you're talking about the computer or the programmer :)
04:35 preflex_ joined 04:36 preflex left, preflex_ is now known as preflex 04:38 kaleem joined
uvtc Ok. So, it sounds then like I'm probably over-thinking this, and can simply start using arrays, and then using brackets when I want them to not flatten. 04:38
TimToady or $@array 04:41
just as you can use @$array to go the other direction 04:42
uvtc Were parens previously required to do that?
TimToady yes
04:42 fgomez left
TimToady well, by the implementation, not by the spec 04:42
at least for a period of time, the spec may have been an underspec before that :) 04:43
uvtc In Perl 5, I tend to most often use the extra curlies when de-reffing, because it seems more readable to be explicit.
In the cheatsheet, it says that `[ ]` is an array composer. I don't know the technical meaning of "composer", but it gave me the impression that if `[ ]` makes arrays, then maybe `( )` makes something else. 04:46
Oh. Wait.
It says that `(, )` is the Parcel composer. So I was wondering if there was something else to make lists. 04:47
TimToady it kinda makes a parcel in the sense that it either makes a real Parcel with , or something that can stand as a parcel as a single item
uvtc Ok. 04:48
TimToady it's a "provisional item" of a parcelish nature
an item context can't look inside a () to see if there are commas or not, in a sense 04:49
but a list context can
uvtc "item context" is new to me, but it sounds analogous to list (array?) context, or boolean context. 04:50
TimToady which is why parens are transparent to list context, but not item context
04:51 SamuraiJack joined
TimToady if you call a function foo((1,2,3),(4,5,6)), it keeps the possibility of treating that as 2 arguements, 4 arguments, or 6 04:51
depending on whether the signature is ($,$), ($,*@), or (*@)
but the $ binding doesn't care whether there's a comma inside or not 04:52
if it were foo((1+3),(4,5,6)) it'd do the same
uvtc Ok.
Thanks, TimToady. 04:53
TimToady it most does what one expects, but sometimes it's hard to describe what people expect :) 04:54
*mostly
05:00 Eddward joined
Eddward Are there any known bugs in rakudo-star-2013.02 in rakudo/t/spec/S02-types/bool.t 05:03
It's not quite 100% but I get an error right after ok 48 - Bool.pick(*) returns two elems 05:04
*** glibc detected *** ./perl6: free(): invalid next size (fast): 0x000000000293bd10 ***
That's followed by a C stack trace. I suppose it could just be my system memory. 05:05
Aside from some TODOs that works I didn't see any other test failures. 05:06
05:07 fgomez joined 05:09 ikrs left, ikrs joined
TimToady huh 05:11
I didn't actually run the tests here; maybe I'll fire off a run now
bool.t passes here, so maybe it's something there 05:15
are you on 32-bit or 64?
(I'm on 64) 05:16
also, is it reproducable? coulda been a cosmic ray... 05:17
uvtc Love that guy's pizza. 05:20
(Cosmic Ray's Famous Pizza) 05:21
Ok then! :)
05:21 uvtc left 05:24 bruges_ joined 05:32 mtk joined 05:40 raiph joined 05:46 drKreso joined 05:50 grondilu joined
grondilu rn: my @a = [<foo bar>], [<1 2>]; say @a»[0]; 05:51
p6eval rakudo 357e60, niecza v24-24-gbdc3343: OUTPUT«foo bar 1 2␤»
grondilu was expecting "foo 1"
pugs: my @a = [<foo bar>], [<1 2>]; say @a»[0]; 05:52
p6eval pugs: OUTPUT«*** ␤ Unexpected "["␤ expecting term postfix␤ at /tmp/N5BRQ1gktS line 1, column 38␤»
drKreso phenny, tell FROGGS This line? my str $line = $PIO.readline(nqp::unbox_s($!input-line-separator)); 05:55
phenny drKreso: I'll pass that on when FROGGS is around.
05:56 drKreso left, grondilu left 05:58 quester joined 06:01 SamuraiJack left
Eddward TimToady: I'm running 64bit. It happens most of the time. I've seen it pass 3 times. When it fails it's always right after 48. 06:05
moritz we've had occasional trouble with bool.t for quite some time now (years) 06:11
Eddward If I have a good recreate, is there anything I should try to debug? 06:12
I've never looked at the internals of parrot or rakudo but I can handle the debugger. 06:13
diakopter Eddward++
quester Eddward, TimToady, moritz: I've also been having problems with t/spec/S02-types/bool.t. I opened a bug report a few days ago, RT #116933.
Eddward Running under valgrind I still get to test 48, but I get a different error from glib. valgrind didn't detect any errors though. 06:19
moritz Eddward: getting a backtrace would be a good first step, as well as anything that looks helpful
Eddward I need to go to bed for now. I'll pop on tomorrow to see if I can help. Thanks. 06:20
moritz ciao
Eddward ok
moritz valgrind and the GC don't love each other
Eddward Do I need an account on rt to add to it?
I assumed that if glibc detected a double free, valgrind might too. 06:21
06:26 Eddward left
quester r: use Test; ok (0 but Bool::True), qq{Bool::True works with "but"} for 1..25; 06:26
p6eval rakudo 357e60: OUTPUT«(signal ABRT)ok 1 - Bool::True works with "but"␤ok 2 - Bool::True works with "but"␤ok 3 - Bool::True works with "but"␤ok 4 - Bool::True works with "but"␤ok 5 - Bool::True works with "but"␤ok 6 - Bool::True works with "but"␤ok 7 - Bool::True works with "but"␤ok 8 - …
quester Hmm... when I try that locally, it usually says ok seven times and crashes on the eighth iteration. 06:28
star: use Test; ok (0 but Bool::True), qq{Bool::True works with "but"} for 1..25; 06:30
p6eval star 2013.02: OUTPUT«(signal ABRT)ok 1 - Bool::True works with "but"␤ok 2 - Bool::True works with "but"␤ok 3 - Bool::True works with "but"␤ok 4 - Bool::True works with "but"␤ok 5 - Bool::True works with "but"␤ok 6 - Bool::True works with "but"␤ok 7 - Bool::True works with "but"␤ok 8 - B…
quester Double hmm... locally, rakudo-star 2013.02 usually runs six iterations and crashes on the seventh. 06:31
diakopter I'll abrt you
quester wonders who is being abrt'ed and whether it is a good thing or a bad thing. 06:32
quester facepalms 06:34
diakopter ? :) 06:36
06:37 drKreso joined
quester r: (0 but Bool::True) and print qq{$_ } for 1..100; 06:38
p6eval rakudo 357e60: OUTPUT«(signal ABRT)1 2 3 4 5 6 7 8 9 10 11 12 13 »
quester star: (0 but Bool::True) and print qq{$_ } for 1..100;
p6eval star 2013.02: OUTPUT«(signal ABRT)1 2 3 4 5 6 7 8 9 10 11 12 13 »
quester Ah. Sigabrt after thirteen iterations... very likely the same way it dies here, with glibc aborting it ("glibc detected *** perl6: munmap_chunk(): invalid pointer: 0x000000000318a410 ***") 06:40
06:47 hypolin left 06:48 rindolf left, rindolf joined
drKreso Hi, I am playing with changing one line in INET.pm in rakudo core. After make && make install I get Missing or wrong version of dependency 'src/gen/CORE.setting' when running scripts. What did I do wrong? 06:53
06:53 xinming_ left 06:54 xinming joined 06:57 rindolf left, ikrs left
moritz drKreso: there's some precompiled .pir files that you haven't regenerated after installing rakudo 06:57
06:57 rindolf joined
drKreso moritz: How can i do that? (Sorry for stupid questions) 06:57
moritz drKreso: delete the 'site' dir in the rakudo installation dir, and then re-install all the necessary modules with panda again 06:58
nwc10 jnthn: all selftests succssful 07:02
07:03 hypolin joined
drKreso moritz: Thanks, it's working now. Is there a better way or it has to be done each time? 07:03
nwc10 jnthn: bother. The VMS system I have an account on only has Java 1.6.0. So I can't test it there :-( 07:04
07:09 ikrs joined, rindolf left 07:10 rindolf joined 07:13 grondilu joined
grondilu rn: say all(^10) ~~ Int; 07:14
p6eval rakudo 357e60, niecza v24-24-gbdc3343: OUTPUT«False␤»
grondilu was expecting True 07:16
07:16 toebu left
sorear rn: say all(^10) ~~ Range 07:16
p6eval niecza v24-24-gbdc3343: OUTPUT«all(Bool::False, Bool::False, Bool::False, Bool::False, Bool::False, Bool::False, Bool::False, Bool::False, Bool::False, Bool::False)␤»
..rakudo 357e60: OUTPUT«False␤»
quester updated RT #116933 with additional stacktraces from different version of Rakudo, FWIW
sorear rn: say map *.WHAT, ^10 07:17
p6eval niecza v24-24-gbdc3343: OUTPUT«(Int) (Int) (Int) (Int) (Int) (Int) (Int) (Int) (Int) (Int)␤»
..rakudo 357e60: OUTPUT«Cannot call 'map'; none of these signatures match:␤:(&code, *@values)␤␤ in block at /tmp/QuiSo3SrLf:1␤␤»
quester wishes everyone in #perl6 a very pleasant evening, or the localtime() of your choice 07:23
07:25 quester left 07:27 tadzik joined 07:32 ikrs left 07:35 ikrs joined, FROGGS joined
FROGGS morning 07:36
phenny FROGGS: 05:55Z <drKreso> tell FROGGS This line? my str $line = $PIO.readline(nqp::unbox_s($!input-line-separator));
FROGGS drKreso: right, this one hangs if it tried to read when there is no more data 07:37
drKreso FROGGS: From yesterday socket blocking problem
07:37 rindolf left, rindolf joined
FROGGS drKreso: I have to research a bit how to fix that 07:38
drKreso FROGGS: I have some sort of workaround github.com/rakudo/rakudo/commit/0b...et/INET.pm but I am unsure what would be right fix, or what is causing it 07:40
FROGGS: Maybe using the old version + binary decoding into string before returning. 07:42
07:44 kaare_ joined
FROGGS drKreso: what exactly is that what you posted? is that a testserver? 07:44
drKreso FROGGS: Yes, from http::easy readme
FROGGS but this is not related to the blacking .readline 07:46
there is a string flying around which is binary encoded, so you could encode/decode it again, to utf8 or so 07:47
that should fix the $key .= uc
drKreso FROGGS: I don't know how to write better test, but when I curl test server it blocks on last header line (after cookie). Something in that last commit is making him block. 07:48
07:48 SamuraiJack joined
drKreso FROGGS: When I revert that commit, it doesn't block any more. But HTTP::Easy breaks on binary string (but thats easy enough to fix). Now I am not sure of what would be proper thing here, since I see that last commit is trying to decode binary string and return regular one. (So HTTP::Easy would just work). 07:51
07:57 ikrs left 08:02 ikrs joined 08:06 kaleem left 08:12 domidumont joined 08:15 kaare__ joined 08:17 kaare_ left
FROGGS drKreso: I'll be afk for a few hours, but I will not forget about your problem :o) 08:21
moritz one could try to explicitly pass an encoding to IO::Socket::INET.new() 08:28
though if .get returns Str, it should never be binary
arnsholt masak: You around? 08:34
08:35 domidumont left, daxim joined, odoacre joined 08:36 ikrs left, domidumont joined 08:37 lizmat left 08:38 fhelmberger joined 08:40 ikrs joined 08:44 sqirrel joined 08:45 ObseLeTe joined 08:50 kaare__ is now known as kaare_ 08:51 wk left 08:58 ObseLeTe left 08:59 ObseLeTe joined
tadzik it should not return string, I think :| 09:04
moritz .get is line orientent, and "line" is string concept 09:10
not a Buf concept
so .get, by definition, must return a Str
tadzik hm, I see
09:15 donaldh joined
jnthn morning o/ 09:24
phenny jnthn: 00:37Z <japhb_> tell jnthn You might want to tag this nqp-jvm milestone, because I would think "can compile itself, but doesn't have the other stuff needed by rakudo" would be an interesting marker for people wishing to port to other platforms.
nwc10 \o/
jnthn: I can't actually find any other machine with Java (1.)7 installed :-(
jnthn
.oO( Write Once, Run Any...wait... )
09:26
tadzik Write once, run for your life :)
jnthn nwc10: I guess time will take care of that, though. 09:28
09:35 ikrs left 09:38 donaldh left
nwc10 jnthn: is NQPJVM able to recompile itself? 09:39
and is it byte-for-byte identical? (Should it be? Or does Java put timestamps into the bytecode) 09:40
jnthn nwc10: Didn't get to trying that yet. :)
diakopter nwc10: javac automatically randomly obfuscates 09:41
jnthn byte-for-byte identical - no, because NQP itself does some naming differently each time
To make sure the different layers of the bootstrap never end up ambiguous. 09:42
nwc10 ah OK.
byte-for-byte identical is (I think) how gcc tries to test its bootstrap
diakopter (kidding, of course)
jnthn Not to mention that we generate Java bytecode, so javac ain't involved :P 09:43
diakopter o yeah :D
diakopter waits for the slow cc 09:46
09:48 labster left 09:54 dakkar joined
drKreso I broke panda :( After reinstalling rakudo when running perl6 bootstrap.pl (for panda) I get t/01-file-find.t ...... Dubious, test returned 1 (wstat 256, 0x100) 09:56
No subtests run
t/02-shell-command.t .. Cannot fetch object from non-existent serialization context . Tried clean, removing site ? What else can I try?
diakopter jnthn: this has to be the funniest error message I've ever seen gist.github.com/diakopter/28e68225461765592ee0 09:57