»ö« | perl6-projects.org/ | nopaste: paste.lisp.org/new/perl6 | evalbot: perl6: say 3;' | irclog: irc.pugscode.org/ | UTF-8 is our friend!
Set by moritz_ on 27 June 2009.
00:03 frew_ joined 00:11 hoelzro joined
pugs_svn r27675 | toyam++ | Add test that @array, * repeats the final element infinitely. 00:17
00:18 toyam left 00:20 alester joined
wayland76 Re: awful vs. awesome. I know that in church circles, a distinction is sometimes made between "awful" (standard dictionary usage), and "aweful" (ie. awe-inspiring) 00:25
There are other similar words. "Terrible", for example, used to mean "terror-inspiring" 00:26
Likewise horrible/horror-inspiring 00:27
00:27 toyam joined
wayland76 I'm under the impression that these have all changed since 1611, but I'm not sure 00:27
[Coke]: I'm fiddling with patches to the ins2 branch 00:35
PerlJam: Can confirm that parrot has not got its installation problems sorted out
00:57 toyam left, hoelzro left 00:59 toyam joined 01:00 alester left 01:03 Chillance left 01:10 orafu left, orafu joined
wayland76 yay 01:16
Rakudo RPM works with slurp now :) 01:17
Oh, on an installed Parrot :)
01:19 molaf_x joined 01:21 skipper left 01:27 Eevee joined
wayland76 pmichaud (and others who care about ins2): The following ticket has my patch attached: rt.perl.org/rt3/Ticket/Display.html?id=63360 01:28
01:30 toyam left 01:35 GabrielVieira_ joined, GabrielVieira left 01:36 GabrielVieira_ is now known as GabrielVieira 01:37 molaf left 01:41 hoelzro joined 01:44 unitxt left, hoelzro_ joined 01:45 frew_ left
hoelzro_ is there an HTTP library for Perl 6 yet? 01:50
01:51 felipe left
missingthepoint_ wayland76: are you saying that you built an RPM which packages Rakudo, and that your RPM'd Rakudo runs on the system Parrot? 01:54
sjohnson hello wayland76 01:55
long time no see
01:56 hoelzro_ left 02:01 hoelzro_ joined
wayland76 missingthepoint_: Yes, that's what I'm saying 02:04
hoelzro_: I don't think so, but if there is one, it'll be connected with the Web.pm project
hoelzro_ wayland76: alright
is there a sockets library?
I'd be happy to try and write an HTTP library 02:05
(keyword: try)
wayland76 sjohnson: Yes, I drop in and out of the perl6 community.
missingthepoint_ hoelzro_: run('curl ...') :)
hoelzro_ missingthepoint_: I was hoping to use the library as a server
it's not necessary for my goal, but it'd be nice 02:06
wayland76 hoelzro_: I'm not the best person to ask, and now that I think about it, I may be wrong -- check Web.pm first
hoelzro_ wayland76: can do, thanks for the information!
wayland76 I believe there's a sockets library: /usr/lib/perl6/rakudo/IO/Socket/INET.pm
02:07 hoelzro left, Whiteknight left 02:08 hoelzro_ is now known as hoelzro 02:09 alester joined, explorer left
missingthepoint_ wayland76: cool :) 02:09
02:09 justatheory joined 02:12 alester left
wayland76 missingthepoint_: Well, I had most of it working ages ago, but pmichaud didn't like part of my solution 02:13
So he did some makefile work, and I reworked my stuff around it
But he still doesn't like some of it
02:14 cottoo left
wayland76 My code: my $grammar = XKB::Grammar.new(); 02:15
My error: too few arguments passed (1) - 3 params expected
Any thoughts anyone?
02:17 tann_ left, colomon left 02:21 cotto joined 02:26 freenose joined, toyam joined 02:27 payload left 02:29 toyam left 02:30 felipe joined 02:33 meppl left 02:35 lucs left 02:36 leedo joined 02:40 KyleHa joined 02:43 freenose left
KyleHa Yay for testing! Ra ra ra!! 02:44
02:44 xinming_ joined 02:49 xinming__ joined 02:54 xinming_1 joined 02:57 xinming left
pmichaud KyleHa: I don't know if you saw the backlog, but we'd like to let you name a rakudo release if you'd like :-) 02:59
KyleHa Awwww, that's sweet!
I'm not yet in the habit of checking the backlog.
Rakudo releases are named for Perlmongers groups? 03:00
pmichaud yes
KyleHa If there's a live one in Chicago, I'll go that route. Otherwise, I'd have to think about it. 03:01
pmichaud Chicago would certainly be a good candidate.
They've done good Perl 6 support in the past, and did two yapc's
03:02 frew_ joined
pmichaud (I was actually thinking about Chicago earlier today as a possibility) 03:02
KyleHa I went to the last YAPC here.
It really made me want to go to PA this year, but I couldn't do it.
03:03 payload joined 03:05 Limbic_Region left 03:06 xinming_ left, christine joined, alester joined 03:08 xinming__ left 03:10 the_irrational_1 joined
araujo christine, place and time, and i am there 03:13
:P
03:15 FurnaceBoy left 03:16 hoelzro left 03:20 donaldh left 03:21 donaldh joined
KyleHa rakudo: my $x; $x ~~ NoSuchClass 03:30
p6eval rakudo 606209: OUTPUT«Could not find non-existent sub NoSuchClass␤»
03:35 tann_ joined 03:46 Sunbeam left 03:50 unitxt joined 03:51 kane_ joined
pugs_svn r27676 | kyle++ | [t/spec] misplaced DOES fudge 03:52
r27677 | kyle++ | [t/spec] Test for RT #67784 03:53
r27678 | kyle++ | pound sign in test description
r27679 | kyle++ | [t/spec] Test for RT #67612
KyleHa pmichaud: I saw a couple of patches come into RT that do the same thing as you were committing at about the same time. Did you see and use those? Should I close/resolve/reject them? 03:56
Oh, I should have looked at the time. Even *I* should be asleep. 03:57
04:03 simcop2387 is now known as her, her is now known as simcop2387, kane_ left 04:04 kane_ joined 04:09 KyleHa left 04:15 alester left 04:16 kane___ joined 04:17 kane_ left, justatheory left, unitxt left 04:19 kane___ left 04:21 kane_ joined
tann_ phenny: tell KyleHa: You can close/resolve those patches/tickets from me....I looked at pmichaud's commits. They applied 'em (and rewrote some for the better of course). 04:24
phenny tann_: I'll pass that on when KyleHa is around.
04:25 tann_ left 04:50 skids left, xinming joined 04:55 xinming_ joined 04:59 GabrielVieira left 05:00 kane_ left 05:04 sri_kraih_ left 05:07 xinming_1 left 05:10 xinming left 05:11 frew_ left 05:28 xinming joined 05:34 rfordinal left 05:35 rfordinal joined 05:44 xinming_ left 05:48 alester joined 06:11 finanalyst joined
moritz_ rakudo: say 3.3.Str 06:12
p6eval rakudo 606209: OUTPUT«3.3␤»
06:18 hudnix left, hudnix joined
pmichaud rakudo: say 3.3.Str.WHAT 06:26
p6eval rakudo 606209: OUTPUT«Str()␤»
pmichaud (just checking :-)0
moritz_: I ended up choosing "Chicago" for this release, unless you want to choose a different name? 06:27
moritz_ pmichaud: that's fine by me
pmichaud KyleHa++ suggested it, and I had been thinking about it earlier.
I've already updated the announcement locally -- I'm just waiting for a spectest run and then I'll push 06:28
moritz_ is there anything else you want to get in before the release (after that push)? 06:29
pmichaud no, I can't think of anything.
moritz_ rakudo: say [+]
p6eval rakudo 606209: OUTPUT«say requires an argument at line 2, near " [+]"␤in Main (src/gen_setting.pm:2447)␤»
moritz_ great, no more Null PMC access 06:30
pmichaud rakudo: say [+]()
p6eval rakudo 606209: OUTPUT«0␤»
pmichaud that's kinda weird, though.
06:30 rfordinal left 06:31 rfordinal joined
moritz_ it used to be like this in STD.pm for a long time 06:31
pmichaud spectest-progress.csv update: 416 files, 11876 passing, 0 failing
moritz_ and only got fixed three weeks ago or so
pmichaud that's a little odd... did we lose a large number of tests?
moritz_ uhm, are you looking at the right column? 06:32
pmichaud sure
(it's an automatic script that generates it)
moritz_ my last run yesterday night was 11875. 06:33
pmichaud hmmm.
06:33 justatheory joined
pmichaud I had us at being at 11992 earlier today 06:33
maybe I added incorrectly.
oops, 11994
ohhh, that's it. I did add incorrectly. 06:34
okay.
no problem.
06:34 zamolxes left
moritz_ sighs with relief 06:34
btw, what do you think of rt.perl.org/rt3/Ticket/Display.html?id=67258 ?
I'm not sure if it's right to add those operators that are supposed to be automatically generated
OTOH both || and // are special in that they are short-circuiting, so it might not be such a bad idea 06:35
Matt-W Morning 06:36
moritz_ good morning
pmichaud I'm wondering if the params for those should be slurpy.
moritz_ should be, yes 06:37
pmichaud (they don't appear to be in the patches I'm seeing)
moritz_ pmichaud: in the last release did you regress on the tests with non-zero wait status? or did you just ignore it?
pmichaud I know I didn't regress. I think I just ignored them. 06:38
moritz_ should I do the same?
pmichaud that's what I would do, yes.
moritz_ ok
pmichaud the tests "pass" except for the exit code.
Matt-W oh of course it's release day
moritz_ then it's even simpler
dalek kudo: ee402a7 | pmichaud++ | docs/ (2 files):
Fill in a release name for #19 (Chicago)
 06:42
kudo: 39d9e10 | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 416 files, 11876 passing, 0 failing
kudo: 9b37b81 | pmichaud++ | docs/announce/2009-07:
Update announcement with spectest results.
pmichaud freel free to tweak my announcement changes
*feel
other than that, I think the repo is ready for release
we might want to check that the announcement text is fine with KyleHa (since he's mentioned in it), but I'm also okay with proceeding directly to release if you want 06:43
06:43 alester left
pmichaud btw, I _really_ appreciate you taking on the release this month. I'm still at OSCON, and frankly I'm a little exhausted. It's definitely freeing up some valuable time for me here. 06:44
moritz_ I have to go now anyway ($university), so I'll have to wait a bit anyway 06:46
if he shows up before the release, I'll ask him
pmichaud okay, great. 06:47
I'm headed off to sleep, tomorrow I'll have to do a bit of work on my presentation.
moritz_ ok; I plan to do the release somewhere during the next 14 hrs, not sure when exactly
pmichaud sounds great, I look forward to it. 06:48
if you run into problems, feel free to delay it until I can help\
(but I don't expect tehre to be problems :-)
06:50 rfordinal left 06:56 NewOrder left 07:03 rfordinal joined 07:08 kPb_in joined 07:20 donaldh left, donaldh joined 07:35 ejs joined
moritz_ moritz.faui2k3.org/tmp/rakudo-2009-07.tar.gz preview of release 07:36
07:37 mikehh_ joined
moritz_ (two non-zero exit status in the spectests are known) 07:37
07:41 kyle_l5l joined 07:44 ejs1 joined 07:52 ejs left, mikehh left 08:09 zamolxes joined 08:14 mikehh_ is now known as mikehh 08:22 Sunbeam joined 08:40 kPb_in left 08:46 pmurias joined 08:48 meppl joined 08:50 justatheory left 09:00 wollmers joined, icwiener joined
wollmers rakudo: say $*PROGRAM_NAME.TEST 09:01
p6eval rakudo 9b37b8: OUTPUT«Method 'TEST' not found for invocant of class 'String'␤»
09:01 masak joined, kyle_l5l left
wollmers rakudo: say $*PROGRAM_NAME 09:02
p6eval rakudo 9b37b8: OUTPUT«/tmp/vMmtxWNY40␤»
wollmers rakudo: say $*PROGRAM_NAME.WHAT
p6eval rakudo 9b37b8: OUTPUT«Str()␤»
wollmers rakudo: say $*NF 09:03
p6eval rakudo 9b37b8: OUTPUT«Use of uninitialized value␤␤»
wollmers rakudo: say $*ENC
p6eval rakudo 9b37b8: OUTPUT«Use of uninitialized value␤␤»
moritz_ rakudo doesn't do the whole Buf/encode/decode business etc 09:04
masak but we're kinda working on it. 09:05
moritz_ masak: if you have some changes, I can make the release so that you can push them :-) 09:06
moritz.faui2k3.org/tmp/rakudo-2009-07.tar.gz (preview)
wollmers Is Spec/S32-setting-library/IO.pod the only place, where the .TEST method is specified? 09:07
masak moritz_: I don't, right know. I'm blocking on PIR-fu and tuits. :/ 09:08
moritz_ wollmers: did you grep?
wollmers Yes, I grepped ./Spec 09:09
09:09 Sunbeam left
moritz_ then you have your answer. 09:09
anyway, that seems to be incorrect
TimToady++ recently decreed that file tests work with $str.IO ~~ :e, which translates to $str.IO.e() 09:10
pugs_svn r27680 | masak++ | [S12] added missing comma 09:11
r27680 | masak++ |
r27680 | masak++ | We generally cannot have two terms in a row, and that goes for the
r27680 | masak++ | inside of hash arguments to a handles trait, too.
wollmers Thus the line 'if $filename.TEST(:e,:x) {...}' should be 'if $filename.IO(:e,:x) {...}' ? 09:12
moritz_ no 09:13
it should be $filename.IO.e() && $filename.IO.x()
wollmers rakudo: say $*PROGRAM_NAME.IO(:e)
p6eval rakudo 9b37b8: OUTPUT«Method 'IO' not found for invocant of class 'String'␤»
moritz_ (but of course the e() is redundant here)
but Rakudo currently implements the auld way
masak or $filename.IO ~~ :e & :x, I guess. 09:14
moritz_ bah, I had local modifications in tools/progress_graph.pl when I did the snapshot. Next try... 09:16
that's what you get for working on two machines at once. 09:17
wollmers rakudo: say $*PROGRAM_NAME.IO.e 09:18
p6eval rakudo 9b37b8: OUTPUT«Method 'IO' not found for invocant of class 'String'␤»
wollmers Ok, not implemented.
moritz_ rakudo: say '/etc/passwd'.":e" 09:19
p6eval rakudo 9b37b8: OUTPUT«1␤»
moritz_ gah, that's even lagging two spec changes behind :-)
wollmers rakudo: say $*PROGRAM_NAME.":e & :f" 09:21
p6eval rakudo 9b37b8: OUTPUT«Method ':e & :f' not found for invocant of class 'String'␤»
moritz_ taht would be a bit too much magic :-)
rakudo: say ?($*PROGRAM_NAME ~~ :e & :d) 09:22
p6eval rakudo 9b37b8: OUTPUT«0␤»
moritz_ rakudo: say ?('/tmp/' ~~ :e & :d)
p6eval rakudo 9b37b8: OUTPUT«1␤»
wollmers say $*PROGRAM_NAME ~~ :e & :f 09:23
moritz_ rakudo: say ?($*PROGRAM_NAME ~~ :e & :f)
p6eval rakudo 9b37b8: OUTPUT«1␤»
jnthn morning 09:24
masak jnthn: o/ 09:25
moritz_ oh hai jnthn 09:26
running (hopefully) last spectest before the release 09:27
wayland76 runs around hoping my running will help the spectest run too :) 09:30
moritz_ lunch& 09:33
jnthn moritz_: Great! :-) 09:34
09:41 mikehh_ joined 09:58 mikehh left, mikehh_ is now known as mikehh
mikehh rakudo (9b37681) builds on parrot r40185 - make test/make spectest(up to r27680) PASS - Ubuntu 9.04 amd64 10:00
10:04 payload left 10:07 explorer joined
pmurias masak: one of the reason for switching to test_done would be that when slowly uncommenting a test i'm working on i would have to care about the test count 10:19
masak pmurias: exactly. the more I think about mst's proposal, the more I like it. 10:25
if we could produce a backwards-compatible solution, that would be so schweet. 10:27
pmurias backwards-compatible?
masak yes, moritz_++ mentioned it yesterday. so people who still want to write 'plan $N' can do that. 10:28
but we're of course free to encourage the plan-less style.
if we could do this in the next few days, I could blog about it as a kind of response to mst's post. :) "Test::More encourages plan-less testing, and now Perl 6 does, too!" 10:29
pmurias masak: if i implement done_testing for mildew will you do it for rakudo? 10:33
masak pmurias: deal. 10:35
pmurias: but I think I'll have to take a peek at Test::More to see how they do it.
somehow they get away with not explicitly having to say 'no plan'. 10:36
pmurias you will be able to peak at the mildew's Test.pm
moritz_ how do I delete a tag on github?
pmurias masak: Test::More has no &plan 10:37
masak pmurias: right, it's incorporated in the 'use Test::More'. I know.
pmurias the plan is in the use line
literal Test::More encourages planless testing?
masak pmurias: but in the newer versions (post-0.88), you don't even have to specify that.
literal: yes, let me dig up the link for you. 10:38
literal: www.shadowcat.co.uk/blog/matt-s-tro...g-no_plan/
moritz_ the release is out. Feel free to commit again to Rakudo 10:40
Matt-W oooh
okay I like his argument
moritz_ I can't send the announcement mails right now, I'll do that from home later on 10:41
Matt-W can we have that in rakudo's test.pm?
moritz_++ 10:42
masak moritz_++
pmurias masak: i figured it out, committing...
masak Matt-W: that was my thought as well.
if we find a good argument, why not incorporate it in Perl 6? :)
Matt-W indeed
Matt-W looks at Test.pm
pugs_svn r27681 | pmurias++ | [re-smop] added &done_testing to Test, added 10:43
r27681 | pmurias++ | t/02-test-pm/6-done_testing.t
Matt-W oh
it already supports that
pmurias masak: something as simple as printing the correct plan at the end wokrs
* works 10:44
masak pmurias: ah, cool.
Matt-W 'plan *'
masak so the plan can appear anywhere in the TAP output?
moritz_ I think it can appear either at start or end
masak Matt-W: but the thing is not having to specify the plan at all...
moritz_ well, the plan can default to *
Matt-W plan * is hardly a plan
masak Matt-W: no, but it's explicit. 10:45
Matt-W masak: we require an explicit plan anyway, so explicitly saying we don't have one isn't a disaster, but as moritz_++ says, we can default it
which seems to be a 3 character insertion
hmm no maybe slightly more complicated than that 10:46
without a plan statement how do you know when testing has started
first time you see a test function I suppose
moritz_ right 10:47
Matt-W there are a lot of those
moritz_ so?
Matt-W this may be a job for wrap
moritz_ there's nothing that Test.pm has to do when testing starts
if it emits the plan afterwards
Matt-W hmm it doesn't really use its testing started flag for much at the moment 10:48
we could live without it
pmurias the END blocks get called when you exit using an exception?
moritz_ rakudo: END { say "foo" }; die
p6eval rakudo 9b37b8: OUTPUT«No applicable methods.␤in Main (/tmp/oOI2eWryxP:2)␤» 10:49
moritz_ rakudo: END { say "foo" }; die(1)
p6eval rakudo 9b37b8: OUTPUT«1␤in Main (/tmp/QeNT4ZgACe:2)␤»
pmurias so plan * is not reliable
10:49 meppl left
masak lunch & 10:50
pmurias we should remove it and use done_testing
Matt-W so done_testing basically does what the END block does now?
moritz_ done_testing emits also a plan. 10:51
Matt-W which the END block does when it's running with no plan
(in the current code)
moritz_ oh, I didn't know that
Matt-W is hacking it at the moment
well first I'm having to update parrot and compile rakudo, but... 10:52
It seems a simple enough set of changes, so I'll make a suggested patch 10:53
once I've tested it of course
how does one test a test module...
pmurias t/02-test-pm 10:55
in the pugs repo 10:56
Matt-W cool 10:57
I think I can do this so that it'll retain compatibility with all existing planned tests
there seem to be tw oversions of 'prove' 11:01
one takes -e for the executable to run
one takes --perl
very, very irritating
my laptop appears to have just switched over to the --perl version
moritz_ 'make spectest' should do the right thing[tm] 11:02
Matt-W hmm I need to nick that code for Form then
if I run make spectest from the rakudo directory, does it use the Test.pm that's found there?
11:03 meppl joined
Matt-W ah yes it does 11:03
excellent
Matt-W checks his modified Test.pm against the spectest suite
moritz_ Matt-W: t/harness in rakudo's repo basically does the same thing as prove does by using the Test::Harness module directly 11:05
pmurias is there a good mergetool that allows me to include *both* lines in case of conflict as opposed to meld--
Matt-W moritz_: aaah
so what we need is a perl 6 test::harness :)
moritz_ Matt-W: feel free... :-) 11:06
Matt-W :)
If you've done six impossible things today, why not round it off by writing Test::Harness and Form.pm?
pmurias Matt-W: what are the impossible things? 11:07
Matt-W Well actually, Form.pm is the impossible thing :) 11:08
I just can't get my brain in the right gear for it these days
anybody who can, you're welcome to have commit bits... (hint hint) 11:09
although I've got an idea I might try later
11:13 finanalyst left 11:15 payload joined
szabgab pmurias, do I remember correctly that you had issues with the Parrot plugin of Padre ? 11:16
tewk Anyone know what the PGE alternative form .item is now that its been removed? 11:18
moritz_ are you talking about $()? 11:19
I don't quite understand the question
11:20 donaldh left
tewk - Deprecations + Removal of PGE rules: + Removal of PGE methods: .text .item .result_object 11:20
Method 'item' not found for invocant of class 'C99;Grammar'
11:20 donaldh joined
moritz_ .ast? 11:21
that might be what you're looking for
tewk Yeah I think your right.
rakudo "Use of uninitialized value" 11:22
Any way I can find out what the value is? 11:23
moritz_ don't think so :(
it should at least include a line number... care to open a ticket for that? 11:24
pmurias szabgab: no 11:27
tewk moritz_: done 11:28
moritz_ tewk: thanks 11:29
szabgab rakudo: my ($x, $y); $x += 2; say $x; $y *= -3; say $y 11:33
p6eval rakudo 9b37b8: OUTPUT«2␤-3␤»
szabgab rakudo: my $x; $x /= 42; say $x
p6eval rakudo 9b37b8: OUTPUT«Use of uninitialized value␤0␤»
szabgab so the default behavior of undef depends not only on Int/Str but also on he operation ? 11:34
11:35 omega left
Matt-W Test.pm patch in RT 11:37
11:38 omega joined
masak Matt-W++ 11:41
moritz_ rakudo: my Int $x; $x *= 3; say $x
p6eval rakudo 9b37b8: OUTPUT«3␤»
Matt-W might have a hole in it the size of an airbus A380
but most of it's there :)
szabgab rakudo: say 2**2, 4**0.5, 1**0.5; -1**0.5
p6eval rakudo 9b37b8: OUTPUT«421␤»
Matt-W -> lunch &
moritz_ szabgab: right. We need default behaviour anyway for empty reductions
szabgab: [*]() is 1, so we can make undef * $num == $num easily too 11:42
szabgab rakudo: say sqrt(-1) 11:43
p6eval rakudo 9b37b8: OUTPUT«NaN␤»
szabgab rakudo: say -1 ** 0.5
p6eval rakudo 9b37b8: OUTPUT«-1␤»
szabgab rakudo: say (-1) ** 0.5
p6eval rakudo 9b37b8: OUTPUT«NaN␤»
moritz_ rakudo: say sqrt(-1+0i) 11:44
p6eval rakudo 9b37b8: OUTPUT«0+1i␤»
masak so in Perl 6, Num is decidedly _not_ a subset of Complex. 11:45
11:45 colomon joined
szabgab shouldn't uniry minus have higher precedence than ** ? 11:45
moritz_ not if you ask liskov
masak szabgab: no, I don't think so. 11:46
szabgab: but let me check.
moritz_ no
jnthn masak: I think s/subset/subtype/ makes that statement a little clearer...
szabgab neither in perl 5
moritz_ it's correct that way, and what you usually expect
masak szabgab: exponentiation is one step tighter than symbolic unary. 11:47
moritz_ -2² is usually read as -4, not as 4
masak jnthn: well, I meant subset in the mathematical sense.
pmurias moritz_: is removing uneed'ed try's from test which test something else (like t/spec/S02-lexical-conventions/unicode.t) ok?
jnthn masak: Ah, OK. In which case it's true, perhaps?
pmurias * unneeded 11:48
masak jnthn: well, in mathematics it is.
moritz_ pmurias: IHMO yes.
szabgab moritz_, I did not have math classes for a long time :-(
jnthn OK, but in Perl 6 isn't a complex composed of two Nums?
masak jnthn: or at least there's a trivial isomorphism between a subset of the complex numbers and the real line.
moritz_ if an implementation has troubles with exceptions being thrown, they should fudge it.
masak jnthn: that's not the point. the point is that Num behaves in a way incompatible with Comples.
s/Comples/Complex/ 11:49
pmurias moritz_: even vim fails t/spec/S02-lexical-conventions/unicode.t ;) 11:50
masak I'm not saying I want it to change, just that it makes Num and Complex a little more dissimilar than they usually are in maths.
moritz_ pmurias: it looks fine here (gvim, utf-8 locale) 11:51
Matt-W blog entry to go with Test.pm patch: alledora.co.uk/wordpress/archives/2009/07/23/394 (need to get on planet perl 6 some time)
pmurias moritz_: gvim works fine, console vim gets crazy on the fudged line 11:52
moritz_ ah, I see what you mean 11:53
pmurias Matt-W: does use Test;ok(1) print out a plan?
moritz_ std: class class { }; class.new; 11:54
p6eval std 27681: OUTPUT«ok 00:02 36m␤»
Matt-W matthew@asymptote:~/rakudo$ ./perl6 -e 'use Test;ok(1);'
ok 1 -
1..1
yes
masak \o/
pmurias Matt-W: it shouldn't
masak so no done_testing? 11:55
Matt-W actually that's a point
it's not meant to
how is it doing that
Matt-W blinks at it
pmurias dunno, it doesn't do that in mildew ;)
Matt-W there's an if condition which specifically checks and shouldn't automatically give you a plan when you're in no_plan mode 11:56
moritz_ in the END block, accidentally?
Matt-W this is the stuff for compatibility with planned tests that don't call done_testing 11:57
moritz_ Matt-W: note that outer variables in END blocks don't work how they should
Matt-W but the condition's a bit weird
aaaaaaah
moritz_: you've hit it right on
moritz_ rakudo: my $x = 4; END { say $x }
Matt-W I forgot to pull that in
p6eval rakudo 9b37b8: OUTPUT«4␤»
moritz_ wait
that worked.
why did I think it wouldn't?
masak :) 11:58
moritz_ rakudo: our $x = 4; END { say $x }
p6eval rakudo 9b37b8: OUTPUT«4␤»
Matt-W matthew@asymptote:~/rakudo$ ./perl6 -e 'use Test;ok(1);'
ok 1 -
moritz_ old habit perhaps :-)
masak rakudo: END { say $x }; my $x = 4
p6eval rakudo 9b37b8: OUTPUT«4␤»
moritz_ bug?
Matt-W moritz_: no it doesn't work properly, I'm having to use 'our' declarations here to make these checks in the END block work
moritz_ known bug.
rakudo: say $x; my $x = 4
p6eval rakudo 9b37b8: OUTPUT«Null PMC access in isa()␤in Main (/tmp/wZzIrlh2xK:2)␤»
11:58 KyleHa joined
Matt-W BAD RAKUDO 11:58
moritz_ you see it doesn't throw a parse time error if the undeclared variable and the declaration are on the same line 11:59
std: $x; my $x 12:00
p6eval std 27681: OUTPUT«Potential difficulties:␤ Variable $x is not predeclared at /tmp/aAYmrVKzkC line 1:␤------> $x; my $x␤ok 00:02 37m␤»
colomon Matt-W: Just read your blog post. The "use Test" / done_testing works in Perl 6 now (or at least will when your patch gets through the system)? 12:01
Matt-W colomon: yes, that's what the patch enables
pmurias: updated patch sent to RT 12:02
I just have to say that git commit --amend is the most useful thing for embarrassing omissions ever
as long sa you spot it before you push to another branch, but since I don't have the permission to push to rakudo, that's fine :)
moritz_ or rebase -i # even more powerful
colomon I hadn't realized Perl 6 testing was so far advanced. Sweet!
KyleHa Good morning #perl6 12:03
phenny KyleHa: 04:24Z <tann_> tell KyleHa You can close/resolve those patches/tickets from me....I looked at pmichaud's commits. They applied 'em (and rewrote some for the better of course).
Matt-W colomon: it's fairly simple code that makes it go, really. Could probably be better.
(patches welcome)
moritz_ KyleHa: I already did tyat :-)
*that
KyleHa I should learn how to use --amend so I can quite having twice as many commits as necessary. 8-)
jnthn rakudo: sub foo() { }; multi bar() { }; say &foo.multi; say &bar.multi;
p6eval rakudo 9b37b8: OUTPUT«Method 'multi' not found for invocant of class 'Perl6Sub'␤»
KyleHa moritz: You took care of what tann is talking about? Awesome! moritz++ 12:04
Matt-W -> lunch for real
not that imaginary lunch
that was deeply unsatisfactory
&
KyleHa I looked back the irclog a little. I love the Rakudo release text. I'll tell all my friends. 8-)
moritz_ KyleHa: I was just typing "I hope you don't mind, because now would be too late" :-) 12:05
KyleHa Heh. I hadn't looked to see if the release happened... 12:06
pugs_svn r27682 | jnthn++ | [t/spec] Unfudge some more introspection tests for Rakudo. 12:07
masak ah! Chicago! 12:08
KyleHa: in case my insistent ++-ing doesn't show it enough, you deserve it. we've been looking for a person to fill the spot you fill for ages. 12:09
moritz_ I just sent out the mail
(and of course I found a place where I had forgotten to s/June/July/) 12:10
KyleHa Thanks masak. I'm surprised it's taken long; it seems like low hanging fruit to me.
moritz_ well, you have to have the courage to do it, and the willingness to understand Perl 6
KyleHa masak++ # copious and coherent bug reports
moritz_ it's not exactly a small language, you know :-)
KyleHa Fools rush in, perhaps. *smirk* 12:11
masak KyleHa: I've streamlined the process of reporting bugs to the point that I hardly think about it anymore. :)
340 more passing tests since last release. that's actually not too bad.
dalek kudo: 1c8c88f | jnthn++ | src/classes/ (2 files):
Make .multi work on Routine/Multi. Gets last fudged tests in S12-introspection/methods.t passing.
masak ho hum. time to do the monthly proto update. 12:12
jnthn This release had some substantial refactors as much as new features. 12:13
(That said, the refactors allowed more features.) 12:14
Some of those are under-tested.
moritz_ ah well, now is a bit too late to complain about the announcement text :-)
eiro hello guys
phenny eiro: 21 Jul 20:46Z <masak> tell eiro if you look at the beginning and end of github.com/masak/web/commit/01cca04...866dfdd0b5 you will find that the request($c) is actually deliberate (and the nasty but cool hack used to make it work) Tene++
eiro: 21 Jul 21:46Z <masak> tell eiro use.perl.org/~masak/journal/39329
eiro good bot phenny :)
jnthn moritz_: Oh, I wasn't. :-) 12:15
moritz_ submitted anounce to use.perl.org
masak eiro: salut mon ami!
eiro masak, i just read some more code, was unable to understand how the caller emulation works ... googled and find your post 12:16
moritz_ jnthn: good :-)
eiro masak, hello mon pote :)
masak eiro: wow, cool!
masak googles 'pote'
ah, 'buddy'. 12:17
12:17 missingthepoint_ left
masak I've learnt a new French word. my day is complete. 12:17
could 'pote' and 'buddy' be etymologically related, even?
that'd be the next thing to find out.
eiro i don't know ... it's familiar way to say "my friend" 12:18
KyleHa Pronounced POEtee ?
masak KyleHa: no, probably 'pot'. 12:19
eiro masak is right
moritz_ silent e would have been my guess too
KyleHa Ah, good.
I know nothing about French, clearly.
eiro i think that it came from 'poteau' => poles = something you can rely on when you fall
KyleHa Now I can mumble "Howdy pod" another way and still claim it makes sense. 12:20
moritz_ wow, we have a volleyball tournament (outdoor, of course) in 40 minutes, and we have rain and thunderstorm... 12:21
eiro damnit! i have to learn more about pir to understand those tricky lines 12:22
masak moritz_: let's hope the volley ball isn't sewn with iron thread or something...
eiro: if there's any way I can help...
eiro masak, what i understand is that getinterp is a function that returns a PMC with the state of the current interpreter. $P0['namespace';1] is the call of the 'namespace' method 12:24
but why %r ... i'm trying to find the doc to figure outt
masak eiro: that's Rakudo-specific.
means 'return this from the PIR block' 12:25
eiro wow!! so %r becames %callerns 12:26
thanks :)
masak np
pugs_svn r27683 | jnthn++ | [t/spec] Tests for :private flag on .^methods. 12:32
moritz_ can I get the value of a private attribute through introspection? 12:34
of a foreign object, I mean
if so, that would make Object.perl possible in Perl 6
pmurias Matt-W: thanks for the --amend hint, used to do it using the git reset... way 12:35
eiro will Bool::True become true some day ? i feel Bool::True breaks the huffman principle
pmurias moritz_: foreign meaning pythonic etc.?
jnthn moritz_: I can't recall anything spec'd to let you do that.
moritz_ pmurias: no, just "not of my class" 12:36
dalek kudo: f5ccc0c | jnthn++ | src/parrot/ClassHOW.pir:
Implement :private flag for .^methods and remove outdated todo comments now those things are todone.
moritz_ eiro: you can call it True as long as there's no other True in scope with which it might conflict
eiro cool! 12:37
pugs_svn r27684 | pmurias++ | [t] changed try to do
r27685 | pmurias++ | [re-smop]
r27685 | pmurias++ | added lives_ok to Test.pm
r27685 | pmurias++ | {say "hi} works
r27685 | pmurias++ | added do {...}
buubot pugs_svn: Error: DBD::SQLite::db do failed: file is encrypted or is not a database at plugins/karma_modify.pm line 57.
pugs_svn r27685 | pmurias++ | the part of t/spec/S02-lexical-conventions/unicode.t which passes is put in
r27685 | pmurias++ | t/worked-on
masak (git commit --amend)++ 12:38
moritz_ pmurias: why don't you just fudge it?
KyleHa I'm reminded of "#t" and "#f" in scheme.
masak may I also humbly recomment rebase to all? it's teh cheezburgerz.
moritz_ KyleHa: what do they do?
KyleHa moritz: They're what we're calling Bool::True and Bool::False but lots shorter. 8-) 12:39
pmurias moritz_: how does fuding work?
* fudgeing
* fudging
moritz_ pmurias: by adding a #?mildew skip 'reason' comment
pmurias: and then running t/spec/fudge mildew $path
which generates a file with s/\.t/.mildew/ 12:40
and you run that instead
KyleHa Maybe we could have $?T and $?F or $*T and $*F depending on which twigil I've misremembered.
moritz_ (maybe you also want the --keep-exit option for fudge)
KyleHa I usually sit in my rakudo dir and type "perl t/harness --fudge --verbosity=1 ../pugs/t/spec/S.../blah.t" when I'm fiddling with blah.t 12:41
I guess (?1) and (?0) aren't that long either. 12:42
moritz_ KyleHa: you can simply type 'make t/spec/your/test.t'
when you're in the rakudo directory
KyleHa Doesn't that require that the test be checked in already? 12:43
pmurias moritz_: why is the exit code changed?
moritz_ pmurias: because a fudged test doesn't pass in TimToady's thinking
KyleHa I guess I'll play with it on the train later this morning.
moritz_ KyleHa: it doesn't ahve to be check in, it's enough if it matches t/spec/*/*.t 12:44
(assuming you're on something unixoid)
KyleHa Cool, thanks. 12:45
moritz++ # testing tips
moritz_ it's even documented (in Rakudo's README) :-)
KyleHa Heh.
People keep telling me about great stuff in that README... 12:46
moritz_ but nobody reads READMEs anyway
maybe we should call it DONTREAD, then people would become curious
KyleHa PERLPR0N 12:47
masak where? where?
oh, the README. well, never mind. :)
KyleHa Gotta get on with my day now, but this has been a great start to it. 8-)
12:47 KyleHa left
masak KyleHa++ 12:47
eiro my %person< fistname lastname>; is cool. is it a way to make a typedef on it ? 12:56
something like type Person = %^hash<firstname lastname>;
my Person %user;
(can be cool to define an array of persons)
masak eiro: you mean like 'enum'? 12:59
I still get a failure in t/spec/S32-io/IO-Socket-INET.t, by the way.
12:59 lumi_ left
masak eiro: ah, I see what it is you want now. I guess you could subtype a hash to only contain a given set of keys. 13:00
rakudo: subset Person of Hash where { .keys.sort ~~ <firstname lastname> }; my Person $p = { :firstname<Kyle>, :lastname<Ha> } 13:01
p6eval rakudo f5ccc0: OUTPUT«Could not build C3 linearization: ambiguous hierarchy␤in sub trait_mod:of (src/gen_setting.pm:142)␤called from Main (/tmp/El7WIy7dsb:2)␤»
masak I think I just found a bug. \o/ 13:02
jnthn: seen something like this before?
jnthn not seen that one before... :-S 13:07
report
I haz Slovak class now.
masak rakudo: subset Person of Hash; my Person $p = {}
p6eval rakudo f5ccc0: OUTPUT«Could not build C3 linearization: ambiguous hierarchy␤in sub trait_mod:of (src/gen_setting.pm:142)␤called from Main (/tmp/3UcdMBeRbE:2)␤»
masak reports
eiro masak, wow! thanks! 13:08
masak jnthn: požívať slovenčina triedy!
eiro this is just amazing!
masak eiro: aye. when it works. :P
rakudo: subset Austria of Array; my Austria $p = [] 13:10
p6eval rakudo f5ccc0: OUTPUT«Could not build C3 linearization: ambiguous hierarchy␤in sub trait_mod:of (src/gen_setting.pm:138)␤called from Main (/tmp/7UFl113pQO:2)␤»
masak rakudo: subset Meercat of Pair; my Meercat $p = :a<b> 13:11
p6eval rakudo f5ccc0: ( no output )
13:12 xinming_ joined
eiro masak, but it doesn't :) 13:12
masak rakudo: subset Austria of List; my Austria $p = []
p6eval rakudo f5ccc0: ( no output )
masak eiro: I know, I'm submitting a bug report about it now.
eiro No applicable candidates found to dispatch to for 'pairs' when i run this code: pastebin.com/f274be584 13:13
masak eiro: that might be because you are getting a Parrot object back, and not a Rakudo object. 13:14
eiro: I'm not sure how to fix that. pmichaud would know, I think. 13:15
pugs_svn r27686 | pmurias++ | [re-smop] add skip to Test.pm, use fudge 13:16
pmurias moritz_: are there any guidlines/rule of good taste for fudging? 13:17
13:18 xinming left
pmurias re your three things you don't like in Perl6 blog one think to take in mind is that commented out code should be visually distinct from real comments 13:21
s/^/masak: / 13:22
masak pmurias: I agree. 13:24
13:24 lumi_ joined
masak one of my many pet peeves concerns commented-out code with no indication as to _why_ it was commented out. 13:24
it's like the pseudogenes of source code.
moritz_ pmurias: guidelines: 1) fudge only files you can run to a resonable amount 2) try not to interfere with other implementaitons 3) profit 13:27
pmurias how can fudging interfere?
moritz_ if you fudge out blocks, you might have to adjust the blocks 13:28
#?rakudo skip 'foo' \n { ok 1; other stuff; ok 2 } 13:29
that fudges out the whole block
now if you fudge for mildew, you can't run ok 1; but skip the rest of the block
pmurias adding a fudge #?mildew skip '...' before ok 2 would be incorrect? 13:31
moritz_ it would still run the other stuff
between ok 1 and ok 2
anyway, just experiment with it; it's unlikely that you break something for rakudo without noticing
13:32 ejs1 left
moritz_ (and if so, I will notice and tell you; forgiveness and so on... :-)) 13:32
13:37 decasm joined 13:43 explorer left, FurnaceBoy joined 13:56 skids joined, szabgab left 13:57 szabgab joined 14:15 Solarion left 14:16 jiing left, jiing_ left 14:19 MaL0 joined, MaL0 left
szabgab rakudo: say 1 < 2 ?? 'y' !! 'n'; 14:26
p6eval rakudo f5ccc0: OUTPUT«y␤»
szabgab rakudo: say 1 !< 2 ?? 'y' !! 'n';
p6eval rakudo f5ccc0: OUTPUT«Statement not terminated properly at line 2, near "!< 2 ?? "␤in Main (src/gen_setting.pm:3343)␤»
szabgab std: say 1 !< 2 ?? 'y' !! 'n'; 14:27
p6eval std 27686: OUTPUT«ok 00:02 36m␤»
szabgab so that should work just not implemented yet, right ?
std: say 1 !== 2 ?? 'y' !! 'n';
p6eval std 27686: OUTPUT«ok 00:02 36m␤»
szabgab rakudo: say 1 !== 2 ?? 'y' !! 'n';
p6eval rakudo f5ccc0: OUTPUT«y␤»
cognominal_ is there a syntax for partial application of operators ala haskell? 14:28
masak cognominal_: .assuming
see S06. 14:29
cognominal_ nothing shorter? and solving the problems of ambiguity between infix, suffix and postfix operators?
masak, I mean for operators, not for functions 14:31
I guess a functions with $_ does the equivalent... 14:32
14:32 cls_bsd joined
masak cognominal_: operators are functions in Perl 6. 14:32
and there's no issue of ambiguity, since infix, prefix, postfix etc are part of the name. 14:33
jnthn back 14:36
cognominal_ discovers the ... operator. so nice! 14:39
14:51 kane___ joined, nihiliad joined 14:57 colomon left 15:01 rfordinal left 15:03 mj41 joined 15:08 kane___ left, felipe left, Matt-W left, nnunley left, rhr left, Aisling left, spinclad left, shachaf left 15:09 [Coke] left, spinclad joined 15:10 Matt-W joined, Aisling joined, rhr joined, kane___ joined 15:11 kane___ left 15:15 shachaf joined 15:17 zostay joined
literal this might be an unusual question, but does Perl 6 allow you to use inheritence without using objects? I'm talking about something like this: pastebin.com/m179f2eea (which doesn't work in Perl 5) 15:18
jnthn I guess that's more about import/export-y things, no? 15:19
masak seems so.
jnthn Damm, how'm I meant to write tests in 34C + 50% humid :-|
jnthn should move to the arctic or something.
15:20 donaldh left
literal I'm talking about the case were you have a bunch of similar packages which might inherit from a common package, yet they aren't objects because they're all singletons 15:20
15:21 donaldh joined
masak literal: still sounds like exports. 15:23
see S11.
15:24 jhorwitz joined
literal I see 15:24
15:25 felipe joined 15:26 nnunley joined
literal the problem with exporting is that I get warnings when I override the imported functions and the originals aren't accessible after doing so, etc 15:28
15:28 M_o_C joined 15:29 jan_ left 15:30 kane___ joined 15:32 hoelzro joined 15:38 KyleHa joined
pugs_svn r27687 | jnthn++ | [t/spec] Tests for .^attributes - don't think we had any before these. 15:40
dalek kudo: 346ae98 | jnthn++ | t/spectest.data:
Add S12-introspection/attributes.t to spectest.data.
 15:44
15:44 masak left
jnthn Right on, masak. 15:44
;-) 15:45
15:45 kane___ left 15:46 kidd_ joined 15:52 payload left 15:55 Jedai joined 15:56 hoelzro left 15:59 hoelzro joined
pugs_svn r27688 | jnthn++ | [t/spec] Some basic tests for .^roles. 16:02
jnthn Phew. That gets me caught up with some of the test writing I needed to do. 16:03
KyleHa jnthn++ # test writing 16:04
jnthn KyleHa: Feel free to review and build upon them.
16:04 rfordinal joined
KyleHa I might be free for that in about ten hours. 8-) 16:04
16:05 sri_kraih joined
pmurias literal: you can re-export functions 16:08
16:09 Juerd_ joined
Juerd_ Just read in the rakudo release announcement that =$handle is deprecated. I'm happy about this. 16:09
jnthn Juerd_: So was I, it was a bit of an odd one from an implementation angle. 16:10
16:10 justatheory joined
dalek kudo: 241545b | jnthn++ | t/spectest.data:
Add S12-introspection/roles.t to spectest.data.
 16:11
Juerd_ jnthn: It was an odd one visually too :)
16:11 justatheory left 16:12 justatheory joined
jnthn Juerd_: Yes, that as well 16:12
I never managed to find it comfortable to look at.
Juerd_ It is typically used together with assignment, and the = = felt uncomfortable. 16:14
Anyway, it's fixed apparently :)
jnthn Aye. :-)
Juerd_ By the way, if anyone cares, I registered rakudo.nl a while ago 16:15
It's been a redirect to rakudo.org so far
It's available for any Rakudo serving purpose if ever needed 16:16
jnthn Cool, good to know.
16:19 justatheory left 16:23 colomon joined 16:29 jdv79 joined
jdv79 is it impossible to have pod directives start anywhere but the left-most column? 16:30
literal yup, impossible
TimToady that is correct
lambdabot TimToady: You have 1 new message. '/msg lambdabot @messages' to read it.
jdv79 that's one of the reasons i avoid formats at all costs..:( 16:31
oh well
Juerd_ Ah, just like heredocs in ancient Perl :)
jdv79 oh, that's what i meant
TimToady heredocs may be indented in p6 16:33
Juerd_ That's why I said "ancient" Perl :) 16:34
jdv79 admittedly i, and i imagine most people, use pod way more than heredocs:)
good to know though 16:35
Juerd_ I still believe in inline documentation for methods/subs 16:36
That is: something in comments
Perl 6 will make it easy to avoid repeating yourself :)
jdv79 maybe comment out the indented pod and then preprocess it - ouch
Juerd_ After having typed "Str $needle" I don't want to type $needle again in the documentation :)
Str $needle, ## Substring to look for 16:37
Something like that perhaps
IIRC it's been discussed on p6l for quite a while, with little to no concensus
16:38 mj41 left, donaldh left 16:42 sri_kraih_ joined 16:45 alester joined 16:48 REPLeffect left 16:50 sri_kraih left, REPLeffect joined, Psyche^ joined 16:51 zamolxes left, alester left 16:53 icwiener_ joined 16:54 alester joined 16:55 jan_ joined
pmurias why pod can't be indented? 16:57
TimToady too ambiguous with interleaved Perl grammar 16:59
colomon jdv79: I use heredocs far more often than pod. They are very handy for C++ code generation, which is one of my main uses for Perl. And I typically don't bother to document the small scripts I write for myself. 17:00
17:02 justatheory joined
pmichaud good morning, #perl6 17:05
jnthn oh hai, pmichaud
TimToady where have you been? :)
pmichaud here and there 17:06
jnthn pmichaud: I'm doing a minor refactor-ish thingy, that lets us rip another file out of src/parrot/
pmichaud: Namely, ClassHOW.pir 17:07
pmichaud jnthn: yay!
jnthn pmichaud: And then subclassing P6metaclass to get our ClassHOW
And adding to that.
17:07 Patterner left, Psyche^ is now known as Patterner
jnthn Sound sane? 17:07
pmichaud jnthn: sounds great to me
17:07 unitxt joined
pmichaud ...but I'll have to see the code :-) :-) 17:07
jnthn OK. I've done some minor change to p6object.
(not comitted yet, wanted to discuss first)
cognominal_ I am reading Any-list.pm and I am impressed 17:08
alester pmichaud: Where are you, I have some ideas.
jnthn Well, two of them. First is in register.
pmichaud I'm in the keynotes right now
alester That i'd like to throw together in person.
Can we chat after he finishes here?
jnthn If we pass in a how option, then it uses that rather than creating a metaclass.
pmichaud sure. meet where?
"how option"?
alester at the doors, pmichaud 17:09
jnthn I think pasting the patch is easier than explaining. ;-)
pmichaud alester: will do
Juerd_
.oO( POD could be indented without grammar clashes if it used a different character... Maybe `? ;) )
pmichaud jnthn: please paste patch, pronto
17:09 hoelzro left, alester left 17:10 icwiener left
jnthn coming 17:10
.oO( why is gist slow when pm says pronto )
gist.github.com/153253
pmichaud i'm walking so i'll be delayed anyway
jnthn kk 17:11
17:11 hercynium joined
TimToady breaking & 17:15
jnthn wonders what TimToady is going to break
;-)
17:15 meppl left
Juerd_ An ampersand 17:18
jnthn poor little ampersand 17:19
huf into what? two ayins?
17:21 justatheory left 17:22 toyam joined 17:24 KyleHa left
jnthn pmichaud: btw gist updated just now, if you already got it up. 17:24
17:25 justatheory joined
jnthn (no hurry at all, I'm going to break for dinner soon) 17:26
cognominal_ gist? 17:27
17:27 icwiener-_- joined
jnthn gist.github.com 17:29
nopaste on good drugs
17:29 icwiener-_- left 17:31 japhb left, icwiener_ left 17:33 justatheory left 17:34 hudnix_ left 17:38 wollmers left, cdarroch joined 17:41 toyam left 17:43 hoelzro joined 17:45 meppl joined, toyam joined 17:54 cls_bsd left 17:56 justatheory joined, hoelzro_ joined 17:57 hoelzro left
pmichaud jnthn: I'm not sure the word "how" grabs me 17:57
even looking at the patch, I'm not _exactly_ sure what you're doing. 17:58
(that can be entirely due to the fact that I'm doing 3-things-at-once)
jnthn pmichaud: Heh, you think I'm exactly sure? ;-) 17:59
pmichaud are you just providing a metaclass object to use instead of having register always create one?
moritz_ re
jnthn That's one half of it.
The register half.
But basically that part of the story is, "don't create a metaclass, just use this one"
The other part is for new_class
So we can say "use this class as the metaclass" 18:00
I don't mind changing how to metaclass.
pmichaud I need some precision
do you mean "use this parrotclass as the metaclass"?
or "use this object as the metaclass"
or ... ?
jnthn For register it means "use this object as the metaclass"
For new_class it means "create a metaclass using this parrotclass"
pmichaud then the option needs different names, I think. 18:01
jnthn Yeah, now I've written that out, I completely agree.
This is why we don't let me name anything. :-)
Any ideas?
pmichaud but if you already have a parrotclass, why would someone be calling new_class ?
18:01 kane___ joined
jnthn You don't already have a parrotclass 18:01
You do know what type you want the metaclass to be 18:02
pmichaud oh
you're meaning you want to create the new class using a metaclass other than P6metaclass
?
jnthn For register though, this assumes you not only have a parrotclass, but also a metaclass.
Correct.
Because I'm moving us from sticking methods into P6metaclass to subclassing it. 18:03
pmichaud ok... so for register, it means "I already have a HOW object, please use it when registering"
jnthn Correct.
pmichaud what else does register do (overall?)
the primary purpose of register was to take an existing parrotclass and put it into the p6object framework
if we already have a HOW object... why do we need to register it? 18:04
jnthn Nothing different. All that changes is that instead of instantiating a p6metaclass for us, it uses the one provided.
register creates the proto
We've been calling register from !meta_compose for a long time to get that done.
pmichaud does it also add parents and the like?
(I don't remember off the top of my head)
jnthn It can do.
pmichaud if we already have a HOW object, do we want to be able to add parents and the like? 18:05
jnthn ## add any needed parent classes
.local pmc parentclass
parentclass = options['parent']
if null parentclass goto parent_done
pmichaud what I'm getting at: is 'register' the correct place for this option, or should 'register' itself be broken down into smaller methods?
and then !meta_compose uses the smaller methods instead of the global 'register' 18:06
jnthn That could be a way to go.
I'm not sure that's a refactor I want to do right now.
This is kinda of a first step.
Basically, what's the smallest change I can do in order for us to have our own meta-class.
pmichaud well, we still want to make sure we don't start building an API we later regret :-) 18:07
jnthn To put it into context, new_class calls register 18:08
So at the moment register is the stuff after you've created a parrotclass.
pmichaud right
jnthn All we're adding is a way to say "and not only do I have a Parrot class for you, but here's a metaclass too".
18:08 jrtayloriv joined
jnthn For now, that's just a different default one for Perl 6. 18:09
In the long run, I guess we'll need to support custom meta-classes.
pmichaud if wonder if this really wants hll_map of some sort
jnthn That'd be helpful for not sufficient.
*but
The other issue I have is 18:10
Later, I want a second refactor, where we won't create the metaclass in !meta_compose, but instead in !meta_create
And that will be the thing that the loadinit has as "metaclass", and what we pass to do trait application. 18:11
pmichaud okay (more)
I'm fine with us passing a :how option to register (more)
jnthn So basically having a default would save a little boilerplate I have added.
pmichaud oh, wait 18:12
no, that's right
I'm fine with us passing a :how option to reigster
that is fine with me
I don't like the 'new_class' approach of passing the class of the metaclass to be created
jnthn OK. What would you like to see instead? 18:13
pmichaud first, do we call new_class from rakudo?
jnthn Yes, in all the stuff in src/classes/*.pir
We want our PIR defined classes to use the Perl 6 metaclass too.
pmichaud okay
agreed.
let me look at something real quick 18:14
(I have to leave in a few minutes to give my pct presentation)
jnthn OK, sure
pmichaud okay.
18:14 jrtayloriv left
pmichaud new_class is already a method on a Perl 6 metaclass 18:15
sorry
wrote that wrong
new_class is already a method on a P6metaclass
we shouldn't need an option saying "use this other thing instead"
jnthn Oh, hmm
pmichaud instead, new_class should use the type of "self"
jnthn oh
I like that.
pmichaud then in Rakudo, $!P6META is our metaclass 18:16
and everything gets created from it.
jnthn *nod*
OK, that is cleaner than what I have now.
pmichaud++
pmichaud and I'm fine with those changes. 18:17
and I'm fine with the option to register being called 'how'
because it's really a 'HOW' object
the only thing I would consider differently is to name it similar to whatever the Perl 6 metaclass API is using 18:18
(if there is such a thing)
jnthn Probably "how"
pmichaud anyway, :how is fine. 18:19
jnthn OK.
pmichaud jnthn++ # I like these changes and the overall way its heading
moritz_ any:how
jnthn pmichaud: Cool, thanks. Tests are looking good with the way I have it now, which means after the now relatively small change to do it better they should also be.
moritz_ when I want to post to rakudo.org, do I have to create a new "Story"? 18:20
pmichaud the one that says it's useful for news items
(looking)
jnthn This also means I can implement can to do the Perl 6 Thing. :-)
moritz_ ok, story then
pmichaud moritz_: yes, "Story"
(just checked)
jnthn Hope PCT talk goes well! :-) 18:21
pmichaud yeah, me too
jnthn pmichaud: You already gave Rakudo one?
pmichaud Yes. I think it went well.
I refactored it heavily yesterday morning before giving it -- I'm very happy with the new talk. 18:22
jnthn Cool. :-)
Great.
pmichaud I think it's a better talk than what I had before.
jnthn Will one at YAPC::EU be similar?
pmichaud yes
TimToady ingy: docs/Perl6/Spec/S32-setting-library/IO.pod
jnthn Nice
pmichaud it'll be the same talk, although I may update the slides quite a bit
like, I may switch out of Spork and into something else
(so I can play with some more pictures, animations, etc.) 18:23
I might need to cut about 5 minutes out of the talk, though.
It ran a little long.
(just a minute or two)
jnthn *nod*
pmichaud anyway, time toset up for talk. bbl. 18:24
18:24 toyam left
moritz_ if you find a good alternative to spork, please let me know 18:25
18:25 kane___ left 18:27 jrtayloriv joined, jrtayloriv left 18:29 hoelzro_ left 18:31 alester joined 18:33 kane___ joined 18:34 tann joined 18:35 KyleHa joined
ingy TimToady: tank yhou 18:36
KyleHa is $x, True, 'truth'; vs. ok $x, 'truth' -- any difference?
18:36 toyam joined, hoelzro joined
moritz_ yes 18:36
ok() explicitly evaluates in boolean context
rakudo: use Test; ok 'foo'; 18:37
rakudo: use Test; is 'foo', True;
ingy is the pugs repo still svn?
p6eval rakudo 241545: OUTPUT«ok 1 - ␤# Looks like you planned tests, but ran 1␤»
rakudo 241545: OUTPUT«not ok 1 - ␤# Looks like you planned tests, but ran 1␤# Looks like you failed 1 tests of 1␤»
moritz_ ingy: yes
KyleHa Oh. First one checks to see if it's literally the True value, but the second just checks if it's not something falsey.
moritz++ # detailed answers to my silly question 18:41
moritz_ but since we haven't yet specified what kind of comparison is() does (eq? or eqv?), it's safer to write 'ok $foo === Bool::True, ...' 18:44
to test if it's actually a boolean 18:45
KyleHa I should read up on the new equivalence operators.
moritz_ I once answered a question regarding them; here: stackoverflow.com/questions/176343/...381#176381 18:46
what it does not mention is that === for value types (Str, Num, Int, Bool etc) is simple identity 18:48
KyleHa Thanks!!
tann has anyone ported Benchmark to p6? it would be nice to use it to benchmark stuff in rakudo
literal there was a port to pugs I think 18:49
moritz_ i've seen a Benchmark.pm in the perl6-examples repo on github, no idea if it actually works
literal there's a ext/Benchmark in the pugs repo
moritz_ github.com/perl6/perl6-examples/blo...nchmark.pm 18:50
a rather basic thing
much less sophisticated than the perl 5 one, which will bump up repetition counts if the values are matched very closely, and similar things
tann and no Time::HiRes :) 18:51
moritz_ rakudo: say time
p6eval rakudo 241545: OUTPUT«1248375107.86179␤»
moritz_ it's fractional already
tann oh nice
moritz_ so better than p5
rakudo: say time() - time() 18:52
tann me like
p6eval rakudo 241545: OUTPUT«-1.50203704833984e-05␤»
18:52 Su-Shee joined
Su-Shee hi 18:52
tann hola
moritz_ hi Su-Shee
Su-Shee the most conventient method of protecting a string containing /, \ and " and <> is...? 18:53
moritz_ protecting? do you mean quoting?
literal a single-quoted string which doesn't use those symbols as delimiters, I suppose
moritz_ rakudo: say Q'a/b\c"d<e>f'
p6eval rakudo 241545: OUTPUT«Statement not terminated properly at line 2, near "\\c\"d<e>f'"␤in Main (src/gen_setting.pm:3343)␤»
literal q[maybe \ like / this " < >] 18:54
moritz_ rakudo: say 'a/b\c"d<e>f'
Su-Shee moritz_: I think so.
p6eval rakudo 241545: OUTPUT«a/b\c"d<e>f␤»
18:54 rindolf joined
moritz_ what literal said - either single quotes, or quotes with custom delimiters 18:54
Su-Shee ah. that simple. I was wildly Q-ing already..
tann should all modules use the underscore separator for identifiers? (as built-ins have all slummed together)... i saw Benchmark in the link that moritz sent...with time_it vs timethis vs timethese etc... just mixed up conventions
moritz_ tann: I prefer them to all use dashes (time-it) for word separators 18:55
literal then again some people hate using dashes precisely for that reason, i.e. a function can't be matched as one word :P
tann moritz: aye aye, sir!
Su-Shee rakudo: say q[foo/bar"?<>baz!'] 18:56
p6eval rakudo 241545: OUTPUT«foo/bar"?<>baz!'␤»
PerlJam moritz++ (Chicago) 18:59
moritz_ I should blog about how easy it was ;-) 19:00
PerlJam you should. 19:01
KyleHa Yeah, that would be a good follow-up to the request for release engineers.
PerlJam blog it and reply to pm's email
moritz_ which mail?
frettled mail.pm? 19:02
(sorry, sorry, where's my coat?)
PerlJam Hrm. I could have sworn he sent an email about this to p6l. Maybe I'm imagining things 19:03
moritz_ anyway, pm reads my blog I think ;-)
19:06 jhorwitz left 19:07 hoelzro left 19:08 M_o_C left
tann i look at Benchmark p5 on cpan...it has lots of p6's constructs...$v = cond ?? $x || $y; sub foo($x, $y) etc... looks like they're already half-ported to p6 !!! :) cpansearch.perl.org/src/TTY/kurila-...nchmark.pm 19:08
19:09 M_o_C joined
moritz_ tann: you're looking at the Kurila version 19:09
and Kurila "stole" more Perl 6ism
PerlJam Well, replying to an email pm sent to p6l would be more about reaching the rest of p6l than reaching pm 19:10
tann moritz_: that's a good thing..looks like they keep it close to the distro version :P
PerlJam tann: I'd be surprised if there were a large group of people using kurila. 19:12
tann: for more perl5-as-perl6isms, see Devel::Declare, Moose, MooseX::Types, MooseX::Declare, etc.
moritz_ how do I link to an always up-to-date file on github? 19:14
tann any module that rakudo can 'use' yet? :)
moritz_ when I just follow the links on the repo homepage I always get URLs to a specific revision
19:14 steffan joined
tann rakudo: use strict 19:14
p6eval rakudo 241545: OUTPUT«Can't find ./strict in @*INC␤in Main (src/gen_setting.pm:445)␤»
moritz_ rakudo: use Test; plan 1; ok 'hi tann', 'this works;
19:14 steffan left
p6eval rakudo 241545: OUTPUT«Statement not terminated properly at line 2, near "'this work"␤in Main (src/gen_setting.pm:3343)␤» 19:14
moritz_ rakudo: use Test; plan 1; ok 'hi tann', 'this works'; 19:15
p6eval rakudo 241545: OUTPUT«1..1␤ok 1 - this works␤»
tann oh hai, rakudo :)
moritz_ the modules in the perl6-examples repo seem to work too 19:16
as do Web.pm, SVG and other modules
tann rakudo: use Benchmark;
p6eval rakudo 241545: OUTPUT«Can't find ./Benchmark in @*INC␤in Main (src/gen_setting.pm:445)␤»
moritz_ tann: they are not installed on the server where p6eval runs on
tann rakudo: use Safe; 19:17
p6eval rakudo 241545: ( no output )
moritz_ this used by default 19:18
in p6eval, that is
19:19 szabgab left, KyleHa left
[particle]1 rakudo: use Test; ok 1; done_testing; 19:19
p6eval rakudo 241545: OUTPUT«ok 1 - ␤Could not find non-existent sub done_testing␤»
moritz_ that's not yet applied
[particle]1 ah
19:20 [particle]1 is now known as [particle]
moritz_ didn't have time to look at it yet more closely 19:20
[particle]: do you like the idea?
[particle] yes, i do.
although, i prefer done-testing;
moritz_ too 19:21
PerlJam how many of the other implementations grok dashes in identifiers?
moritz_ pmurias: would you object to rename done_testing to done-testing in mildew's Test.pm?
[particle] not pugs, but std does
moritz_ PerlJam: mildew, rakudo, maybe elf
elf: sub a-b($x) { say $x }; a-b(3) 19:22
p6eval elf 27688: OUTPUT«Parse error in: /tmp/1o5PjlOdM4␤panic at line 1 column 0 (pos 0): Can't understand next input--giving up␤WHERE: sub a-b($x) { say $x }; a-b(3)␤WHERE:/\<-- HERE␤ STD_red/prelude.rb:99:in `panic'␤ STD_red/std.rb:76:in `scan_unitstopper'␤ STD_red/std.rb:224:in `comp_unit'␤
..STD_red…
moritz_ ah, its port of STD.pm is too old
PerlJam If we're trying to keep Test.pm such that pugs can run it, then done-testing is out.
moritz_ PerlJam: I'm not going to compromise for dead implementations.
PerlJam moritz++ good deal :)
moritz_ PerlJam: if somebody resumes work on pugs, it's not very hard to fix 19:23
[particle] having multiple implementations will make it harder to say 'perl 6 is dead' 19:24
19:24 kane___ left 19:25 toyam left
moritz_ right. I'll be very happy when somebody revives pugs. 19:25
19:26 justatheory left, alester left
moritz_ and using pugs to argue against the "perl 6 is dead" meme is dead wrong, too 19:28
[particle] looks around for lamdacamelias
Su-Shee I think one really working and released perl 6 1.0 looks more alive than 4 half finished ones. 19:29
PerlJam Su-Shee: What does that mean really?
19:29 solarion joined
PerlJam Su-Shee: what is "finished"? 19:29
tann Su-Shee: 1.0? :P 19:30
methinks, it should be 6.0, no? :)
[particle] rakudo "1.0" will be released well before the perl 6 spec is complete
PerlJam I mean rakudo works and is released monthly. It doesn't do all the things we'd like perl 6 to do, but it's a good start. 19:31
moritz_ it doesn't even install ;-)
Su-Shee that I think in terms of "perception" perl 6 looks more alive with one implementation steadily being updated and reaching version 1.0 some day.
PerlJam I could say something similar about perl 5 though too. Perl 5 isn't "finished"
moritz_ it will be, in a few years or decades :/
PerlJam python isn't "finished"
java isn't "finished"
C isn't "finished" 19:32
Su-Shee PerlJam: well, I have a more loose definition of "finished". ;)
PerlJam I wish COBOL was finished though :)
moritz_ do you have to use it?
cognominal_ well, the real question is ossification
Perl 5, C, C++ are clearly ossified 19:33
tann if there are a dozen moritzes, rakudo will see christmas soon :P
PerlJam moritz_: I don't *have* to use it, but it keeps showing up around here. There's still a class at TAMUCC that is a COBOL class.
cognominal_ Perl 6 design is mostly againt ossification
moritz_ tann: actually I don't add many features to Rakudo - we need more jnthns and pmichauds ,-)
Su-Shee anyway. the question of "alive" or not has more to do with how many people use the language, what projects exists, if the language evolves and so on. at least in my opinion.
cognominal_ so by design Perl 6 should be never "finished" 19:34
PerlJam cognominal_: let's hope not!
moritz_ well, but at some point we have to start to take snapshots
and labeling them
perl 6.0, perl 6.0.1 etc.
tann is c finished? last i checked they stopped at c99? :)
moritz_ I think that's more what Su-Shee is talking about 19:35
PerlJam wishes he had some good swiss chocolate right now
moritz_: we do that every month :)
moritz_: you did so yourself today!
cognominal_ some say the C++ next spec will be finished next year, that is 200A
Su-Shee moritz_: actually I was supporting your "not compromise dead implementations".. :)
moritz_ PerlJam: we take snapshots of Rakudo, not of Perl 6. 19:36
perlgeek.de/blog-en/perl-6/releasin...-easy.html btw
PerlJam moritz_: rakudo isa Perl 6
moritz_ PerlJam: it's still not what I mean
Su-Shee PerlJam: yeah, but which one of the perl 6s' is going to be released as perl-6.0.0.tar.gz?
tann morirtz_: *ra*kudos to you for the chicago *bow* 19:37
PerlJam moritz_: I know what you mean, I'm just being an extremist for a little bit.
cognominal_ hum, what is the prefix pipe?
Su-Shee calls homeland security, department of version numbers for programming languages. 19:38
moritz_ cognominal_: interpolation into captures/argument lists
rakudo: sub a($x, $y) { say $x - $y }; my @l = 3, 4; a(|@l) 19:39
p6eval rakudo 241545: OUTPUT«-1␤»
moritz_ rakudo: sub a($x, $y) { say $x - $y }; my @l = 3, 4; a(@l) # without pipe
cognominal_ should learn about jquery and make the Perl 6 documentation more navigable :(
p6eval rakudo 241545: OUTPUT«too few arguments passed (1) - 2 params expected␤in sub a (/tmp/VMNYqYoqme:1)␤called from Main (/tmp/VMNYqYoqme:2)␤»
Su-Shee cognominal_: which documentation do you use?
PerlJam cognominal_: coordinate with literal 19:40
cognominal_ perlcabal.org/syn/
cognominal_ fail to undestand the difference between the splat and the prefix | 19:41
maybe some rest will help
19:41 synth left, synth joined
cognominal_ PerlJam, noted. 19:41
moritz_ cognominal_: "splat" is the old name for "slurpy"... 19:42
cognominal_: and the difference is that prefix:<|> is used on the caller side
PerlJam cognominal_: well, you don't need to necessarily coordinate with him now that I see you're talking specifically about the perlcabal.org website :)
moritz_ cognominal_: and * (as in sub foo(*@a)) is used in the parameter list
cognominal_ I have not done Perl 6 in a long time and it shows
Juerd_ I recently read something about Perl 6 somewhere and was confused by the whateverish *.
PerlJam cognominal_: but making the docs more navigable is part of the mission of perl6's version of perldoc 19:43
Juerd_ I missed the part where * became whatever instead of splat
19:43 rfordinal left
moritz_ Juerd_: it has both meanings; context an dall 19:43
Juerd_ Is the mnemonic for slurping | that it looks like a straw?
moritz_ just like it is also multiplication, and a regex quantifier
PerlJam Juerd_: that's a good mnemonic as any I suppose 19:44
cognominal_ to paraphrase McLuhan, the medium is the passage ... to general conscousness
the langage is only part of the medium 19:45
19:46 mikehh left
cognominal_ seing how much of Perl 6 is used in src/setting gives me a lot faith in Perl 19:47
19:47 zamolxes joined 19:52 japhb joined 20:02 hoelzro joined
jnthn back from nom...and back to hacking 20:02
20:04 icwiener joined 20:18 explorer joined
eiro pastebin.com/d317cb15c 20:19
i don't know but i think it's a bug :)
moritz_ it's know 20:20
eiro ok
moritz_ but I think you want .kv, not .pairs
eiro :(
is there an easy way to know known bugs in order to not disturb ? 20:21
moritz_ .pairs returns a list of pairs, and your block uses two variables up
eiro moritz_, in kv, values are empty :(
moritz_ if there's an odd number of pairs, you get that error
rakudo: for %*ENV.pairs -> { say .key, ': ', .value } 20:22
p6eval rakudo 241545: OUTPUT«Method 'key' not found for invocant of class 'Failure'␤»
jnthn -> is not wanted
moritz_ ouch
right
jnthn rakudo: for %*ENV.pairs { say .key, ': ', .value }
p6eval rakudo 241545: OUTPUT«PERL6LIB: ␤TERM: ␤SHELL: ␤USER: ␤LD_LIBRARY_PATH: ␤LS_COLORS: ␤PATH: ␤MAIL: ␤PWD: ␤EDITOR: ␤LANG: ␤HISTCONTROL: ␤HOME: ␤SHLVL: ␤LOGNAME: ␤LESSOPEN: ␤LESSCLOSE: ␤_: ␤»
eiro oh? is Pair une object ?
moritz_ everything is an object.
jnthn Yes, instance of Pair class 20:23
eiro ok
jnthn But even so...I'd expect .value to work there.
moritz_ but it seems that way the values are also empty
another bug
jnthn say %*ENV.perl;
rakudo: say %*ENV.perl;
p6eval rakudo 241545: OUTPUT«{"PERL6LIB" => "", "TERM" => "", "SHELL" => "", "USER" => "", "LD_LIBRARY_PATH" => "", "LS_COLORS" => "", "PATH" => "", "MAIL" => "", "PWD" => "", "EDITOR" => "", "LANG" => "", "HISTCONTROL" => "", "HOME" => "", "SHLVL" => "", "LOGNAME" => "", "LESSOPEN" => "", "LESSCLOSE" =>
.."", "…
jnthn ...wow.
:-|
moritz_ rakudo: say %*ENV<PWD>
jnthn rakudo: say %*ENV<PWD>
p6eval rakudo 241545: OUTPUT«/home/evalenv/rakudo␤» 20:24
moritz_ lol
jnthn ...scary...
pmichaud hello again
M_o_C well, it was the shortest one to type...
jnthn OK, so if that works... :-/
pmichaud: hi! How was the talk?
pmichaud I think it went okay.
moritz_ I think I should open a ticket.
pmichaud what are we needing a ticket for? 20:25
M_o_C Nevermind my last comment, I think I completely missed the point :x
jnthn pmichaud: %*EVN oddness.
moritz_ pmichaud: %*ENV even less awesome than though
*thought
pmichaud well, part of it is that %*ENV isn't set up as a true Perl6Hash
which particular oddness are we after now? ;-)
jnthn .pairs and .kv lose the values. 20:26
pmichaud just need to fix .pairs then
(thinking)
rakudo: say %*ENV.pairs 20:27
p6eval rakudo 241545: OUTPUT«PERL6LIB TERM SHELL USER LD_LIBRARY_PATH LS_COLORS PATH MAIL PWD EDITOR LANG HISTCONTROL HOME SHLVL LOGNAME LESSOPEN LESSCLOSE _ ␤»
pmichaud oh, I might know the problem. 20:28
pmurias cognominal_: concepts which a pretty radical thing (very similiar to haskell typeclasses) nearly got into C++ 20:29
cognominal_: so it's hardly not evolving
pmichaud rakudo: say %*ENV.pairs.perl 20:30
p6eval rakudo 241545: OUTPUT«["PERL6LIB" => "", "TERM" => "", "SHELL" => "", "USER" => "", "LD_LIBRARY_PATH" => "", "LS_COLORS" => "", "PATH" => "", "MAIL" => "", "PWD" => "", "EDITOR" => "", "LANG" => "", "HISTCONTROL" => "", "HOME" => "", "SHLVL" => "", "LOGNAME" => "", "LESSOPEN" => "", "LESSCLOSE" =>
.."", "…
pmichaud ...those all look correct to me.
jnthn pmichaud: But the values (to empty strings)? 20:31
moritz_ correct?
jnthn pmichaud: Contrast with:
pmurias HOME shouldn't be
jnthn rakudo: say %*ENV;
p6eval rakudo 241545: OUTPUT«PERL6LIB TERM SHELL USER LD_LIBRARY_PATH LS_COLORS PATH MAIL PWD EDITOR LANG HISTCONTROL HOME SHLVL LOGNAME LESSOPEN LESSCLOSE _ ␤»
pmichaud oh
moritz_ HOME should be '/home/evalenv/' or so
jnthn I mean
rakudo: say %*ENV<HOME>;
pmichaud you're right... those are wrong.
p6eval rakudo 241545: OUTPUT«/home/evalenv␤»
pmichaud looking.
pmurias rakudo: say %*ENV.values 20:32
p6eval rakudo 241545: OUTPUT«␤»
pmichaud I'm guessing it's a problem with Associative
pmurias rakudo: say %*ENV.keys
p6eval rakudo 241545: OUTPUT«PERL6LIBTERMSHELLUSERLD_LIBRARY_PATHLS_COLORSPATHMAILPWDEDITORLANGHISTCONTROLHOMESHLVLLOGNAMELESSOPENLESSCLOSE_␤»
cognominal_ pmurias, indeed I have seen some twits about C++ concepts being rejected.
20:32 hoelzro left, toyam joined
pmichaud ugh, spotty connectivity here. 20:33
cognominal_ Perl 5 is evolving too but I maintain that both are ossified because they were not designed to evolve and they are constrained by past choices.
pmurias all the new things in C++ have horribly ugly syntax 20:34
cognominal_ I bet. 20:35
20:36 hercynium left
moritz_ all the easy syntax is already taken, I suppose 20:36
pmichaud oh, I think I see the trouble 20:37
pmurias it maybe also a matter of bad taste ;)
20:38 kane___ joined 20:43 nihiliad1 joined 20:44 hoelzro joined, nihiliad left
pmurias perl6: say "{}" 20:47
p6eval elf 27688: OUTPUT«{}␤»
..pugs, rakudo 241545: OUTPUT«␤»
20:48 justatheory joined 20:49 bionoid left, bionoid joined, [particle] left 20:50 bionoid is now known as Guest77012 20:51 rfordinal joined
pugs_svn r27689 | pmurias++ | [t] fix test description 20:52
r27690 | pmurias++ | [t] use '' so {} doesn't get interpolated
jnthn Does anyone else see an unexpected win in S02-builtin_data_types\nan.t? 20:58
(been getting it for a while, but don't know if it's a platform specific one...)
pmurias unexpected win? 20:59
jnthn oh
pass
.oO( win is the opposite of fail, no? ;-) )
 21:00
moritz_ no, it's all "clean" here
jnthn ok 7 - truncate(NaN) ~~ NaN# TODO truncate(NaN) 21:01
nok for you?
21:01 Su-Shee left 21:02 jiing joined
moritz_ rakudo: say tuncate(NaN) 21:02
p6eval rakudo 241545: OUTPUT«Could not find non-existent sub tuncate␤»
moritz_ rakudo: say truncate(NaN)
p6eval rakudo 241545: OUTPUT«-2147483648␤»
moritz_ not OK for me.
jnthn > say truncate(NaN) 21:03
0
moritz_ there's a parrot related test that passes on windows but not on linux? wow, I'm duely impressed
lambdabot Not in scope: `say'Not in scope: data constructor `NaN'
moritz_ rakudo: say NaN
jnthn Heh, the pass may be bogus.
p6eval rakudo 241545: OUTPUT«NaN␤»
jnthn lolz
I get here
> say NaN
Inf
moritz_ pmurias++ # test fixes 21:04
21:04 solarion left
colomon Hi all. Just wanted to report Chicago builds and passes basic tests on my MacBook Pro (OS X 10.5.6), but fails make spectest. 21:13
21:13 icwiener left
pmichaud how much of make spectest does it fail? 21:15
moritz_ two "Non-zero wait status" are to be expected.
pmichaud do ti just failhmmm.
colomon t/spec/S32-io/IO-Socket-INET.t (Wstat: 256 Tests: 0 Failed: 0)
Non-zero exit status: 1
Parse errors: Bad plan. You planned 3 tests but ran 0.
pmichaud that's often a timeout error. 21:16
21:16 kane___ left
pmichaud moritz_: I hate to say this... but now that I'm awake I think that in the June release I *did* comment the expected failures out of the spectest.data for the release (and then put them back in right after the release) 21:16
moritz_ pmichaud: might have been better that way, yes 21:17
pmichaud My bad.
colomon Is there an easy way to re-run it?
pmichaud colomon: make t/spec/S32-io/IO-Socket-INET.t 21:18
21:19 Whiteknight joined, [particle] joined
lisppaste3 colomon pasted "make spectest failure" at paste.lisp.org/display/84094 21:20
moritz_ ouch 21:21
I now see what happened
21:21 solarion joined
moritz_ regex sp was removed from the core, so to say 21:21
but it onoly gets called on darwin 21:22
what did it used to match? a single blank?
pmichaud ...?
yes, it would match a single space
moritz_ given $*OS { ... when 'darwin' { $netstat_pat = regex { [ ^^ .+? '.' (\d+) <.sp> .+? ]+ $ }; ...
pugs_svn r27691 | moritz++ | [t/spec] remove old regexp <sp>, colomon++ 21:23
moritz_ colomon: thanks for the report 21:24
colomon No prob, thanks for all your hard work.
21:24 szabgab joined
colomon I can svn update that file and test the patch? 21:25
moritz_ that would be great
colomon "All tests successful." 21:26
21:27 kane___ joined 21:28 meppl left
moritz_ pmichaud: the last release does contain S12-attributes/class.t in t/spectest.data, as well as S14-roles/basic.t. So either you used a different mechanism of regressing (changed tests?), or it was a different set of files back then, or you mis-remembered 21:28
I also can't find any related commit 21:29
pmichaud then I guess I *didn't* remove them.
moritz_ git checkout 2009-06; git log -p
pmichaud my brane is a little fried today :-| 21:30
moritz_ (I don't try to be picky, I'm just curious, and want to do it right next time)
PerlJam pmichaud: stop dipping it in hot oil!
pmichaud: long day?
pmichaud day*s*
moritz_: you did right.
21:31 hoelzro left
moritz_ speaking of long days, I should probably end mine right now. 21:31
PerlJam moritz++ again for Chicago
21:31 pmurias left 21:32 toyam left
jnthn pmichaud: In PARROT_REVISION 21:33
Do I remove the 1.4.0?
If I'm bumping it?
moritz_ yes.
jnthn Danke.
21:33 pmichaud_ joined
pmichaud_ I think I'll change the release guide to indicate that unresolvable failing spectests should be commented out of spectest.data before release and then immediately commented back in 21:34
but that'll be a future change
21:34 pmichaud is now known as pmichaud__, pmichaud_ is now known as pmichaud, pmichaud__ is now known as pmichaud_ 21:39 hoelzro joined
jnthn oh f**k 21:39
Rakudo doesn't even build against latest Parrot. 21:40
pmichaud yes, lots of Parrot deprecations landed on tuesday.
We have to apply some patches to catch up.
jnthn gah. :-/
Were the patches supplied?
Or we gotta write them? 21:41
pmichaud there are a few available here and there
but I don't know that they're complete
jnthn ugh
pmichaud one of the biggest changes is the .pmc files, though
jnthn Well, it's quarter to midnight and I can't be arsed. So much for landing this metaclass refactor tonight. :-(
21:42 maja_ joined
pmichaud I think cotto provided a patch for that 21:42
21:42 toyam joined
jnthn OK. I'll probably sort it out tomorrow, if nobody beats me to it. 21:42
jnthn doesn't see the patches in RT anyway 21:45
Do you know where they were posted?
21:54 maja left 21:56 NewOrder joined 21:57 justatheory left 22:00 Jedai left 22:01 justatheory joined 22:02 decasm left, M_o_C left 22:06 rfordinal left, rfordinal3643 joined, toyam left 22:07 hoelzro left 22:08 rfordinal joined 22:09 hoelzro joined, rfordinal3643 left, rfordinal3643 joined, nihiliad1 left 22:10 rfordinal left 22:12 hoelzro left 22:13 alester joined 22:15 hoelzro joined 22:16 hoelzro left 22:17 meppl joined 22:18 kane___ left 22:20 alester left 22:29 justatheory left 22:30 pmichaud left 22:31 rfordinal joined, rfordinal3643 left 22:32 cls_bsd joined, rindolf left 22:39 toyam joined 22:47 explorer left 22:50 ihrd joined 22:51 explorer joined
jnthn gets some rest, will try and fix up Rakudo build under latest Parrot tomorrow. 22:52
23:00 justatheory joined 23:12 nihiliad joined 23:16 frew_ joined 23:17 justatheory left 23:20 pyrimidine joined 23:23 kane___ joined 23:24 Limbic_Region joined, hoelzro joined 23:25 pyrimidine left 23:27 kane___ left 23:31 kidd_ left 23:34 hoelzro left 23:36 toyam left 23:38 frew_ left, frew_ joined 23:42 kane___ joined 23:45 japhb left 23:47 hoelzro joined 23:49 rfordinal left 23:51 hoelzro left 23:56 alester joined 23:58 unitxt 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!