»ö« | perl6.org/ | nopaste: paste.lisp.org/new/perl6 | evalbot usage: 'perl6: say 3;' or rakudo: / pugs: / std: , or /msg p6eval perl6: ... | irclog: irc.pugscode.org/ | UTF-8 is our friend!
Set by diakopter on 25 January 2010.
pugs_svn r29610 | colomon++ | [t/spec] Fudge for ng1. 00:18
r29611 | colomon++ | [t/spec] Fudge for ng1. 00:20
lulu_ anyone know about the latest programming languages? 00:22
hellloooooooo?
diakopter lulu_: hi
lulu_ can u help me? 00:23
diakopter with what
lulu_ about the latest programming languages
lulu_ helo people. 00:24
help me. duhh
go|dfish you are probably beyond help. 00:25
lulu_ why is that ? 00:26
Tene lulu_: What proramming language do you want help with?
lulu_ the latest version. anything
helooo 00:28
Tene Hi! 00:34
Oh, it left.
diakopter wow. 00:52
the latest version of *any* programming language.
Juerd We haz it.
diakopter "as long as its the latest [and greatest] version, it doesn't matter which language you choose [for your homework project]."
it 00:53
's
_ash i wish it could be that easy, i haven't ever had a programming assignment i could do in 'any language'
even if xxxx language was the latest and greatest versions 00:54
Juerd I only have programming assignments I can do in any language. 00:56
diakopter I figured it was a "research a programming language" assignment 01:04
for a computer-skills or intro to programming for high schoolers class. 01:05
colomon pmichaud: tests like my @a; is @a.end, -1 are failing left and right. ng1 bug or spec issues? 01:10
colomon pmichaud: IMO we should make ng1 ng ASAP (tomorrow morning, maybe), and then schedule some time next week to make it master. 02:39
seems like that would avoid having a mostly broken master while giving us something to push for. 02:40
Hmmm... looks like passing @ parameters is broken in ng1? 02:43
> sub a(@b) { @b.perl.say; }; a(1...10);
Null PMC access in get_string()
> sub a(@b) { @b.perl.say; }; a(1,2,3);
Null PMC access in get_string()
ash_ > sub a(@a) { say 'hi'; }; a(1, 2); 02:44
Null PMC access in get_string()
its not even calling it, i think 02:45
pugs_svn r29612 | colomon++ | [t/spec] Fudge for ng1, plus fix typo. 02:46
colomon ash_: I'm pretty sure it's dying in the parameter binding, or something like that.
certainly 1...10 works.
colomon ng: sub foo(*@args, *%named) { 1 }; &foo\ .() 03:00
p6eval ng 58d896: OUTPUT«Confused at line 1, near "&foo\\ .()"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (src/stage0/HLL-s0.pir:328)␤»
colomon > my @a = 1..10; @a[2] = 10000; say @a; 03:13
Cannot assign to readonly value
So, I've patched some tests and turned others off. There are still 12 or so failing test files in the ng1 spectest. (That's as compared to standard ng.) I'm hoping that number can be zero 24 hours from now.... 03:20
but I've got to get to bed before I fall asleep at the computer.
tushar hi 03:55
need some documentations on perl6. noob for perl6, developer using perl5. 03:56
tushar Waiting for response 03:57
ash_ well, there is a good book thats being worked on, you can download the pdf of it at github.com/perl6/book , or you can read the spec
tushar thanks ash_ ! 03:58
tushar wgeting github.com/perl6/book
ash_ ummm thats not a direct link to the pdf 03:59
cloud.github.com/downloads/perl6/bo...009-12.pdf is a direct link
tushar ash_: ya sound's like it's a git hub repository and found lots of directories but books uh!. looking for the pdf link now 04:02
ash_ its in the downloads section 04:03
tushar thanks ash_ ! 04:04
sjohnson failblog.files.wordpress.com/2010/0...e-fail.jpg 06:38
Su-Shee good morning 07:21
masak o/ 07:38
phenny masak: 04 Jan 13:38Z <moritz_> tell masak maybe perlgeek.de/blog-en/perl-6/perl-6-i....writeback inspires you while writing spec changes summary ... though it's very coarse and selective
masak: 09 Jan 21:23Z <mberends> tell masak is this you? irclog.perlgeek.de/perl6/2010-01-09#i_1894733
masak: 25 Jan 03:27Z <japhb> tell masak What is the current status of proto's installed-modules branch?
masak moritz_: thanks. maybe I'll get around to it. 07:39
masak mberends: funnily enough, I found out about that translation days prior. :) 07:39
masak japhb: very close to being merged. though it may not run on ng as such. 07:40
japhb masak, wow, that's one hell of a ping RTT. ;-) 08:06
masak, but thank you, that's excellent news.
masak sorry 'bout the lag. :)
it's good to be back. 08:07
masak has a look at proto right now
japhb masak, you merge installed-modules, and I'll mine shiny new tuits.
np, welcome back
masak as far as I recall, there was only one module collision left. 08:08
something about URL.pm in both november and url.
masak checks
ah. URI.pm in november and uri. 08:09
well, I'll just rename the former to November::URI and add a comment about eventually moving over to the latter. 08:10
slues YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa 08:22
warily YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa 08:22
smocked YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa 08:22
chuckle YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa 08:23
masak oh, I heard about your spam problem.
thought that'd be solved by now.
hyped YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa 08:25
belts YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa 08:25
hyped YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa
belts YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa 08:25
masak Tene++ 08:26
beastly YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa
masak thanks.
beastly YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa
altered YO oN Da ReaLz I Be A StrAIght Up DowN aSS HooOD NiGGAaa. I BEEn on My SlANG sINce I was TwO. PuSHIn Dat ShIEt BAbYA gOTza HuTle Fo DAT ChedDaAAAAA. i gO TO dA CLUB WIT mAH BOIZ AnD gETS dAT pUSsSAAAAAYY. nAw meAN hAHAHA. BUT cHEAY fO yO LADiEZ HOLLA AT yO bOIIIII I TREAT YOUu fO REALZ lIKE A bLACk PRiNCEsS. irc.hardchats.com #gnaa 08:26
masak surely kickbanning on a case-by-case basis isn't a viable long-term solution, though? 08:27
Tene Right.
Tene I'm looking... 08:27
Tene There, I think that does it. 08:28
Tene That's going to only allow registered people to talk in the channel. 08:29
masak should do it. Tene++
moritz_ good morning 08:50
masak o/
dalek vember: c5fd6a4 | masak++ | (2 files):
[Makefile] added to version control
 09:24
vember: 95134d4 | masak++ | (5 files):
[URI] renamed to November::URI
moritz_ masak: good morning and welcome back 09:25
masak thanks. :)
masak phenny: tell carlin that I finally addressed irclog.perlgeek.de/perl6/2009-12-12#i_1828479 09:31
moritz_ oh wow, now the IRC logger is also a bug tracker :-) 09:32
masak phenny seems to have been muted along with the spambots.
according to carlin, that should be the last conflict. at least it was in December, and I don't think the landscape has changed that much. 09:33
masak there's been a slight slump in new modules. :) 09:33
or projects, rather.
moritz_ hugme: hug masak 09:34
masak hugme is muted, too :)
moritz_ :(
and p6eval
masak and all the black princesses. 09:35
mathw Morning 09:38
masak mathw: \o
mathw masak!! 09:39
masak mathw!!
frettled yayitsmasak! 10:05
masak \o/ 10:05
masak does a couple of free-wheels on the floor of #perl6
I can't do that IRL, so I might as well do it on IRC. 10:06
frettled is all "ooh" "aah"
masak bows
but now, let's get to work. we need to restore peace and justice to the galaxy. there's much on the agenda. 10:07
masak has another look at proto 10:10
moritz_ is always happy when a new feature is implemented by a diff that has only inserts 10:13
it's a bit superstition
masak :) 10:14
I'm concerned about this commit: github.com/masak/proto/commit/13484...b8c54a3731
moritz_ but it looks to me like I was careful enough previously so that the new feature is not very intrusive
masak I'm thinking I might invoke my dictatorial powers and revert it.
moritz_ hugme: register 10:16
masak Marc Chantreux, if you're out there: sorry -- I liked the intent of your commit, but proto is the wrong vessel for something like command flags. maybe in its next incarnation, it'll have them. 10:16
moritz_ hugme: hug masak 10:19
it should register itself with freenode now
jnthn omg masak?! 10:20
masak a while ago, I read (a fictional account) about an IT workplace where they installed a hugging machine. it quickly became very popular.
jnthn! dear old chap. :) 10:21
frettled jnthn: it's an optical illusion
jnthn frettled: Oh...
frettled masak: for some reason, that reminds me of JPod
jnthn deflates and goes to make coffee
masak frettled: that's the one.
frettled I'll go submit my lottery ticket now. 10:22
moritz_ hugme: register 10:23
moritz_ hugme: verify_register 10:25
moritz_ hugme: verify_register 10:26
hugme: hug frettled
hugme hugs frettled
moritz_ \o/
masak :)
moritz_++
moritz_ that's a PITA to do non-interactively 10:27
setting up the account, verifiying it etc.
masak the things we do for bots...
moritz_ which reminds me, I wanted to add an interactive console to hugme
providing the password to the POE::Component::IRC->spawn() method doesn't log it into the freenode account 10:28
frettled hugme: hug moritz_
hugme hugs moritz_
frettled :D
moritz_: my IRC client automatically registers after a successful login, so that ought to be possible for the bot :) 10:29
moritz_ frettled: it's possible, but not the way I originally thought it would 10:29
frettled ah 10:30
moritz_ hugme: list projects 10:31
moritz_ hugme: list projects 10:32
hugme moritz_: I know about book, gge, hugme, ilbot, json, november, nqp-rx, nqpbook, perl6-examples, proto, svg-matchdumper, svg-plot, temporal-flux-perl6syn, tufte, web
masak by the way, GGE is really shaping up. I'm almost at 100% conformance with PGE. things like captures work now: gist.github.com/284959 10:39
masak and the tree spider is gone. 10:40
jnthn masak: Nice. 10:41
masak yes. I've learned loads even since BPW. 10:42
jnthn masak: Ooh! code-gen!
masak yes!
masak after it completely melted my brain, it was actually quite easy. 10:43
moritz_ that's how I always figured that regexes engines must be 10:43
masak code-generating? or brain-melting-but-then-simple? 10:44
moritz_ the latter
masak yes. that's why I think the 7-wonders tutorial will be important. 10:45
it's a way to unload the things I've learned in a graspable format.
jnthn Does your brain re-solidify after the melting process is over?
masak yes, but it's forever changed by the knowledge of code-generating regex engines. 10:46
it remains to be seen whether that knowledge can be faithfully transferred to an unmelted mind.
masak kinda liked this comment: perlmonks.org/?node_id=817738 10:53
colomon o/ 11:27
masak colomon: \o
colomon masak: Good to see you back. I think we're about have to the mother of all target-rich environments for bug-hunting... 11:28
masak we are. 11:29
we'll be the eye of a mightly storm of RT activity.
pugs_svn r29613 | colomon++ | [t/spec] Fudge for ng1. 11:48
r29614 | colomon++ | [t/spec] Fudge for ng1. 11:55
pugs_svn r29615 | colomon++ | [t/spec] Fudge for ng1. 11:58
pugs_svn r29616 | colomon++ | [t/spec] Fudge for ng1. 12:01
colomon Slowly but surely....
jnthn colomon: How much has ng1 lost over ng? 12:02
colomon jnthn: depends on ow you look at it
there were 20-something failing spectest files when I started 12:03
most seem to be only a couple of tests failing
colomon but they are kind of catastrophic failures. 12:04
things like my @a; not working
or @a[0] = 1 not working.
moritz_ is it really worth to fudge such things out? 12:05
or would it be better to wait until somebody fixes it?
colomon well, I've been working around the first with my @a = () 12:06
but the problem is that spectest was just too noisy to be useful, which is a bad state of affairs 12:07
so I think it's worthwhile, even though I hope most of the fudges are undone this week. 12:09
colomon s/this week/today/ :) 12:18
colomon okay, ng1 spectests clean here, and I think I've got everything pushed / checked in. 12:34
big issues seem to be the ones mentioned above plus trying to use iterators like they are old-fashioned lists. 12:35
Method 'elems' not found for invocant of class 'GatherIterator' is a typical example. 12:36
colomon I'm pretty sure I can fix the latter, but don't know how kosher my fixes would be considered. 12:37
afk...
colomon ...btk 13:13
masak ok, so proto's installed-modules branch is actually looking pretty good. 13:15
colomon \o/
masak I'm thinking of merging it.
colomon that's huge news.
jnthn Woo. 13:16
masak most of the work has been done by mberends++ and carlin++
colomon Merge it! and then I propose we make getting ng to work for it a high priority. 13:17
masak sounds like a plan.
still only at step 13 of 17 in the mberends++ TODO, but I don't think that matters. 13:18
takadonet morning all 13:18
masak takadonet: \o 13:19
colomon o/
btw, does anyone know if p6 is spec'ed to have a function which takes a list and returns the first $n elements of it? 13:20
seems to me like that is highly desirable in a world of potentially infinite lazy lists... 13:21
masak @list[^$n]
colomon hmmm, makes sense. and completely fails in ng1. :) 13:23
colomon no array slices, I guess. 13:24
masak at one point it was like that in master, as well :) 13:25
colomon still kind of likes the idea of a .first-n method... 13:25
moritz_ colomon: like 'take' in Haskell? 13:26
colomon moritz_: I don't remember (long time since I last messed around with Haskell), but probably.
moritz_ multi sub infix:<take>(@LHS, Int $n) { @LHS[^$n] } 13:27
Prelude> take 5 [1..100] 13:28
[1,2,3,4,5]
colomon will also be happy when sink context is properly eager in ng1... 13:32
pmichaud good morning, #perl6 13:53
moritz_ good morning pmichaud
takadonet morning pmichaud
frettled lolitspmichaud :) 13:54
mathw good localtime pmichaud 14:03
colomon pmichaud: morning! 14:13
pmichaud: the good news is the ng1 spectest now tests clean. the bad news is I've got a pile of bugs for you. 14:14
takadonet hehe
masak morning, pmichaud. 14:37
pmichaud (pile of bugs) excellent! squash bugs, squash bugs! 14:44
frettled there's hardly anything better than mashed bugs, rich in protein 14:50
masak boil 'em, mash 'em, stick 'em in a stew. 14:52
I'm all for squashing things, but usually I squash features. 14:53
masak proto is now merged. 15:11
colomon \o/ 15:17
jnthn back 15:23
takadonet welcome back jnthn
jnthn :-) 15:24
mberends pulls proto, and wonders how it would go with ng/ng1 15:32
pmichaud (squash features) -- isn't that what microsoft calls "patch Tuesday" ? ;-) 15:39
masak I'm weekday-independent.
and I prefer to deflate features by discovering flaws that are already there, not by contributing new ones. 15:40
TimToady we should rename getitems to a more user-friendly name if it will be used like Haskell's take 15:58
frettled and the debugging function could be called "masak" 15:59
frettled TimToady: gettake? ;) 16:00
jnthn .gimme
TimToady gimme ain't half bad 16:01
masak C<temp> is defined (by S06) as a macro which replaces an existing thing temporarily within a given scope.
I understand the mechanism by which the macro affects the statement with C<temp> in it. By which mechanism does the macro affect the end of the scope?
TimToady usually by installing a LEVAE 16:02
LEAVE even
masak ah. cunning.
TimToady let uses KEEP or UNDO
masak that would even stack correctly.
pmichaud ...and one of the reasons rakudo doesn't have 'temp' yet is because we don't have good leave semantics in parrot 16:03
TimToady being able to fire off actions after a longjmp is one of the most important features of Perl 5, internally 16:03
TimToady I'm quite willing to put off generalized temp till after 6.0 though, as long as we have dynamic vars (including the special 'temp $*foo' form) 16:07
(which doesn't need a LEAVE)
PerlJam frettled: "gettake" makes me think it should be pronounced like "bukake" for some reason. 16:08
masak sounds reasonable to me. 16:10
doesn't 'bukkake' have a double 'k', by the way? not that I would know.
hudnix transliteration of japanese words to Roman characters is rigorous. 16:12
er, *not* rigorous.
PerlJam masak: Im pour at speling.
hudnix Like perl, tmtowtdi 16:13
masak :)
TimToady now I'm trying to figure out why you'd want to hang a Buddha :) 16:14
or borrow a hit
frettled heh
But yes, the normal transliteration is «bukkake».
PerlJam indeed, the doubled "tt" may have even been the trigger that made me think of bukkake to begin with. 16:15
TimToady or maybe pull a punch
pmichaud (yes, we have dynamic vars in rakudo. I'm still working on the special 'temp $*foo' form)
TimToady it's just 'my $*foo = CALLER:<$*foo>' really 16:16
jnthn Oh great, I think I'll glance here for light debugging relief and the subject is bukkake...
TimToady it's possible we shouldn't really reuse 'temp' for that
masak TimToady: is 'to hang a Buddha' a euphemism in Japan? :)
jnthn TimToady: temp kinda made logical sense to me there, but it does feel a tiny bit odd.
frettled Wikipedia is exceptionally helpful regarding «bukkake», but perhaps NSFW and NSFK. 16:17
TimToady the primary damage is that it seems to prevent temporizing some caller's dynamic var
maybe that's a feature :)
frettled "Daddy, daddy," said the little girl, "can you explain why ..." 16:18
and you'd better hope that daddy's being asked to explain features of Perl 6 :)
TimToady okay, it's related to 'hit' and 'hang', but with metaphorical overloaded meanings 16:20
masak seems so. 16:21
pmichaud "Ours not to explain why. Ours but to 'say' and 'die'"
TimToady no wonder it wasn't in my Japanese dictionary :)
TimToady so, um, PerlJam, how are you acquainted with that word? Or were you thinking of some other word? (TimToady gives PerlJam a way to save face.) 16:23
pmichaud (PerlJam: When already in a hole, sometimes it's best to stop digging :-) 16:25
japhb "Save face" takes a whole new meaning in that context"
pmichaud, what, before you hit magma?
TimToady (PerlJam: here's your cue to say you meant 'bukkaku', yeah, that's the ticket) 16:26
japhb masak, if I read the backlog properly, it sounds like installed_modules has been merged, but proto won't run on ng yet?
pmichaud it wouldn't surprise me if proto doesn't run on ng yet. ng is missing some features 16:27
masak japhb: correct. haven't tried to run proto under ng, but cf. snowballs in warm places.
japhb roger that.
pmichaud okay, so yesterday I postulated that perhaps we should simply bite-the-bullet and make ng1 into the new master tomorrow morning. Any thoughts yay/nay ? 16:28
japhb I've been away a while ... what is the difference between ng and ng1?
PerlJam pmichaud: +1
masak pmichaud: it's your call. 16:29
pmichaud ng1 is the branch of ng where we've prototyped iterators and lazy lists (and seems to be the direction we'll go)
one important thing to note is that I'm not likely to be around much Thu-Mon (I'll be online some, but at odd hours)
masak now that proto's installed-modules is merged, I thought I'd focus a bit on ecosystem-wide testing. I think that'll wash out tons of regressions not caught by the test suite. 16:30
japhb Has master been touched at all since the release? 16:31
masak but I don't want ng progress or merge status to stall on that.
pmichaud japhb: not significantly, no
PerlJam TimToady: btw, I don't see a need for face-saving. I'm familiar with that word like many other bizarre things because I hang out on IRC :) 16:31
masak .oO( wonder if Buddha hangs out on IRC ) 16:32
pmichaud I believe the Buddha nature is everywhere :-)
masak Mu.
PerlJam masak++ 16:33
hudnix ugh
japhb pmichaud, then I'd say might as well bring ng1 to the forefront sooner rather than later, to maximize time with everyone working on it, and just put something in the readme about using the January release if they can't spend tuits working with ng problems yet.
pmichaud japhb: vote noted, thanks! 16:33
japhb: I'm in some agreement there (which is why I proposed this) -- I think we should just commit to the new version even though it's a significant regression, and hopefully that can spur us on a bit more 16:34
pmichaud and also remove any questions about where our focus lies :) 16:34
japhb idly wonders if any of his Perl 6 code has a chance of running after the changeover. :-)
exactly.
PerlJam pmichaud: yep. Rakudo has been pregnant with ng since Oct and people have been itching for it to give birth given that Rakudo Star looms on the horizon 16:36
pmichaud okay, so should I do it tomorrow or Tue of next week?
or delegate to someone else? ;-)
PerlJam sooner rather than later IMHO
pmichaud okay, plan is to switch tomorrow. note objections in the backlog :) 16:37
(I'll check the logs before committing)
masak just to be clear here: as soon as ng1 is called master, I can report whatever I want to RT, right? :) 16:38
pmichaud masak: right
masak \o/
pmichaud masak: consider it a birthday present or something :)
masak this is the bestest... ever! *sniff*
pmichaud although we know there will be a lot of bugs. I might recommend holding off on an rt onslaught until mid next week :)
mberends
.oO( remember to upgrade RT's storage capacity )
pmichaud it'd be like shooting fish in a barrel :-) 16:39
actually, more like shooting fish in a barrel filled with nothing but fish
colomon japhb: the answer is almost certainly no.
masak locks and loads
colomon japhb: I'd be surprised if any randomly chosen Perl 6 code of over ten lines works in ng1.
pmichaud so, I'd say after Tue of next week, rt bug reporting is fair game :-) 16:40
on the other hand, we may also be able to eliminate some active bugs in rt :)
PerlJam wait ... so what happens Tue of next week? 16:41
masak very well. I'll hold off a week.
japhb colomon: noted. Especially since I tend to write fringe code in the first place. :-)
masak the goal, obviously, is to minimize the failures in the cheese by Feb release. 16:41
colomon japhb: I mean, unless pmichaud's hacked since I last checked, simple things like my @a; and @b[0] = 1 don't work right now. 16:42
masak either by fixing rakudo or by workarounds in the cheese.
colomon masak: the goal is that Feb is full of awesome.
pmichaud Tue of next week is when I get back from my absence this weekend :)
and can more actively work on removing bugs
masak colomon: that sounds like the lyrics of a song. :)
PerlJam pmichaud: but how will you know which bugs exist withoug a nice fat RT queue for you to tick off? ;) 16:43
pmichaud PerlJam: ummm, "make spectest" ?
"my @a;" doesn't work? 16:44
PerlJam is just being goofy 16:44
jnthn ng: my @a;
p6eval ng 58d896: ( no output ) 16:45
PerlJam Besides, jnthn will be working, yes?
colomon pmichaud: well, it doesn't complain when you do it, but you can't do anything with @a afterward.
pmichaud colomon: okay, I'll fix that.
masak ng: my @a; say @a.perl 16:45
pmichaud anyway, depending on other events around here between today and tomorrow, I should be able to get lists and arrays working
p6eval ng 58d896: OUTPUT«[]␤»
colomon at least, there were a zillion tests like my @a; @a.push(1); which failed.
jnthn PerlJam: I'll have a decent bit of Rakudo time from a day or two's time, and over the weekend. :-)
colomon > my @a; say @a.perl
Null PMC access in elements()
(that's what I get in ng1) 16:46
pmichaud if someone wanted to tackel s/IterDone/EMPTY/g that'd be useful
*tackle
colomon is there more to it than just search and replace?
PerlJam can't help but mentally insert a "G" before IterDone every time
pmichaud also, change RangeIterator to RangeIter and SeqIterator to SeqIter
of course, when a SeqIter is finished it becomes a non-seqiter :-)
masak groans 16:47
pmichaud afk, fetching kid lunch 16:48
colomon working on RangeIterator...
PerlJam colomon++
colomon Oooo, filenames need to be changed too. 16:53
afk, lunch
TimToady oh, it's getsome that needs renaming, not getitem 17:05
though getitem might be usefully renamed to getparcel 17:07
or maybe getslice
TimToady except it would be odd to use getslice for a positional parameter 17:08
TimToady getparcel is also strange if the next thing isn't a parcel, but I suppose anything can be considered a degenerate parcel 17:10
getitem unfortunately sounds like you're getting one item from the next parcel, which is wrong
that's what get does
TimToady pity the nibble/bite metaphor won't work 17:12
masak phenny: tell colomon that I pulled some guerilla commits and did the replacements in ng1. 17:14
or don't. whatever.
TimToady isn't registered, prolly
masak 'aight.
colomon wouldn't it be easier to just tell me, anyway?
:)
masak thought you were at lunch. 17:15
colomon I was.
groovy. 17:16
masak++ 17:17
masak it even compiles afterward. 17:18
but for some reason I can't run the spectests. thinking of doing a clean checkout.
colomon ah, so you haven't run the spectests yet? 17:19
woah, make test just blew up.
masak so maybe it's not just me :/
colomon I'm doing a make clean; make now. 17:20
nope, still massive fail. 17:22
masak :/
colomon looks like "say 1..24" is failing. 17:23
(for instance, that's from 00-parrot/01-literals.t
) 17:24
masak and that worked before?
colomon yes.
masak I tried 'say "OH HAI"' on the interactive prompt before cleaning.
got a Null PMC access.
colomon same here.
also with "hello".say 17:25
masak conclusion: my s///g changes caused something to blow up.
shall I revert them?
colomon let's try reverting them just on our own machines and see what happens.
masak nod.
masak hm. after a clean download, 'make test' passes. 17:28
colomon Okay, I'm trying to ... really?
masak trying 'make spectest'.
oh wait. belay that. :/ 17:29
I was still on master.
colomon It's the EMPTY patch. I reverted it (and tweaked Makefile.in to use IterDone again) and things seem to work. 17:30
make test passes.
masak ok.
feel free to push the revert commit.
colomon I'm really more concerned with how something so simple could break the universe. 17:31
masak that is the question.
pugs_svn r29617 | colomon++ | [t/spec] Change RangeIterator to RangeIter. 17:33
colomon should I push the revert commit or not? indecision here.... 17:34
masak please do. 17:34
I'm leaving soon for $NOM, and it feels better to leave the branch in a working state. 17:35
colomon okay, on it.
done. 17:36
pmichaud EMPTY didn't work? 17:47
masak apparently not.
pmichaud ....weird.
masak nature abhors a vacuum? :) 17:48
jnthn That'd suck.
masak :P
pmichaud I'll give it a shot 17:51
(shot in the dark?)
(fish in a barrel?) 17:52
colomon Seriously, there's nothing apparently wrong to the eye about masak's patch, but it blew up the world.
pugs_svn r29618 | lwall++ | [S07] pick slightly less misleading names for iteration primitives
r29618 | stop using "item" to mean contradictory things
r29618 | remove confusing discussion of outer flat/slice context; these methods don't
r29618 | care because that's why there are different method names for get vs getobj
pmichaud github down :-( 17:57
colomon nooooooo........ :( 17:58
pmichaud hmmm, my s/IterDone/EMPTY/g seems to have worked 17:59
at least enough to pass 'make test'
checking 'make spectest'
so far so good
masak huh. 18:00
maybe I missed a spot.
masak food & 18:02
pugs_svn r29619 | pmichaud++ | [t/spec]: Update range-iterator.t with IterDone=>EMPTY . 18:16
pugs_svn r29620 | lwall++ | [S05] more @@ whackage 18:20
r29620 | [S07] define flat and slice methods in terms of get and getobj
r29621 | lwall++ | [S05] forgot to change date 18:22
Tene It's not using &fail to return the Empty yet, is it? 18:36
pmichaud no. 18:37
Tene I might look at that tonight.
Considering my recent history... it's unlikely I'll do anything tonight, though.
sjohnson hi masak long time no see 18:54
colomon pmichaud: did you have a scheme in mind for handling the use of the old standard List methods on iterators? things like .elems? 19:58
pmichaud I'm not sure if Iterators does Positional or not. If so, then the Iterator probably needs to convert to a Seq and do that. 20:00
Under that interpretation, .elems should also be processed on a Seq, I guess.
Either that or .elems could be directly on the Iterator, simply doing .get(item) repeatedly. 20:01
For now I'd put .elems directly on the Iterator -- we can update later.
colomon okay, I'll go ahead with that...
pugs_svn r29622 | lwall++ | [S06] more @@ clobberation. 20:25
r29622 | also remove * as a feed target, just use a variable.
r29622 | conjecture a **() special form for interpolating into slice lists
TimToady Iterator doesn't do Positional, but method requiring a Positional object can probably do Cool-like autocoercions to Seq 20:26
colomon pmichaud: I've pushed Iterator.elems. Should I do the same thing to create an Iterator.Str? 20:27
TimToady: oooo.
pmichaud well, for .elems I'd prefer to "optimize" it a bit and not keep track of the reified elements.
TimToady what does say $it.elems; say $it.get; do? 20:28
pmichaud I would presume the number of items left in the iterator, followed by a thrown exception
TimToady that is, does .elems throw away the values?
colomon TimToady: it has to, doesn't it? 20:29
pmichaud it doesn't have to, it could convert the iterator into a Seq
TimToady that's not so good
TimToady Iterators should just iterate, and let the type system coerce, I expect 20:29
colomon pmichaud: Even if it converts the iterator into a Seq, the iterator is going to be consumed, isn't it?
TimToady a la Cool
pmichaud that's kinda weird, though
I mean: @foo.map({...}).elems 20:30
TimToady well, if you throw away a Seq, it gets GC'd 20:30
pmichaud right, but if the .map produces a 1-million element Seq...
TimToady I guess that argues that the user wants to throw them away 20:31
pmichaud I'm somewhat going on the "iterators tend to be invisible" principal here
and that if you somehow got a hold of a visible iterator, you know what you're doing with it
I'm fine if we auto-coerce iterators to Seq in all instances... it just seems sub-optimal in some common cases
TimToady so maybe it's okay if .elems consumes presuming we don't have another alias around
pmichaud that's why I started with the idea that perhaps .elems is okay on Iterator directly, but other Positional type things make it into a Seq 20:33
TimToady well, upgrading to Seq isn't gonna help my example anyway
pmichaud right
TimToady unless we do it in-place, which is antisocial
colomon oh, good, I'm glad that wasn't just me being dense.
TimToady it was all of us :) 20:34
pmichaud do parcels still tend to upgrade to Seq ?
or just not in-place ?
(in older rakudo, parcels tended to become Seq in-place)
TimToady well, my brand-new spec probably contradicts itself on that
pmichaud okay
pmichaud in general I'm heavily biased against any sort of in-place operations until it's clear that it's the only way to do something :) 20:35
TimToady currently I think explicit .slice does, but not binding via .getobj
pmichaud oh, I would think that .slice would return a new Parcel :-)
TimToady but maybe it does there
I mean, what .slice returns instead of internal parcels
ruoso TimToady++ # last changes are just great
TimToady ruoso: I'm trying to simplify... 20:36
just haven't got up the nerve to attach @@ in S09 yet, so I'm attacking it everywhere else first
*attack
pmichaud colomon: anyway, .elems on Iterator is fine for now. I'd go ahead and have .Str pass through Seq for the time being 20:37
eh, or maybe that should be the other way around... Seq's Str passes through Iterator.Str ? ;-)
ruoso TimToady, I'm not sure we'll need @@ sigil at all 20:38
if I understood right
colomon pmichaud: I don't have a clear notion how to handle passing through Seq? (your most recent comment has me even more confused, alas.)
ruoso my @a = ((1,2),(3,4))
still doesn't flatten or unflatten it
TimToady ruoso: I've been destroying @@ for several days now
pmichaud colomon: method Str() { self.Str.Str }
sorry
method Str() { self.Seq.Str }
colomon pmichaud: oh!
pmichaud an Iterator needs to know how to turn itself into a Seq :-)
ruoso TimToady, yeah... I saw it... but I understood from your saying that you were trying to find an alternative to it for S09 20:39
colomon pmichaud: do we have Iterator.Seq yet?
pmichaud colomon: don't think we do
colomon yeah, I don't see it in there.
TimToady no direct replacement for @@, but using ** in a couple spots to indicate sliceyness
in contrast to * which indicates flatness 20:40
colomon pmichaud: I need to run another load of boxes to the storage facility, back in about an hour or so.
pmichaud colomon: okay
ruoso @@ was required before because list assignment was eagerly flattening
since that's no longer the case
we can simply drop it
and use plain @
TimToady we haven't change list assignment that I know of. That's still flattening
ruoso ah... 20:41
right...
TimToady binding is another matter 20:41
ruoso ahm...
TimToady and item assignment certainly upgrades to Seq 20:41
ruoso and so are feeds
diakopter +1 simplify 20:42
simplify |= maxvalue 20:43
TimToady it's a good bet that the spots where nobody has tried to implement them yet might be indicative of spots where simplification is warranted
ruoso TimToady, I'm not sure it's something spec-wise... but I've been thinking of iterators in the same sense of unix pipes... 20:44
in a sense that each end of the iterator will run a different implicit thread
TimToady certainly I think of feeds that way
ruoso only feeds?
TimToady to the point of possibly requiring braces around
just to keep things honest lexically 20:45
ruoso so...
map { ... }, grep { ... }, map { ... }, @stream
won't paralellize unless feeds are used
TimToady they can if there's no side effects, but ==> promises good behavior there to a greater extent, I suspect 20:46
ruoso right... so, conceptually, regular map and grep can only paralellize as an optimization...
while feed operators do paralellize by default 20:47
TimToady anyway, as we develop feeds, we'll have to think very carefully about how to limit variables being shared accidentally
nodnod
it's a matter of capturing programmer intent again
darn it, I've forgotten to eat lunch again...my brain is starting to issue warnings... 20:49
avocado & 20:50
sjohnson yum yum
ruoso TimToady, anyway... I think feed operators need to be defined in terms of Iterators... 20:51
diakopter mmm... avo... 20:51
ruoso TimToady, with two control exceptions "BufferEmpty" and "BufferFull" that would control the scheduling of the threads... 20:54
PerlJam We've already got EMPTY, so we just need FULL :) 21:00
sjohnson :) 21:15
szbalint anyone heading to FOSDEM from here, btw? :) 22:05
jonasbn szbalint: good evening 22:07
szbalint: good work on the event front
szbalint heh, I'm still not gabor ;-) 22:29
PerlJam szbalint: you and he should have made a picture where you are dressed identically and mirror each other's pose. 22:33
szbalint PerlJam: that could be arranged =) 22:43
colomon pmichaud: Iterator.Seq and Iterator.Str checked in. I'm a bit uncertain about how it creates the Seq, but it does seem to work. 22:44

Powered by Cro and the Raku® Programming Language.

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

Thank you!