6.2.7 released! | pugscode.org/ <Overview Journal Logs> | pugs.kwiki.org | paste: sial.org/pbot/perl6
Set by autrijus on 13 June 2005.
svnbot6 r4930, mugwump++ | Don't ship Date yet - tests not even failing gracefully 00:03
r4931, theorbtwo++ | Better errors for cast failures (via the joys of copy-and-paste -- I'm sure
r4931, theorbtwo++ | there's a better way, but this way worked.)
clkao oom issues? 00:07
autrijus: Agent Tang's ticket has be handed over stig 00:12
autrijus clkao: yup, understuud. thanks! 00:13
clkao have fun!
and produce something useful before i come back townm
autrijus I try, I try :) 00:19
theorbtwo Autrijus: give me a good todo that doesn't involve hacking makefiles? 00:25
svnbot6 r4932, luqui++ | Fixed a very small bug.
autrijus theorbtwo: sure... investigate failing cases from smoke -- :todo<feature> or :todo<bug> the non-regressions, and see if you can repair some obvious hardfails 00:27
192.168.0.123/smoke.html
mugwump nice url 00:28
autrijus is a sort of up to date graph -- you can ignore the exists.t and array/has/h tests for now; I'm working on that
autrijus shuts down dereferencing and suddenly get unexpected successes 00:31
Darren_Duncan back later 01:28
mugwump is it just me, or are all the links in this fortnight's perl6 summary broken? 01:35
svnbot6 r4933, luqui++ | Ze estrippa vorks.
autrijus mugwump: all Date tests are currently broken 02:02
svnbot6 r4934, autrijus++ | * Instead of implicitly derefercing and evaluating in rvalue
r4934, autrijus++ | context when deriving reference types, we now simply look
r4934, autrijus++ | at the refType for constant scalar references. This means that
r4934, autrijus++ | sub foo ($x) { $x.ref } my %y; foo(%y);
r4934, autrijus++ | no longer evaluates to List, but to Hash now.
r4934, autrijus++ | * Also cleaned up related tests.
autrijus mugwump: can you look into it? I'd like to enter preflight if possible
mugwump autrijus: I added ext/Date to MANIFEST.SKIP 02:07
is that ok? I can tidy them up to fail gracefully if that's not good enough... 02:09
I'm getting a weird error building pugs w/perl5 embedding atm, too
/usr/lib/perl/5.8/CORE/reentr.h:611: error: field `_crypt_struct' has incomplete type
I'm sure this is something simple... 02:10
Do you think it's work shipping a module that only consists of an untested API? 02:11
theorbtw1 mugwump: Is that the only error?
Also, what distro are you on?
mugwump theorbtw1: that's the middle of a big long failure ... I'll nopaste it
jabbot: nopaste?
jabbot mugwump: nopaste is irc.csie.org:8888/
theorbtw1 mugwump: I recognoze it. 02:12
It's nasty, and AFAIK, nobody has figured out WTF is up with it.
mugwump it's very strange, actually - I only got it after purging my wc and re-building
theorbtw1 Odd. 02:13
autrijus mugwump: if you have some cycles, please mark them so they fails gracefully 02:15
(just so ext/ don't become the old modules/ again)
mugwump: maybe perl and ghc were built with different gccs? 02:16
mugwump hmms
the errors' irc.csie.org:8888/110 if you haven't seen it already
theorbtw1 autririjus: I'd expect very different errors there. 02:17
Link errors or segfaults, possibly...
autrijus I seem to remember it has something to do with cc/dlflags from EU::Embed 02:18
theorbtw1 That seems mildly possible. 02:19
The odd part is that if you try to force the ifdef that makes _crypt_struct be defined, it still doesn't work. 02:20
svnbot6 r4935, autrijus++ | * lvalue comma was broken from the vCast change; it's now fixed. 02:48
r4936, autrijus++ | * Chinese (traditional) for Main.hs poem for this release. 03:25
r4937, autrijus++ | * fix .isa to use .ref; unTODO a passing test
Darren_Duncan there and back again 03:26
03:37 theorbtw1 is now known as theorbtwo
autrijus theorbtwo: 192.168.0.123/smoke.html 03:49
theorbtwo: please download and cache it locally :)
jql laf 03:50
jql admits he clicked it
jql puts on the dunce cap
theorbtwo Being behind the same firewall)++ 03:51
(Having a cell phone on the other end)--
svnbot6 r4938, theorbtwo++ | t/builtins/arrays_and_hashes/pairs.t: correct buggy test. 04:24
sili_ unf 05:26
mugwump right, I'm off ... should get a chance to get those Date tests into line before Sunday my time (GMT+12), probably committing Saturday night. Otherwise the MANIFEST.SKIP will withold them from the distribution and they can come another release... 05:37
pasteling "p]" at 210.83.203.100 pasted "sial.org/pbot/perl6" (1 line, 28B) at sial.org/pbot/11317 05:57
gaal_ inflammable pizza huh? 06:18
sili_ mmm 06:20
svnbot6 r4939, Darren_Duncan++ | updated test arg_ref_type_preservation.t : commented out all the 'control' tests that used a temp var (most failing as-is as they may be wrong now), leaving just the core 8 tests that don't use a temp var; 6 succeed, 2 fail 06:24
Darren_Duncan I notice that referencing private attributes is now broken in the current build of plain no-Parrot Pugs 06:27
I assume this is due to work in progress so I'll not mention it further now
but if you weren't expecting it, well now you know : example death is seen when running ext/Locale-KeyedText/t/LKT_20_TrnProps.t 06:29
good night 06:30
steveq Anyone know what to do to clean up the wiki? 08:40
It seems to have been spammed badly on the frontpage
gaal_ *sigh* happens a lot :-(
steveq yeah
gaal_ go to Revisions, find the last good version, and Save it
steveq gaal_: done 08:43
Looks like they only spammed the frontpage
gaal_ cool 08:46
well, not cool
but you get my drift :/
08:47 gaal_ is now known as gaal
wilx Compiling Emit.PIR ( src/Emit/PIR.hs, src/Emit/PIR.o ) 10:49
ghc-6.4: could not execute: /usr/local/lib/ghc-6.4/ghc-asm
*** Error code 1
nothingmuch good morning 12:11
coral e l , v t m n o 12:25
oops, off by one error
oh well
autrijus for ('a') -> $letter { is( $_ ,'a', :todo<clarification>) } 12:32
I seem to remember damian saying something about methods no longer set $_
as invocant
I wonder if the same applies to pointy now 12:33
autrijus decides to ask lwall/luqui when they're up
broquaint IFCP in perl6? 12:36
Nattfodd broquaint: in parrot assembly 12:37
broquaint Isn't the F in IFCP "functional"? 12:38
autrijus it's icfp
and perl6 is functional :)
(but, to answer your question, icfp does not place a limit on the type of language)
broquaint Madness. 12:39
How does one "win" this contest? Isn't it subject to the judges opinions? 12:40
autrijus indeed. if the problem this year is unfit for p6, then the opportunity cost is too high and maybe I'll give up and just continue the hackathon on p6
broquaint: there's always a quantifieable metric. see previous year's.
google for icfp 2004 contest etc
malaire icfpc.plt-scheme.org/past-contests.html 12:41
broquaint Hmm, not many PHP entries, I wonder why that is ... 12:44
Surely nothing to do with its quality as a language and the community that surrounds it ...
awwaiid hey don't knock PHP... as they continue to adopt more features of perl they are improving :) 12:45
besides... would you really want all those folks of whom you speak hanging out in perlmonks? 12:46
broquaint They wouldn't even get that far. 12:47
QtPlatypus awwaiid: Though it is in urgent need to make its builtins consistent. 12:48
awwaiid indeed
broquaint It is in urgent need to stop sucking in general.
But I guess it's not really relevant to the discerning hacker.
awwaiid :)
cognominal do Perl6 and its parrot backend supports non boxed types? 12:49
QtPlatypus broquaint: I don't think thats a helpfull comment. Understanding why it sucks, and how it can be improved.
cognominal: That is what is intended.
cognominal I note that most signatures use Int where int would do
autrijus cognominal: perl6 does. parrot does. pugs doesn't :)
broquaint This is IRC, QtPlatypus, hardly the place for helpful comments.
cognominal hu, I meant Pugs,
coral QtPlatypus: while people on irc may often be helpful, it is not guaranteed. 12:51
for instance i'm usually indirectly helpful 12:52
wolverian autrijus: methods don't alias $_ to invocant? 12:53
autrijus wolverian: damian said so in a p6l post last week. I'm not sure whether it's normative or not. 12:55
will ask.
QtPlatypus coral: Just because there are unhelpfull people doesn't mean that we shouldn't strive to be helpfull ourselves. 12:56
wolverian autrijus: thanks! 12:57
QtPlatypus: helpful
cognominal howdy, BooK 12:59
coral QtPlatypus: striving to be helpful all of the time results in burnout and a lack of focus on the essential needs of self 13:02
see also #perl topic (efnet variant)
wolverian (efnet #perl)++ 13:03
coral with a bunch of people there's often at least one person in the mood to be helpful, which can be enough to carry the day
QtPlatypus coral: True.
coral and for those unable for whatever reason to be helpful (lack of clue, lack of coherency, lack of energy, ...?) there is still reason to participate 13:04
<--
wolverian coral: lack of time
coral i replaced time with coherency and subsumed that into energy
mod at will
wolverian well, there's still some limit to how much energy you can spend per second 13:05
:)
but now it's time for me to go to the beach with girlfriend and maybe get some tan
(read: burn myself)
cheers
coral 'second' is defined partly by velocity
wolverian no, it's defined by the speed of light
coral relative to those being helped
autrijus nothingmuch: icfpc.plt-scheme.org/spec.html 14:02
svnbot6 r4940, autrijus++ | * more regenging... this time we hack &return() so it won't
r4940, autrijus++ | flatten hash/arrayrefs ever again, even at slurpy context.
autrijus very... complex 14:06
Nattfodd autrijus: yes 14:10
and very tricky, too
wolverian wow. that is pretty hard. 14:15
obra 'morning
wolverian morning 14:17
PerlJam so ... is anyone attempting the icfp contest in perl6? 14:41
QtPlatypus is pondering it. 14:42
PerlJam bets the future modification will be something like the robbers act together rather than the cops 14:45
svnbot6 r4941, autrijus++ | * pairify const pair-derived arrayref. 14:46
theorbtwo I think the modification will be adding extra cop headquarters, or possibly making pedestrian-only streets. 14:50
Adding multiple robbers that act together would be a large change. 14:51
PerlJam or some combination of all of the above
autrijus right. 14:53
it looks like somewhat trivial to come up with a reasonable implementation that can survive the regular season 14:54
but then will probably be mercilessly murdered during playoff
Limbic_Region theorbtwo - I don't see pedestrian only streets as being a difficult minor change - cars already have to check if they can enter a street because some are 1-way 14:55
you wouldn't need 24 hours to change that
theorbtwo True.
We do have 1/3rd as long to addapt to the change as to write the initial implementation.... in theory. 14:56
obra w/in 4
sorry
Limbic_Region I am guessing the change(s) will be more substantive in how you adapt to game play 14:58
IOW - cops can now smell you 3 blocks away - how does that affect your strategy as a robber
theorbtwo Having multiple robbers that do not communicate seems about the right level.
malaire in first stage, map seems to be allways the same - so change could also be a different map 15:09
BooK well, the rules say that the change is subtantial enough that the code must be well-written enough to survive it (i.e. not recode the whole game logic because of this change) 15:13
so I guess it's bigger than making pedestrian-only streets 15:14
broquaint I'm guessing the addition of a love triangle involving a beautiful seductress-bot which both the cop-bots and robber-bots love but neither sure of her true identity which is that she is a secret-agent-bot working for an external power seeking to gain advantage from the situation, whatever the outcome. 15:17
malaire Bots could be written to take advantage of the fact that map is fixed, but these would suffer horribly should the map be changed. On the otherhand it's harder to give bots better AI, but that probably would be easier to adapt to a new map.
Odin-LAP Hmm. 15:20
What about a vigilante-bot?
With some of the copper features, and some of the robber features? :>
That'd maybe be "too much". :p 15:21
Arathorn mixins... :)
Odin-LAP Arathorn: ;) 15:22
svnbot6 r4942, autrijus++ | * pointy binds $_ again... at least now... 15:28
r4942, autrijus++ | * more test skips
spinclad [06/23 20:27] <autrijus> 192.168.0.123/smoke.html 15:43
autrijus, would you have another address for that? 192.168.x.x is local... 15:44
theorbtwo spinclad: It also denies my connection. 15:50
svnbot6 r4943, autrijus++ | * make the usual things that takes Pair arguments recognise pairs.
spinclad it's an unroutable address range, no access outside of local net 15:51
theorbtwo I know. He's sittingon the coutch next to me.
It's last night's smoke, anyway. 15:52
spinclad ah, i thought it was a site for regular reports. NM, then. 15:53
wolverian why is it that I'm so tempted to write an RPG engine in perl6? 16:11
wolverian suspects it's the QBasic coder in him
pdcawley_ is tempted to write all sorts of things in perl 6.
wolverian well. me too. 16:12
it's just that the role-based OO is so directly compatible with RPGs, I guess. :)
Limbic_Region mmmm - I love the smell of smoldering monitors in the early afternoon
wolverian ouch.
pdcawley_ I keep wondering about writing a debugger.
Limbic_Region guess I won't be using this machine to be doing any data analysis ;-)
kungfuftr pdcawley_: yar... should be easier with things like .wrap 16:13
pdcawley_ Which means I should probably take a look at Deve::ebug
wolverian you don't have a neural cannula?
pdcawley_ Is .wrap implemented yet? 16:14
theorbtwo Check prims.hs?
pdcawley_ grins, get in early and rejig sub/method/class/etc to do instrumentation tricks and you're laughing. Or at least smiling giddily to yourself. 16:15
theorbtwo Coolo, pdc. 16:23
(That was a typo, not an attempt to be 80s, BTW.) 16:24
sili i thought it was spanish.
pdcawley_ Have we got macros that I can use to rewrite sub yet? 16:31
svnbot6 r4944, luqui++ | A little backslashy test.
Arathorn gah 16:32
pdcawley_ Actually, a better bet might be to extend ::Block
Arathorn talk about smouldering monitors
the EHT just sparked over in this thing
pdcawley_ ?eval 1
theorbtwo pdc: Not implemented, and their existance is somewhat under question.
EHT?
Arathorn the line output voltage 16:33
theorbtwo Ouch.
Arathorn thunderstorm just started so humidity's gone crazy, and so the voltage arced somewhere and collapsed
lightning outside & inside the office
pdcawley_ What are we calling them now? I understand they're just a special case of 'pointy', but I can't remember what class a pointy is.
Arathorn returns to writing an incremental cvs-mirror script in p6 :)
theorbtwo PDC: Read autrijus' last journal entry. 16:37
pdcawley_ Ooh... '-> $x {1}()' is actually parsed..
I think I already did. 16:38
theorbtwo I think it's "Block" unless it has a spelled-out (multi)?(sub|method)
pdcawley_ Yeah.
I just found that '::Block.new' doesn't seem to return the same kind of referency thing as -> {...} 16:39
theorbtwo Oh, you want the name, not the actuall class? I think it's still named "pointy".
Hm. That's possibly a bug.
pdcawley_ When you do '::Block.new' running in a 'bare' pugs you get {obj:Block} as your output. 16:40
When you do -> {...}, you get sub {...}
theorbtwo Possibly it's not... what would you expect it to do when called?
Oh. That's a bug, I think, as they don't have the same signature, I think.
pdcawley_ Well, I expect 'em both to give you the same sort of thing.
theorbtwo -> {}.perl should be -> {}, ans sub {} should perlify to sub {}... I think. 16:41
pdcawley_ Point is, I want to be able to wrap ::Block::BUILD or ::Block::new or whatever.
theorbtwo I'm somewhat confused by it.
pdcawley_ Yeah, blocks perlifying to subs is definitely wrong.
theorbtwo Hm, that shoulsounds like it should work, but i'd certianly expect it to not work in current pugs. 16:42
sili hmm
i thought blocks are supposed to perlify to subs
pdcawley_ Well, if we can get it to work, it enables a whole bunch of interesting debugger possibilities.
theorbtwo Yeah, it does. 16:43
pdcawley_ sili: But blocks and subs have different return semantics.
sili ah.
castaway_ thunk pointys and subs are both just ways to write subs
pdcawley_ It might be possible to *promote* blocks to subs (and get sub return semantics), but vice versa could well be difficult.
broquaint use Enable :mode('I am an evil coder and I shall do as I please');
s/evil/smart/whatever/ 16:44
pdcawley_ castaway: Down that line of thought lies interesting confusion when you look at for 0..10 -> {return "Found it!" when 4}
castaway_ s/thunk/remembers from reading S04 16:45
theorbtwo Possibly sub foo (...) is Pointy would do. 16:46
pdcawley_ Yikes! &?CALLER_CONTINUATION perlifies to sub {...} does it...
'does it'? Don't ask me what I meant. 16:47
Is there any way of inspecting a running pugs to find out what classes it knows about? 16:48
svnbot6 r4945, autrijus++ | * there's no List type anymore... let's call it Array.
theorbtwo Try ./pugs -e 'my $Foo::Bar=1; sub foo{ %CALLER::.perl}; say foo' 16:51
geoffb autrijus, safe to pull pugs now, without tests exploding? 16:59
pdcawley_ Is %:: still our symbol table? 17:00
theorbtwo Yes, but I'm not sure you can mention it directly, without a %MY::, %CALLER:, etc. 17:01
(%MY:: is not implemented.
pdcawley_ So you have to look at %CALLER::? 17:02
theorbtwo Yeah.
pdcawley_ Doesn't do what I want.
theorbtwo Also, the %:: code seems to miss things.
pdcawley_ Yeah. In perl5 you can see every package loaded by inspecting %:: 17:03
Which is global of course... what's a global variable look like? %*:: ?
theorbtwo Possibly. 17:05
* is a twigil, not a (psudo)package, though, so my %:: code won't work with it. 17:06
pdcawley_ grins. Right, I shall go read the appropriate apocalypses and take a look back at my perl 5 magic and write some tests shall I?
Because if I can have symbol table introspection, a working wrap and working continuations, I can get a good long way to writing a debugger. I think.
theorbtwo Would be nice.
Great! 17:07
(I need to figure out a better way to work...)
pdcawley_ And if I can show a proof of concept debugger, I can maybe get some TPF funding to build one :)
Limbic_Region pdcawley_ - just hope that no Klingons sit on the TPF board of approvals 17:21
www.klingon.org/resources/klingon_code.html
pdcawley_ Nice! 17:23
(And not Mice! as I initially typoed)
Limbic_Region that was a link in the interview Damian recently did with brian d foy 17:24
pdcawley_ Ah. Not read that yet.
Limbic_Region want that link too?
pdcawley_ Right. I'm going to search for food.
Arathorn anyone know where you report pm spam to on freenode?
Limbic_Region theperlreview.com/Interviews/damian...22.html?up
Arathorn (if anywhere?) 17:25
Limbic_Region heads off to a meeting now
theorbtwo Ara: There's instructions on the web site, IIRC.
Arathorn checks it out 17:26
18:01 Arathorn is now known as Aragone
nothingmuch_ autrijus: colabti.de/irclogger/irclogger_log/...8,Thu#l753 18:36
18:37 nothingmuch_ is now known as nothingmuch_____
Darren_Duncan it appears that the openfoundry went down in the last half-hour 18:53
I can't update or commit ... but could update a half hour or hour ago
but I am pleased to say that last night's Pugs updates fixed the bug exposed by t/subroutines/arg_ref_type_preservation.t ... 18:54
and the bug exposed by t/builtins/arrays_and_hashes/pairs.t 18:55
meanwhile, the bug exposed in exists.t is not broken and moved TODO, but it occurs to me that it should be possible to work around this bug meanwhile, by copying an entire hash to another temp hash -- then reading from the second one shouldn't erroneously add a key to the first 18:58
so I'll try employing said work-around to get my module tests to pass
openfoundry is still down 19:01
acme_ autrijus: icfp-ing? 19:04
Darren_Duncan seen autrijus 19:10
jabbot Darren_Duncan: autrijus was seen 4 hours 15 minutes 55 seconds ago
castaway_ he's here somewheres.. 19:15
(that is, sat on couch across the room from me)
PerlJam castaway_: who all is at the hackathon? 19:16
19:16 castaway_ is now known as castaway
castaway Autrijus, Luke, Ingy, Larry, Stevan, theorbtwo, John MacDonald and me 19:16
(so far.. apparently Patrick is turning up today)
PerlJam castaway: ah, excellent. You guys should get alot accomplished. 19:18
castaway well not so much me, but the rest seem to be doing well.. see Autrijus' journal from yesterday
PerlJam If nothing else, a face-to-face should get people fired up and get some ideas flowing
castaway indeed, its quite enlightening 19:19
PerlJam whenever patrick shows up tell him I said "hi and happy hacking" :) 19:20
castaway nods
Darren_Duncan that's great to hear 19:23
castaway oops, I missed out nothingmuch, who is also here 19:24
Darren_Duncan so do you do a lot of 'over the shoulder' programming there, several people on one screen? 19:26
also, how do you get by without openfoundry? 19:27
do you have your own repository you can sync to openfoundry later?
castaway you're asking the wrong person, Darren ,)
they were certainly using svn earlier to check stuff in 19:28
Darren_Duncan it went down a good hour ago
are you resting now
castaway most of it seems to get discussed across the room, with sometimes resorting to paper or laptop.. (Which makes it hard to follow at times)
no, I dont hack pugs/haskell
castaway is more of a taxidriver here 19:29
Darren_Duncan I will try to participate in a hackathon in the future ; minor regret at missing this one
castaway were discussing chompedness just now, tho I missed the conclusion
afaik, autrijus cant fix of even if he notices when its down 19:30
Darren_Duncan are some of you taking photos and/or video of the hackathon ... I want to watch it some time
too bad
castaway no photos that Ive noticed at all (I would, but my camera didnt get brought, its dead)
I'm not sure it would be all that entertaining, just in pictures 19:31
Darren_Duncan should help with good memories for the participants, though
castaway hmm
PerlJam When openfoundry goes down the typical reaction is to switch repositories to the perl.org one temporarily. 19:32
Darren_Duncan openfoundry seems to go down frequently ... might it be reasonable to switch semi-permanently? 19:33
PerlJam (the only thing is that not everyone has commit access to the perl.org one)
I'm sure though that it would be reasonable for the hackathon to have a local repository that they commit to and then periodically sync that to openfoundry or perl.org or whatever
Darren_Duncan while autrijus can't fix openfoundry, is there someone he can contact who can, or does he have to wait for them to notice on their own? 19:34
PerlJam dunno. Past outages have been noticed rather quickly (within a couple of hours) but still slow to fix (probably due to the nature of the outage) 19:35
castaway afafir, one just has to wait
Darren_Duncan fortunately I also have perl 5 stuff to do, and that isn't dependent on openfoundry
castaway PJ, yeah, it might have been clever.. ,) (also a proxy server, given we're all conncted over the same cell phone ,) 19:36
whiteg it's not easy to find someone dealing with openfoundry's outage.
it's weekend now...
PerlJam castaway: the only network connection there is via cellphone?
castaway correct 19:37
PerlJam Bummer.
Darren_Duncan too bad ... I know some hosts have procedures to maintain 24/7 uptime
castaway (its running at about 11kbit/s I think
oops, 115kbit
its not that bad
PerlJam well that's good then 19:38
Shillo Uh.. svn.openfoundry.org is gone x.x ? 21:25
Darren_Duncan yup 21:30
down for 3-4 hours now 21:31
integral SVK to the rescue! :-) 21:45
sili integral: noob 23:08
Limbic_Region is the normal depot out of commision ATM? 23:47
Darren_Duncan yes it is 23:49
and has been for about 5 hours now
castaway eep, nm is taking pictures 23:53
castaway hides
Odin-LAP Do so. Or alternatively, destroy the camera. 23:55
23:55 knewt__ is now known as knewt_
Darren_Duncan oh come on ... embrace the camera ... you don't get these chances very often 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!