-Ofun: xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by apple-gunkies on 11 November 2005.
autrijus xinming: huh? it was always like that in pugs 00:00
?eval (1,2,3).say
evalbot_8119 OUTPUT[123 ] bool::true
autrijus ?eval [1,2,3].say
evalbot_8119 OUTPUT[123 ] bool::true
autrijus wait... it is "1 2 3" on my terminal
obra autrijus: different sequences make sense for different people
autrijus something's not right
obra perldoc Perl6::Doc should suggest sequences 00:01
autrijus xinming: I'll look into it in a minute.
obra: okay, that works for me... so you think Perl6::Doc is better than P6Doc:: ?
obra P6Doc is obscure.
xinming autrijus: make a file to test. :-)
autrijus Perl6::Spec::Operators # what about this?
xinming xinming@HomePc:~/tmp$ cat asdf.p6
#!/usr/bin/pugs
my $a = ~list( =$*IN );
print $a;
autrijus xinming: how about you do that and I'll use the test :)
xinming :-) 00:02
obra autrijus: what would be in ::Operators? for numbered docs (like the existing specs), I sort of like the numbers...but they make them harder to find
xinming and run that file, You will get what It shouldn't be. :-P
obra of course, if the specs ARE hard to understand, making them harder to get at than the ::Doc is a win
autrijus obra: I'm more oriented in moving AES into a single Spec 00:03
because it's very hard to chase AE as they are
obra are you going to try to get @larry to bless it as normative? 00:04
autrijus I'm probably only going to rename Perl6::Bible at this moment
obra wants to understand intent
autrijus the intent is to make python people who want to learn about perl 6 actually able to start learning it with minimum distraction.
obra the equivalent docs in perl5 are perldoc perl*? 00:05
autrijus the immediate thing is to make irrelevant documents out of the way, and name the relevant documents sensibly (i.e. by name)
yes. perl*is divided into: Overview, Tutorial, Reference, Internals, Platform
and FAQ
and I think the Reference part sorta matches Synopses 00:06
obra ok. then yeah. I'd rename Perl6::Bible as Perl6::Spec
autrijus and ::Spec is maybe better than ::Reference (more normative sounding)
xinming autrijus: hmm, did you try that?
obra and then start cannibalizing content from them into Perl6::Doc::{O,T,R,I,P}
autrijus xinming: yes, it joins the list with spaces for me 00:07
xinming autrijus: So, I don't think this should be happen
autrijus xinming: what do you think should happen? 00:08
xinming though, list( ... ).as("%s", '') will resolve the problem, But for default behaviour. It shouldn't add extra space.
autrijus it's not unlike the p5:
$a = "@{[<STDIN>]}";
print $a;
xinming don't add space, what it slurped, should be printed as the same as what It get. 00:09
autrijus S02: In string contexts container references automatically dereference to appropri?
ate (white?space separated) string values.
xinming: but you can do 00:10
@a = list(...);
say @a
which won't add spaces
xinming autrijus: hmm, In fact, this is what fglock asked yesterday, "how to slurp stdin", 00:11
autrijus: I think it's not a bug now. :-/
autrijus xinming: you do it like 00:12
my $a = $*IN.slurp;
or slurp($*IN)
xinming autrijus: hmm, I tried this, but It will print each line after you press enter 00:13
instead of slurp.
autrijus xinming: yeah because slurp is lazy... 00:14
xinming: what's wrong with that though :) 00:15
xinming for $a, you won't slurpy all the input.
autrijus but it's slurping all the input... 00:16
just not all at once
xinming hmm, what I mean is, It will slurp oneline, and process it, then slurp another line,
autrijus yes 00:17
but try this
my $a = $*IN.slurp;
say substr($a, 0, -1);
xinming what does -1 mean here? :-/ 00:21
autrijus never mind...
it means (end of string)
so try
my $a = substr($*IN.slurp, 0);
which forces evaluation until end of slurp 00:22
there should be better ways to do this.
xinming maybe a force... :-)
autrijus nod
or make slurp strict by default. I'm not sure.
it's p6l domain :)
xinming my $a = {$*IN.slurp}; 00:26
print $a();
:-P
xinming don't wish the () here.
autrijus *nod* and it's bogus anyway... once we get lazy arrays it will also be lazy 00:28
so the real problem is .slurp's laziness.
Khisanth lazy slurp? isn't that contradictory? 00:31
xinming autrijus: doesn't pugs implemented the ** operator? 00:35
You have to use unary ** to get a non-lazy flattening list context (that is, to flatten immediately like Perl 5).
from S02
if so, then, I think we can use **$*IN.slurp
autrijus ok 00:41
impliementing it 00:44
asavige: hi! 01:01
asavige hi autrijus! just got the coke in a brown bottle from supermarket as specified in VICTUALS 01:03
but what do you eat for breakfast?
autrijus carbonhydates and protein 01:04
with some water 01:05
(i.e. anything goes)
bread is fine, cereal is fine... :)
clkao autrijus 01:06
asavige ok, thanks, no need to buy fruit loops or anything. might get some quiona in honour of Schwern (never tried it myself)
autrijus k.
obra Coke in a brown bottle?
autrijus obra: caffeine free coke
clkao i've got piljs's runjs.pl using perl5-capable js
asavige here in australia the diet caffeine free coke (as specified in VICTUALS) comes in a brown bottle 01:07
autrijus asavige: also, just recently (during OSDC) I came out under the new real-world name Audrey (autrijus is still around as a net handle), and I'd prefer female pronouns if that's okay with you :)
asavige no worries mate
clkao now i can implmeent preludes for io-ish stuff using perl5 for js backend
autrijus cool :)
clkao: yay, that will give iblech some more tests to pass 01:08
clkao r:D
tell me i am insane
autrijus clkao: you are insane
xinming: **... is implemented. committing
clkao XD
autrijus my $a = **($*IN.slurp);
print $a;
this now works
clkao ?? 01:09
what's ** doing?
autrijus clkao: force deep sequential eager evaluation
deepSeq in haskell
it's in as r8695 01:17
I'm going to revive svnbot6 in a sec
xinming autrijus: :-) Sorry, I was out for breakfast. 01:34
autrijus: how about the $?CLASS bug? :-P does it have to be fixed after the meta-model finished? 01:35
autrijus xinming: it will be automagically fixed with objspace, which I'm working on at this moment 01:36
xinming bbl 01:38
meppl gute nacht 02:20
autrijus gute nacht, meppl :) 02:21
xinming: so my Chinese name is now finally public information :) 02:22
xinming autrijus: :-/ Sorry, I still don't know. 02:23
svnbot6 r8122 | autrijus++ | * Update the AUTHORS file by running util/fixauthors.pl. 02:25
autrijus xinming: it's in the AUTHORS files... also zh.wikipedia.org/wiki/%E5%94%90%E9%B3%B3 02:27
autrijus goes back to objspace 02:28
PJF Strange. Compiling Pugs.Parser croaks with a q{GHC's heap exhausted...} error, but when I remake it handles it fine. Does GHC do something special like dynamically allocate heap based upon memory size?
autrijus PJF: yes I think 02:29
I sometimes needed incremental build too
PJF Everything seems to be running smoothly so far. The only changes I've had to do is patch ExtUtils::MM_Cygwin, manually run 'setup configure', and restart make after an out-of-heap error. 02:30
svnbot6 r8123 | autrijus++ | * s/Autrijus Tang/Audrey Tang/g in all places that may have legal
r8123 | autrijus++ | meanings (copyrights, author names, etc.)
meppl ;) 02:31
autrijus wait... Imention "patch" ;)
s/I//
what's the patch about?
PJF rt.cpan.org/NoAuth/Bug.html?id=16375
MM_Cygwin doesn't seem to think that anything in /cygdrive/* is executable. So maybe_command fails to identify my ghc.
autrijus looks at the paatch 02:32
PJF It's not a patch to pugs, but it does allow it to install on my cygwin laptop, which has GHC installed in 'c:/ghc/..'
Or /cygdrive/c/ghc/... from Cygwin space. 02:33
xinming autrijus: wikipedia is banned here. :-/ 02:34
hcchien ...... 02:35
xinming Most people knows this in China,
autrijus: but I wish to know, if that character is 凤 in Simplified Chinese. 02:36
autrijus Alias: svn.openfoundry.org/pugs/docs/AES/S22draft.pod 02:37
Alias_ ta
autrijus xinming: yes it is
PJF: I promoted it to a patch to pugs anyway 02:38
PJF: welcome to AUTHORS file -- a commit bit is on your way 02:40
PJF laughs!
Thanks Audrey. ;) Clearly my mumble at OSDc wasn't enough to dodge the commit bit. ;)
autrijus actually, why not commit yourself to the AUTHORS file
PJF I will once I've got a real patch. Pugs is giving me some silly error about not being able to find Test.pm. I'm investigating why. 02:41
autrijus okay then 02:42
svnbot6 r8124 | autrijus++ | * add PJF to AUTHORS file for the cygwin patch. 02:45
PJF Hey, that's not letting me add myself. ;) 02:46
Alias_ Then remove the J before you become known as an initialled person 02:48
autrijus the jfdi policy mandates that if something needs to be done and some other committer is delaying, other people just go ahead :)
Alias_ The brian_d_foy problem 02:49
PJF I believe that brian_d_foy quite likes his name that way, provided you use the right font. 02:50
Alias_ Nope 03:00
He's decided it was folly, and now wishing to be only known as Brian Foy
Mainly because people were continuing having thinkos of "de Foy"
i.e. a french-like version
PJF laughs. 03:01
Well, I don't mind PJF. I've been enjoying it for about twelve years now.
PJF discovers that pugs doesn't like cygwin absolute paths, quite probably because his ghc lives in windows land. 03:02
autrijus aye. 03:05
svnbot6 r8125 | autrijus++ | * refresh M::I to 0.40 03:06
autrijus I think it boils down to that need cygpath support
PJF I'm currently looking at PugsConfig::add_path to try and kludge something in to run the tests. Is there a better place I should be looking?
autrijus no I think it's the right place 03:07
for some value of right place
PJF Thanks! 03:08
(AFK a bit)
autrijus thank _you_ :)
svnbot6 r8126 | autrijus++ | * bring more stuff forward from M::I 0.40. 03:09
autrijus anyone running native win32 here?
(non cygwin)
would be cool to test the last few changes re Makefile.PL 03:11
(should still all work)
Alias_ PJF: I meant more Paul J Fenwick :) 03:13
autrijus: Silly, I'm 4 feet away and running native Win32
of course, I write this knowing he's just left the room :/
PJF I'm back, Adam W.D. Kennedy. 03:16
autrijus lol 03:18
autrijus goes to lunch w/ Alias &
PJF discovers that OSSF doesn't seem to remember project invites if the invitee decides to use a different address to register. 04:24
ingy_ autrijus: hola 04:32
autrijus ingy: still there? 05:43
PJF: right. you have another account now?
PJF Welcome back, Audrey A. Tang. I registered [email@hidden.address] not [email@hidden.address] 05:44
The username is still 'pjf', but it doesn't recognise the invite because I used a different e-mail address.
But I now have tests under cygwin running happily. Working on getting the smoke kit installed.
I hope Adam W. D. Kennedy provided you with a nice lunch. 05:45
ingy autrijus: yo
autrijus ingy: so I have this idea that makes Perl6::Spec::Operators instead of Perl6::Bible::S03 happen 05:46
ingy: do you think it's sane(ish)? and if yes, are you willing to hack on it, or mark me as comaintainer of Perl6::Bible so I can refactor the things away, and let me know where the repo for that is? or should we just move the spec fetcher into pugs/docs/? (which may be better)? 05:47
xah autrijus: are there many hackers in taiwan? 05:55
autrijus xah: plenty of, I think 05:57
xah i went back to taiwan for 6 months in 2003... 05:58
didn't have much opportunity to meet any...
and apparently didn't come across them in my US IT industry life
them -> programers of Taiwan 05:59
in fact, you are the only one i know.
and a celebrity too.
autrijus xah: well... there's plenty of them in IRCNet 06:00
#bsdchat, #dot, #perl.tw, etc etc
xah mmm...
would that be ircnet.org? 06:01
autrijus irc.seed.net.tw if you are in taiwan
irc.xs4all.nl in .eu
not sure about other places
PJF Do we have many cygwin pugs hackers? I'm happy to tweak the build process to improve cygwin support, but if we have anyone else who's been working on that I'd like to coordinate efforts. 06:02
autrijus PJF: welcome aboard :)
PJF: no, I think cygwin was never fully working.
(you should be able to commit now)
PJF Oh good (regarding both cygwin and commits). All the problems I've hit so far with cygwin have been dealing with cygwin paths vs windows paths. 06:04
autrijus I reckon it'd be the case.
PJF It's not hard to win32ify the paths for pugs under cygwin, which is quick and ugly, but seems to be working. In theory I'd expect it all to work with a cygwin-native build of Haskell, which I haven't looked for yet. 06:05
autrijus not sure there is one. 06:09
svnbot6 r8127 | pjf++ | Pugs doesn't yet deal nicely with absolute paths under cygwin. This
r8127 | pjf++ | kludge allows the test suite to run by also providing win32ified paths
r8127 | pjf++ | under cygwin.
r8127 | pjf++ | This does not fix general cygwin pathing issues. It only helps the
r8127 | pjf++ | test suite.
autrijus ooh, a Real Commit
obra heh 06:10
pjf++ # More/Eaiser testing is always a win
PJF Thanks obra. I've still got a few tweaks to make before the smoke framework runs happily. 06:11
svnbot6 r8128 | pjf++ | Corrected speeling of 'PJF'.
obra PJF, sorry ;)
Alias_ W. D. Kennedy?
Like WD40?
PJF Currently we still require 'setup' to be run manually during the build process under cygwin. That shouldn't be hard to correct.
W.D. as in "World Domination"
Alias_ ah 06:12
I prefer just D for Danger 06:13
Especially when registering for conferences
dduncan well, openfoundry is back online ... if anyone noticed it was down a half-hour ago
autrijus it was, I restarted apache 06:14
dduncan that's a little different ... as I recall you used to have to wait for someone else to fix that server
autrijus well, the machine wasn't down, that helped :) 06:16
I can't physically reboot the machine from abroad
dduncan autrijus, why change your name? ... or is there an explanation posted somewhere? 06:22
autrijus dduncan: nah... I'll journal about it sometime somewhere, probably.
dduncan: in short, my gender role did not match my self image (and how my brain was wired since birth) for a very long time; I couldn't go on living like this so had to make the change. 06:23
dduncan so the old name didn't describe you then
autrijus the english one is sorta okay. the Chinese one totally doesn't work. 06:24
svnbot6 r8129 | pjf++ | Under Cygwin with a Win32-native GHC we find ourselves with a 'setup' that
r8129 | pjf++ | doesn't understand cygwin-flavoured paths. This patch turns '/cygdrive/x/'
r8129 | pjf++ | style paths into 'x:/' paths on cygwin systems, causing setup to run cleanly
r8129 | pjf++ | under cygwin.
dduncan okay 06:25
PJF goes AFK while he runs 'make smoke'. Happy hacking. 06:26
wolverian for some reason, Audrey reminds me of The Neverending Story. 06:28
autrijus wolverian: good, because it's the origin of the "autrijus" name too :) 06:29
wolverian autrijus, yeah, maybe I'm thinking of that, really, not Audrey. but the Childlike Empress was played by Audrey Gardiner.. 06:31
dduncan do you plan to update your account names with cpan and perl.org etc as well, and IRC handles, etc?
if the former are possible
autrijus dduncan: no... I'm fine with the net handle "autrijus" 06:32
half of the people in AUTHORS has net handles that bears no relationship with their passport name
obra waves
autrijus yeah, like this obra here
obra doesn't really ever use 'obra' off IRC
Except with a set of friends who know me primarily from online 06:33
autrijus and that's how I plan to use 'autrijus'
dduncan sounds good to me
xah i think online handles are a plague 06:35
ingy autrijus: let's discuss when we meet next. the repo is svn.kwiki.org/ingy/Perl6-Bible/ iirc 06:36
I can give you update access
obra xah: why? 06:37
xah no offense, but they are vestiges of childish behavior
think of teens and script kiddies
ingy autrijus: access granted (to all my code)
obra xah: I disagree
a handle is a persona. 06:38
They don't have to be childish personas
autrijus ingy: cool
Alias_ waves as well
obra Also, they tend to more uniquely identify individuals
Alias_ except in my case... :)
obra There are other Jesse's involved in perl.
But no other obra ;)
ingy autrijus: hsinchu is lovely today. wish you were here :p
PJF *click* Oh! Hi Jesse! 06:39
GeJ morgen
Alias_ thinks about other curtises
obra Also, there are some cases, like "autrijus" and "clkao" where a handle is shared by a team of 20 or 30 hackers
Hi PJF. Do I know you?
autrijus not "team"
the term is "consortium"
obra apologies
was typing too fast ;)
autrijus like the "unicode" handle which is also shared by a consortium
Khisanth there are 20 or 30 autrijuses hoping around the world?
xah online handles are a side-effect of tech-geeking 06:40
Khisanth (that would explain a lot actually)
autrijus Khisanth: it's an old joke, on the same vein as Mark Jason-Dominus is actually an infinite amount of monkeys who also keep octopus as pets.
PJF obra: Only in passing, and through other people. I'm that crazy guy who runs Perl Training Australia.
obra PJF: ah. Hi :)
xah few simple indicators: 06:41
* the younger they are, the more wild is their handle
* the more tech geeker they are, the more arcane is their handle
PJF obra: We know each other primarily through Skud and Robert Spier.
obra Ahh. Ok :)
xah * the more renowned, reputable, professional aperson is, the less his assumes a handle 06:42
* the more sincere a person is, the less he'll use a handle
autrijus xah: on the other hand... people didn't choose their name either.
xah that's some general, worldly, all encompasing, indicators.
obra xah: I respectfully disagree
Khisanth autrijus: only till we figure out this time travel business! 06:43
PJF Xah: So an old, sincere, tech-savvy person should change their name? ;)
I've always wanted to be known as "Bogomip Fenwick".
GeJ I never met the guy, but chromatic doesn't seem to match all the criterias
xah autrijus: not clear about your point.
obra GeJ: 'chromatic' isn't a handle ;)
xah as for name change, yes. If people wanted to change their name, many do, and actually do so legally.
GeJ obra: would you care to elaborate? is it his _real_ name? 06:44
;)
obra It's what he goes by pretty much universally. It appears on his book
Khisanth now would be a good time to quote Shakespeare :) 06:45
autrijus xah: the point is that for some people (like me) who started using computer at 8 and internet at 12, we can't really change our realworld name 06:47
but yet we didn't choose it ourselves
so something that aligns more with my self image is actually a form of sincerity
obra coughs at audrey
"CAn't really change our realworld name" 06:48
autrijus obra: at age 12?
xah autrijus: that may be so. But look at the world, history, perspectives.
obra Oh. parsefail.
autrijus obra: so my point is that people usually do when they got the chance, but there are circumstances where it may not be feasible for some time.
obra nods 06:49
xah autrijus: that's right. The point i'm making about online handles is that, it is special affinity to tech geeking.
obra xah: it's also common with artists
xah for instance, there's are pennames, pseudonyms etc.,
obra and musicians
xah each are used for a purpose. But, 06:50
obra DJs
xah they are different than online handles.,
obra howso?
dduncan xah, personally I find handles do lead to more things one has to remember, so I just use my real name or abbreviation thereof online ... never invented a handle
never plan to
xah dduncan: good point. 06:51
dduncan: and a point for my point. :)
obra xah: What's your real name? 06:52
xah heh. Xah. :D
dduncan I'll also say ... both ways ... I know for a fact that people of all ages use handles, but they seem more common the younger someone is
I thought that
also, the use of handles I find more common in the groups I talk to about fiction, eg X-Men, whereas programming forums there is a much greater fraction that use real names and eschew handles ... 06:53
xah dduncan: exactly.
dduncan I consider it more "professional" to use your real name
I also consider it more professional to have your own domain name, rather than having your email at geocities or hotmail etc 06:54
autrijus seriously ponders /nick audrey
autrijus goes back hacking objspace
xah not that online handles are intrinsically bad, or indicate some bad about it's user, it's just that by the nature of things it's things tech geekers or teens would do.
dduncan and I practice what I preach
yes, they're not a bad thing, but you will look more professional ... to the general public anyway ... using real names 06:55
xah it's like dress code, a person's mannerism, his diction... gives an indicate of what the person is or is like. Online handles is just a indicator of the million. 06:56
dduncan that's more the case if people looking at you are strangers ... it matters less if they know you
xah yes.
online handles can give a hint of geekdom, for instance,
dduncan for example, around here I know you guys enough to know you are pros, so having nicks doesn't speak unprofessional to me 06:57
xah and if it some batman or vampire, then you know its author is perhaps young, or in a jesting mood.
dduncan that's a good point ...
xah and of course there are exception geeks who stick with fancy handles.
dduncan *what* people use for nicks says a lot about them
xah exception -> exceptional.
dduncan you can identify young children easily by nicks based on stories that mainly pre-teens would like 06:58
obra Like the neverending story?
xah hey, we make a rap duo.
obra ducks
xah :)
autrijus obra: not sure it's a preteen book :)
a preteen movie perhaps. 06:59
obra Fair enough
obra admits to not having read the book, just seen the movie
autrijus fglock++ # speccing corner cases meticulously in docs/notes/laziness.txt
dduncan what I mean is ... 07:00
you can identify young children as they use CURRENT FADS as nicks
and fads change every few years
xah ninja turtle!
dduncan exactly ... but that's a decade out of date 07:01
plus
15 years ago
now its like pokemon and such
xah umm.. pokemon!
dduncan or in an X-Men fan group, the younger ones are more likely to be limited to a few fad-hyped characters like Wolvie 07:02
wolverian I would like to state at this point that my nick has nothing to do with x-men :)
dduncan older people would probably use chars that have been around for awhile and aren't the most hyped ... and there are a thousand to choose from
actually, I wasn't thinking of you
about half of the lists I'm on are X-Men related, and the other half Perl related 07:03
autrijus wolverian: I didn't make the link before you made the connectino to wolverine :)
GeJ So you're telling me that leo@#parrot used to run in the streets with a blue mask when he was a boy? naaah. I don't buy it 07:03
wolverian agh! stupid me :)
autrijus GeJ: Leo is actually named Leo.
(sort for Leopold)
GeJ I knew... I just had to say to something. It was either some stupid/fun comment or "Does anyone live in Virginia and is skilled enough tou build an explosive device cause I could use some help..." 07:05
:)
dduncan fyi, if you want a sampling of X-Men fan nicknames, look no further than: xday.info/archives/xday2003/default...view_prtcp
some are real names (including mine), some not 07:06
GeJ if you ever consider changing your job. DO NOT EVER go to work for a company that registers domains.
worst... job... ever...
PJF wonders if ext/FA-DFA/t/DFA.t is supposed to take a very very long time. 07:06
dduncan fyi, getting the coding done for xday2004 is the main thing competing with my time to work on Rosetta et al 07:07
yes, I'm that far behind
oh, one more X-Men related thing ... 07:09
darren duncan. net / xmen 3 funeral. jpg 07:10
remove the spaces, then go there (spaces added so google doesn't index it from irc archives)
I may have mentioned on IRC that I got a bit part; well I'm in that scene 07:11
those of you who saw me at OSCON may be able to pick me out ... for the rest, I'll defer adding an arrow until the movie comes out, to be on the safe side 07:12
one hint: left side
that clip came from the trailer, released 2 days ago 07:13
GeJ Ha! gotcha! the guy with the blue face? ;)
GeJ ducks
dduncan no, that's the Kelsey Grammar guy
I'm in the second row; all the main cast are in the first row, or standing 07:14
or should I say 'a' second row
anyway, that was just a fun little thing to do ... my career aspirations are entirely in the computer field 07:15
Alias_ second row on the left, 4 people in? 07:16
dduncan if you must know, in the left most group, second row, closest to the isle shared by the second left most group 07:18
I'm 27
soon to be 28
Alias_ next to the lady sitting down right? 07:19
Thought so 07:20
(only one with glasses I could spot actually)
dduncan Alias_, yes
so then, I guess that for those of you that haven't met me, you know a bit of what I look like 07:21
Aankhen`` LMAO... go.reuters.com/newsArticle.jhtml?ty...EnoughNews 07:26
PJF Compiling Pugs.Parser ( src/Pugs/Parser.hs, dist\build\src/Pugs/Parser.o ) 07:50
GHC's heap exhausted: current limit is 268435456 bytes;
Use the `-M<size>' option to increase the total heap size.
Build failed: 64256 at util/build_pugs.pl line 87.
Should I be passing -M to ghc when it compiles setup, or does setup understand the -M switch?
I'm suspecting the former.
autrijus line 32 config.yml
PJF Oh! That's very easy, thank-you. 07:51
autrijus were you using unoptimized?
no prob... but we should intelligiently incrementally rebuild and/or extend heap size
("make unoptimized" that is)
PJF Nope. Just a plain "make smoke"
However I expect that we want smoke to run happily without the user having to change any options. Am I correct in that expectation? 07:52
autrijus yes. 07:53
definitely
and so far only win32 ha
s the heap flow problem
PJF So we may want to conditionally increase heap memory based upon us having a win32 ghc? 07:54
I have noticed that restarting setup (without increasing heap size) causes it to complete the process. So it looks like it's not completely releasing the heap in between each file compilation. It certainly compiles a lot of files before getting to Parser.hs 07:55
Changing the heap size for Win32/Cygwin looks very easy, and it can still be overridden using the configuration file. Audrey, are you happy for me to increase the default heap for building pugs under Win32/Cygwin? 08:01
autrijus sure. 08:02
I think heap problems are getting better for ghc 6.4.2 08:03
but for the time being, adjust it as neccessary for 6.4.1
(as 6.4.2 is still some time away)
PJF Okay!
autrijus PJF++ 08:04
svnbot6 r8130 | pjf++ | Win32/CygWin GHC is a memory hog. This change gives it a memory sty. 08:13
r8130 | pjf++ | config.yml continues to be respected if ghc_heap_size is set.
r8131 | pjf++ | Fixed Win32 GHC memory adjustment commited in 8130 08:22
gaal hello! 08:40
autrijus greetings gaal! 08:42
Alias_ Hail Baal! 08:43
oh wait...
gaal hoezzit?
s/e/w/; mwahaha();
autrijus gaal: excellent weather and getting lotsa things done
gaal great! 08:44
when'd be a good time to revisit prag stuff?
autrijus Alias_: no chance of us getting net access tonight in $home?
(also, I thought we should be heading home for christmas tree and stuff) 08:45
Alias_ erm...
You went out today and bought a switch during business hours right?
WRONG!
</chewbacca>
autrijus aw.
gaal: so, tomorrow morning.
Alias_ I mean, you have net access on THAT computer
Which has X-Chat I imagine
Alias_ chuckles
autrijus eh...
Alias_ So the english channels you can have 08:46
autrijus Alias_: do we have an extra realtek network card?
random PCI ethernet card
Alias_ We aren't changing the hardware of my parent's computer :)
autrijus well, we are already typing away on IRC when we are at arm's length from each other
Alias_ But then security won't care about us until 2am
I figured you wanted others to evesdrop
gaal autrijus: great
autrijus no... it's just my interpersonal skills needs work 08:47
Alias_ ah
autrijus gaal: anyway, so 15 hours from now
gaal that's a little early on my timezone, will you be free a little later that day? 08:48
by this rate, it might be ready for yapc :- 08:50
{/)(} 08:51
autrijus sure, I'll be around for ~10hrs from that on
gaal okay :)
PJF Gaal, you with with Alias, I take it? 08:52
autrijus I'm so happy to be back to regularly-scheduled hacking :)
PJF s/with/work with/;
Alias_ no
If he did, he wouldn't need to be scheduling with autrijus in IRC :) 08:53
PJF Sorry misparsed. I shouldn't try multitasking so much.
I dunno. Audrey and yourself seem to be doing a lot of communication via IRC. ;)
Alias_ Only for your benefit 08:54
PJF Mind you, I do exactly the same with jarich. I've usually got noise-reduction headphones and thumping techno on.
Alias_ finally parses the meaning from "jarich"
eep
PJF Sorry. Jarich = jacinta_a_richardson. It's french. 08:55
Like Brian de'Foy. ;)
Alias_ You listen to techno...
I really should have transfered the 30gig of it I had on my laptop while I was at your house 08:56
oh well...
PJF Mind you, I'm much too lazy to actually pick songs myself. I usually listen to di.fm. 08:57
Alias_ Ditto, but with a disk dump on :)
PJF Goodness me! The RIAA may not appreciate that. Perhaps we'll need to give you a pirate name, like PeuceBeard. 09:00
Although wikipedia tells me that Puece is an island located near the Danube river, so obviously my spelling is poor today.
autrijus bbiab... will be back after dinner (I think) 09:05
gaal PJF: oops, sorry, missed you. Nope, I'm in .il.
PJF What's our recommended way of altering tests that seem to tickle endless loops in bugs?
I've found that FA-DFA/t/DFA.t appears to churn on 100% cpu after the first test.
I want to get this smoke report finished. I guess I can tweak it locally to a SKIP... 09:06
gaal ideally fix the harness to portably catch that from without :-)
but yeah - skip is what we do today.
PJF So use Acme::HaltingProblem to determine if the code supplied will halt, and then skip it if we know that it doesn't? 09:07
gaal heh, no, use some kind of timer or OS limit
demo of the idea:
?eval 1..
evalbot_8131 Error: unexpected end of input expecting end of input or term 09:07
gaal ?eval 1..Inf
evalbot_8131 pugs: out of memory (requested 1048576 bytes)
gaal ?eval 1 while 1 09:08
evalbot_8131 (no output)
gaal if the harness could do that (on all platforms), swell. 09:09
PJF Well, on *nix systems we can use ulimit to impose a limit on each test.
gaal yes
PJF On Windows systems I'm not quite as sure. We can use alarm() to schedule a signal, although that could cause some tests to fail on slow systems. 09:10
gaal if you want to impose a wall clock limit you always face that problem. 09:11
xinming1983 ?eval role A { method t ( *@a ) { @a }}; class B does A; class C does A { has B $.b; method BUILD { $.b = B.new; $.b.t( 'haha','hehe','xixi');}} C.new
evalbot_8131 Error: No such method: "&t"
gaal (and you do)
PJF So far nobody's objected to me tweaking the build system. I'll fiddle a bit and see if I get any complaints from altering test.
gaal ?eval sleep 60*60*24
evalbot_8131 Error: No such method: "&sleep"
gaal grrr, it's unsafe. but if it weren't, evalbot would not wait forever.
although obviously cpu time is very low for that computation.
PJF: that's the spirit :-) 09:12
gaal frustratingly can't think of a topic for a talk for YAPC 09:13
PJF How long are the YAPC slots?
gaal "variable"
from lightning to 3 hours, if I want.
PJF That should make it much easier then.
gaal I'm just out of ideas :( 09:14
PJF I recently gave a lightning talk at OSDc on using a few Win32 modules to automatically play minesweeper. There seems to be an absence of good information on Win32 GUI testing, so that's something that could easily be expanded outwards. 09:15
gaal two areas I don't particularly like. :-)
PJF Minesweeper and testing? 09:16
gaal though the talk must have been plenty fun
heh. no, gui and win32.
minesweeper's not that hot as distractions go, either
PJF The talk was fantastic, although I probably did too many lightning talks at OSDc.
Although I think some people are expecting me to automate solitaire now. :/ 09:17
gaal :) 09:18
afk 09:22
PJF finds it amusing that set-infinite-recurrence.t appears to have hit an infinite loop during his smoke test. ;) 09:52
r0nny yo 10:10
rafl Huh.. why s/Autrijus Tang/Audrey Tang/g? 10:47
PJF Gaal: I note that you were working with a demo version of alarm() some time ago. Did that make it ever near pugs? 11:51
gaal PJF: Not really. For sure we need %*SIG, but while we're at it a variant of alarm that doesn't use signals (and is possibly thread-specific) would be nice to have. 13:07
%*SIG btw would be a little hard to get on win32, because the usual GHC there doesn't have the posix libs. 13:08
PJF Gaal: Thanks Gaal. I didn't think it was in PUGS yet, but thought it best I check. 13:19
Looks like compiling a POSIX-flavoured GHC is reasonably easy with Cygwin, but I've put the code away for the evening. No doubt I'll have a little more of a play tomorrow. 13:20
I'll probably end up writing some wrapper that terminates tests that are running too long. At least on my system, some of the infinitely lazy tests seem to evaluate code that's infinitely eager. 13:21
gaal PJF: If you do a real cygwin port (then, great!) some refactoring of Compat.hs will be necessary 13:31
PJF++; # POSIX stuff on win32 would rock 13:32
wilx Hnnm.
PJF I don't have a cygwin port of GHC yet. However I do have a cygwin environment, and from glancing over the GHC docs it used to be built with cygwin. 13:33
wilx I was not able to build GHC 6.4.1 out of the box with Cygwin.
PJF It appears they stopped to provide compatibility across all systems, not just those with cygwin installed.
wilx: I haven't tried yet. I'm expecting there to be some hiccups.
wilx I tried that unregistered or whatever it is called build.
gaal PJF: I meant pugs' Compat.hs, which kinda assumes you're either on unix or on native win32. 13:34
PJF I've got pugs working under cygwin with a win32 ghc. However it acts more or less like it's on a win32 platform.
There's a bunch of patches I added today to fix up some pathing oddities. All I really want is to be able to run a smoke to completion. ;) 13:35
gaal PJF: yes, cygwin as a work environment with natve win32 tools worked most of the time and broke occasionally. 13:36
I'm talking about pugs' internals though, which provide compat functions for things GHC doesn't provide on Windows, eg stuff to access the environment. 13:37
Corion wrote most of the compat code I think
see src/Pugs/Compat.hs
that file is under PUGS_HAVE_POSIX, which with luck will be true on your build; but still some things may need to be taken from there. 13:38
(I imagine.)
meppl guten morgen 13:45
rafl kane_: ping 14:25
kane_: I'm about to buy a new root server and it still needs a domain. How about sixpan.org? 14:27
PJF Thanks Gaal! I'll be sure to check Compat.hs in the morning. ;) 14:31
nemux hi all 15:52
svnbot6 r8132 | fglock++ | * misc/Parser-Mini/pil2_json_emit_p6.p6 - it can now compile small programs, 16:09
r8132 | fglock++ | except for a few syntax errors like:
r8132 | fglock++ | '($x,$y)=(1,2);' compiles to
r8132 | fglock++ | '(&infix:<,>($x, $y)) = (&infix:<,>(1, 2));'
fglock autrijus: isn't this supposed to work? '(&infix:<,>($x, $y)) = (&infix:<,>(1, 2));' 16:26
btw, how about: '&assign((&infix:<,>($x, $y)), (&infix:<,>(1, 2)));' 16:28
or better: '&infix:<=>((&infix:<,>($x, $y)), (&infix:<,>(1, 2)));' 16:30
fglock wishes more of pugs was implemented in the Prelude 16:48
PJF: re set-infinite-recurrence.t - something is seriously broken - even ext/Span t/span.t enters an infinite loop. It looks like there is a problem with autogenerated methods vs. actual methods 17:17
Khisanth so ... we were having an argument about writing code in multiple languages and this came up 17:27
<Khisanth> ParityBit: it only starts being fun when you mix ltr and rtl languages ;)
how does perl6 handle that?
integral Khisanth++ 17:29
Khisanth sure that isn't likely to happen but Perl hackers seems to like doing things just because Perl allows them to :P 17:31
PerlJam Khisanth: it does some really good guessing about where the end "blocks" should be and allows you to say "parse this way for the following block" 17:46
Khisanth: as far as the actual rtl part, that's just an implementation detail ;)
jez Anyone here good at Sudoku? :-) 17:54
Dazhbog depends what is good ;) 17:55
jez www.game-point.net/misc/puzzle.pdf 17:56
how does the system determine that that's 9?
i can see 9 or 8
Dazhbog well you see the 9 has to be in either two squares in the left? 17:57
then if you look at the leftmost vertical line..
it is missing 2,7,8,9 17:58
so there has to be a 9 in some of the 4 empty squares
jez ohh
how did i miss that
damn
Dazhbog it happens =) 17:59
jez here's another i cant figure out 18:05
www.game-point.net/misc/puzzle.pdf
why does 5 have to go there
eric256 wonders what suduko is. and guesses that one labeled VERY HARD probably isn't the place to start. ;) 18:08
integral the hardness is just a measure of how much backtracking you have to do
jez ive been playing it a while now
still cant get this
there doesnt seem to be a logical way of getting that number there 18:10
eyt there must be as that computer system doesnt do guesses/trial and error
eric256 hmmm rules seem simple enough.
possibles for that square are 3,4,5 right? 18:12
jez apparently only 5 18:14
they're the obvious only possibilities
pd en.wikipedia.org/wiki/Sudoku <-- they go through some reasoning and algorithms here 18:15
eric256 i would guess you need to work out more of the board to get the point wher eyou find out why that has to be five. i know the first two after the 1 in that row can't be 3 ;) so proud of myself. ;) 18:16
seems like you could get a long ways on a board only to find out you have an impossible setup and have to back track.
pd that was implied in one of the pages I saw... apparently at some point you try out a number and see where it goes, and sometimes you hit a dead end etc. 18:17
eric256 jez why so caught up on that specific spot? or are you just curious? 18:18
jez eric@ eh? 18:19
eric: no, that's a hint given to me by an online puzzle 18:20
it should be possible to figure out the red number from the black numbersw, logically
without revealing any more numbers
eric256 right. then you can't expect to know all the reasons behind the hint until you finish more of the puzzle.
jez eric: that's why you never guess or use backtracking
erm
yes you can
eric256 sure. if you can hold all your conclusions in your head. 18:21
the number in each square is affected by the numbers in its row, col, square whether they are given or not, so you will have to deduce some other locations before the reasoning behind that one becomes apparent
jez yeah, i've tried 18:26
:-)
i cant figure it out
i was hoping someone more clever could
eric256 the only other square in that colum that can be five is the second from the top, so i would assume there is some relation that means it can't be 5 18:27
jez i wouldnt assume that 18:32
there's a ton of ways of determining a number, the most ostensively easy isnt always the way 18:33
Dazhbog jez: it's a "bit" hard but yes it can be determined from that
eric256 i didn't assume anything. well i assumed that the 5 is chosen because it is the only possiblye sqyare in that col/row that can be five.
Dazhbog want me to explain that? =) 18:34
eric256 of course it could be something different, but then jez thats what i was saying that you would have to solve more of the puzzle before it became apparent
jez dazh: sorry? 18:35
care to explain how?
Dazhbog you don't have to solve any number for sure
jez even "what if?" doesnt solve it 18:36
Dazhbog jez: you can deduce that in that column where the number 5 is..
jez that would allow 4
#
Dazhbog jez: the top empty squares must be populated by 2 and 9
you don't know for sure which one is in which but you don't there must be 2 and 9 in them 18:37
hrm.. second don't = know
still with me?
jez hmm yes i see
i'd only marked up 2 for those top ones
not done a comprehensive markup
starts to get very messy if you do markups for everything :-)_ 18:38
Dazhbog well if you mark that 2 and 9 have to be in either.. then you know there can't be any others
eric256 shit. dashog is a genius ;)
jez there's no reasonably easy way to determine that
i guess when i get stuck it's time to start comprehensiuvely marking up EVERY damn cell
and it's gonna get messy
Dazhbog no..
jez but it's the only way
eric256 so the five can't be the second from the top, therefore the five is below (and my assumption holds)
Dazhbog yes.. the five which is missing in that vertical row.. can't be in either of the top two 18:39
jez yeah
Dazhbog it cannot be in the middle bigger square because there is a 5 already
jez yeah if i'd comprehensively marked those 2 up i'd have fiogured it out
but i dont do that for all cells as it gets very messy 18:40
heh
Dazhbog um I did it in my head =)
jez yes, but you knew which cell to concentrate on
Dazhbog sure
jez remember at this point i had the whole board and that red 5 wasnt there
i got the computer to give a hint
:-)
Dazhbog well.. you asked why =)
jez yeah
thanks
Dazhbog no problem =)
jez i'd call it "very hard" or "very messy"
one of the two
doing that in your head is extremely tough 18:41
eric256 jez. why would you do it in your head? thats why man invinvted the pencil. and yes you have ot pick a square at random to start working on. I would probably pick one that had the most givens in its col/row or square 18:42
jez eric: im working on this generally on a bumpy train :-) 18:44
that's why
:-)
Dazhbog heh 18:45
eric256 then build your self a perl program that you can type the numbers into ;) or just let it solve them .haha 18:46
Dazhbog get a little pda or console ;)
like gp2x =)
jez eric: www.dailysudoku.co.uk/cgi-bin/sudok...nscribe.pl
eric256 using someone elses, and making your own are different prospects ;) 18:48
jez bugger it 18:49
if i cant figure out 1 cell in half an hour i just throw it away
eric256 lol. i think you are missing the patiences gene ;) them japanese and there damn puzzles of patience. ;)
jez the paper i get (free) is a bugger because they dont tell you the puzzle difficulty beforehand 18:50
:-)
they just sometimes have tough and sometimes very easy
eric256 so go pick one of those books full of them
jez it's not as fun when you cant solve it in 30 mins
you just end up thinking 'what a fucking waste of time, i could've done something more productive' 18:51
or your journey ends
Limbic_Region idly wonders if there are any hackathons planned for the Philippines in the next 3 weeks 19:01
otherwise - everyone who observes a holiday in this time frame - happy holidays
Khisanth you are in the Philippines? 19:04
fglock autrijus: would it be possible to emit <quoted> subroutine names in PIL2-JSON? It would fix cases like 'infix:<:>', which is output like "&infix::" 19:06
nevermind, I found a workaroud 19:08
(around)
oops - the workaround didn't work 19:12
Khisanth my $workaround is lazy; # :P 19:13
fglock I tried this: '$s ~~ s:perl5{\&(.+fix:)([^:].*)}{&$0:<$1>}'
Khisanth isn't that greedy match a bit dangerous? 19:14
err those
fglock the problem is that it doesn't mix well with names that contain a namespace spec (::) 19:15
Khisanth what sort of thing are you trying to match? 19:16
fglock I'm trying to fix "&infix::", which is how pugs compile 'infix:<:>'. This is a worst-case situation - other cases work ok 19:17
Limbic_Region fglock - no, I will be soon though 19:18
fglock Limbic_Region: I guess you are talking to Khisanth 19:20
Limbic_Region um no 19:21
I was talking to you but I should have been talking to Khisanth
fglock :)
Limbic_Region my mind is already on vacation I guess
Khisanth fglock: so everything after the first : is the operator?
fglock happy holidays to you too - I'll be on vacation on dec-19 :) 19:22
Khisanth: yes, unless you have ::, which means it is a namespace
Khisanth but that would mean infix:: is completely ambigous! 19:23
fglock &main::infix:x is a valid name too
I guess &infix::infix:infix is valid too :) 19:24
Khisanth so /&[^:]+fix:(?!:).+/ ?
or does foo:infix:x mean something too? 19:25
stevan autrijus: ping 19:45
fglock Khisanth: sorry, I had a long phone call - I'll test your regex now 19:47
hi stevan!
stevan hey fglock
fglock I'm working on p6->p6 conversion - it would be nice to test this with MM 19:48
it compiles to "baby p6"
stevan p6->p6??
fglock which is closer to the native thing
stevan fglock: where are the files located? 19:49
fglock which means, all calls are function calls - there are no predefined primitives
r0nny yo
fglock see misc/Parser-Mini/pil2_json_emit_p6.p6
r0nny i got a problem with pugs 6.2.10 (r8093)
i call a new method in a BUILD submethod, should pass a object, but instead it saves the ref to the new method 19:50
fglock stevan: please try this: ../../pugs -Cpil2-json -e 'say "hi"' | ../../pugs pil2_json_emit_p6.p6 19:51
stevan fglock: cant, I dont have a recent compile of pugs
but I am looking over hte source
stevan suddenly remembers feather :)
r0nny anyone got an idea what to do ? ome rev'S earlyer it worked fine 19:53
stevan fglock: very interesting 19:54
I assume that is the PIL2 tree
fglock your version have a long set of comments in the beginning? (the version in svn may be a bit old now) 19:55
s/have/emit/ 19:56
gaal hey all
stevan yes it has a long set of comment at the begining
he gall
hey gaal
fglock ok, I'll commit a new one in a while
gaal yo stevan
r0nny darn 19:57
stevan so fglock, is the purpose of this just to round trip p6? or do you have a more sinister purpose in mind ;)
fglock ok, comitted 19:58
r0nny im allways shocked, a few days incrase the rev by 39
svnbot6 r8133 | fglock++ | * misc/Parser-Mini/pil2_json_emit_p6.p6 - minor fixes; debugging turned off by default
fglock the new version is easily retargettable - I want to emit Parrot, but it can be used for other languages (Forth comes to mind) 19:59
stevan fglock: very nice 20:01
how would you want to integrate the MM into this?
fglock it can be integrated by writing the Prelude in p6, and then using it as the infrastructure for Classes (maybe precompiling it to PIL) - so that pugs's Classes are not used at runtime 20:03
it means that MM will be available to naive runtimes which don't implement objects 20:04
and we can work out some of the ideas in Blondie, maybe
stevan_ hmm
this is very similar to the work on the object space 20:05
and what autrijus is doing currently
so you want to have the mm bootstrapping to be inside the prelude 20:06
so that your generated p6 uses the metamodel,.,. and not native (aka pugs) classes
right?
fglock yes
stevan_ very nice 20:07
you are bringing it around full circle :)
p6 on p6 :)
fglock this is a short program - I think it can even be integrated with pugs itself, such that the classes used at compile time can be written in p6! 20:09
stevan_ fglock: check in with autrijus about this
I think maybe he will want to get the objspace in HS first 20:10
but this sounds like a fairly logical next step after that
fglock I was just thinking aloud :)
stevan_ fglock: that is always a good thing :) 20:11
stevan_ has to get back to $work ... cursed $client wants me to work on friday ;)
fglock bye stevan
stevan_ later fglock 20:12
stevan_ &
r0nny submethod BUILD() 20:17
{$.populator=new Builder::Populators::Simple::Compiled;}
fails 20:18
svnbot6 r8134 | fglock++ | * misc/Parser-Mini/pil2_json_emit_p6.p6 - fixed some corner cases of subroutine name quoting 20:20
fglock Khisanth: thanks for the help. I ended up using 3 separate regexes, but it worked
r0nny omg 20:21
i just realized something awfull
its not a problem with refs, but a problem with inheritance - and i tihnk i just made it whorse 20:23
darn the output of the perl method :/
found a serious issue 20:38
while the perl method returns \Builder::Builders::Simple::Compiled.new(("populator" => Builder::Populators::Simple::Compiled.new()),) the defined sub returns false 20:39
(for the attr populator) 20:40
anyone got an idea ? 20:41
fglock r0nny: sorry, no 20:51
r0nny darn it 20:52
also it fails on sone other stuff again (like the endless loops on method dispatch to undef) 20:53
fglock I'm getting infinite loops too, in Span.pm and other places 20:55
which used to work
r0nny darn it :/ 20:58
fglock autrijus++ re PIL2: 'Class x { has .$a }' compiles to '(&Class((&postcircumfix:<{}>((&has(($a())))))));' 21:59
err, '(&Class((&postcircumfix:<{}>((&has($a))))));'
'Class x { method b { 1 } }' doesn't emit code correctly yet - I'll check this later 22:06
&
Alias_ I have kidnapped your leader 22:13
He is now trapped in my volcano lair
If you want him back, it will cost you one MEEEEEEEEEEEELLION dollars! 22:14
And for your more important leader, well... she will cost one BIIIIIIILION dollars! 22:19
autrijus any rumour of my hostage situation is greatly exaggerated. 22:20
Alias_ YOU ESCAPED! 22:21
RELEASE THE FLYING MONKEYS!
obra Alias_: one meelion australian? I have that in my couch 22:23
Alias_ hahahahah
Well it is a very very large couch
autrijus stevan_: ping 22:58
kane-xs_: ping
gaal: ping
stevan_ autrijus: pong 23:05
autrijus stevan_: yay
stevan_ can't talk really,.. just wanted to see how objspace is going
autrijus stevan_: okay, a whole slew of questions...
aw. :/
stevan_ doh
hmm
how long will you be awake?
autrijus it's doing fine but had some design questions :) when's your next slice of time?
10hr from now on
stevan_ you will be awake for 10 hours? or awake again in 10 hours 23:06
Alias_ She'll vanish in about 2-3 hours for 1-2 hours, but otherwise around for next 12+ hours
stevan_ ah 23:07
Alias_ Having escaped the volcano lair, I must relocate to my rainforest lair
stevan_ :)
stevan_ always suspected Alias_ was the type to have a "lair" or two
Alias_ When one tries to take over the world, one has to deal with this sort of thing 23:08
And do you have any idea how complex occupational health and safety issues are in regards to lairs?
The shark tank alone! :(
stevan_ a true evil genious would ignore OSHA regulations
:P
autrijus: I need to disappear to meatspace for about 2 hours,.. but after that I will be around 23:09
Alias_ stevan_: That's a stereotypical view of evil geniuses!
autrijus stevan_: ok, got it 23:10
stevan_ Alias_: you got me there,.. I suppose you are the kindler gentler evil genius type :)
stevan_ has gotta go before he gets in trouble &
Alias_ stevan_: Think of it as smart evil... knowing who not to mess with ( OH&S Inspectors)
PJF Morning everyone. 23:45
I'm trying to determine my expectation for some tests in pugs. ext/FA-DFA/t/DFA.t appears to go into an infinite loop after the first test. Is my system special, or is anyone else able to reproduce this? 23:46
autrijus I can reproduce this. 23:48
sec.
PJF Thanks Audrey. How should I respond to infinitely-looping tests? Obviously I should fix them when possible, or mark them to be skipped. However should I be submitting code that causes tests to be skipped? I'm sure we want the test suite to run without human intervention, but I feel quite uneasy committing code that would effectively remove tests. 23:50
autrijus PJF: first report them here. then put a flunk() in place of the test 23:51
comment out that test
PJF What I really want is something that zaps tests after they've been running for a specified period without any output. I just haven't had the time to write something that will work in a cross-platform fashion.
autrijus and say fluck("infinite loop")
s/fluck/flunk/
PJF And commit the flunk?
autrijus yes.
PJF Thanks Audrey. Shall do. 23:52
autrijus I'll bbiab... gotta get objspace running to a respectable degree today 23:53
&
luqui is going to use fluck as his new expletive 23:54