Parrot 0.9.0 | parrot.org/ | 566 RTs left.
Set by moderator on 31 January 2009.
rg1 No discussions yet. 00:00
or were you refering to the visualizations?
particle yes, and datasets.
rg1 seems to work without a login 00:01
particle fab, thanks!
00:09 AndyA joined
Whiteknight Does anything use the get_pointer and set_pointer vtables of the parrotinterpreter PMC? 00:18
because I want to try to repurpose them 00:22
particle change them and see! 00:26
purl particle: that doesn't look right
Whiteknight purl you don't look right
purl Whiteknight: i'm not following you...
Whiteknight well, if I comment them out I don't get any test failures
and I can build perl6 without problem, so that's a good bellweather 00:27
dalek rrot: r36369 | whiteknight++ | trunk/src:
[PMC] changing ParrotInterpreter set_pointer and get_pointer to point to PMC_data() instead of PMC_struct_val. Should help with later UnionVal deprecation
 00:49
Whiteknight yeah. put that in your local checkout and smoke it
I think we can do the UnionVal switch over now for ParrotInterpreter and ParrotThread PMCs now 00:57
I just need to find all the places they are accessed 01:00
nopaste "cotto" at 96.26.202.243 pasted "ParrotInterpreter and ParrotThread info" (45 lines) at nopaste.snit.ch/15497 01:05
cotto there you go
01:06 Fayland joined
cotto That's everything that gets hit during make test. 01:06
The line numbers may be a bit off. I don't remember what revision I used to generate that list. 01:07
Whiteknight++ 01:08
Whiteknight thanks cotto+! 01:09
thanks cotto++!
cotto I'll attach the whole file to the wiki page. 01:10
dalek tracwiki: v | cotto++ | 01:12
tracwiki: trac.parrot.org/parrot/attachment/..._pmc.2.log
shorten dalek's url is at xrl.us/beeyms
cotto Cool. I don't even have to post a link. 01:13
Whiteknight urg, the pmc_reuse function is going to be a mess 01:15
that's going to have to be one of the last things updated
hey cotto, how is inheritence happening now? Where does the parents data struct get stored? 01:17
cotto Since I can assume single inheritance, a parent's ATTRs are stored before the child's. 01:24
That way if a child is accessed as a parent, nothing breaks.
01:24 Woody4286 joined 01:25 DietCoke joined 01:26 silug joined
Coke . 01:26
Whiteknight hola coke
s1n Coke: let me know when you have that write up, i might be willing to help out with the old rt tickets 01:27
01:33 szabgab joined
Coke s1n: due sunday. 01:38
but basically, the plan is to close out the old tickets in place.
verifying that old reports still generate the same broken output is helpful. 01:39
making sure patches still apply is helpful.
that's probably the limit of things that can be done without fixing things. 01:41
chromatic Adding TODO tests.... 01:43
s1n Coke: verifying the tickets are still valid is all i can do at this point, so just say when and i'll do what i can to help 01:46
01:51 kid51 joined
Coke I don't know what you mean, "say when." 02:02
when?
I mean, I know what the /phrase/ means... =-)
s1n Coke: are you not a native speaker? (not an insult) 02:09
by that i mean, post the article, let me know, and i'll do what i can when i can to help move the tickets over 02:10
Coke no, I'm a native speaker; I'm just confused. 02:11
you don't need a go ahead; feel free to do those things right now.
Thank you for any cleanup you can provide on the queue. It'll be a help. 02:12
02:57 ask_ joined
dalek rrot: r36370 | jkeenan++ | trunk/docs/book:
Apply patch submitted by tuxdna in rt.perl.org/rt3/Ticket/Display.html?id=63036: Correct spelling/grammatical errors in POD.
 03:21
03:36 janus joined 03:51 ask_ joined
dalek rrot: r36371 | chromatic++ | trunk/t/op/interp.t:
[t] Made a TODO test pass by avoiding buffering order issues. See RT #52354.
 03:57
Coke kid51: I just tested the original hang in that RT you closed; it doesn't hang, so we're all cool. 03:59
(it's invalid pir, so you have to change it up slightly.) 04:00
kid51 must sleep 04:05
purl $kid51->sleep(8 * 3600);
04:15 mberends joined 05:08 Tene_ joined
dalek rrot: r36372 | Infinoid++ | trunk:
[cage] Fix some codetest failures.
 06:24
rrot: r36373 | Infinoid++ | trunk/compilers/pirc/src:
[pirc] Fix the -o flag in -b mode, for specifying the .pbc filename.
Infinoid r36372 contains changes to imcc.y which I can't regenerate. Does anyone have the bison/flex secret decoder ring, so they can tell me which combination of software versions successfully builds in --maintainer mode? 06:28
chromatic bison 2.3 here, flex 2.5.35 06:34
dalek rrot: r36374 | chromatic++ | trunk:
[docs] Added initial version of the Parrot Support Policy (TT #244).
 06:39
06:39 alinbsp joined
dalek tracwiki: v109 | chromatic++ | ParrotRoadmap 06:39
tracwiki: trac.parrot.org/parrot/wiki/Parrot...ersion=109
shorten dalek's url is at xrl.us/beezjp
06:42 rurban_ joined 07:01 uniejo joined
dalek tracwiki: v8 | jimmy++ | WhereIsIt 07:29
tracwiki: trac.parrot.org/parrot/wiki/WhereIsIt?version=8 07:30
moritz msg pmichaud when I checkout parrot below the rakudo dir 'make test' throws some errors because the t/pmc/*.t files want to run ../../parrot 07:47
purl Message for pmichaud stored.
moritz oh, and the generated makefile contains this line: 07:53
t/spec : svn co //svn.pugscode.org//pugs//t//spec t/spec
that's not working, ETOOMANYSLASHES
janus blog.t14g.de/wp-content/uploads/200...minute.jpg :) 08:00
shorten janus's url is at xrl.us/beezoc
08:00 davidfetter joined 08:17 workbench joined 08:19 iblechbot joined 08:44 mj41 joined
moritz rakudo: say "test" 08:56
polyglotbot OUTPUT[testā¤]
08:56 bacek joined 09:17 bacek joined 09:18 alvar joined 09:25 masak joined 09:29 Eevee joined
dalek rrot: r36375 | cotto++ | trunk/src/pmc/resizableintegerarray.pmc:
[pmc] fix bug from UnionVal to ATTR conversion
 09:43
rrot: r36376 | cotto++ | trunk/src/pmc/resizableintegerarray.pmc:
[pmc] undo accidental commit
 09:53
10:01 kj joined
dalek rrot: r36377 | rurban++ | trunk/src/packfile.c:
[docs] add missing function documentation
 10:12
rrot: r36378 | rurban++ | trunk/config/inter/libparrot.pm:
Checked cygwin dll: An old libparrot dll does no harm to cygwin. To be checked on mingw/msvc also
 10:22
10:23 braceta joined 11:03 iblechbot_ joined
dalek rrot: r36379 | rurban++ | trunk/src/exec_save.c:
[cage] use correctly signed index vars
 11:08
11:10 elmex joined 12:31 rg joined 12:36 TiMBuS joined 12:39 kid51 joined 13:24 rob joined 13:46 AndyA joined 13:47 bkuhn joined 13:48 AndyA joined 13:55 AndyA joined
dalek rrot: r36380 | Infinoid++ | trunk/compilers:
[imcc] Regenerate recent changes from imcc.y.
 14:22
szbalint so far so good
rrot: r36381 | Infinoid++ | trunk/config/gen/makefiles/languages.in:
[pod] languages/Makefile needs a realclean hook to remove languages/pod/Makefile.
rrot: r36382 | Infinoid++ | trunk/languages/pod/pod.pir:
[pod] Get languages/pod compiling. (One of the includes was failing on case-sensitive filesystems.)
rrot: r36383 | Infinoid++ | trunk/src/packfile.c:
[cage] Fix some trailing whitespace.
14:26 Whiteknight joined
kj Infinoid++ # fixing languages/pod 14:28
sjn jonathan: ping 14:30
14:33 gryphon joined
Infinoid cotto++ # finding bugs in dalek's rss parser 14:40
14:42 dalek joined, rurban_ joined
dalek allison@perl.org | Debian/Ubuntu chroot Environment Setup: 14:42
link: www.perlfoundation.org/parrot/index...ment_setup
shorten dalek's url is at xrl.us/beesjm
sjn jonathan, pmichaud, masak: I've updated the hackathon wiki page with a couple ideas... What do you think of them? (especially #3): www.perlfoundation.org/perl6/index....katon_2009 14:46
shorten sjn's url is at xrl.us/bebpnx
masak looks
I'm all for it. the more hackathon, the better. 14:47
sjn good
sjn is still a bit unsure of how to pull off a "pre-hackathon" at the conference though 14:48
masak: how would you have done that?
masak sjn: not sure what you mean. :) a hackathon is a couple people sitting in the same room, hacking.
so, you need a room, and a time slot. 14:49
sjn yeah, but this one is at the conference
which means it might be more a kind of "hackathon demonstration"
at the conference venue, hopefully 14:50
PerlJam sjn: demonstrating a bunch of people sitting around with computers isn't much of a demonstration ;-)
masak pretending to hack?
sjn dunno
masak I don't want that. :)
sounds highly frustrating.
sjn that's why I'm asking...
how can we make it into something that's interesting and relevant so people at the conference who just "drop in" want to come to the proper hackathon? 14:51
PerlJam sjn: have pizza!
masak aye!
sjn yeah, food works :)
masak food and soft drinks. 14:52
sjn but food maketh a hackathon not
masak people demo-ing exciting things they wrote helps, I guess.
sjn PerlJam: what would you have loved to see at something like that?
demos.... hm.. 14:53
masak preferably demos of things that are still ongoing projects and that are hackable right away.
sjn like an "update on what I've been hacking on the last few weeks"?
PerlJam sjn: A sign saying "Parrot Hackthon" with food and drinks and people talking about parroty things makes enough of a hackathon.
(for starters)
sjn hm 14:54
PerlJam People that want to join will see the sign (hopefully) and ask questions and such and hopefully join.
sjn yep 14:55
masak I like that.
sjn sure, but shouldn't there be.... more?
masak Perl 6?
purl Perl 6 is probably the spec, rakudo and pugs are two of the implementations.
masak purl: no, Perl 6 is amazing. 14:56
purl okay, masak.
PerlJam sjn: you could even do something like "the food and drinks cost one RT ticket closure" or something :)
sjn haha
with a "we'll help you close a ticket" stand :D
PerlJam yeah
:-)
masak that's a great idea.
sjn and they get their name on the ticket and in the source code 14:57
PerlJam you're free to hack and hacking buys you food :)
sjn that's an awesome idea :D
PerlJam: would you mind adding that idea to the wiki page? :) 14:58
(say under 3.1 :)
PerlJam well, I don't understand "pre-hackathon" really. To me, it's just a hackathon. 14:59
Like ... what's a pre-lightning talk? :)
sjn something that comes before the "main" lightening talk? :)
kj it's a "warm-up" session ;-)
masak pre-lightning never strikes the same spot twice. 15:00
sjn (as opposed to "post-lightening talk")
15:03 dalek joined 15:04 Gerd joined 15:07 Gerd left
sjn PerlJam: what would you call it instead? 15:08
particle hackathon.
purl hackathon is in the dorms' common area
particle pre-conference hackathon?
sjn it's not pre-conference... it's _at_ the conference 15:09
PerlJam then it's just a hackathon.
sjn (well, strictly speaking, it's after the regular schedule on thursday)
particle -athon connotes something ongoing, like a marathon.
is it ongoing, or a hacking session during regular sessions?
is it a hack track, where there are different hacking sessions throughout the day(s) in the same room? 15:10
sjn it's a hacking session after the talks are done, for those who want to stick around (there'll be other things happening at the venue too)
perhaps for a few hours
PerlJam A hack-track would be awesome. 15:11
masak hack track hackathon.
sjn it's a one-off, as part of the Go Open conference "Meet the experts" thing
sjn is guessing there'll be some other projects and things happening at the same time 15:12
stuff that hopefully makes people want to hang around :) 15:13
PerlJam: feel free to change the wiki text to something more understandable ;-) 15:14
PerlJam sjn: yeah, but then I'd have to actually login and I know I've forgotten my password :) 15:16
particle sjn++ # conference organizing is *hard*
sjn particle: not really :-P
particle: especially if someone else does the work :)
PerlJam sjn: careful ... that sounds like you're volunteering to do it all the time to me :)
sjn PerlJam: what a wonderful opportunity to rectify that problem! :D 15:17
sjn isn't volunteering at all :-P 15:18
15:18 alinbsp joined
PerlJam is still waiting for the email 15:18
sjn it's called "wasting my employer's time" :-P
sjn toddles off to get a cup of coffee
jonathan is back 15:38
15:39 jq- joined
masak jonathan: vitajte! 15:40
sjn heya, jonathan :) 15:41
jonathan "hackathon demonstration" - could be like, a drama. "masak: I HAZ A BUG! WE CAN HAZ RT TICKET!" "jnthn: OMG I WRITEZ THE FIX...OMFG I CAN'T USE GIT...AH FINALLY" 15:42
PerlJam jonathan: heh 15:43
jonathan But then people might think they need to learn lolspeak in order to attend...
masak jonathan: somehow, the first steps gets repeated 100 times.
kj well, that's easy with pmichaud's any-to-lolcode converter :-) 15:44
masak s/steps/step/
jonathan masak: It's a demonstration, not a complete history. :-P
masak :P
sjn jonathan: on a scene, with dramatic music, and people holding signs "FIX BUG!" shouting slogans "Christmas in april!" 15:45
moritz you also misssed OH HAI I CAN HAZ TEZT? part
masak moritz: you coming to NPW?
jonathan sjn: Hackathon page looks good.
sjn: Though later, details on how to get to it etc would be good.
PerlJam jonathan, sjn: that would make an awesome lightning talk of sorts . Get a bunch of people with appropriate signs and run through a hackathon demonstration by coordinated display of those signs. 15:46
sjn hehe
yep 15:47
jonathan: what do you think of PerlJam's idea to have "fix an RT bug for food" stand?
have a* 15:48
moritz masak: no :(
PerlJam or make it a carnival game. "Step right up! Fix a bug and win a prize!" :-) 15:49
jonathan sjn: That means I'll come out REALLY fat and masak will come out with weight loss. ;-)
sjn hehe
masak :)
sjn jonathan: or you just hang around and help people fix bugs (they get the commit and the name in the source code)
you and whoever is around to help 15:50
use it as a recruiting thing
get people involved :)
PerlJam sjn: that's where I thin the hack-track would be useful
s/thin/think/ 15:51
sjn PerlJam: hack-track, as in a full track, during the conference?
PerlJam yeah.
sjn not sure we have room at the conference for that. (it's easier to make things like that happen in the evening) 15:52
I'd love a hack-track, but I'd also love to see people at the Perl 6 talks
and the other talks 15:53
purl the other talks are japanese
jonathan sjn: I think that last year it was good that we had the morning "planning" meeting and end of day summary.
PerlJam sjn: if you could have a few long sessions with various parrot/rakudo/etc hackers shepherding people towards making contributions, maybe they'd be more likely to "get into it" on their own.
It's just an idea. 15:54
sjn jonathan: what do you mean? (you lost me there, what are we talking about?)
jonathan At the testing hackathon in Oslo last time, in the mroning everyone got together and said what they planned to do for the day
sjn PerlJam: that's part of my ambition with the hackathon-proper after the conference
jonathan: ah, yes
we'll do that this time too :) 15:55
jonathan sjn: Was quite good. But also I don't remember if at the end of the day we had a "what did we achieve" bit too. But I think it's nice.
sjn we did
jonathan Doesn't need to be long (even time-limit people lightning-talk style if needed).
But was good.
I don't think people rambled on for too long last time though.
So probably no need for explicit time limit.
sjn but right now I'm more concerned about the hackathon-demo-at-thursday-evening thing 15:56
jonathan What's the goal of the demo?
sjn get people involved
jonathan To give people an idea of what a hackathon is like?
Hmm, OK.
sjn and that yes
jonathan Would it not be better as a kind of section in the intro talks? 15:57
PerlJam sjn: well, the hack-track would be ideal during the conference to snag new people and the post-conference hackathon woudl be for the serious hackers :)
jonathan While people are paying attention?
sjn and show people outside the Perl community what we're doing
jonathan I'd expect the normal talks would show people what kinda stuff is happening as well, though?
sjn yeah, probably
purl well, yeah, probably is something that needs doin/working on
sjn smacks purl 15:58
purl rubs her red behind.
sjn jonathan: we have the option of doing something on thursday evening, as part of the Go Open "meet the experts" thing 15:59
for a couple of hours at least
jonathan Oh, OK, I didn't know that this section existed.
sjn jonathan: how would you use that time?
jonathan Is there URL aobut it?
*about
sjn it's not a "section"
jonathan OK, well, "thingy". ;-) 16:00
sjn it's something I'm considering
jonathan Ah, OK.
"Meet the experts" sounds a bit...maybe...like, "come meet celebrity programmers" or something though.
sjn we have the option, my first idea is to make it into an early start of the hackathon
jonathan: yeah, it's a stupid name
jonathan A name more indicating it's a way to find out ways to get involved, is better. 16:01
sjn never mind that
jonathan But yes, I like the idea.
Like a ProjectFest
sjn something like that
jonathan So different projects can promote themselves and have things for newbies who might be interested.
sjn demos, stands, open source projects
recruiting
purl somebody said recruiting was suspiciously out of place on IRC. Go to www.itknowledge.com/tpj/jobs.html
16:02 Tene joined
sjn and community visibility (which I'm mainly thinking about) 16:02
jonathan *nod*
Yeah, something like that.
purl it has been said that yeah, something like that is what I'd do as well
sjn you know... "Perl is dead, nyah, nyah, nyah" 16:03
jonathan A way for people already involved in projects and people curious about them to get together.
sjn exactly
jonathan Without having to have the luck to catch a free moment with them around the conference.
sjn and I'm not expecting everyone to show up at the hackathon, so I'd like to make something like it available to the people right there at the conference venue
jonathan Sure 16:04
I guess you want to make more the project discovery part there rather than the actual writing code etc part of what happens at hackathons...
PerlJam a hack-track carnival with games and flashy lights sounds like the way to go ;-)
jonathan Yeah!
sjn and if we get 120+ Perl people, plus perhaps 100-200 Go Open attendees aware of what's going on, then that might work out very well visibility-wise 16:05
jonathan Whichever project has free beer gets my interest! ;-)
sjn www.freebeer.org/blog/
(we were considering getting some of that brewed for the conference :) 16:06
doesn't look like it'll happen though :-(
jonathan :-(
Oslo beer = expensive
sjn mm
PerlJam Make a big sign that says "Perl 6 Microbrewery" :-) 16:07
sjn "Fix an RT bug, get a Free Beer"
PerlJam you probably don't want to be handing out a pint each time though. 16:08
sjn anyhoo, if you have any ideas on how to use thursday evening at NPW, write a few words in the wiki
PerlJam: yeah, let's make it a thimble-full per bug :) 16:09
16:18 Theory joined
pmichaud beer for bugs? 16:20
moritz beer *against* bugs please :-)
pmichaud rakudo: say 'hello-world'.split(/<ws>/).perl; 16:21
polyglotbot OUTPUT[["hello", "-", "world", ""]ā¤]
pmichaud that seems wrongish -- seems like there either should be a "" at the front or no "" at the end. 16:22
rakudo: say 'hello-world'.split(/<wb>/).perl; 16:23
polyglotbot OUTPUT[["hello", "-", "world", ""]ā¤]
pmichaud invites masak to submit a rakudobug. :-)
masak submits
particle rakudo: say 'hello world'.comb(/\\S+/).perl; 16:24
polyglotbot OUTPUT[["hello", "world"]ā¤]
moritz pmichaud: what's wrong with it? the trailing item?
pmichaud moritz: I can't decide if it's missing a leading item or has an extra trailing one. 16:25
moritz maybe TimToady can comment on that?
pmichaud rakudo: say 'h w'.split(/\\s*/).perl;
polyglotbot OUTPUT[["h", "w", ""]ā¤]
pmichaud I'm guessing "missing a leading item", but my guesses seem to be wrong more often than not. 16:26
afk for a bit 16:30
dalek rrot: r36384 | NotFound++ | trunk/compilers/imcc/imcc.y:
[imcc] fix a bison problem, TT #275
 16:33
Infinoid NotFound++ # that's been a problem for me for *months*! 16:35
NotFound Infinoid: not so fast, I don't tested it with 2.4 16:37
Whiteknight pmichaud, can you take a look at RT #37578 and see if this is still an issue for you? 16:38
I'm pretty sure :unique_reg fixes it
Infinoid NotFound: I'm going to test with 2.3 and then 2.4.1, one moment
NotFound I have 2.3 16:39
16:40 Andy joined
kj Infinoid: could you also run bison 2.4.1 on compilers/pirc/src/pir.y? 16:42
(i don't have 2.4.1)
sjn pmichaud: Tell me what you think of <xrl.us/bebpnx>, and especially point #3? :) 16:43
Infinoid strange, that resulted in a linker error. trying again in a fresh checkout 16:45
Confirmed in a fresh checkout. (though I had to touch imcc.y to convince make to regenerate) 16:50
/home/infinoid/parrot-dev/bison-test/blib/lib/libparrot.so: undefined reference to `_namespace'
I also tested with bison 2.3; that worked fine
NotFound _namespace? 16:51
kj not sure, is there a version of bison after 2.4?
ISTR that 2.4 had issues...
NotFound We must not have any global symbol with that name-
Infinoid there's 2.4.1 16:52
kj yes, but I think there's a 2.53 or so
Whiteknight particle: ping
NotFound symreg.h
111:EXTERN Namespace * _namespace;
WTF?
16:54 davidfetter joined
Infinoid it's referenced in push_namespace() in symreg.c 16:55
kj it seems, from a google search, that many projects have problems with 2.4.1
Infinoid I can just blacklist that version so gentoo stops updating me to it 16:56
NotFound Some objection to rename it as pesky_global__namespace ? 16:57
Infinoid I have no problem with that. Though I'm not sure why I get the error with bison 2.4.1 but not with bison 2.3 16:58
kj preferably it would go into IMCC_STATE or whatever
the IMCC structure with global stuff
NotFound Infinoid: is declared as extern but never defined.
Infinoid does imcc 2.3 define it?
symreg.c is code we wrote, and the reference to _namespace is dated 2007-04-18, so apparently it was there before 16:59
...but I don't see it anywhere in my local diff, so I'm confused
NotFound Don't know why it doesn't always fail
kj are we sure that EXTERN actually expands as 'extern'? 17:01
and there seems to be something in imcc/unit.h (checking..)
Infinoid kj: it's referenced on compilers/imcc/symreg.c line 133, regardless of what EXTERN does
kj oh that's not it (unit.h) 17:02
Infinoid oh, I see, if EXTERN defines to "", it'll declare it and then life is good?
kj well, if EXTERN is really expanded as 'extern', then it's not a declaration; otherwise it is
well, life is better ;-)
Infinoid heh :)
NotFound #ifdef PARSER_MAIN 17:03
# define EXTERN
#else
# define EXTERN extern
#endif
kj which in this case is really a paradox... it's not "better" (than good). Anyhoo...
NotFound: exactly, that's what I thought
Infinoid "It's better than bad, it's good."
NotFound This sucks
purl Nothing sucks, it only BLOWS.
kj (prototype code)--
Infinoid log++
kj but apparently EXTERN is expanded differently based on the bison version?! 17:04
particle git fetch origin && git rebase master ## get changes from github's master, and merge with my local changes?? 17:05
Infinoid itym "git rebase origin"
it's git rebase <from>, not git rebase <to> 17:06
particle sorry, i got the wrong command from the cheatsheet
Infinoid err. I think I failed to make it any clearer.
particle git rebase master is for a new branch
git fetch origin && git rebase origin
NotFound The funny thing is that is used only in symreg.c, there is no need to make it extern
Infinoid git rebase master means "take my local diffs and stack them on top of the stuff in the master branch"
particle i want the equivalent of "svn up" 17:08
Infinoid yeah, "git fetch && git rebase origin"
or if you like, git pull --rebase
(same thing.)
I think we can make this simple 17:11
in your ~/.gitconfig or /etc/gitconfig or whatever:
[alias]
update = pull --rebase
now "git update" works.
17:12 hercynium joined, elmex joined 17:14 particle1 joined
particle1 git doesn't seem to respect my .gitconfig, i need to modify one in every project root 17:16
Infinoid urk.
msysgit, right?
particle1 in any case, my rebuild rakudo script (rbr.cmd) is now working 17:17
yes
Infinoid I'm gonna test msysgit a little, and bang out some bug reports if I can't find a better way :) 17:18
particle1 thanks!
i do have HOME defined, btw
so it's not that simple :)
Infinoid ok. but they might be preferring "gitconfig.ini" or somesuch on that platform 17:19
particle1 oh, and i'm using it via windows command shell, not git bash
hrmm.
Infinoid I don't know. if all else fails, I'll just run "strings" on their binary :)
Whiteknight particle1: Has your Inf/NaN work recently resolved RT #38887?
particle1 yes, will close now 17:22
dalek rrot: r36385 | NotFound++ | trunk/compilers/imcc:
[imcc] drop EXTERN qualifier from the _namespace symbol, make it static and rename it creatively
NotFound Many moons ago, there was a ticket about killing globals from imcc 17:24
particle1 kill them all!
purl www.monkeydyne.com/rmcs/dbcomic.phtml?rowid=7480
kj I think the _namespace is one of the last few... 17:25
(hint hint ;-)
Whiteknight particle1++ 17:26
particle stupid net.
davidfetter anybody @ FOSDEM?
NotFound Infinoid: can you try it again now? 17:27
Infinoid NotFound: linked successfully, running testsuite now 17:33
NotFound: All tests successful on linux-x86-64 with bison 2.4.1 17:41
NotFound Good :)
Infinoid particle: C:\\Program Files\\Git\\etc\\gitconfig works for me 17:46
particle with no .?
hrmm
Infinoid yeah. There's a 3 level config inheritance scheme... first /etc/gitconfig, then overwritten with values from ~/.gitconfig, then overwritten with values from .git/config 17:47
I'm still not sure where the $HOME one looks.
particle ok. i have c:\\usr\\git\\etc\\gitconfig, so at least i can modify that
Infinoid try adding [test] foo = bar, and see if it shows up in "git config -l"
particle wtf. 17:49
added [foo2] bar2 to %HOME%/.gitconfig and it worked
Infinoid I didn't do it, I swear
particle can aliases refer to each other? 17:50
alias.up=update ; alias.update=pull --rebase
Infinoid I don't know. But you can just stick whatever you want in a .cmd or .bat file and run that, if you prefer 17:51
prefixing your command with ! means it's treated like a shell command
more details are in 'git help config', for the interested.
particle C:\\Users\\particle>git up 17:52
Expansion of alias 'up' failed; 'update' is not a git-command
no worries though, fixed
infinoid++ 17:53
Infinoid Great. Next thing this box needs is ccache-win32. 17:54
particle seems git pull --master gives a non-zero exit if you're up to date already :(
annoying, but worked around.
ccache was easy to install config, now that i fixed parrot 17:55
install/config
particle wants pmccache
Infinoid I was thinking of starting with strcache. But they're both useful ideas 17:56
17:59 chromatic joined 18:03 mberends joined
Infinoid Ooh, doublequoting error on mingw 18:03
nopaste "Infinoid" at 96.238.213.50 pasted "mingw build failure" (47 lines) at nopaste.snit.ch/15502 18:05
particle ugly 18:06
18:07 elmex joined
rurban -Wl,-L blib/lib can be removed from mingw 18:09
cotto NotFound++ #pesky_global__namespace 18:11
particle 6629/145.86 18:14
purl 45.4476895653366
particle 6629/88.78 18:15
purl 74.6677179544943
particle rakudo passes 45% of the official test suite, and 74% of its regression suite
wow. 18:16
chromatic 7250 tests is halfway? 18:17
particle t/spec has 14586 tests
chromatic 7293 18:19
rurban Infinoid: do you fix this mingw double-quote or should I take it?
chromatic We could hit that in a month.
particle yeah, i think we can 18:20
Infinoid rurban: I'm at work, I won't have any tuits until this evening, please feel free :) 18:21
rurban hints: libparrot_ldflags => "\\"$build_dir\\\\libparrot.dll\\"", => dynpmc: # Paths need quoting as they may contain spaces. 18:23
but the path is already quoted of course
nopaste "rurban" at 212.183.50.172 pasted "fix "mingw build failure" (47 lines) at nopaste.snit.ch/15502" (21 lines) at nopaste.snit.ch/15503 18:25
rurban I'm running this through my "buildfarm" now but it should only affect mingw 18:28
nopaste "rurban" at 212.183.50.172 pasted "fix "mingw build failure" at nopaste.snit.ch/15502 #2" (22 lines) at nopaste.snit.ch/15504 18:36
Infinoid Isn't it possible (but less likely) that someone may be building parrot on unix in a directory with spaces? 18:38
Infinoid isn't really sure this needs to be a special case 18:39
chromatic It's possible.
Infinoid Though I suppose using quotes everywhere makes the build logs ugly 18:40
rurban pwd is space poisened? hmm, unusual, but should be thought of. 18:41
We could check it in advance and quote the space with \\ 18:42
only cmd.exe cannot properly quote space with \\, but there we have it quoted. 18:43
Infinoid Or just detect spaces in pwd, and only bother setting $PATHQUOTE in that case
I dunno.
I think that'd be nicer than just doing it in the $^O eq 'MSWin32' case 18:44
rurban true 18:45
but what other chars need to be shell quoted then?
Infinoid space is by far the most common one 18:46
rurban perl5 has some qx logic to detect these chars and quote it corerctly then
I'll have to leave in 15min so I won't do it either :) 18:48
nopaste "rurban" at 212.183.50.172 pasted "mingw msvcrt failures" (12 lines) at nopaste.snit.ch/15505 18:57
rurban mingw32-make realclean => make -C languages clean process_begin: CreateProcess(NULL, make -C languages clean, ...) failed. 19:03
make <=> mingw32-make
But I'm not sure if this is a kosher error, or not a wrong rsync inbetween 19:04
Oh, how I love my daily solaris: "No space left on device" ZFS is awsome 19:13
Infinoid heh 19:21
particle i thought zfs fixed everything! 19:26
rurban I came up now with this solution: zfs list -t snapshot | perl -ane 'system("zfs destroy $F[0]") unless /^NAME/' 19:30
ZFS is a dream if you have enough space left 19:32
if not you cannot even delete files anymore, boot and login neither
Coke wonders if he can afford to go to any parrot cons this year. 19:34
rurban rm -rf /usr/X11R6 helps, hopefully 19:36
19:42 barney joined 19:43 ask_ joined 19:44 ask- joined
Whiteknight how do I add a new core PMC to the build? 19:51
rurban makefiles/root.in
config/gen/makefiles/root.in exactly
Whiteknight where in root.in? I just looked through there and couldn't find any obvious place to add it?
Coke src/pmc/pmc.num ? 19:55
rurban config/auto/pmc.pm 19:56
right: src/pmc/pmc.num
Whiteknight urg, is there a way to automatically renumber this list? 19:57
particle ack -af tools/ 19:58
or maybe ack -ag pmc tools/
Whiteknight i found the "pmcrenumber" make target
yeah, that worked 20:00
Coke (renumber the list) hell, that would have been useful when I was removing pmcs.
20:00 lu_zero joined
Whiteknight I've got a prototype BigNum PMC I'm playing with 20:01
it's mostly te BigInt PMC run through a few regexes, but it's not bad
chromatic Why do we need a separate BigInt and BigNum?
Whiteknight GMP has two completely separate sets of functions for the two types 20:02
plus, one doesn't have decimal points
rurban well, one is exact, the other not. 20:03
chromatic I'm not sure GMP's model is right. 20:04
Whiteknight that sounds like a problem for the GMP people
Coke no; we can use external libraries, but we should still decide what /our/ API is. 20:05
Whiteknight We have an RT ticket that requests the creation of a BigNum PMC. ergo, I am writing one 20:06
and when we decide that it's a stupid thing to have, we can open a second ticket to have it removed
...and I'll rake in the karma on both transactions :)
chromatic I can see having Float and an Integer types, but I don't see a lot of benefit to having more than one arbitrary precision unlimited size type. 20:07
Whiteknight In which case, we would keep BigNum, since it's more general and useful then BigInt is 20:08
either way, we do need to write BigNum. The question is whether we keep a separate BigInt after the fact
chromatic Fair enough.
rurban BigInt is a special faster case of BigNum 20:09
Whiteknight questions like this are for the architect, I'm just a mindless grunt
barney BigInt can be abused as a BitArray 20:10
Whiteknight barney: STRINGs can do that too, and ResizableBooleanArray 20:12
barney Yes, I always thought the BigInt and ResizableBooleanArray should be merged 20:13
NotFound Please don't abuse STRINGs
Whiteknight NotFound: we have a binary: charset that can be used to turn strings into dumb buffers
NotFound We have it? Last time I asked was considered a bad idea. 20:14
Whiteknight BigInt seems to be more infinitely growable: BigNums have machine-dependent limits on manitissa size
NotFound: I thought we had it
chromatic Is that a GMP limitation?
Whiteknight chromatic: yes, based on the documentation I am seeing
looks like the mantissa of a BigNum can be +/-INT_MAX 20:15
NotFound Whiteknight: some people seem to think that latin-1 is a sort of plain binary charset.
chromatic What about IBM's fixed decimal math library? 20:16
Whiteknight I dont know anything about IBM's library 20:17
you have some kind of a link for it? 20:19
chromatic speleotrove.com/decimal/
Whiteknight ideally, I think the BigNum and Bigint PMCs could both be wrappers over any number of math libraries, depending on what the user has on their system 20:21
chromatic Ugh. 20:22
I should elaborate. 20:23
"Ugh, installation, configuration, and run-time behavior differences problems leading to unpredictable and difficult-to-debug behavior."
20:24 IceGuest_7 joined 20:25 kj joined
Coke external libraries are ok IMO, but only if we're consistent about it. 20:26
NotFound For example, we are consistent about we use icu for unicode things, but people still complains about things that does not work without icu ;) 20:29
And when we add some quick fix to make it work, they use it to exemplarize that we have fallbacks. 20:30
20:37 szabgab joined
rurban speleotrove.com/decimal/#summary looks weird to me. 20:41
use decimal; should be tried out as perl5 pragma first. for financial stuff 20:43
dalek rrot: r36386 | kjs++ | trunk/languages/pod:
[pod] add lists for blocks and =over lists. + add 'say' builtin
 20:51
kudo: f990451 | (Moritz Lenz)++ | tools/autounfudge.pl:
[tools/autounfudge.pl] various small updates

   * allow finer graded control over what we want to have fudged
   * a bit more documentation		 21:21
shorten dalek's url is at xrl.us/bee3zs
GeJ Good morning everyone 21:24
21:40 szabgab joined
NotFound Hey, stupid bots, you've stolen my commit message 21:41
21:52 AndyA joined
cotto rrot: r36387 | NotFound++ | trunk/compilers/imcc:
[imcc] kill EXTERN macros

Rename is_def symbol and declare and define it without macros		 21:54
all better now
NotFound But you're not a bot, aren't you? 21:58
Tene I am! 22:04
22:05 confound joined
chromatic Hm, instead of a pesky global, how about more flags on IMCC_INFO(interp)? 22:07
Infinoid No, see, that would be smart.
NotFound chromatic: the intention of the names is that someone gets rid of them. 22:08
chromatic Refactoring via verifiable baby steps is fine with me.
Hah, that's like the comment in the GC a while back that said "If you use this, it'll probably segfault."
NotFound I like better executable comments 22:09
chromatic It took less time to fix than it would have to write the comment.
NotFound Not to me, I fear touching imcc ;)
22:09 Whiteknight joined
rg would someone mind telling the newbie, why you're still fiddling with imcc when pirc is supposedly the future? 22:10
Infinoid Because pirc isn't the present yet, apparently. 22:11
NotFound rg: becuse it brokes somenone's build
Infinoid Though pirc seems to be working fine, here 22:12
(At least, examples/pir/substr.pir gets through it and runs fine)
particle compile rakudo with it. 22:14
22:14 khatar joined
khatar hello world 22:14
purl i guess hello world is www.roesler-ac.de/wolfram/hello.htm
Infinoid hello khatar
22:15 szabgab joined
khatar i hope my question isn't too stupid: is it possible to write functions in haskell in order to link it to parrot through NCI ? 22:36
moritz haskell is said to have a pretty good NCI 22:37
so you could probably embed parrot into haskell apps
don't know how hard/easy it is the other way round
PerlJam moritz: not just probably; definitely. audrey did it. 22:38
khatar isn't NCI here to do the opposite ? (write in haskell, use in parrot)
PerlJam (but, yeah, I'm not sure about the other way around either. Not enouhg haskell-fu)
moritz khatar: NCI is for "write in C, use in parrot"
PerlJam khatar: if you can compile haskell code to a library with the C calling conventions, you can probably use NCI to get at it from parrot. 22:39
NotFound khatar: NCI is intended to interface with system libraries, that is, things that have a C interface.
khatar ok... so the question i have to answer isn't can NCI use haskell code but can ghc compile haskell with C calling conventions 22:41
thanks guies
22:42 rurban_ joined
NotFound Alternatively, you can write a Haskell compiler in parrot 22:42
moritz ;-) 22:43
khatar NotFound, haha i'm not skilled enought fot these kind of work
NotFound And then, compile pugs with it
khatar my first attempt is to write my wiki syntax grammar 22:44
22:55 szabgab joined
dalek rrot: r36388 | NotFound++ | trunk/languages/ecmascript/src/classes/Boolean.pir:
[ecmascript] fix Boolean stringify
 23:02
23:03 szabgab joined 23:05 Whiteknight joined
Whiteknight urg, 31 active tickets outstanding for the 0.9.1 release 23:12
moritz which is... when?
in 9 days?
Whiteknight 12 days 23:13
still, not much
23:25 szabgab joined
Whiteknight has Rakudo moved to it's own repo yet? 23:34
seems like TT #267 could be closed when that happen 23:36
GeJ Whiteknight: the last commit in Parrot's rakudo is 4 days old. 23:38
I believe that development moved to github. 23:39
github.com/rakudo/rakudo/commits/master
chromatic If every committer fixed two or three tickets.... 23:40
Infinoid I'm too busy making new tickets 23:41
(Translation: too busy being part of the problem.) 23:42
23:43 ron joined
Whiteknight I'm closing tickets as fast as I can 23:44
23:45 bacek_ joined
Whiteknight I can close TT #178 if I can add an extra pointer to PObj and all isomorphic objects 23:46
the tradeoff is that eventually we will be taking the UnionVal structure out of them, which is a net win 23:47
plus, that extra pointer can be added conditionally only when the MS GC is active 23:48
chromatic I'm not sure that's the right approach.
Whiteknight I don't think it is either, which is why I mention it here and not just do it and commit it 23:49
23:49 TiMBuS joined, kid51 joined
ron Does anyone know why the object pmc forwards bitwise (and/or) and compare mmd vtable functions through the proxy pmc but won't use proxy on math vtable mmd functions? 23:51
chromatic Probably an oversight.
ron Should I post a bug note including the vtable mmd functions that seem to have been left out of lib/Parrot/Pmc2c/PMC.pm (sub vtable_method_does_multi)? 23:54
chromatic If you like. A test case would be even better. 23:55
ron Thanks - have a test case with cmp_num and hope to get to it today/tomorrow. 23:56
Whiteknight With the UnionVal gone, there won't be any pointers in a PObj at all 23:57
chromatic Evil thought:
a union containing a PObj and a GC pointer. 23:58
Whiteknight I had thought about that
probably going to run into all sorts of alignment problems
chromatic I wouldn't think so.
The compiler ought to pad any pointer sufficient that you can store it wherever it has to land. 23:59
23:59 ron left
chromatic As long as PObj is big enough to contain a single pointer (and it already contains one, right?) then it should work. 23:59

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!