pugscode.org/ | nopaste: sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by diakopter on 11 July 2007.
00:00 justatheory left 00:16 sunnavy joined 00:21 elmex left 00:38 polettix left 00:50 weinig_ left 00:51 silug joined 01:04 devogon left 01:08 nipotaway is now known as nipotan 01:22 Siggy_SS joined 01:23 Limbic_Region left 01:25 weinig joined 01:42 thorat left 01:49 REPLeffect left 01:50 REPLeffect joined 02:04 kolibrie joined 02:12 jhorwitz left 02:15 Lectus joined 02:18 kolibrie_ left 02:20 weinig_ joined 02:22 weinig left 02:44 Siggy_SS left 03:14 Lectus left 03:26 cmarcelo left 03:27 pro21ms4 joined 03:49 Siggy_SS joined 03:52 justatheory joined 04:04 torz joined 04:23 Aankhen`` joined 04:48 f0rth left 04:53 penk left 05:01 f0rth joined 05:06 f0rth_ joined 05:08 f0rth left 05:10 Eidolos_ joined 05:20 REPLeffect left 05:21 Eidolos left 05:26 devogon joined 05:29 weinig_ left 05:39 weinig joined 05:50 jisom joined 05:52 drrho joined 06:04 justatheory left
bloonix morning 06:12
dduncan soon 06:24
06:26 amnesiac left 06:29 Siggy_SS left 06:35 elmex joined 06:37 daxim_ joined 06:42 pro21ms4 left 06:52 daxim left, audreyt joined 06:59 polettix joined 07:02 franck___ joined 07:08 daxim_ left
meppl good morning 07:10
07:11 daxim joined 07:13 iblechbot joined 07:24 marmic left 07:27 flokuehn joined 07:42 marmic joined 08:02 weinig left 08:04 mj41 left 08:16 IllvilJa left 08:18 weinig joined 08:27 buetow joined 08:32 daxim_ joined 08:40 BooK_ joined 08:50 daxim left 08:57 dduncan left 08:58 BooK left 09:03 weinig left 09:06 ruoso joined 09:08 dalecooper joined 09:09 dalecooper left, dalecooper joined 09:21 xinming joined 09:29 masak joined 09:33 Aankhen`` left
masak moritz_: I'm on another channel, #bioclipse, on freenode. they are looking for an IRC logging solution. what would you recommend? 09:34
09:36 jisom left
pugs_svnbot r17184 | masak++ | [t/syntax/interpolation/strings.t] 10:04
r17184 | masak++ | * some tests in eval() where not run, adding // ok( "" )
diff: dev.pugscode.org/changeset/17184
lambdabot Title: Changeset 17184 - Pugs - Trac
masak I still think it would be very possible to develop a little Emacs/vi tool that auto-calculated the number after 'plan' 10:05
I'd like to write such a tool, but I would need some real-world data to test on
then I realized that the Pugs test suite is perfect for this
& # lunch 10:07
10:12 franck____ joined 10:29 franck___ left 10:47 thoughtpolice left
wolverian masak, the one we use seems ok 10:48
10:48 iblechbot left
wolverian hm, now that I look at it, it aligns the nicks higher than the text lines 10:48
that looks weird
omg it uses Net::IRC 10:49
10:57 lisppaste3 left 11:05 cmarcelo joined 11:08 lisppaste3 joined
masak wolverian: do you mean this one? colabti.de/irclogger//irclogger_logs 11:19
lambdabot Title: Logs of IRC channels
masak appears that it's the #irclogger people that handle that one 11:20
wolverian masak, I meant irc.pugscode.org, as in topic
masak wolverian: ah, yes. that's why I asked moritz_++ 11:21
wolverian but that looks good too, and maybe it doesn't use Net::IRC ;)
masak it uses something called irclogger2
don't know if that uses Net::IRC
wolverian it's python, so no..
masak ah 11:22
ilogger2: so it's you who keeps our logs at colabti... 11:23
ilogger2 masak: I am a logger bot. Lines starting with [off] won't be logged. Extra help available at colabti.de/irclogger/irclogger_logs.
lambdabot Title: Logs of IRC channels
masak right. I knew that
11:23 chris2 joined
masak but thx :) 11:23
is it "it's you who keeps" in modern english? 11:24
"it's you who keep" sounds wrong to me
but in archaic english you say "Father, [you] who art in Heaven" 11:25
yuck, verb endings that change depending on grammatic person are stupid 11:26
diakopter masak: but English might regain such endings, if the trend towards dropping nominative pronouns continues. 11:29
masak there's such a trend? 11:30
didn't know
diakopter at least on IRC/IM/txt :P 11:31
masak now that I think about it, I do it in Swedish quite a bit when writing emails to friends
wolverian interesting, search.cpan.org/~philcrow/Java-Java...java2perl6 11:32
lambdabot Title: java2perl6 - a Java to Perl 6 API translator - search.cpan.org
diakopter yeah... much of the time, the context or even (*gasp*) `lilt' can indicate person. 11:33
masak "lilt" ? 11:36
diakopter intonation, I meant. 11:38
masak ah
it's a bit of an art to convey intonation correctly over the webbernet, though 11:39
I sometimes cringe at people who use exclamation marks liberally
diakopter I guess for someone like me (who's not learned Japanese or similar) considers all intonation as sounding like a "lilt". :)
er, how'd that "for" creep into that sentence? 11:40
masak diakopter: not sure I follow. how does a "lilt" sound?
what's the best way today to generate an AST from p6 source? 11:41
diakopter I'm using it to mean several of these definitions: dictionary.reference.com/browse/lilt 11:42
lambdabot Title: lilt - Definitions from Dictionary.com
masak I just realised that this would be far superior to parsing test files directly in a script
diakopter: wow, there really is such a word. I thought you made it up, or borrowed it from some other language
diakopter heh... as if there are words that don't arise in those ways... ;) 11:43
masak diakopter: 'course, but in this particular case...
diakopter masak: the implication being that if I use such a word, there is such a word. 11:44
(I or anyone else)
masak diakopter: I'd agree, except there's still the extra authority of finding a word in a dictionary 11:45
that means at least two people have used it :)
hm, maybe I mean something like "authoricity" instead of "authority"
diakopter: you don't happen to know how to generate an AST from perl6 source, do you? 11:46
diakopter masak: Hey! Authoricity isn't a word! Someone already coined a term for your usage! ;) 11:47
masak diakopter: :)
diakopter (joking...)
masak yes
diakopter hmm, well, actually I was just studying pugs' src/ to find exactly that. 11:48
funny you ask
masak heh
diakopter: as long as you have a certain familiarity with src/, maybe you can answer some other questions ;)
where is string interpolation being done? I've looked, but turned up zilch 11:49
also, what's the difference between a an anon black and the common space outside all blocks? 11:51
(muddy question, maybe I need to clarify that)
it's pugs t/magicals/subname.t, the last test 11:52
it wants &?ROUTINE.name to be undef, and I can do that
diakopter looks for a long while
ahah!
masak but right now it's '<anon>', and I need to test for something that distinguishes it from test #3, otherwise that will fail
masak loves discussing wanted features based on failing tests 11:53
diakopter masak: pronoun nominatives can disappear today because the imperative mood (implied 2nd person) has all but disappeared!
masak diakopter: yes, are right about that
er 11:54
I mean "nod"
(which could be imperative, but isn't)
diakopter LOL 11:55
"diakopter, yes, are right about that"
(I assume that was a typo)
masak diakopter: no, just testing writing w/o nominative pronouns 11:56
diakopter k; gotcha :D
polettix lol
masak it would have been a bit too coincidental to have been a typo, wouldn't it?
diakopter no.... I drop (random) all the time
masak ("gotcha" is fun too... it's the only time "you" is spelled "cha") 11:57
diakopter "whatcha" doing
masak ah, true
diakopter "captcha" this! ;)
masak no ;)
for one thing, "capt" is not a verb 11:58
diakopter lol; sorry, just taking the opportunity to exercise my visual word index. 11:59
masak diakopter++ # for exercising
diakopter k, so, gimme a while to research some of your questions. 12:00
masak you forgot "cha-cha" in that case
diakopter: sure
diakopter ... "k, so, " being a plea for tolerance of the forthcoming imperative verb (because otherwise someone might take it as a demand or command, as if I think I have authority or influence). Pray suffer me a time to research. 12:04
masak that's how I read it
12:08 polettix left 12:16 buetow left 12:20 buetow joined
allbery_b the archaic meaning of "pray" is supplication; the modern equivalent is "please" 12:20
in modern English "pray" has all but lost its secular usage
masak likes the new linguistic touch to #perl6 12:21
diakopter but first, 'authoricity'. From the (extremely few) usages indexed by Google, it can denote "[intended or effective] authoritative-ness or definitive-ness"... is that your usage?
allbery_b (where's TreyHarris, he can probably give you a complete rundown on it... :)
masak diakopter: that's my usage
diakopter: "authority" felt like something else when I tried it 12:22
even if it can probably also be used to denote what "authoricity" denotes
12:23 iblechbot joined
diakopter allbery_b: I was trying to be archaic, hence "suffer" for "allow" or "let me have". 12:25
albery_b: it sounded like you were correcting me, but now I'll assume you weren't. :) 12:29
12:35 Gothmog_ left 12:36 nipotan is now known as nipotaway
allbery_b no, I dropped in late and didn't spot any context in a quick scan of backscroll 12:36
what I did see implied someone trying to figure out unusual English usage so I tossed that in 12:37
diakopter ah. :) 12:40
msft is BSDish-licensing IronRuby's source...
oh, I'm behind the times. 12:41
masak diakopter: they are? cool. kudos to msft 12:43
12:45 TJCRI joined
diakopter TextMate doesn't like audrey's nested .hs block comments 12:48
masak well, it's open source, isn't it? should be able to change 12:49
diakopter the syntax highlighting grammar is, yeah 12:51
masak yes. maybe not the editor itself
would be difficult to sell for money then
diakopter thoughtpolice: ping 12:52
12:55 dwave left 12:57 jhorwitz joined 12:59 Schwern left 13:00 Schwern joined, Schwern left 13:02 Schwern joined, drrho left 13:05 jhorwitz left, jhorwitz joined 13:06 jhorwitz left 13:07 jhorwitz joined 13:13 Aankhen`` joined, xinming_ joined 13:19 mncharity joined
pugs_svnbot r17185 | diakopter++ | minor corrections; still haven't determined whether the draft Concurrency.POD should be linked from here... any suggestions, anyone? 13:23
diff: dev.pugscode.org/changeset/17185
13:25 xinming left
masak speaking of which, I'd like to remind anyone who has the required bits to change it, 13:32
that perlcabal.org/syn/ has a broken link
namely A07, dev.perl.org/perl6/doc/design/apo/A07.html
now I know the background of A07, and that it doesn't contain any essential text 13:33
but still, I think it's a shame that it gives a 404
the official docs of a product contribute to much of the initial impression about the product 13:34
and perlcabal.org/syn are as much official docs for perl6 as you can get
people will get the 404 and go "oh well, guess there's not enough people reading those docs to have noticed yet" 13:35
"guess there's not that much interest for perl6, then" 13:36
...which I think would be sad, when it's very easy to patch
masak winds down
13:36 thepler joined 13:37 hkb_zz left 13:38 hkb_zz joined 14:00 BooK_ left 14:07 BooK joined 14:25 amnesiac joined
diakopter masak: is there an A07 anywhere? 14:32
masak I should think so
but it's very small
basically says that it didn't need to be written 14:33
because formats are not core anymore
I don't know where or if it can be found on the net
but I remember reading it at some point 14:34
and there's a (broken) link to it, but not to many of the other Axx
diakopter: how's it going with my questions? :) 14:36
don't want to rush you in any way, but I'm very curious 14:37
14:40 IllvilJa joined
diakopter masak: :P slowly... was afk for a while. building my new 1GB vps for slightly-more-private-than-feather development... www.vpsfarm.com/ 14:40
lambdabot Title: vpsFarm
masak diakopter: ok, don't let me slow you down further :P
14:42 qmole left
diakopter (a darn good deal if you ask me... and I've tried several vps companies - unixshell.com(1 too many lost-all-data crashes) prgmr.com(yikes! 3 too many lost-all-data crashes) and spry.com/vpslink.com (1 too many lost-all-data crashes). Haven't lost any from vpsfarm.com yet. 14:42
avar likes slicehost.com 14:43
masak agrees with the 'yikes', in general
diakopter haven't tried that one yet - it's always been somewhat more expensive than those others I mentioned. If vpsfarm fails me... I'll try slicehost next.
masak I haven't needed such solutions yet. diakopter, what do you use them for? 14:44
diakopter masak: clients/employers need commodity offsite backup locations (that are more customizable than the ASP offerings), email filtering/holding/hosting, vpn/proxy endpoints... and I like to toy around on them. it's like having vmware or parallels, except much easier (less laptop hdd intensive, I mean) to reload OS images while "learning" tricky system admin skills. :) 14:48
vpsfarm.com is the best I've seen, so far... and of course, by far, the cheapest. $40/month for 1GB ram and 5Mb/s throttled/unmetered. Oh, and I also use them for file distribution/hosting for clients/employers. 14:50
you don't get a VPS for the CPU... you get one for the memory and bandwidth. 14:51
masak ok
14:54 rdice joined
diakopter btw, all those companies advertised RAID1/5 storage for their vps' filesystems. But in each case of failure, it was the RAID controller that failed, and the arrays couldn't be rebuilt on a replacement controller, either because "couldn't get an exact-model replacement in time" or "the dying controller ruined the array". 14:54
(so they said) 14:55
Juerd That's why I always use software raid.
BTW, my company has VPS'es too
See www.rootaccess.nl </spam>
diakopter so... for clients I get them in pairs, and make the vps host insist they're on distinct machines.
Juerd (We still haven't built a real site, because we're doing well enough without even) 14:56
diakopter Juerd: can you beat vpsfarm.com's prices?
Juerd I have no idea
I'm not interested in beating prices, to be honest.
No, I can't
diakopter Juerd: I will definitely let you know as soon as vpsfarm has a crash or goes out of business or something :) 14:57
RAID-10 in software?
14:59 moritz joined
Juerd Yes. 14:59
moritz masak: do you still need an irc log solution?
masak moritz: yes. I've asked on #irclogger, but they have not responded yet
Juerd diakopter: I've stress tested that extensively and it works very well
We have great I/O on our servers 15:00
diakopter oh, interesting, it can even do RAID1+0ish on odd numbers of disks.
15:00 irclogbot left
moritz masak: my server is rebooting atm (new kernel or something), after that I can start logging 15:00
15:00 irclogbot joined, moritz_ left
Juerd diakopter: We use it on 4 disks, the old fashioned way 15:00
masak moritz: cool
moritz masak: if you have old logs (irssi or something), so can mail me and I'll try to parse and import them
masak moritz: I don't 15:01
not sure if Emacs' erc saves my logs, but it's nothing I've enabled
anyway, even if it does, it's not much and it doesn't really matter
15:02 Casan left
moritz allright, don't worry 15:02
masak ok
15:03 Casan joined
diakopter @tell TimToady in case you missed the hilite.... perlDreamer?> TimToady: In S29, there are only 3 argument forms of substr defined. Does that mean that the 2 and 4 argument forms from perl5 do not exist in perl6, or that since they are unchanged they remain? 15:03
lambdabot Consider it noted.
moritz waits for his server to come up again 15:04
diakopter open question for all: remember we discussed funding Perl 6 work previously? Please help me get an idea of how much anyone here would "pledge" (commit to fund on a monthly-for-12-months or prepaid annual basis) if 200% matching funds to your pledges were promised. Would anyone here pledge, and how much. 15:07
in other words, some (possibly anonymous) benefactors promise to pledge $40/month for every $20/month that anyone pledges/commits. 15:08
masak senses an infinite loop there somewhere 15:09
diakopter (assuming full transparency and consensus for the identity, goals, and activities of the supported person, but not assuming US tax-deductible donations).
integral And what's the money used for? www.perlfoundation.org/parrot_grant_from_nlnet shows lots of unclaimed money for parrot 15:10
diakopter that anyone *else* pledges/commits.
moritz integral++ # that's my question as well
masak diakopter: that's better :)
15:10 franck____ left
moritz diakopter: I think the microgrants took a long time to be consumed... 15:10
[particle] integral: that money is free for the taking. feel free to attack one of the milestones listed there
integral There's also TPF's microgrants, of which, only 4 have been allocated 15:11
[particle] moritz: there are still 5 microgrants that hasn't been consumed
moritz diakopter: and audreyt++ might want to get paid for pugs hacking, but not at the moment...
[particle]: good to know ;)
[particle] it'd be a shame to come out of conference season without any new microgrant proposals 15:12
diakopter [particle]: do you get to see the ones that are rejected? Have there been (m)any that were rejected? 15:14
[particle] diakopter: i don't get to see them, that's for jesse and leon. but i talked to jesse, and there haven't been many proposed 15:15
i'm not aware of any that have been rejected -- but there may have been a few
moritz the java2perl6 module that was mentioned earlier - is that the (intermediate) result of one of the micro grants? 15:17
diakopter seems to me that anyone who knows alot about the status of the project wouldn't propose something that would get rejected... so it makes sense that there haven't been (m)any rejected. I suppose an "outsider" could submit a non-serious proposal, though.
moritz aye 15:18
masak allright, gotta go 15:20
diakopter moritz: yeah, see perlfoundation.org/microgrants - Phil Crow...
masak see you all
diakopter masak: I'll be online in your morning
masak diakopter: good. see you then
moritz masak: see you ;) 15:21
masak moritz: yes. I'll be back sometime tonight
15:21 masak left
moritz hopefully my server is back then ;) 15:22
diakopter moritz: regarding where the funding/matching funds would go - I hadn't thought about that, specifically. I was assuming consensus among the benefactors could be reached about that. I was more interested in finding out who would be willing to pledge the small amounts, if they knew that their small amounts would be multiplied by 3.
moritz diakopter: I'd be willing to donate a bit if I knew _in advance_ where it went... 15:23
diakopter: and given the fact that there _are_ unused funds (though not too much, I guess) that's the question that has to be answered first
diakopter oh, that's what I meant; sorry. The pledges would all be tentative, pending assignment/consensus. 15:24
moritz diakopter: but I do appreciate your effort
diakopter I'm not looking for "pledges" here on the channel, I'm just asking how much anyone would "consider" contributing.
moritz I think I'd donate up to 300$ once
you can divide that by 12 for your self ;) 15:25
diakopter heh
moritz ?eval ~ 300/12 15:26
diakopter um, evalbot is currently dead. 15:27
I, uh, was fixing it.... :( but...
gimme a second; thanks for the reminder.
moritz reminds evalbot to raise from the dead 15:29
diakopter I created a fresh .svk (2.0.2) mirror/local of pugs in the process, btw... someone should .bz2/post that somewhere.
I mean, b/c audreyt's .svk is ... in an unknown state of usability. 15:31
moritz and I should publish my git mirror
15:31 orafu left
moritz I first did it on my server, but the differnt versions of git-svn caused that not to update correctly :( 15:31
so I guess I'll just install a cron job on feather 15:32
diakopter oh... yeah... git-svn on feather is like, bleeding edge. 15:33
moritz 1.5.3~rc1-1 15:34
as opposed to 1.4.4.4-2 on debian etch
15:35 Gothmog_ joined 15:36 irclogbot left, irclogbot joined, moritz left 15:37 Blwood is now known as dingotux, dingotux is now known as Blwood, moritz joined 15:38 perlDreamer joined 15:41 evalbot6_r17176 joined
diakopter whoa; it actually worked. 15:41
?eval 300*3/12 # 25, duh 15:42
75
lol
hm, well 15:43
<sigh>
(at least the revision number is correctly appearing now)
moritz ?eval 1 15:44
15:44 nekokak_ joined, evalbot6_r17176 left
diakopter tries again 15:44
boy..... it's slow to svk update -s ; svk info ; make clean ; make soon 15:52
15:56 pugsbot_r17185 joined
Juerd wonders when TPF will wire the money 15:56
I hope it'll happen before thursday, but probably not
I'll be gone for a week from friday afternoon
[particle] Juerd: i suggest you follow up with email. i've found that pestering helps speed things up 15:57
diakopter ?eval 1
15:57 unobe joined
diakopter argh. 15:57
Juerd [particle]: Did that, obra helped too.
15:58 sunnavy left 16:00 Eidolos_ is now known as Eidolos 16:02 nekokak left 16:04 xinming_ is now known as xinming
obra [particle]: we're already fully pestering 16:13
[particle] great :) there are others waiting on tpf money, too 16:16
Juerd I wonder what is slowing this small organization down 16:20
Perhaps lack of tuits, perhaps fundamental problems. It's so opaque so we can only guess :|
16:21 orafu joined 16:22 franck___ joined 16:23 Blwood is now known as Ivanlef0u, Ivanlef0u is now known as Blwood 16:24 weinig joined
diakopter Juerd: it appears there is someone named as Treasurer, but who is Ex officio 16:25
Juerd Aha 16:27
Maybe phones would work
It's a relatively new invention that helps people communicate over arbitrary distance ;)
diakopter in realtime!!! :) 16:29
rgs in _constant_ time
diakopter rgs: thanks, yes.
SamB eh.
you guys didn't get very far in physics, did you?
Juerd rgs++ 16:31
samb++
SamB: Sufficiently arbitrary distance, then :)
amnesiac same SamB as ##C?
SamB yeah 16:32
Juerd It's not like the Treasurer is likely to be on another planet :)
amnesiac I see
same Juerd as ##perl? :P
Juerd There's a ##perl? Where?
diakopter SamB: that's kind of a snide remark. :/ Some people stop a course of education for reasons other than not being able to "get" further.
amnesiac at undernet
and dalnet!
Juerd I'm on undernet? Dalnet?
Hm
Could be :)
As far as I know, I'm the only Juerd who writes Perl.
There are a few other Juerds, but I've never seen anyone with this name do anything with Perl 16:33
amnesiac I'm the only Marco that writes Perl!
SamB diakopter: I meant, if they'd taken it they ought to remember that we don't know how to make things go faster than the speed of light
Juerd amnesiac: No :)
amnesiac Juerd, dang!
Juerd, well, the only mexican Marco
better? :P
Juerd Can't verify that :)
amnesiac :/
Juerd I think it's funny that approx 99% of all Google hits on "Juerd" are related to me. 16:34
16:34 justatheory joined
Juerd Yay for CPAN and its gazillion mirrors. 16:34
SamB diakopter: I wasn't trying to insult their intelligence
amnesiac with my name, Google shows my CPAN account too.
and all the stuff I've been into
(including drugs and pr0n movies) 16:35
Juerd diakopter: He didn't mention any reasons for stopping...
diakopter Juerd: not explicitly, no.
Juerd did not get very far in physics 16:38
I got far enough in Perl though
moritz is much further in physics than in perl
I which I knew my perl nearly as well as my physics ;)
Juerd I guess what I'm trying to say is that if that remark offended you, you might be a bit over sensitive 16:39
In my humble opinion, of course.
SamB or maybe I'm insensitive! 16:41
meppl gugu
SamB meppl: what you do that for?
diakopter Juerd: I got the message; I appreciate the sentiment. I was [and remain] unoffended, but I did find it annoying that SamB didn't read my mind and see that I said "realtime" when I meant "synchronous". 16:42
meppl good evening samb
SamB, just for greeting
diakopter :D
SamB heh
ah.
good, um, lunchtime ;-)
meppl bon appetit samb 16:43
moritz Juerd: the changeset r17181 changed index.html on www.pugscode.org 21 hours ago, and still isn't showing on the front page...
Juerd: could you please check if the synchronisation works? 16:44
or am I missing something stupid here?
16:44 Aankhen`` left 16:48 weinig is now known as weinig|brb 16:52 dalecooper left
Juerd moritz: ENOTUIT 16:52
diakopter moritz: I'll look at it after the evalbot is fixed 16:53
I'm not sure what's wrong. it's setup nearly exactly the same way as it is in audreyt's tree.
16:54 pugsbot_r17185 left
moritz ok 16:54
diakopter: does the no-autobuild evalbot work for you? 16:55
diakopter haven't tried running it without the wrapper. 16:56
16:59 weinig|brb is now known as weinig 17:00 pugsbot_r17185 joined
diakopter ?eval "I'm a bot" 17:00
moritz \"I'm a bot", Bool::True 17:01
17:02 pugsbot_r17185 left 17:08 hkb_zz is now known as hakobe 17:13 pugsbot_r17185 joined 17:14 vel joined, beppu joined, pugsbot_r17185 left, TJCRI left
moritz ?eval 1 17:14
diakopter ahah 17:15
it's b/c multiple dispatch is broken.
(eye roll).
17:16 flokuehn left 17:19 Psyche^ joined 17:26 qmole joined 17:31 Patterner left, Psyche^ is now known as Patterner 17:32 pugsbot_r17185 joined
diakopter ?eval 1 17:34
pugsbot_r17185 1
diakopter oh look, I can work around pugs' brokenness. ran across an interesting question, though... one minute... 17:35
17:37 pugsbot_r17185 left 17:38 vel left 17:40 amnesiac left 17:43 amnesiac joined, TJCRI joined 17:56 moritz left, moritz joined, moritz left 17:59 jisom joined 18:00 fglock joined
fglock hi 18:00
lambdabot fglock: You have 1 new message. '/msg lambdabot @messages' to read it.
18:01 moritz_ joined
diakopter fglock: hi 18:01
fglock diakopter: hey 18:02
obra hey fglock
how's life?
fglock I'm doing the tourist thing :) 18:03
obra nice
moritz_ where?
fglock and a bit of $work - I'm implementing a language-to-java translation - with several parts reusable for kp6 18:04
moritz_: Lisbon
obra what language? ;)
moritz_ fglock: cool ;) 18:05
fglock it's called "tea", but there seems to be several languages named like that
I'm finishing the type analysis
the grammar is written in Perl 6 18:06
moritz_ cool ;)
fglock ruoso is implementing the Java emitter 18:10
avar fglock: what $work allows you to hack p6?:) 18:11
ruoso actually I'm doing tree transformations in tea itself now...
refactoring the syntax tree
so that if { func(); func() } { then() } { else () } can be written more sanely in java 18:12
18:12 rindolf joined, dduncan joined
ruoso it would be nice if there was a way to manipulate the tree in a more readable way 18:13
fglock avar: it's a code base migration - we are the R&D crew :)
ruoso I think that applies to kp6, don't it fglock?
currently the tree transformation is weird Perl 5 code
rindolf ruoso: what is kp6, BTW?
ruoso kinda-perl-6 18:14
a subset of Perl 6 and a supper set of mini-perl-6
fglock hmm - we could have a tree transformation library, or a tree transformation language
ruoso I need to take a better look in TGE, but at the first sight, it seems very parrot specific
avar fglock: Take a lesson from lisp, trees are naturally lists of lists:) 18:15
[particle] forget tge (it's parrot-specific). look at nqp
rindolf Hi all, BTW.
avar So I guess if you have a easily modifyable ast everything leads from thereon:)
ruoso actually, I figured out that the match should be more like a tree
fglock nqp?
ruoso actually implementing a tree-traverser of the match is kind of always case-by-case 18:16
as the match object is today
obra nqp is patrick's "perl6 like language for tree transformation"
I believe it's a proper subset.
[particle] not-quite-perl
ruoso nice... any links?
[particle] sure!
svn.perl.org/parrot/trunk/languages/nqp/ 18:17
lambdabot Title: Revision 20381: /trunk/languages/nqp
18:19 buetow left
ruoso hmmm 18:19
18:19 buetow joined
ruoso doesn't seems to be much of help in our case... 18:19
at least I didn't see anything tree-specific 18:20
[particle], did I miss something or it is just a subset of perl 6? 18:23
moritz_ rindolf: hi, see www.pugscode.org/kp6.html ;) 18:24
lambdabot Title: kp6 - KindaPerl6
rindolf moritz_: I see. 18:25
moritz_ was too dump to really help with kp6, so he wrote this homepage instead 18:26
are io layers supposed to work with STDOUT in p5? 18:28
open(STDOUT, '>:utf8', $file) 18:29
that doesn't work for me
[particle] ruoso: it's a subset of perl 6, but will be used for tree transformation in parrot
rindolf moritz_: I think so, yeah.
ruoso hm.. I see, it will be used with TGE, turning TGE less parrot-specific (is it?) 18:30
[particle] well, no, nqp uses tge to build itself.
moritz_ ok, then I'll write a bug report
ruoso but anyway... It really would be nice if the match object had a tree interface
[particle] until nqp is bootstrapped (possible but not high-priority) we'll keep tge around
ruoso but, nqp, being a subset of perl 6, won't help that much with tree transformation 18:31
ruoso saying that from this specific experience
we need something that looks like XSLT to transform a match into something else 18:32
[particle] ruoso: it will certainly help us in parrot, we'll be replacing the parse -> PAST transformations with nqp
ruoso [particle], I see, but I'm talking about yet another abstraction layer 18:33
maybe a pluggable tree interface that could keep that logic.. 18:36
and a tree transformation engine that understands that tree interface 18:37
fglock we could apply a 'Tree' role to the Match class # ??? 18:38
ruoso that would work
actually, it should be two different objects 18:39
so that the match object doesn't have any info of where it is in the tree
what makes it possible to change the "data" for a node
without losing the children
eventually the tree could have all the "data" slots changed, but still have the same strucutre 18:42
fglock my $tree = MyTree.from_match( $/ )
ruoso or have all the "dat" slot unchanged, but with a complete different structure
that seems sane... 18:43
Considering the match is turned into a tree
we need a language to specify the transformations
like a Grammar is to text 18:44
fglock $/.xml 18:45
ruoso fglock, $/.dom would be better, actually
fglock # xslt
ruoso $/.xml_dom 18:46
actually, thinking about xml as a way to represent a match is very much straight-forward 18:47
18:47 franck___ left, franck___ joined
fglock PCR::Runtime::Match::XMLDOM ? 18:48
ruoso specially in terms of debugging the grammar
analysing a XML file is easier than the output of Data::Dumper ;)
fglock how about the compiler environment # implemented as closures 18:49
ruoso that should be part of the transformation language 18:50
and stored inside the tree as appropriate
fglock for example - a BEGIN block that modifies the grammar 18:51
in a specific scope
ruoso hmmm...
ruoso in crazy mode...
so the entire Perl 6 grammar is part of the compiler env, rihgt? 18:52
fglock yes
ruoso this is actually a sad thing, because this way we can't have multi-stage compilers...
fglock BEGIN only executes at the first stage 18:53
ruoso but it runs arbitrary code that can change the compiler environment
so, the first stage must have available all the other stages 18:54
which, in the end turns everything into a single stage again
18:54 stevan_ left
fglock kp6 BEGIN finishes with a tree transformation 18:56
which copies the side-effects back to the AST under construction
what if BEGIN returned xslt ?
ruoso we would need to know exactly what happens inside BEGIN... which we don't 18:57
the idea of a simple copy
seems simpler
fglock hmm - after BEGIN finishes, we can get to the side-effects 18:58
such as, a global was modified 18:59
ruoso we copy the env to the new compiler instance that will handle BEGIN, and then copy it back to the tree representing that this new tree is available after that statement
19:00 stevan_ joined 19:02 bcmb_ joined 19:03 franck____ joined 19:05 masak joined 19:06 jettero left, jettero joined 19:07 weinig_ joined, weinig left 19:11 franck___ left 19:12 hakobe is now known as hkb_zz 19:14 weinig joined, Blwood left, weinig_ left
fglock [from the offline discussion] compile-time pads must be serializable, so that we can represent them in the tree 19:15
19:15 bcmb left
ruoso TimToady, ping 19:19
one of the big things of the BEGIN in Perl 6 is that it can change the grammar, right?
and this change will affect the compilation from that point on 19:20
moritz_ aye
ruoso how exactly is the syntax for that?
and how is the compiler supposed to handle a parse tree with a different structure?
moritz_ I don't know... one way to change the syntax is to define new/override old operators 19:21
the other thing is you can just define a token the STD namespace, and per mmd it will be used instead of the original one 19:22
19:22 JarJarBinks left
ruoso would that be a substitution rule that generates parseable code? 19:22
moritz_ at least that's how I think it works
19:23 ilogger2 joined
ruoso moritz_, actually, that's what macro are 19:24
TimToady a macro is just a convenient way for a user to write a new grammar rule
lambdabot TimToady: You have 1 new message. '/msg lambdabot @messages' to read it.
ruoso TimToady, yes... and that's not what I'm talking about... I'm talking about that COMPILING::<$?PARSER> thing... 19:25
TimToady there is some speculation in the comments at the top of STD about that, but obviously they are not entirely fleshed out yet. 19:26
ruoso reading that, and getting no clue about how it should work ;)
TimToady certainly you are correct that it doesn't make much sense to swap in grammar rules without also adding corresponding semantic rules
unless you're just testing for syntactic correctness... 19:27
but basically you just never return something from a grammar reduction that you don't also have a way to handle downstream
ruoso but the problem is that the code accepted as correct by a new rule (not counting macros) will be completely unknown to the compiler in the second stage
TimToady there is no "the compiler"
rindolf Hi TimToady 19:28
TimToady Perl 6 isn't a single language
ruoso s/the/one/
moritz_ ruoso: no necessarily.. if you just overwrite a rule, and insert the same {*} tokens, the semantics might not see the changes at all
s/no\b/not/
ruoso moritz_, and that's a mine field ;)
moritz_ ruoso: aye. most fiddling with internals is ;) 19:29
TimToady {*} is just a kludge currently to avoid speccing the AST transforms in STD.pm
ruoso TimToady, we stop then in the need for a language for the second compiling stage, where the parse tree is transformed in a semantic tree
TimToady the important point about {*} is that it is bound to a semantic identity 19:30
ruoso which is where this discussion started, btw...
heh
TimToady if you change the semantics of {*}, then you have to give it a different identity
moritz_ TimToady: right. Which makes it possbile to attach documentation instead of a compiler to the grammar ;)
TimToady p6 is all about maintaining identity of the "current language"
moritz_ TimToady: that's what I want to do for perlhints
TimToady yeah, though I think they are kinda tending to clutter STD in the same way that explicit reductions would 19:31
maybe we can get those indirect like the reductions
ruoso the problem is that getting to the AST is a bigger problem than it looks at first sight 19:32
even if you can parse correctly the code
parsing is now (thanks to Perl 6 rules) the easiest step
but transforming that into something actually usefull is another story
TimToady well, of course, minimally you can just interpret the AST like pugs. everything else is optimization. :) 19:34
which, of course, explains pugs blazing speed ;) 19:35
*pugs's
*not* interpreting the AST directly is one way that p5 got faster than p4 19:36
but I completely agree that it's a fair amount of work, or maybe an unfair amount of work 19:37
ruoso yep... the experience from this "tea" compiler shows that...
I'm quite sure that we need a "tree grammar language" to make this job acceptable 19:38
TimToady we were working on something like that earlier in the design of p6 19:39
mostly based on binding tree nodes to signatures
ruoso votes for getting to work on that again ;)
heh
TimToady with a sufficiently restricted subset of returns you can even make it reversible 19:40
[particle] timtoady: pmichaud and i discussed a possible #{*} for a hook that didn't affect longest-token analysis
ruoso feels the smell of functionall programming again...
[particle] ruoso: see txl for a tree transformation language i happen to like 19:41
ruoso [particle], url?
TimToady after having spent months teaching the p5 compiler not to forget things just so I could do p5-to-p6, I am a little more pleased with the idea of reversibility than I was 15 years ago...
[particle] www.txl.ca/ 19:42
lambdabot Title: TXL Home Page
TimToady [particle]: well, by definition a comment wouldn't affect anything, but it also doesn't give you anything that actually gets called.
19:42 sunnavy joined
TimToady a <?{*}> assertion doesn't affect longest token and still gets called (eventually) 19:43
[particle] TimToady: yeah, i know. it would mean smart comments
TimToady I don't like smart comments, actually
[particle] what about reserving <*>
TimToady the #= stuff notwithstanding
I think we even mentioned <*> a couple of weeks ago. It's certainly available.
[particle] in general, i agree, but the door opened with #= :) 19:44
TimToady #= was aimed at enabling a preprocessor kludge
I wasn't expecting pm to turn it into real smart comments. :)
diakopter TimToady: it's in its pugs form, but check this out: rafb.net/p/JtCYyZ42.html 19:45
lambdabot Title: Nopaste - foobar
[particle] you ivory tower designers just don't know a good implementation hack when you spec one :)
diakopter chromatic clarified the question: <@chromatic> I think the question is "Why does binding a code block to a module-global name occur at the point of compilation, not at the point of execution within a surrounding context?"
then, my next question was: Is Perl 6 supposed to change this behavior (and why [not])? 19:46
masak TimToady: do you happen to know what today's best method for getting p6 source -> AST is? 19:48
ruoso masak, actually that's the question of the day ;)
TimToady masak: depends on which subset of p6 you want to run. 19:49
masak ruoso: woot
TimToady: the tests, which would be all of p6, I guess
I aim to do statistics on the tests
ruoso it would be better if we had a p6 AST definition
:)
TimToady I was hoping the implementors would just all get together and agree on an AST, just like they did a grammar, oh wait... 19:50
pasteling "pugsbot_r17185" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26583
19:51 pugsbot_r17185 joined
moritz_ ?eval 1 19:51
pugsbot_r17185 1
masak ?eval $?VERSION
pugsbot_r17185 \"6.2.13"
diakopter yeah, it seems ok now 19:52
moritz_ diakopter++ # being *bot SLAVE ;-)
fglock svn.pugscode.org/pugs/src/Pugs/AST/...ternals.hs inspired kp6 AST, but it's not exactly as-is
diakopter b/c mmd is broken in pugs currently, it was always calling the {...} placeholder subs.
which led to my question on #parrot... and me repeating it here. :)
19:52 thoughtpolice joined
TimToady I'm going to have to spend a lot of time backlogging after lunch... 19:53
19:54 weinig joined 19:55 diotalevi_ joined
masak `pugs -CParse-Pretty` seems to do what I want 19:58
except I would prefer not to parse text at all
can I get the same as p5 or p6 objects?
fglock masak: pugs -CPIL1-Perl5 -e ' 1+1 ' 20:00
masak yay 20:01
just so
20:05 [particle1 joined 20:15 jrockway joined 20:17 szbalint joined 20:22 Lectus joined 20:43 TJCRI joined 20:47 Lectus left 21:21 TJCRI left 21:22 thorat joined 21:30 vel joined
cj TimToady: is it ready for production yet? :) 21:35
and did I miss anything exciting at OSCON this year? 21:36
21:47 cmarcelo joined 21:49 Limbic_Region joined 21:56 dmq joined 21:58 weinig left 22:06 statico joined 22:12 weinig joined 22:13 vel left, xinming_ joined 22:15 vel joined 22:28 vel left, vel_ joined 22:29 cmarcelo left 22:36 weinig_ joined 22:37 weinig left 22:38 Blwood joined
meppl good night 22:47
22:51 polettix joined 23:05 vel_ left 23:24 weinig_ left 23:33 wilx left, wilx` joined
TimToady the irc logger seems to be broken 23:36
<p><a href="/irclog/out.pl?channel=perl6;date=2007-07-29">&larr; Previous
doesn't like the &, apparently
Limbic_Region colabti.de/irclogger/irclogger_log/...-07-31,Tue
lambdabot Title: #perl6 2007-07-31,Tue, tinyurl.com/ynk7kv 23:37
Limbic_Region are they all broke?
perlDreamer irc.pugscode.org seems to work okay
TimToady false alarm; did a yum upgrade and it confused firefox till I restarted 23:39
23:48 weinig joined 23:50 drupek12 joined 23:54 vel joined