Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2006-07-10

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:04 * nothingmuch ponders tagging http://www.flickr.com/photos/clkao/185926028/ as "may offend"
00:04 nothingmuch awesome! http://www.flickr.com/photos/clkao/185928599/
00:13 Khisanth joined perl6
00:14 svnbot6 r11386 | fglock++ | v6 - fixed infinite loop parsing invalid expression
00:16 sanug joined perl6
00:19 TimToady perl5/Sub-Multi/inc seems to be both under svn and clobbered by "perl Makefile.PL".
00:19 TimToady anybody know what's going on there?
00:19 frederico joined perl6
00:19 obra TimToady: if there's a .author directory in inc/ it's blown away at runtime
00:19 obra likely the .author inside shouldn't be versioned
00:19 audreyt fixed.
00:19 audreyt dinner &
00:19 TimToady thanks
00:20 TimToady btw, still blowing up on my smokes after upgrade to Task::Smokes
00:20 svnbot6 r11387 | audreyt++ | * TimToady noted Sub::Multi clobberness.
00:21 TimToady next step is blow the whole thing away and re svn up I guess...
00:24 TimToady actually, now that I look at it, it's pulling in files from 5.8.6!  That can't be right...
00:29 sanug ping pmichaud
00:35 svnbot6 r11388 | fglock++ | v6 - clear other warnings when reporting syntax errors
00:54 TimToady It took me half an hour to upgrade Test::Smoke to 5.8.8.  Gah...we've got to do better on this...
00:55 TimToady my cpanp was stuck in 5.8.6 land for some reason, so I had to learn cpan shell syntax all over again.
00:56 TimToady we need to translate these things to Perl 6 so we can translate them to JavaScript and run 'em under Firefox.  :)
00:59 TimToady we just need a refactoring translator that's smart enough to remove stupid shell prompts and replace them with nice buttons.  :)
01:00 TimToady maybe 6PAN should just be an svn repostory.  :)
01:00 TimToady s/st/sit/
01:01 TreyHarris TimToady: I feel your pain.  I've been fighting for the past several hours because too many dist packagers don't make allowances for the possibility that I might want to install somewhere other than /usr or /usr/local
01:01 TreyHarris (or wherever perl itself is installed)
01:01 scook0 joined perl6
01:07 TreyHarris so where does the most up to date S29 draft live now?
01:09 TreyHarris is it docs/Perl6/Spec/Functions.pod in the repository?
01:11 TimToady *nod*
01:11 ludan joined perl6
01:18 bennymack maybe 6PAN should jusit be an svn repostory.  :)
01:18 bennymack huuh?
01:20 TreyHarris s/sit/st/; s:2nd/st/sit/;
01:33 xdg joined perl6
01:35 xdg FYI: Strawberry Perl Alpha 1 for Win32 is available at http://vanillaperl.com.  It should make experiments with v6.pm even easier than Vanilla Perl.
01:35 lambdabot joined perl6
01:36 xdg Also Perl::Dist::Builder is up on CPAN.  While hacky and poorly documented, it can be used to create custom variations with other bundled modules.
01:40 markstos joined perl6
01:41 lollan joined perl6
01:49 datasieve left perl6
02:24 revision17_ joined perl6
03:07 kanru joined perl6
03:26 revision17_ is now known as Revision17
03:29 kanru joined perl6
03:30 neoesque joined perl6
03:40 vel joined perl6
03:49 ajs_home TreyHarris: latest S29 is now docs/Perl6/Spec/Functions.pod in pugs, yes.
03:54 scook0 joined perl6
04:14 tup joined perl6
04:24 TreyHarris ajs_home: thanks.  so what were you thinking about for system (run, whatever) as far as its return behavior?  that success will be booleanized true is a given.  but are you thinking it will return an int with a different boolean role, or an object you can call into for exit status, signal, maybe output & error, etc.?
04:25 ajs_home It will almost certainly return a Conc::Proc::Status object, just like wait
04:26 ajs_home Getting that to do the right thing in boolean context should be trivial enough
04:29 TreyHarris a Conc::Proc::Status will numify to the old wait return value, I assume
04:29 ajs_home I specifically wrote wait to try to match both the Perl 5 programmer's expectations and S17, so it should serve as a good model for anything that needs to think about an external process's status.
04:29 ajs_home TreyHarris: Correct
04:30 ajs_home That's stated in the section for wait, but in an unfortunately obscure way. I'll fix that.
04:30 TreyHarris is there a draft S17 somewhere?
04:32 ajs_home There is...
04:32 ajs_home I'm hard pressed to recall where, and I have the bookmark at work
04:32 audreyt docs/Perl6/Spec/Concurrency.pod
04:32 ajs_home heh, oh right :)
04:33 svnbot6 r11389 | ajs++ | Make it clear that wait needs a numifyable return value
04:40 TreyHarris interesting.  what would happen under "Co-Routines" in Concurrency.pod if you wrote: "for all(@list) -> $x { produce $x }" ?
04:41 audreyt evaluation order of junctions and hyperops are undefined.
04:41 audreyt (intentionally)
04:43 TreyHarris audreyt: right, but I thought C<for all(@list)> would ordinarily cause parallel execution.  would the presence of a produce inside suppress that behavior, and just cause the order to no longer be defined?
04:43 audreyt actually, it _can_ cause parallel execution
04:44 audreyt but the semantic is the same as if it's randomized and run separately sequentually.
04:45 ajs_home It might do the same thing in the general sense as "for(@foo) { fork || return }", which is probably not what you wanted.
04:46 ajs_home Best to combine straighforward looping with produce in order to be sure of the results.
04:46 ajs_home A warning might be in order....
04:47 TreyHarris the same issue would present with produce as return
04:47 ajs_home Yes, I was simplifying for demonstration purposes.
04:48 vel joined perl6
04:48 ajs_home The fact that produce sets up a co-routine is beside the point I think... combining it with process creation is probably more volatile than you wanted to deal with in a sane world.
04:49 ajs_home This is all off the top of my head, having only skimmed the co-routine docs.
04:50 TreyHarris yes, but parallelization has always been more the domain of ninjas than pirates.  i still hold out hope that P6 can bring it to pirates too
04:53 dduncan seen obra
04:53 jabbot dduncan: obra was seen 4 hours 33 minutes 20 seconds ago
04:54 dduncan for those of you going to OSCON 2006, for whatever reason, I will now probably be joining you ... following obra's suggestion yesterday, I have just registered for the hallway track, aka exhibit hall only
04:54 dduncan so I look forward to meeting you in person and/or working together at whatever available opportunities
04:56 dduncan the main hurdles I have yet to overcome is to find 1-3 people to share a lodging with, and getting things in order at home so my $job can get along without me for that time
04:56 dduncan if you, or anyone you know is looking for room-mates, please let me know
04:56 dduncan and yes, I saw the terse room-sharing wiki
04:56 dduncan but wanted to ask closer to the community first
04:57 dduncan s/looking for/is willing to accommodate more/
04:57 dduncan Red Lion is preferred as it is near by, but I can take other options
04:57 * clkao smokes -BPerl5
04:58 audreyt ooh!
04:58 clkao test::harness hates everything
04:58 audreyt dduncan++ # prospect of joining the OSCONathon
04:59 dduncan audreyt, I seem to recall you specifically asked if I was going ...
04:59 dduncan I don't suppose you have space in your hotel room that I could share?
04:59 sanug audreyt: Need help with properties in svn - can you help?
04:59 audreyt dduncan: sadly not.
04:59 audreyt sanug: sure
05:00 sanug I have read about them in a book just now
05:00 sanug I would like to setup the html and jpg files in perl6_compilation properly
05:00 sanug Could you review my current properties and explain what I need to do so I will remember in the future
05:01 dduncan as far as dates ... I figured that if I arrived late monday and left early saturday, that should fit all the important stuff in for tues-fri
05:01 sanug Also could you send me you defaults as you set them up in your svn config file
05:01 dduncan bus travel will take about 11-12 hours each way from home
05:01 sanug 2nd you s/you/your/
05:01 dduncan unless I can find a carpool
05:02 dduncan but that's not a concern to carry on here
05:02 sanug audreyt: Also please explain what you do to review the properties and what you look for...
05:03 dduncan anyway, if someone knows of a space to stay, let me know ... meanwhile, I'll ask those few on the roomsharing wiki
05:03 audreyt sanug: I use the "pl" command
05:03 audreyt svk pl *
05:03 audreyt (same as svn I believe)
05:03 audreyt to list props
05:03 audreyt then I see you have both mime-type and svn:mime-type
05:03 sanug Do you like svk better than svn?
05:03 audreyt the mime-type one is bogus
05:04 audreyt well, seeing I'm svk's #2 committer...
05:04 audreyt ...yes :)
05:04 sanug What is best method to acquire svk?
05:04 audreyt svk.elixus.org has links to .pkg
05:04 sanug I use  OS X 10.4
05:04 dduncan as do I
05:05 dduncan in my experience, the way of getting the svk package was a bit indirect ...
05:05 dduncan I'm not sure a current one is linked on the svk homepage ... will look ...
05:05 sanug Assume that clkao herein is Chia-Liang Kao
05:07 svnbot6 r11390 | audreyt++ | * remove the mime-type prop
05:08 dduncan sanug, the OS X packages I have been using of svk were by Michael Brouwer <michael@tlaloc.net>
05:08 dduncan not sure if that specifically was linked on the svk homepage, or whether I found it by Google or an IRC link
05:09 clkao just click download and there's the same macosx package on openfoundry
05:10 clkao why get rid of the mietypes
05:10 clkao ?
05:10 audreyt I'm seeing application/octet-stream for some .jpg files
05:10 audreyt changing it
05:10 audreyt clkao: sanug had it as mime-type and also svn:mime-type
05:10 audreyt so I removed the first variant
05:10 clkao sodes
05:10 audreyt sanug: this is what I'm doing:
05:10 audreyt svk ps svn:mime-type image/jpeg *.jpg
05:10 sanug Please say how incorrect modes were detected and removed
05:11 audreyt I detect it with
05:11 audreyt svk pg svn:mime-type *.jpg
05:11 audreyt (again, should be same with svn)
05:11 sanug Now  the html files are they setup properly
05:12 dduncan clkao, yes, the downloads section seems to have it
05:12 sanug They were not opening properly in browser
05:12 svnbot6 r11391 | audreyt++ | * change jpg files to image/jpeg
05:12 dduncan I just recall going somewhere else once ... maybe it was for a special version for 10.3
05:12 dduncan no matter
05:12 audreyt sanug: they look correct.
05:14 sanug We setup svn when I downloaded pugs - does switch to svk require any addition?
05:19 audreyt yes. it involves
05:19 audreyt mv pugs pugs-old-svn
05:19 audreyt svk co http://svn.openfoundry.org/pugs
05:19 lambdabot Title: "Revision 11391: /"
05:19 audreyt hit <enter> multiple times
05:19 audreyt and then wait a long, long time.
05:20 audreyt preferably do it before a meal, or before sleep
05:20 audreyt but, you can full revision history of Pugs, and so can find any file's content in any timepoint, even without network.
05:21 audreyt well, unless that timepoint is in the future, in which case you need an additional "sleep" command and then "svk pull".
05:21 audreyt after you get the checkout, "svk ci" is the same; instead of "svn up" you say "svk pull"
05:21 audreyt otherwise it's identical. except suddenly it feels much faster.
05:22 sanug So this co does a mirror the first time
05:22 audreyt aye.
05:22 sanug I see that the html files - some of them have extra carriage returns in the properties:
05:23 sanug Properties on 'perl6_compilation/Perl_6_C​ompilation_Overview.html':
05:23 sanug  PHEME.html : text/html; charset=UTF-8
05:23 sanug  svn:mime-type : text/html; charset=UTF-8
05:23 sanug Properties on 'perl6_compilation/Perl_6_​on_Haskell_is_Pugs.html':
05:23 sanug  PHEME.html : text/html; charset=UTF-8
05:23 sanug  svn:mime-type : text/html; charset=UTF-8
05:23 sanug the second file has PHEME.html property even though it is a Perl_6_on_Haskell_is_Pugs.html'
05:23 sanug Perhaps this is in just my version?
05:24 ivanfrey Why is ranlib not working any more.
05:24 sanug There also seems to be 2 carriage returns after that propert and the correct svn:mime-type
05:25 audreyt that is svn bug
05:25 audreyt I mean, feature.
05:25 audreyt svk doesn't display 2 carriage returns
05:25 sanug Do you see PHEME.html as property?
05:25 audreyt yes and I've just pd'ed it
05:26 audreyt svk pd PHEME.html *html
05:26 sanug Good all gone...
05:26 svnbot6 r11392 | audreyt++ | * remove PHEME.html property
05:27 sanug Language_AddressBook.html STILL opens up the text file though - I just checked out the recent update in r11392
05:27 ajs_home Because of a very long discussion in late 2005 around Time, I'm going to mostly punt time in S29. But if anyone cares, I'm moving time down one namespace hop from Time::Local to Time, and pushing off any documentationn on the use of a Time object to an external document.
05:28 audreyt ajs_home: oh, btw.
05:28 ajs_home yes?
05:28 audreyt ajs_home: I wonder if it actually makes some sense to list only the core (as in, have no qualified form) functions in Functions.pod
05:28 audreyt and move everything else to docs/Perl6/API/*
05:28 sanug audreyt: PHEME_Parrot_based_Scheme.html opens up correctly as an image-map file.
05:28 audreyt as in Math.pod
05:28 audreyt sanug: good
05:29 audreyt ajs_home: may make it more manageable in the long run
05:29 audreyt esp. with the new P6AST, there's a lot of API on core objects that we need to list
05:29 audreyt and Functions.pod is already too long
05:29 audreyt (see also the doc layout for builtin prelude of all other OO languages)
05:30 audreyt and the Functions.pod can just say
05:30 ajs_home I would tend to agree... however it sort of breaks the stride of the original purpose of sectioning the Ap/Sy documents the way Larry did, doesn't it?
05:30 audreyt use Math :EXPORT;
05:31 audreyt which imports from Math and then reexport
05:31 audreyt ajs_home: I think it ceased to make sense a while ago.
05:31 nothingmuch http://en.wikipedia.org/wiki/Ingy_döt_Net
05:31 nothingmuch =)
05:31 ajs_home Plus EVERYTHING is prefixed in S29... we  haven't yet defined which things are auto-use/exported.
05:31 audreyt (which is why in copying it to docs/perl6/ I removed the numbering)
05:32 audreyt ajs_home: esp. for the API, layout it out in Perl6/API/ I think provides more room for edits
05:32 audreyt as in, hierarchical
05:32 ajs_home Ok, let me continue as I am for a bit, and then at some point, I'll take on the brain surgurey that that suggests.
05:35 audreyt okay!
05:35 audreyt ajs_home++
05:37 sanug audreyt: I moved directories around as you suggested and started svk, can I still do svn pe on moved directory and update Language_AddressBook.html so it has property
05:37 svnbot6 r11393 | ajs++ | Cleaned up word and added docs for time/localtime/gmtime -- in 2005 there was a big discussion about time that I want to avoid for now....
05:37 sanug svn pe svn:mime-type Language_AddressBook.html
05:39 sanug audreyt: Or should it be svn ps svn:mime-type 'text/html; charset=UTF-8'  Language_AddressBook.html
05:40 svnbot6 r11394 | ajs++ | Corrected sprintf signature
05:40 audreyt sure, you can
05:44 stevan joined perl6
05:46 svnbot6 r11395 | sanug++ | Updated svn:mime-type properties using this command so that each html file would open up in a browser instead of showing the image map to the user:  svn ps svn:mime-type 'text/html; charset=UTF-8' *.html
05:49 buetow joined perl6
05:50 sanug audreyt: HTML files now all properly set up. Thanks for lesson in svk, svn - switching to svk.
05:50 audreyt yay. :)
05:51 audreyt I need to sleep now. bbiab tomorrow :)
05:51 obra audreyt: I'm hoping to leave the house between 9 and 10 tomorrow
05:51 obra so that yuval gets f00d before he has to fl33
05:56 ajs_home must go sleep as well. night all
05:57 audreyt obra: ok
05:58 nothingmuch night guys, i'm also off to bed
06:04 clkao audreyt: should fix -c with v6. the first time (compiling) it doesn't exec, just check. the second time it does run the program
06:05 audreyt clkao: okay, sounds like a great idea! :)
06:05 audreyt (it's just the body of v6.pm)
06:05 audreyt sorry, really gotta sleep.
07:11 Aankhen`` joined perl6
07:14 iblechbot joined perl6
07:19 drrho joined perl6
07:36 elmex joined perl6
08:12 dduncan left perl6
08:52 gaal joined perl6
08:52 gaal greetings from Tel-Aviv!
09:04 miyagawa gaal: hi
09:06 gaal hey :)
09:08 revision17_ joined perl6
09:10 clkao miyagawa!
09:11 cmatei joined perl6
09:19 gaal clkao: isn't it like, very early where you are?
09:20 miyagawa clkao: still in boston?
09:22 clkao miyagawa: in london now
09:22 clkao gaal: it's already 10am
09:22 miyagawa great. so everyone's back home now
09:24 rw-rw-r-- joined perl6
09:29 * clkao is kind of homeless
09:34 xinming joined perl6
09:35 nnunley clkao: Awww.
09:36 nnunley miyagawa, gaal: Hey.
09:37 miyagawa nnunley: I just blogged about Algorithm::RabinKarp (in japanese) http://blog.bulknews.net/mt/archives/002057.html
09:37 lambdabot Title: "Rabin Karp アルゴリズムでコード重複の検出: blog.bulknews.net"
09:39 nnunley Typo in the module name in that post.
09:40 miyagawa fixing
09:40 xerox Japanese...
09:41 miyagawa xerox: babelfish it :)
09:41 miyagawa nnunley: hm, where's the typo?
09:42 nnunley Last line, cpan Algorithm::RabinKarpe
09:42 miyagawa mk
09:42 miyagawa fixed.
09:43 gaal nnunley: heya! I didn't get a chance to use the data gathered yet. but now that the hackathon's over maybe it's time :)
09:43 * nnunley grins.
09:43 nnunley Golf time.
09:43 paultcochrane joined perl6
09:46 nnunley Audrey: I've extracted possible prerequisite module containing all the calls to the P(ugs|erl6)::Internals namespace. Had two issues: 1. There's two namespaces for Internals. Should there be? 2. there's an *::Internals::eval ... all the other evals have a suffix of the evaluator to be used.  Is this intentional?
09:46 nnunley gaal: I still have to improve the dup collation code.
09:48 gaal nnunley: cool, I'll be happy to try it out when you do
09:49 gaal what's in Perl6::Internals? Can't see anything that uses that namespace
09:49 Aankh|Clone joined perl6
09:50 gaal As for the evals, maybe I'm not understanding the question, but the idea is that eval is a fontend because in usermode you say "eval 'str' :lang<foo>"
09:53 nnunley Pugs::Internals::eval is a direct hook to the Haskell runtime.
09:53 gaal ahh, you're asking should that be renamed to eval_perl6?
09:53 nnunley I'm assuming that everything in the Internals is primative, and is supposed to be supplied by the perl6 runtime.
09:53 nnunley Right.
09:54 nnunley For consistency.
09:54 gaal probably it should. fixing.
09:54 cognominal joined perl6
09:54 nnunley gaal: Look at src/perl6/INTERNALS.txt
09:55 nnunley That's the map of all functions called from the Prelude in the Internals namespace.
09:57 gaal yes, fixing that too :)
10:04 nnunley Cool.
10:05 svnbot6 r11396 | gaal++ | * Rename Pugs::Internals::eval to Pugs::Internals::eval_perl6
10:05 svnbot6 r11396 | gaal++ |   for consistency, as advised by nnunley++
10:06 gaal Looking again I wasn't sure about the Perl6:: issue, looks wrong to me but I'm not familiar with that code.
10:07 stephanepayrard_ joined perl6
10:07 gaal bbiab &
10:07 nnunley k
10:08 kanru joined perl6
10:09 stephanepayrard_ joined perl6
10:27 Aankh|Clone is now known as Aankhen``
10:40 szbalint joined perl6
10:47 myavuzselim joined perl6
10:51 myavuzselim joined perl6
10:51 nothingmuch joined perl6
10:56 Jedai joined perl6
11:14 kanru joined perl6
11:17 HXanadu joined perl6
11:19 orafu joined perl6
11:23 myavuzselim joined perl6
11:33 mauke_ joined perl6
11:36 kanru joined perl6
11:40 iblechbot joined perl6
11:48 mauke_ is now known as mauke
11:58 tifo joined perl6
12:03 ludan joined perl6
12:10 bpphillips joined perl6
12:32 Limbic_Region joined perl6
12:33 Limbic_Region audreyt ping
12:35 * Limbic_Region is playing catchup as he was (and still is) under the weather
12:35 Limbic_Region did someone already point out that the p6 summary has links to discussions from 3 years ago
12:35 Limbic_Region ?
12:36 * Limbic_Region knew something was fishy when he saw Dan S. in the summary
12:36 nnunley There's a bit of a backlog. ;)
12:36 Limbic_Region I wasn't sure if it was a mistake, a joke, or a test to see who was paying attention
12:40 nnunley It does look like a mistake.
12:42 nnunley the p6-(internals|porters) stuff is from 2003.  The p6-language summary is from 2006.
12:42 christopher joined perl6
12:42 nnunley Right month, though.
12:43 Limbic_Region rafl ping
12:43 ivanfrey I have forgotten how to run the command ranlib. What options am I supposed to pass it.
12:44 prefiks joined perl6
12:55 marcus_ if you run the parrot test suite backwards, it says 'ONE RUNTIME TO RULE THEM ALL, AND IN THE DARKNESS BIND THEM'
12:58 buetow joined perl6
13:02 fglock joined perl6
13:12 myavuzselim joined perl6
13:13 myavuzselim joined perl6
13:14 chris2 joined perl6
13:20 lollan joined perl6
13:20 xinming joined perl6
13:32 Limbic_Region fglock - thanks for the starter example - I have been under the weather for several days now (and still am) but I should get to it sometime this week
13:42 vel joined perl6
13:46 kanru joined perl6
13:49 fglock Limbic_Region: ok!
13:50 myavuzselim_ joined perl6
13:50 myavuzselim joined perl6
13:51 fglock cofee &
13:56 justatheory joined perl6
14:01 ivanfrey In the INSTALL file this address " http://wiki.kn.vutbr.cz/mj​/?Perl%206%20and%20Parrot" doesn't work
14:07 ajs_ Looks like the site is down... might be transient
14:08 nothingmuch audreyt: wakey wakey
14:08 ivanfrey I get errors about libHSpugs-HsSyck-0.1.a and libHSpugs-fps-0.7.a not having a table of contents. I run ranlib on them, but the make process just recreates the archives.
14:12 weinig joined perl6
14:24 Limbic_Region joined perl6
14:24 gaal nothingmuch: moosey moosey
14:25 clkao nothingmuch: i can't send my smoke output. server saz invalid format
14:26 ajs_ Oh great... -Barry- is trolling the Perl 6 Wikipedia page now... I live in hell.
14:26 ajs_ is now known as ajs
14:26 gaal ajs: did the bit arrived untoggled?
14:27 ajs gaal: I was able to create an account, and check in, if that's what you're asking
14:27 gaal yup. welcome aboard :)
14:27 ajs thanks. Felt good to be able to check in where others would actually read it ;)
14:27 myavuzselim_ joined perl6
14:28 gaal ah, you already checked in? haven
14:28 gaal 't blacklogged yet.
14:29 ajs Yeah, I checked in my batch of updates, and then did a few small things last night
14:29 ajs Heh, just panicked when I updated my at-work copy of the tree and saw all of the examples being deleted ;)
14:30 coderhasan joined perl6
14:31 gaal heh, something similar happened to me when I ran 'svk sw' on a wifiless train and watched in horror as CapInternals.hs was D'ed. Happily another sw revived it :)
14:31 svnbot6 r11397 | fglock++ | Pugs-Compiler-Rule 0.10 - added Pugs::Runtime::Common.pm
14:31 svnbot6 r11397 | fglock++ | - testing on the svn dir works, but it fails all tests in a clean installation
14:31 svnbot6 r11397 | fglock++ | (it looks like a Makefile.PL problem)
14:32 coderhasan joined perl6
14:36 ajs Does MMD dispatch on named parameters? If I have "multi foo(Int $x){...}" and "multi foo(Int $y)", is "foo :$y" going to do what I expect?
14:37 coderhasan left perl6
14:37 myavuzselim__ joined perl6
14:43 fglock gaal: where is the current definition of the AST?
14:43 svnbot6 r11398 | fglock++ | v6 - emitter uses Pugs::Runtime::Common
14:45 clkao fglock: do we want the smoke output
14:46 clkao i can't upload it for some reason. and i have a full output of the failed compile etc
14:48 fglock clkao: it would be good to have it for comparison with later versions
14:48 clkao not sure where to put them though.
14:48 clkao mm feather
14:50 clkao i had to kill two tests i remember. one is the operator/listquote
14:51 ludan joined perl6
14:51 clkao that seems also infinite loop when compiling
14:53 TimToady clkao: I had to upgrade to latest Task::Smoke to get it to upload my smokes.  It was pulling something from 5.8.6 sitelib instead of 5.8.8...
14:53 TimToady same error as you're getting, so probably same problem.
14:53 fglock clkao: ok - I'm looking operators/listquote.t
14:54 clkao TimToady: ah.
14:54 TimToady ajs: that is specced to not be required in 6.0.0.
14:55 clkao it'd be nice if the smoke system also captures stderr
14:55 ajs Ok, I'll try to not rely on it in Functions
14:55 clkao but see http://feather.perl6.nl/~clkao/ for the smoke output and the result
14:55 lambdabot Title: "Index of /~clkao"
14:55 clkao the output is more interesting because you can find things not compiled yet.
14:55 TimToady what we do require in 6.0.0 is that if you have proto ($x) then both variants must use $x (which would be a collision).
14:56 TimToady so all the infix builtins will have proto of ($x,$y), for example
14:56 TimToady at least the numeric ones...
14:57 ajs Right
14:58 TimToady now if you have multi (; :$x) vs multi (; :$y), then it probably can differentiate them, but on the basis of weeding out completely unacceptable candidates in the first pass, not on the basis of trying to bind to a positional arg.
14:59 TimToady to MMD those look like multi () with extra constraints
14:59 r0nny joined perl6
14:59 ajs Oh, that might be useful for run() then...
14:59 clkao fglock: also, it seems 08-test now requires 400M ram to finish
15:00 TimToady one other factoid is that proto not only limits what you can use for the name, it forces rewrite of the call.
15:00 r0nny joined perl6
15:00 TimToady so if you have proto($x,$y), then foo(:$y,$:x) gets rewritten to positional.
15:00 TimToady (just for any recognized args)
15:01 TimToady in other words, proto allows us to recognize some subset of named args as positionals at compile time.
15:02 ajs Hmm... ok, lots to digest.
15:02 TimToady (but only within the scope of the proto declaration, so different callers can have different protos)
15:02 ivanfrey Does pugs require Perl 5.8.8 or is 5.8.6 good enough?
15:03 snowstalker joined perl6
15:03 ivanfrey In light of TimToady's comment about Task::Smoke.
15:03 TimToady it's possible that the smoke mechanism now requires 5.8.8, though likely it's just a change to the smoker modules.
15:04 TimToady I did not attempt to install new Task::Smoke on 5.8.6.
15:04 TimToady since I try to run my machines on the latest and greatest.
15:04 snowstalker left perl6
15:04 TimToady aka bleeding edge
15:05 TimToady but we need people who will do the other thing too...
15:06 fglock 35% !!!
15:07 clkao fglock: well, i'd say 2%
15:07 snowstalker joined perl6
15:07 clkao because many tests aren't compiled at all, so the total test count is wrong
15:07 clkao so look at v6-smoke.output and fix them ;)
15:14 SamB joined perl6
15:16 fglock is postfix:{'<','>'} disallowed after circumfix:{'(',')'} ? I'm trying to parse '(1 | 3)<3'
15:16 hexmode joined perl6
15:18 TimToady that's a syntax error.
15:18 TimToady you must have a space before infix:{'<'}
15:19 TimToady otherwise it will always be taken as the postfix.
15:19 fglock is t/operators/listquote.t wrong?
15:20 elmex joined perl6
15:21 TimToady yes, it's wrong.
15:21 TimToady that should parsefail.
15:22 fglock ok - thanks!
15:23 merlyn joined perl6
15:23 nothingmuch gaal: moosy
15:23 nothingmuch clkao: i don't know iblech's code very well
15:23 nothingmuch it wants an HTML report
15:23 nothingmuch i don't know how it could be invalid
15:24 nothingmuch except for if compression is involved
15:26 fglock bbiab &
15:28 svnbot6 r11399 | fglock++ | v6 - Expression.pm - added comments confirming '<' parsing behaviour
15:28 svnbot6 r11399 | fglock++ | - note: a parsing error on '<' currently causes an infinite loop
15:29 gaal nothingmuch: why does the lopsided picture of you possibly be tagged as "may offend"?
15:30 nothingmuch because it's very very ugly ;-)
15:30 gaal haha
15:30 * nothingmuch odes *not* photograph well
15:30 clkao you know Nosing!
15:30 gaal fglock: src/Pugs/AST/CapInternals.hs - but not finished yet :)
15:30 nothingmuch nossing!
15:31 gaal look for 'data Exp', 'data Stmt', etc, but please note that Exp is no longer necessarily the root of an AST
15:31 gaal fglock: though a compilation unit, it looks like, desugars to something like:
15:32 gaal COMPUNIT := module MMMMooo { our $se; sub elk {...} };
15:32 gaal so at present we have data CompUnit's ast field be an Exp representing that binding.
15:33 amnesiac joined perl6
15:35 Limbic_Region can someone nopaste me their third-party/installed/packages.conf file?
15:42 svnbot6 r11400 | ajs++ | Took a stab at run and runinstead as proposed by Larry. Suggested :bg mode as well, but that might not make sense
15:55 mj41_ joined perl6
15:56 hexmode joined perl6
15:58 gaal wow, that was longer than expected
15:59 gaal (p6-c post on testing)
16:07 ryanus joined perl6
16:08 ruoso joined perl6
16:10 nothingmuch ee/w 26
16:12 svnbot6 r11401 | lwall++ | more tests for < postfix vs list policy
16:13 ivanfrey what's the pugs mailing list? I'm trying to subscribe to the perl6-compiler list and I'm having trouble there.
16:16 gaal ivanfrey: there's no official pugs list, but p6-c is the closest, I suppose. what trouble are you seeing there?
16:20 fglock gaal: thanks
16:22 fglock ?eval ( {1=>2,}|{3=>4,} )<3>
16:23 TimToady you killed it!
16:23 ivanfrey gaal: maybe the mailto subscribe link is wrong I'm getting back an undeliverable mail.
16:24 gaal ivanfrey: perl6-compiler-subscribe@perl.org I believe.
16:24 TimToady in the larrybot it gives undef.  should probably be undef|4, methinks
16:25 ryanus left perl6
16:25 fglock larrybot: thanks :)
16:26 gaal undef? junction? why?
16:26 svnbot6 r11402 | ajs++ | Fleshed out OS section with chroot, getlogin and kill -- kill suggests an OS::Signal class with string or integer initializers
16:30 fglock gaal: it's an operation on a junction of pairs; it returns a junction of values
16:30 TimToady subscripting should carry through junctions just like any other normal operator.  It's only boolean context that forces resolution
16:31 gaal ooh, sorry, I didn't see the pipe in the expression.
16:31 TimToady "pipe", he sez.  mutter, mutter...
16:31 svnbot6 r11403 | ajs++ | Corrected signature of method style kill
16:32 gaal I didn't see the elognated quantum particle either...
16:33 buu TimToady: Mind a quick question? Do you prefer lexical variable filehandles, eg $filehandle, or FILEHANDLES?
16:34 fglock gaal: it's not vertically elongated - it's horizontally compressed
16:34 gaal a compressed quantum particle? jeepers!
16:34 TimToady given that Perl 6 doesn't have the latter, I think you can guess my opinion.
16:34 svnbot6 r11404 | ajs++ | Concurrency.pod suggests a zero-argument method variant of kill. Document that here.
16:36 theorbtwo Are we considering a method with no noninvocant arguments to have zero-arguments (and arity zero) now?
16:36 yango joined perl6
16:37 theorbtwo ?eval (method bar () {42}).arity;
16:37 ajs Don't read my checkin comment as anything so sweeping. I may have used the wrong terminology.
16:38 ajs But, I would think that that would be the case.
16:40 nothingmuch see you at stevan's house
16:40 theorbtwo OK.
16:40 * stevan looks around for nothingmuch, but cant seem to find him
16:41 * kolibrie looks around for stevan's house, but can't seem to see it
16:41 stevan kolibrie: its behind all those buildings,.. to your left
16:42 kolibrie stevan: oh, right, okay.
16:44 kolibrie s/right/left/
16:45 PerlJam good morroow perl6 people
16:46 kolibrie good morrow, PerlJam
16:50 weinig is now known as weinig|away
16:53 stephanepayrard_ where is the repository for the synopsis?
16:54 svnbot6 r11405 | fglock++ | v6 - implemented ' enum day <Sun Mon Tue Wed Thu Fri Sat>; '
16:55 gaal stephanepayrard_: http://svn.perl.org/perl6/doc/trunk
16:55 lambdabot Title: "Revision 10067: /doc/trunk"
16:55 gaal but RO
16:55 stephanepayrard_ thx
16:55 gaal wow, fglock++!
17:14 vel joined perl6
17:24 svnbot6 r11406 | fglock++ | v6 - enum importing: ' enum bool <False True>; say True; say bool::False; '
17:26 fglock t/oo/enums.t - "Segmentation fault" :(
17:30 kudra joined perl6
17:31 kudra hi
17:31 fglock hi!
17:32 Limbic_Region kudra - if you didn't notice, audreyt didn't set a very good example for the p6 summary
17:32 Limbic_Region some of the lists that were summarized were from 3 years ago
17:32 kudra That is my fault.
17:33 Limbic_Region oh
17:33 kudra I am not sure precisely how they ended up among the feb 2006 in a chronological list
17:33 kudra but next time I will be sure to check the date of every post.
17:34 kudra audrey can, at most, be blamed for not catching my mistakes ;)
17:34 rgs good job anyway !
17:34 Limbic_Region yes
17:34 Limbic_Region kind of hard to complain about volunteer work
17:34 rgs for people like me that don't have time to read p6-all
17:34 Limbic_Region especially from someone who already does soooo much
17:35 * Limbic_Region was only teasing anyway
17:35 kudra well, I prefer suggestions to complaints, but that didn't sound like a complaint anyway.
17:35 kudra but I was already informed of my error by email
17:35 kudra I just thought I'd stop by to see if there were other suggestions.
17:36 rgs I have one
17:36 rgs be sure to include keywords
17:36 rgs so summaries are greppable
17:37 rgs for example dont say "S29", but "S29 about this and that"
17:38 kudra Okay.
17:38 kudra I made an effort to come up with descriptive titles early on, but I think that didn't go so well toward the end.
17:38 Limbic_Region did I read that the summary will also cover discussions here and/or on #parrot?
17:39 rgs this will even help you in the end, to track what was written
17:39 rgs (speaking as a former summarizer)
17:39 kudra rgs: My biggest problem is that I don't know anything about what I'm writing about ;)
17:39 kudra Limbic: after August, yes.
17:39 rgs not a problem, because you'll learn :)
17:39 Limbic_Region gotcha
17:39 kudra I was not warned of this occupational hazard.
17:40 rgs I began to summarize P5P without knowing anything about the P5 internals, and look at me, I'm pumpking now.
17:40 TimToady that's okay, it's also perfectly fine to say "this degenerated into a bunch of technobabble I couldn't be bothered to follow." :)
17:40 PerlJam rgs: pumpkingship is your pennance ?  :)
17:40 kudra Limbic: right now I just can't devote any more time to it.
17:40 TimToady just pace yourself.
17:41 TimToady don't let other people put requirements on you that don't fit.
17:41 Limbic_Region kudra - no worries, I am very appreciative of your initiave
17:41 * Limbic_Region can't speel
17:41 * fglock started writing a compiler before knowing the syntax
17:41 kudra one a week seems reasonable.  Now I know how long it will take me, so I will try to work on it on my commute instead of saving it for Saturday/Sunday.
17:41 TimToady hey, I started designing a language without knowing the syntax.  :)
17:42 TimToady still don't, in fact...
17:42 kudra it's hard to know syntax that doesn't exist yet.
17:42 PerlJam kudra++ in any case for summarizing :)
17:42 TimToady in fact, I'm thinking about making listop<foo> illegal.
17:42 SamB that reminds me of english, except for the design part...
17:43 TimToady basically, if you want to put a postfix right after a listop, you're required to say listop()<foo>
17:43 kudra well, thanks for the suggestions.
17:43 TimToady otherwise we have the counterintuitive result that listop <foo> binds tighter than listop<foo>
17:45 TimToady long term commitments are hard, so just try to look for things that are burnouty and avoid 'em occasionally.
17:45 TimToady (speaking as the one volunteer in the community who is never, ever allowed to burn out. :)
17:47 kudra I think with what I'm working on it will be busy until September, then settle down to a reasonable rate.
17:48 kolibrie kudra: you did a great job
17:48 PerlJam TimToady: People are always talking about -Ofun here and I don't think I've seen anyone ask you if you're having fun, so ...   Are you having fun?
17:48 kolibrie that is most important
17:48 TimToady from time to time. :)
17:49 TimToady though I still hate writing speeches...
17:49 svnbot6 r11407 | fglock++ | v6 - enums.t passes 2/58
17:49 Limbic_Region oh, can someone nopaste me their third-party/installed/packages.conf file?
17:49 Limbic_Region s/someone/anyone/
17:50 PerlJam TimToady: Get Damian or Audrey to ghostwrite them for you ;-)
17:50 kolibrie Limbic_Region: audreyt just told me to delete mine, when I had a build error that mentioned it
17:50 Limbic_Region kolibrie - what platform are you on?
17:50 kolibrie but maybe that was specific to my build
17:51 kolibrie Limbic_Region: debian
17:51 Limbic_Region ok - anyone running ghc 6.2.4 on Win32 that works?
17:51 * Limbic_Region blames rafl
17:51 TimToady Unfortunately (or rather, fortunately) neither Damian nor Audrey thinks like me...
17:52 Limbic_Region the only way I can get it to build is by copying the gcc-lib from my ghc directory to the pugs directory
17:52 Limbic_Region this has to do with a change rafl made to the build process to use utils/runcompiler instead of ghc
17:52 PerlJam TimToady: Ah, then the meta problem is to find (or create) someone that thinks like you.  How much do your children think like you?  :)
17:53 TimToady they is all crazy different from me.
17:53 PerlJam TimToady: Also, I've noticed TimToady-ish tendencies in pmichaud on occassion.  But that's probably only superficial as seen through his/your writings.
17:54 TimToady well, $.commute(), alas.
17:55 TimToady paying for a family is also not always fun...
17:55 TimToady $job &
17:56 Limbic_Region ok - so does anyone on Win32 (regardless of ghc version) have pugs compiling?
17:57 kudra bye everyone.
17:57 kolibrie kudra: bye
17:57 mj41_ r11097 - Successfully built!
17:57 mj41_ r11211 - Could not find module `Data.ByteString':.... Build failed: 256 at util\build_pugs.pl line 126.
17:58 Limbic_Region mj41 - ok, so you are having the same problem I am
17:58 Limbic_Region there is a work-around
17:59 Limbic_Region ugly hack of copying the gcc-lib dir from your ghc root to the pugs root dir
17:59 Limbic_Region but I want to figure out why that is necessary - but can't seem to ping rafl
17:59 Limbic_Region who audreyt says is responsible for the change
18:00 mj41_ r11211 to r11392 - 70 attempts - still same problem :-(
18:01 Limbic_Region yeah, it is because build_util.pl was changed to use util/runcompiler.exe instead of ghc.exe
18:02 buubot joined perl6
18:03 * Limbic_Region is wondering if setting a env var might fix it
18:03 svnbot6 r11408 | fglock++ | v6 - parser: added 'cmp', 'does', 'but'
18:07 buubot joined perl6
18:09 Limbic_Region mj41_ I think I am getting somewehre
18:09 Limbic_Region somewhere even
18:11 buubot joined perl6
18:13 buubot joined perl6
18:19 iblechbot joined perl6
18:20 Limbic_Region nope - dead end
18:27 lollan joined perl6
18:30 Limbic_Region mj41_ does your nmake realclean also fail?
18:30 svnbot6 r11409 | fglock++ | v6 - added prefix:<->
18:30 svnbot6 r11409 | fglock++ | - added file Pugs/Runtime/Perl6Prelude.pm
18:30 svnbot6 r11409 | fglock++ | -- written in perl 6
18:30 svnbot6 r11409 | fglock++ | -- implements enum bool <False True>; true()
18:30 svnbot6 r11410 | rootmj++ | * PaP6 ( Parrot and Perl 6 ) testing
18:30 svnbot6 r11410 | rootmj++ | ** "svn up *-src dir (svn co first time), recursive copy to *-temp, configure, make, make smoke" test loop
18:30 svnbot6 r11410 | rootmj++ | ** redirect output to *-results
18:30 svnbot6 r11410 | rootmj++ | ** plus win32 watchdog
18:30 svnbot6 r11410 | rootmj++ | ** no warranty and no free time to work on it now
18:30 Limbic_Region wow
18:30 Limbic_Region mj41_ does your nmake realclean also fail
18:31 Limbic_Region and if yes, perhaps the reason why nmake is failing is because some file that needs to be deleted isn't
18:31 * Limbic_Region is grasping at straws
18:31 Jedai My Win32 pugs is compiling
18:32 Jedai r11332
18:32 Limbic_Region Jedai - what version of GHC?
18:33 Jedai ghc 6.4.2
18:33 mj41_ i am compiling with PaP6 ... svn up without any modifications
18:33 Limbic_Region ok - and you don't have a gcc-lib dir in your pugs root directory right?
18:34 Jedai no
18:34 Limbic_Region Jedai?
18:34 Jedai I don't have gcc-lib
18:34 Jedai I use mingw and nmake
18:34 Limbic_Region ok - so what's different
18:34 Limbic_Region how exactly do you use mingw?
18:35 Jedai I don't know if there is anything in pugs that use mingw though
18:35 Limbic_Region oh, so you aren't specially using mingw in the pugs build process - it's just there
18:35 cognominal joined perl6
18:36 Jedai Yes I said taht thinking of Parrot (which use mingw)
18:36 Jedai that
18:36 Jedai My compilation use runcompiler.exe and not ghc as you said
18:36 svnbot6 r11411 | ajs++ | Added getpw. OS::PW and PW::PWEnt are tentative names.
18:36 svnbot6 r11412 | fglock++ | v6 - added 'use Perl6Prelude' to compiled code
18:37 Jedai I don't know what's the difference... o_O
18:37 Limbic_Region would you mind doing a 'set' from the command line and nopasting me your env variables
18:37 Limbic_Region removing of course any that aren't appropriate to share
18:38 pasteling "Jedai" at 86.205.173.169 pasted "env win32" (36 lines, 1.4K) at http://sial.org/pbot/18296
18:40 Limbic_Region ok - now can you nopaste me the results of Perl Makefile.PL
18:40 Limbic_Region er s/Perl/perl/
18:41 b_jonas joined perl6
18:42 pasteling "Jedai" at 86.205.173.169 pasted "perl Makefile.PL" (88 lines, 3.3K) at http://sial.org/pbot/18297
18:42 * Limbic_Region does a little dance as he thinks he figured out the problem
18:43 Jedai What was it ?
18:44 Limbic_Region don't know for sure - but I got the behavior to change when I switched around %PATH
18:44 svnbot6 r11413 | fglock++ | v6 - fixed default package name
18:44 Limbic_Region don't yet know if it actually fixes anything
18:45 Jedai I'm using ActivePerl build 817
18:45 mj41_ ActiveState perl? and nmake from cmd.exe (no from msys)?
18:46 Jedai from Visual Studio Express (free)
18:47 Limbic_Region hrm - while "different", not fixed
18:47 Jedai it's the latest version, 8.00.50727.42 (I copy it from VS to my Path because having VS on the path disturb ActivePerl Config)
18:47 * Limbic_Region continues to thrash around hoping something works
18:51 r0nny joined perl6
18:54 Limbic_Region ok, I don't know how, but I fixed it
18:54 Limbic_Region which is the worst kind of fix
18:55 Jedai That's always better than no fix at all ??
18:55 Limbic_Region oh, no, it is still b0rk
18:55 Limbic_Region or rather - I figured out how I "fixed it"
18:56 Limbic_Region a batch file I use to automate some things had included the workaround of copying gcc-lib back into pugs
18:56 Limbic_Region so back to square one - I cry uncle
18:56 Jedai What's the content of this gcc-lib ?
18:56 Limbic_Region mj41_ are you subscribed to p6.compiler?
18:57 Limbic_Region Jedai - ld.exe and friends
18:57 Limbic_Region Jedai - in your ghc root directory, you should have a gcc-lib directory
18:58 Jedai It might be that my mingw provide all of that stuff
18:58 Jedai I have it in my %PATH%
18:58 Limbic_Region as do I
18:59 Jedai strange :(
18:59 Limbic_Region re-arranging my PATH to mirror yours produced different results - but still b0rk
18:59 Limbic_Region not having the first clue as to how this stuff works doesn't really help much
18:59 * Limbic_Region gives up for the day
19:00 Jedai Maybe you should try with a fresh svn checkout ?
19:00 Jedai (Well if you didn't already tried it...)
19:03 Limbic_Region Jedai - any time I have any hiccups I do a nmake realclean; svn up; perl Makefile.PL; nmake
19:03 Limbic_Region I am not the only one this is affecting btw
19:03 Jedai but if nmake realclean is broken... (mine is)
19:03 Limbic_Region but everyone else (but you) who had it working on Win32 indicated a different version of GHC
19:04 Limbic_Region Jedai - AFAICT, it is only broke at the very end while wiping out Makefile.PL
19:04 Limbic_Region err Makefile that is
19:05 Limbic_Region though, that was one of my earlier suspicions - something isn't being deleted that should be
19:05 Limbic_Region the easy fix is to sit back and wait
19:06 Limbic_Region eventually, audreyt will want to release a new version
19:06 * Limbic_Region will time his restatement of the failure accordingly
19:07 Limbic_Region till then - I just stick gcc-lib in Pugs
19:08 Jedai it seems my doesn't use config.yml anymore
19:09 Jedai Was it superceded by anything ?
19:09 Jedai Limbic_Region: as long as you can "fix" it like that I guess that's ok. :)
19:11 Jedai just noticed this line : Setup.exe: gcc-lib\ld.exe: runProcess: does not exist (No such file or directory)
19:11 Jedai Guess my build has broken bits (but I get my pugs in the end and it works so I can hack on it)
19:12 Limbic_Region interesting
19:13 Jedai this was r11413
19:15 Limbic_Region clean build or incremental?
19:15 Jedai this one was clean
19:15 Jedai (almost since ghc ran out of memory in the middle... didn't read my config.yml apparently)
19:16 Limbic_Region oh, yeah - heap problems
19:17 Jedai ?eval method bar () {42};
19:18 Jedai I get arity = 2 for this method, is this normal ?
19:18 * Limbic_Region wouldn't think so but he is likely not the right person to be asking
19:19 * Limbic_Region wanders off hoping that rafl will ping him tomorrow
19:19 Jedai Well method signature seems to be broken (or rather TODO I guess ?) anyway
19:22 ludan joined perl6
19:29 myavuzselim joined perl6
19:37 rindolf joined perl6
19:37 rindolf Hi all.
19:37 rindolf ingy: here?
19:42 b_jonas left perl6
19:43 jkanter joined perl6
19:46 Daveman joined perl6
19:51 ingy hi rindolf
19:52 rindolf ingy: hi.
19:52 rindolf ingy: I saw your pictures in nothingmuch' Flickr Moose stream.
19:52 rindolf ingy: you're not too fat.
19:52 rindolf ingy: in fact, I'm probably as fat as you are.
19:53 rindolf ingy: did you lose weight?
19:53 jkanter is now known as jikanter
19:54 Daveman rofl
19:55 rindolf http://www.flickr.com/photos/nuffin/tags/moose/
19:55 lambdabot Title: "Flickr: nothingmuch's photos tagged with moose"
19:56 rindolf http://www.flickr.com/photos/nuffin/179250879/ # She's pretty cute.
19:56 lambdabot Title: "DSC_3407.jpg on Flickr - Photo Sharing!"
19:59 rindolf OK, Firefox 1.5's text editor when editing bidi text is FUBAR. Guess I'll have to use Konqi.
20:00 rindolf DILLIGAD == Do I look like I give a damn. I found it amusing.
20:02 rindolf http://en.wikiquote.org/wi​ki/Hebraic_proverbs/Errata
20:03 rindolf http://en.wikiquote.org/w/index.php?title=Hebr​aic_proverbs&amp;diff=231588&amp;oldid=224844 # Funny how many people don't know the wikipedia ettiquette.
20:04 rindolf There was also something like that with the tchrist entry, that got fixed by someone else.
20:04 weinig|away is now known as weinig
20:07 SageLT joined perl6
20:16 bpphillips left perl6
20:24 svnbot6 r11414 | Sage++ | Parser now has rough handling for multiple lines in a quoted yaml scalar block. \nRegexs in quotes in a split now translate (split('\s+', $foo) -> split(/\s+/, $foo)). \nSplit on a space now translates to .words (but doesn't handle a third arguement, split(/ /, $foo, 3) just drops the 3 at this point).
20:37 weinig is now known as weinig|bbl
20:51 svnbot6 r11415 | daxim++ | naming docs/notes/context_coer[cs]ion.pod to something more reasonable
20:54 svnbot6 r11416 | Sage++ | Updating to reflect completed translations.
21:00 fglock sent Pugs-Compiler-Rule-0.10 to cpan
21:01 * nothingmuch returns with great dramatic effect
21:01 ajs Should I expect that :$foo always expands to :foo<$foo>, or only in captures?
21:02 ajs In other words, if I "return(:$foo, :$bar)" do I return a list of two pairs?
21:02 audreyt it is in a capture.
21:03 audreyt so you make a named return
21:03 audreyt if it's used as a list instead of being bound to
21:03 audreyt then it becomes two Pair objects
21:03 audreyt compare
21:03 ajs Nice. Thanks
21:03 audreyt sub f { return(:$foo, :$bar) }
21:04 audreyt my ($foo, $bar) := f();
21:04 audreyt with
21:04 audreyt my ($foo, $bar) = f();
21:04 audreyt note however there is a way to unambiguously return two pairs instead of named returns
21:04 audreyt and it's perhaps surprising:
21:04 fglock audreyt: did you see the new Pugs::Runtime::Perl6Prelude.pm ?
21:04 audreyt return (:$foo, :$bar);
21:04 audreyt fglock: no... not yet
21:04 clkao fglock: oh? did you talk to nnunley?
21:05 fglock clkao: no
21:05 audreyt fglock: btw, what do you think about making dirname-distname agree
21:05 audreyt svk mv perl5/Pugs-Compiler-Perl6/ perl5/v6-alpha/
21:05 clkao summon nnunley
21:05 audreyt nnunley: *summon*
21:05 fglock audreyt: sure - split into 2 modules, even
21:06 fglock mm - no, it's easier to maintain in one place
21:06 audreyt fglock: may make sense to bundle them until we have a second backend for v6.pm to dispatch to
21:06 nnunley Yes?
21:06 audreyt which = some time in the future I'm sure
21:06 nnunley audreyt: You rang?
21:06 audreyt (but not in the forseeable months)
21:07 clkao use should reuse the src/perl6/Prelude.pm
21:07 audreyt nnunley: clkao summoned you; I'll bbiab :)
21:07 clkao audreyt: second backend - i think we should hook up a revised emitter in perl6, and self-bootstrap
21:07 * nnunley glances across his flat.
21:07 clkao nnunley: fglock is wokring on pugs::runtime::perl6prelude
21:09 nxu7 joined perl6
21:09 nnunley fglock: Ok.  I have a stub implementation of the Internals that need to be present in order to compile Prelude.pm in perl6.
21:09 nnunley There were some namespace issues (Pugs vs Perl6)
21:10 nnunley But other than that, there weren't many things that couldn't be hoisted into the perl5 backend.
21:10 fglock nnunley: the Prelude can export the subs, so the actual namespace doesn't make much difference
21:11 nnunley True.
21:12 nnunley fglock: Have you seen perl6/INTERNALS.txt? That is the list of functions that the old Prelude expects to be exported.
21:12 fglock how about packaging the Prelude and Test.pm into a cpan package that would be common to all implementations
21:14 fglock nnunley: my Perl6Prelude is mostly a placeholder - it would be nice to reuse Prelude.pm
21:15 * nnunley nods.
21:18 clkao fglock: can you see why 08test is takin 400M ram?
21:18 svnbot6 r11417 | daxim++ | explain use of "(history of this directory)" in openfoundry.org browsable VCS
21:18 fglock otoh, the prelude is a bit low level, and implementation-dependent
21:19 clkao there are many bits implementation-indepedent, i think the idea is to have those shared and define a list of required things to be implemented by backend
21:19 fglock clkao: compiling it with non-backtracking rules will fix that - but PCR is not ready for the transition yet (missing a few functions)
21:20 fglock clkao: each character in the source code requires creating a closure :(
21:20 clkao *nod*
21:21 nnunley Ieeee.
21:21 clkao fglock: did you get a chance to look at the failed compile smoke output?
21:22 clkao subroutine/multidimensional_arglists is also infinite loop
21:22 fglock yes - that's why I started implementing enum - it's the test with more failures :)
21:23 clkao ;)
21:24 fglock TimToady fixed the listquote test
21:30 fglock some subs in Prelude.pm are already implemented - like eval, caller
21:33 nnunley fglock: There's the multi eval, and then there's Pugs::Internals::eval_perl6 (which was eval in the INTERNALS doc when I wrote it)
21:34 weinig|bbl is now known as weinig
21:34 larsen joined perl6
21:36 nnunley fglock: But you're right. Things that are implemented in the Prelude should be trimmed from that list.
21:40 fglock having a common Prelude is a good idea - nothingmuch was working on an idea for using a base Prelude and additional implementation-specific ones
21:41 fglock the base Prelude would "depend on itself", so the implementation-specific Prelude would be used to break circularity
21:41 svnbot6 r11418 | lwall++ | hash subscript no longer allowed directly after listop.
21:44 fglock nnunley: where is INTERNALS.txt ?
21:45 nnunley src/perl6/INTERNALS.txt
21:46 nnunley fglock: I can also commit my stub implementation. I was going to inject it into the Runtime directory of Pugs-Compiler-Perl6
21:48 fglock nnunley: how about merging into Runtime::Perl6.pm (or Perl6Prelude.pm, if it is v6) ?
21:49 fglock just create new 'package' sections
21:53 nnunley Yeap.
21:54 fglock nnunley: thanks!
21:55 fglock we can rule a common rule for internal namespaces with audreyt later
21:55 fglock s/rule rule/rule/ whatever
21:57 b_jonas joined perl6
21:57 fglock clkao: how about bundling Test.pmc with v6 in cpan?
21:59 clkao sure
22:03 fglock clkao: I already submitted the changes I've made to PCR - if you are ok with Data::Bind in cpan we could do a new v6 release
22:07 fglock s/we/you/ :)
22:07 svnbot6 r11419 | fglock++ | v6 - updated MANIFEST; ChangeLog
22:09 clkao fglock: surething
22:10 nnunley fglock: Hrm. What I have causes a parsefail. Do you want me to commit anyway?
22:12 fglock nnunley: does it fail the v6 t/* ?
22:14 clkao fglock: did you fix the hash test?
22:16 Jedai Did someone manage to get Devel::Caller 0.10 to pass the tests on win32 ?
22:16 fglock clkao: I implemented some of the syntax, but it's not complete yet
22:17 nnunley fglock: Hrm. Doesn't much like me using 'caller' as a function name.
22:17 nnunley Ok. I've commented out the offending line.
22:18 fglock Jedai: I'm reinstalling - current is 0.11
22:18 fglock nnunley: we can put 'caller' in the identifier mangling list if necessary
22:19 prefiks left perl6
22:20 nnunley fglock: Committed.
22:20 nnunley The signatures and documentation for the arguments are a bit lacking.
22:20 fglock nnunley: thanks!
22:21 nnunley Hrm. Current Prelude dies with a parse failure on 'class Control::Basic'
22:21 fglock Jedai: Devel::Caller 0.11 installed fine with vanilla-perl on windows XP
22:21 svnbot6 r11422 | nnunley++ | Importing stubs implementations of internal functions required by /src/perl6/Prelude.pm
22:22 fglock nnunley: given/when is not implemented yet in the parser
22:24 * nnunley nods.
22:26 fglock ' eval("sleep($seconds)", lang<perl5>) ' is compiled at runtime - is there an option in perl 6 to precompile?
22:26 Jedai fglock: Thanks it works fine here too
22:26 Jedai On ActivePerl build 817 with mingw and nmake
22:27 TimToady BEGIN {...}
22:28 fglock TimToady: the sentence is: sub sleep($seconds) { eval("sleep($seconds)", lang<perl5>); }
22:29 fglock we need something like: sub sleep($seconds) { BEGIN { eval("sleep($seconds)", lang<perl5>) } }
22:30 nnunley Or macro notation for the backend...
22:31 fglock or intercept the calls in the emitter, which we've been doing so far
22:32 fglock is macros the way to go?
22:32 TimToady or are you just wanting to inline Perl 5 code?
22:32 fglock TimToady: yes
22:34 fglock TimToady: that's Prelude code for v6.pm
22:38 clkao in the js prelude, it's using special javascript_inline
22:39 fglock TimToady: can we use :P5 like in regexes? ' :P5 { sleep $seconds } '
22:42 svnbot6 r11423 | clkao++ | Data::Bind 0.27.
22:45 svnbot6 r11424 | fglock++ | v6 - Perl6Prelude.pm s/lang<perl5>/:lang<perl5>/
22:45 svnbot6 r11425 | clkao++ | Sub::Multi 0.001_2
22:46 nnunley Doh.
22:47 fglock clkao: can it be Sub::Multi 0.002 ? cpan doesn't install _x versions easily
22:48 clkao orz
22:52 fglock home &
22:53 fglock left perl6
23:04 froh-doh joined perl6
23:19 ivanfrey Pr
23:24 svnbot6 r11426 | lwall++ | Discouraging constructs that look like subscripts but aren't.
23:50 frederico joined perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo