% chromatic has left chromatic!~chromatic@sub17-30.member.dsl-only.net % iblechbot has joined #parrot % contingencyplan has left contingencyplan!~contingen@cpe-76-186-27-146.tx.res.rr.com % contingencyplan has joined #parrot % cosimo has joined #parrot % mire__ has left mire__!~Frodo@41-169-222-85.adsl.verat.net % kj has joined #parrot % jjore has left jjore!~jjore@c-24-16-241-176.hsd1.mn.comcast.net % jjore has joined #parrot % DarkWolf84 has joined #parrot % lbr has left #parrot % AndyA has left AndyA!~andy@82.152.157.85 % AndyA has joined #parrot r26088 | chromatic++ | trunk: : [Lua] Constant String PMCs need to contain constant STRINGs. Hopefully this is just a temporary workaround until we get constants sorted. : Lua still doesn't quite run with GC enabled, but it's close. r26089 | kjs++ | trunk: : [ecmascript] % marmic has left marmic!~chatzilla@89-253-66-101.customers.ownit.se % marmic has joined #parrot % gabriel has left gabriel!~gabriel@c-76-126-210-9.hsd1.ca.comcast.net r26090 | kjs++ | trunk: : [docs] improve operator table documentation a bit. diff: http://parrotvm.org/svn/parrot/revision/?rev=26090 % kid51 has joined #parrot % iblechbot has left iblechbot!~iblechbot@135.16-dial.augustakom.net % davidfetter has joined #parrot r26091 | jkeenan++ | tcif: : Until Test-Harness-3.11 is available, will need to make a system call to : prove to run pre- and postconfiguration tests, as App::Prove::run() : exits prematurely. diff: http://parrotvm.org/svn/parrot/revision/?rev=26091 % dwave has left dwave!~asksolem@pat-tdc.opera.com % dwave has joined #parrot r26092 | jkeenan++ | tcif: : Correct overly strict test #2; use 'ok' instead of 'is' in 01.t tests. diff: http://parrotvm.org/svn/parrot/revision/?rev=26092 % Coke has joined #parrot % wknight8111 has joined #parrot % svnbotl has left svnbotl!diakopter@193.200.132.135 % svnbotl has joined #parrot % davidfetter has left davidfetter!~chatzilla@start.fetter.org % mire__ has joined #parrot % iblechbot has joined #parrot % Andy has joined #parrot % Andy has left Andy!~Andy@64.81.227.163 % DarkWolf84 has left DarkWolf84!~dwolf@89.215.234.147 % liona29 has joined #parrot % liona29 has left liona29!~liona29@d033.dhcp212-198-248.noos.fr % kid51 has left kid51!~jkeen@pool-71-247-55-230.nycmny.east.verizon.net % kj has left kj!~IceChat7@ip565fd420.direct-adsl.nl % gryphon has joined #parrot % Coke has changed the topic of #parrot to: Parrot 0.5.3 Released | http://parrotcode.org/ | see http://www.parrotcode.org/misc/parrotsketch-logs/ for logs gmail is giving me ads for "bird diaper flight suits" when viewing parrot tickets. % jq has left jq!~jquelin@merlin.mongueurs.net It's a very, very mad world. % cout has left cout!~cout@c-68-58-247-49.hsd1.sc.comcast.net % cout has joined #parrot % jq has joined #parrot lathos: hai. Coke, parrot is still leaking after all? ... hence the diaper? yea =-) does that make chromatic and andy the diaper team? Not going there. is there an easy way to reverse an array pmc? (fixedpmcarray) % jhorwitz has joined #parrot I have experience with diapers. % AndyAway is now known as Andy We know there is plenty of leakage. % camgirl29 has joined #parrot % camgirl29 has left camgirl29!~camgirl29@d033.dhcp212-198-248.noos.fr % cj has left cj!~cjcollier@66.152.65.2 Current Parrot leakage from "perl6 -e'say "Hello, World!"' http://pastebin.com/m2d682b4 Andy: If you run Rakudo in interactive mode, do "class Foo { }" and press enter, do you get an error? Just want to check someone other than me is seeing this... can't look right now ask me in 5 rebuilding, checking pmichaud: I was working on a Parrot guts patch when I ran into this...thought for ages it was my patch, then tested again in a clean tree and it's still here... what error are you getting? uniqua:~/parrot $ ./parrot - class Foo { } it has been said that class Foo { } is okay error:imcc:syntax error, unexpected PARROT_OP ('class') in file '-' line 1 I know that this worked up until relatively recently, because otherwise I'd never have been able to do/test 26081 andy: forgot perl6.pbc jonathan: did you try running with -G ? class Foo { } class Foo { } is okay too few arguments passed (0) - 4 params expected c:\usr\local\parrot\trunk\languages\perl6>perl6 class Foo {} too few arguments passed (0) - 4 params expected pmichaud: Same with -G ayep But when I did, say, 26081 I was not svn up'd to latest Parrot core. iwbni the "too few arguments" message also said what sub was being called iwbni? * particle kicks purl NEWSFLASH: LAMER ATTACKS INFOBOT. FILM AT 11 The odd thing is, it says one sub in the stack trace, and the line number is that of the next sub. cognominal_: it would be nice if... purl, iwbni is "it would be nice if" OK, particle. in case of syntax error is that possible to get the parser stack? I went back to 26074 and the bug is there too, so it wasn't related to the unicode improvements. i'm going through -t 1 now ...parser stack? No, call stack at the point of error. but "point of error" is the point at which we detected the error it doesn't say anything about how far the parser got If we get to execution, I mean. Note that --target=pir works. in case of "syntax error" we don't get to execution, right? -e fails, though (my "parser stack" note was in response to cognominal_'s q, sorry) pmichaud: Ah, sory, OK. :_) "particle" at 24.19.3.148 pasted "tail of output from ../../parrot perl6.pbc -t 4 -e "class Foo { }"" (1062 lines) at http://nopaste.snit.ch/12409 running with -t1 gives me a segfault I mean, what the parser has commited so far and how or something like that it's in HLLCompiler's evalpmc cognominal_: that needs to come from the error message yes particle: Yup, that's where I saw it. But the line number matches up with HLLCompiler's eval. It's funny that this code fails, but "say 'hi';" works oh there's something wrong with the PIR oh, nm it's okay hmmm r26094 | coke++ | trunk: : [tcl] : Simplify [lsort] - one of our pmc ancestors has a sort; use that one, : delete our own version. add in a way to reverse a list, to implement : --decreasing (since the builtin sort doesn't let us flip the sense what is "$def" ? Andy: I'm vaguely curious as to whether or not ./tclsh -e "puts {Hello, world.}" leaks more. The currently being defined class. but why is it a lexical? Coke: Remind me later today and I'll run it for you. Because I can't see a way in PAST to just say "use a register". but why do you need to store it in a register in the first place? It's a temporary, but I need to refer to it in many places, passing it off to various methods... ahhh Like, for add_attribue, add_role, etc. I'd like a cleaner solution, if there is I'd be happy. But it was the best I could think of without resorting to inline PIR. ok 26049 works 26074 does not. So the change is somewhre between those. my thinking was to collect all of the things to be added into a single call to create the class as opposed to separate calls for each attribute, each role, etc. I guess that could work too. so, each action method for 'has', etc. simply stores entries in the symbol table, then when we're done with the block for the class we collect all of the attributes together then That could work too. Maybe cleaner. I don't feel strongly either way. anyway, that's not what is at issue here :-) Sure. Trying to work out which revision it is... Spent ages trying to figure out what on earth I'd done wrong with my "does" patch... 26052 is not working % peeps[work] has joined #parrot http://parrotvm.org/svn/parrot/revision/?rev=26052 Other patches between this one and the one that works were to Pheme, so I suspect this one. I wondered about this patch when I saw it that's just converting to tailcalls. *shouldn't* be a problem. I'm trying to spot why it's wrong, and it's not obvious to me... (if it is, it's exposing a parrot bug.) a tail call is not exactly the same especially if the called function returns something other than a single value if the original assignment was to a PMC, it doesn't matter. if the original assigment was to a base type, yes, you're then missing an autobox. huh ? % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com the tailcall can still be missing an autobox but lets look at arg mismatch Didn't I just say that? if sub 'foo' does .return ($P0, $P1, $P2) then there's a difference between $P1 = 'foo'() .return ($P1) and .return 'foo'() % peepsalot has joined #parrot ... what's the difference? I'm not saying that's the error in this case, I'm just saying that they aren't equivalent the non-tail call version returns a single PMC the tailcall version returns three PMCs seems reasonable, ok. (in this case, though, that's not happening. =-) similarly, if sub foo does .return ($I0) % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com then there's a difference between $P1 = 'foo'() .return ($P1) and .return 'foo'() not really. the first does autoboxing to a PMC % peepsalot has joined #parrot the second may or may not since all the autoboxing paths I can think of will result in the same answer. assuming that the tail call is within sub 'bar', then there would be a difference with $S0 = 'bar'() because the tailcall would coerce directly from int to string without going through PMC % uniejo has left uniejo!~uniejo@langebro.adapt.dk OK, found it It's the second hunk in the HLLCompiler patch 426 $P1 = $P0(source) 427 .return ($P1) 425 .return $P0(source) ugh % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com - $P1 = $P0(source) - .return ($P1) + .return $P0(source) wtf You gt the idea. :-) % peepsalot has joined #parrot yes, from the nopaste :-) so, the PIR compiler doesn't like being called via a tailcall? Basically, tail-calling into what we get back from compreg doesn't work. That's my analysis. I can believe it. The PIR compiler PMC is probably somewhat odd Yeah. Seems a likely culprit. OK, guess for now we revert that bit of the patch, then it's a ticket. yes nice catch I'll do the fix. Anyone for the ticket? ;-) Who wanted the tcl help? coke did. that's the guy that isn't me. [RESEND] Andy: I'm vaguely curious as to whether or not ./tclsh -e "puts {Hello, world.}" leaks more. Coke, YOU want me to try some tcl stuff (than the equivalent perl6 version) ok, making. what exact commandline do you want? % peeps[work] has left peeps[work]!~peepsalot@bwext.kpimdp.com I don't see a tclsh ahhh, it's a target Reverted that one tail call. ../../parrot tcl.pbc -e "puts {Hello, world.}" would also work. % mj41_ has left mj41_!chatzilla@pc-jurosz.ro.vutbr.cz MANY more leaks 800k % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com Writing the ticket for RT now... % peepsalot has joined #parrot okm mailed 'em to you, Coke. back to work I go. % particl1 has joined #parrot r26095 | jonathan++ | trunk: : [pct] Undo a tailcall optimization; tail-calling into the PIR compiler sometimes seems to not work. diff: http://parrotvm.org/svn/parrot/revision/?rev=26095 tcl back down to two failing tests if i disable GC. Andy++ thanks. you're welcome. % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com % peepsalot has joined #parrot % particle has left particle!~particle@c-24-19-3-148.hsd1.wa.comcast.net % particl1 is now known as particle r26096 | jonathan++ | trunk: : [core] Add pmc variant of the does vtable method and a related op. diff: http://parrotvm.org/svn/parrot/revision/?rev=26096 % camgirl29 has joined #parrot % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com % peepsalot has joined #parrot % camgirl29 has left camgirl29!~camgirl29@d033.dhcp212-198-248.noos.fr % cj has joined #parrot % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com % peepsalot has joined #parrot % peeps[work] has joined #parrot % jjore is now known as jjore_away % jjore_away is now known as jjore % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com % peepsalot has joined #parrot % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com % peepsalot has joined #parrot % Theory has joined #parrot % cosimo has left cosimo!~cosimo@pat-tdc.opera.com % mj41 has joined #parrot r26097 | jonathan++ | trunk: : [core] Implement does_pmc and do some other bits of refactoring of the does code. diff: http://parrotvm.org/svn/parrot/revision/?rev=26097 % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com % peepsalot has joined #parrot % davidfetter has joined #parrot is there a perl based tool for finding cut and pasted code? Something like PMD's CPD that works on arbitrary text files? % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com % peepsalot has joined #parrot % Bill_Ruppert has joined #parrot % Bill_Ruppert is now known as Bill % parrot-poke has joined #parrot PMD? CPD? * davidfetter has gone and forgotten who's supposed to be in .pt at the moment you are! http://pmd.sourceforge.net/cpd.html i meant apart from yours truly * davidfetter touching up his slides for a lisbon.pm meeting... % silug has joined #parrot Hmm. So...what next on Rakudo... :-) r26098 | jonathan++ | trunk: : [rakudo] Implement smart-matching against a role name to check if the class does the role, plus a little clean-up to the role creation code. r26099 | jonathan++ | trunk: : [rakudo] Add a file missed out in the previous ci. Oops. diff: http://parrotvm.org/svn/parrot/revision/?rev=26099 jonathan: try make test i'm getting a sanity test failure, 07-simple-multisubs.t with r26099. same for you? push_pmc() not implemented in class 'Closure' jonathan, I want all the pair syntaxes :) I am curious how you will do it because so far I fail. % sjansen has joined #parrot particle: Is r26099 latest? yep particle: Here: t\01-sanity\07-simple-multisubs....ok did realclean after your does_pmc patch Right. And those that came after it shouldn't have needed that. Hmm. ok, well i'll step away from rakudo for a bit, anyway. i have too many patches in the air, gotta commit some if you could get 'is export' working, i'd love it. Rakudo ones or elsewhere? elsewhere OK. i've got a config step for gettext Let me know if the multisub issue goes on; since I'm throwing so much into Rakudo I'm likely to blame for problems. wilco is export - yeah, now I've got and wabbed on about how it's supposedly easy... I only meant the "put a multi in the global namespace" bit is easy, though. Is that the bit you're wanting? hrmm. i want to replace the pir classes with perl 6 classes in so doing, i'd like to have 'is export' working cognominal_: All of them? All of them? Sheesh! :-) so List.max can be called with max List cognominal_: But yes, I want them too. :-) particle: Right, that's the multi-creation bit. OK, looks like I've got my next tasks cut out for me. or, you could work with me on getting zev's patch applied How far off working is it? it's close. lemme nopaste what i've got jonathan, pairs are everywhere in Perl6 if you want to have a perl6 open function, you need pairs cognominal_: Sure, they're important. "particle" at 24.19.3.148 pasted "zev++'s patch, modified" (103 lines) at http://nopaste.snit.ch/12410 I have been looking into doing them. Just getting it straight(ish) in my head before digging into the code. particle: Will look into this later, having a chat with Liz now about concurrency stuff. because I can't help you I am translating som pir in Perl6 and pairs would make it clean right-o Liz++ OK, will try and do some work on paris soon. back in a bit we'll always have pairs % Psyche^ has joined #parrot purl: seen kid51? kid51 was last seen on #parrot 13 hours and 44 minutes ago, saying: must sleep % particl1 has joined #parrot my stupid internet is broken. % peepsalot has left peepsalot!~peeps@cpe-67-9-161-48.austin.res.rr.com particl1: did you set svn props on those new files before checkin? probably not, can't recall. what's the incantation for that? wondering if that explains the odd doublespacing. prove t/distro/file_metadata.t : do what it tells you. that's annoying. we should have a script that just does it kick particle, would you? % peepsalot has joined #parrot % particle has left particle!~particle@c-24-19-3-148.hsd1.mn.comcast.net % particl1 is now known as particle thx % Patterner has left Patterner!~Psyche@f054002071.adsl.alicedsl.de % Psyche^ is now known as Patterner fixing r26100 | particle++ | trunk: : [config] add gettext detection : ~ if gettext libs detected, -DHAS_GETTEXT is added to ccflags : ~ added simple tests, more needed for full coverage diff: http://parrotvm.org/svn/parrot/revision/?rev=26100 r26101 | particle++ | trunk: : don't use magic numbers r26102 | particle++ | trunk: : fix svn metadata r26103 | particle++ | trunk: could somebody update to 26105, realclean, and configure? % barney has joined #parrot i wanna make sure i didn't break non-windows platforms with that gettext config patch it'd be nice to have results with and without gettext r26104 | particle++ | trunk: : [t] fix file metadata test failure output so commands work on windows, too r26105 | particle++ | trunk: : [config t] fix metadata for real this time Seeing if your configuration includes gettext...........................no. so far, so good on linux-x86 particle: OOC, why did you remove the semicolon on that map? where? you mean before \n" semicolons are not command separators on windows if they're there, i can't copy the output from the failed tests and run it And now *i* can't do that. =-) what is a command separator on windows? ok, looks like we'll need && i wonder if foo && will work. eh. I can always cut and paste separately. I am now inconvenienced, but less then you were before the patch. if we prefix && on the second - last lines, it'll work for both that's easy enough. moment. % stephen has joined #parrot % stephen has left #parrot ... meh. nevermind. feel free to fix it if you like, else I'll do it later. =-) % mire__ has left mire__!~Frodo@41-169-222-85.adsl.verat.net % IllvilJa has left IllvilJa!~jilves@emea-netcache1.oracle.co.uk % mire__ has joined #parrot % IllvilJa has joined #parrot r26106 | bernhard++ | trunk: : [Plumhead] : Get rid of '"' in double quoted strings. diff: http://parrotvm.org/svn/parrot/revision/?rev=26106 % davidfetter has left davidfetter!~chatzilla@start.fetter.org cotto++ for work on PHPArray r26107 | bernhard++ | trunk: : #51178: [PATCH] exists_* and delete_* for PHPArrays : Courtesy of Christoph Otto. r26108 | particle++ | trunk: : [docs] typo and ws fixes pmichaud: ping barney, thanks r26109 | bernhard++ | trunk: : Remove trailing whitespace in C-code. diff: http://parrotvm.org/svn/parrot/revision/?rev=26109 * barney hates Perl scripts with three BEGIN blocks, t/perl/Parrot_Test.t which does not compile here the third begin can likely be removed Is anybody looking at the patches from Andrew Whitworth ? 50884, 50882, 50768 '84 doesn't seem to have a patch attached particle: pong It's a natural language patch :=) "particle" at 24.19.3.148 pasted "pmichaud: not-quite-working perl6-in-perl6 patch, modified from a submission by zev" (103 lines) at http://nopaste.snit.ch/12411 I was hoping to avoid the perl6-bootstrap.pmc target *.pbc I'd rather not call it 'bootstrap' at any rate -- i.e., I'd prefer BOOTSTRAP_SOURCES to be PIR_SOURCES and RUNTIME_SOURCES to be PERL6_SOURCES see, for example, the way that PGE's makefile uses PGE to build a component of itself without having to have a separate PGE.pbc target s/target/bootstrap target/ what's the not-quite-working part? pmichaud: yes, i was planning to use pir/perl6 hang on, phone % peeps[work] has left peeps[work]!~peepsalot@bwext.kpimdp.com ok, i'm back too few arguments passed (0) - 1 params expected now that jonathan found tha other tailcall problem, i wonder if it's related current instr.: 'foo' pc 100441 (src\gen_runtime.pir:19) called from Sub '_block10' pc 31 (EVAL_7:15) called from Sub 'parrot;PCT::HLLCompiler;eval' pc 785 (src\PCT\HLLCompiler.pir:458) oh. Dummy.pm is generating code that is being treated as main? or perhaps it's because we end up with multiple _block10's hrmm. there's no :main but, yeah, i think it's multiple _block10 yes, that's it. let me svn up and see what I get ok, you'll need svn up and apply patch patching... building phone TimToady++ % ruoso has left ruoso!~ruoso@mail.verticalone.pt r26110 | bernhard++ | trunk: : #50768: [PATCH] Update Win32 platform documentation. : Courtesy of Andrew Whitworth diff: http://parrotvm.org/svn/parrot/revision/?rev=26110 % schmalbe has joined #parrot % flora29 has joined #parrot % barney has left barney!~bernhard@dslb-084-058-165-149.pools.arcor-ip.net % flora29 has left flora29!~flora29@ANantes-257-1-17-244.w90-25.abo.wanadoo.fr particle: RT#44333 can probably be closed done, thanks OK, what is up kids? I can breathe for the afternoon. you gonna hunt leaks in src/dod.c? I don't understand it. so no. :-) I've picked off all the leaks I can. at this point I'm just tracking updates in the valgrind output to see what's what how's splint looking? it spews a lot of stuff. % jmcadams has joined #parrot hey, it's that perlcast guy! r26111 | bernhard++ | trunk: : #44439: [TODO] Skip files copied wholesale into repository when running perlcritic.t : Shorten the excemption list, : as some external modules habe been removed. : RT#44439 is implemented hello :) jmcadams: did you get my email about yapc banners? just looking at my email now.... i'm guessing that i screwed up the html :( the banners are missing a colon... YAPC:NA i think pointblack took poetic license in the design :/ ah, yeah, just read it oversight it has been said that oversight is Schuyler's guess, but a thorough explanation would require more internals wisdom than I posess. * jmcadams will get some replacement ones up great, thanks that's funny though... we caught a few things in the proofing, but something so obvious slipped by one of the edits even involved removing a hurricane from the image in the background.... it looks Louisiana bound yet we didn't see the colon i wonder why the western usa is red heh I think I'm going to make some bash autocompletion on prove. so it only finds .t files r26113 | chromatic++ | trunk: : [config] Removed trailing spaces and replaced tabs with spaces. This cleans up : some codingstd test failures.