% magnet.llarian.net has changed the topic of  to: Piper
<bacek> shit... I can't reply on #55000
<bacek> any parrot commiters alive here?
<bacek> seen chromatic
<purl> chromatic was last seen on #parrot 4 days and 11 hours ago, saying: It'll be a PMC.  [May 27 17:24:52 2008]
<nopaste> "bacek" at 202.7.166.165 pasted "Fix for #55000" (23 lines) at http://nopaste.snit.ch/13135
<bacek> msg chromatic http://nopaste.snit.ch/13135 fix for #55000
<purl> Message for chromatic stored.
<bacek> purl, good bot
<purl> :)
% kid51 has joined #parrot
% masak has left masak!~user@130.238.45.242
% braceta has joined #parrot
<dalek> r28006 | jkeenan++ | codetests:
<dalek> : Per suggestion by pmichaud, change t/harness option to '--code-tests' for
<dalek> : consistency with other options.  In addition, change 'make' option to
<dalek> : 'code_tests' for consistency with other 'make' targets.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28006
<dalek> allison@perl.org | Articles of Incorporation:
<dalek> link: http://www.perlfoundation.org/parrot/index.cgi?articles_of_incorporation
<shorten> dalek's url is at http://xrl.us/bkyfj
<dalek> allison@perl.org | A foundation for Parrot:
<dalek> link: http://www.perlfoundation.org/parrot/index.cgi?a_foundation_for_parrot
<shorten> dalek's url is at http://xrl.us/bkxq5
<dalek> r28007 | jkeenan++ | trunk:
<dalek> : Applying final version of patch submitted in http://rt.perl.org/rt3/Ticket/Display.html?id=54992.  This (a) creates a 'make code_tests' target to run the file metadata test and basic coding standards tests; (b) creates a '--code-tests' option to t/harness, which runs the same tests; and (c) eliminates all instances of '' from config/gen/makefiles/root.in.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28007
<dalek> r28008 | jkeenan++ | codetests:
<dalek> : Branch has been merged into trunk; no need to retain.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28008
<dalek> r28009 | jkeenan++ | codetests-27904:
<dalek> : Branch associated with this tag has been removed from HEAD; remove tag as well.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28009
% ambs has joined #parrot
% ambs has left ambs!~ambs@255.92.54.77.rev.vodafone.pt
% ank has left ank!~ank@ppp121-44-210-24.lns1.hba1.internode.on.net
% IllvilJa has left IllvilJa!~jilves@emea-netcache1.oracle.co.uk
<dalek> r28010 | jkeenan++ | trunk:
<dalek> : Provide more data so that test operates without warnings.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28010
% braceta has left braceta!~Braceta@bl10-34-59.dsl.telepac.pt
% IllvilJa has joined #parrot
% rdice has joined #parrot
% luca_ has joined #parrot
<luca_> hi parroteers
<luca_> i would like to start to contribute to parrot
<NotFound> luca_: Just do it!
<luca_> notfound may i ask you some questions in private?
<luca_> are you currently involved in parrot development?
<NotFound> luca_: I'm not very experienced, better do it here, so anyone can correct me if wrong.
<luca_> just to summarize my experience
<luca_> i mostly develop server systems for my work
<luca_> lately i have been writing apache and openldap modules
<NotFound> If sending patches and get some of them accpeted is to be involved, yes ;)
<luca_> i know from experience that taking on a project from scratch
<luca_> looking at the source code and docs takes a while
<luca_> it tooks for example about working 20 days to start to produce something on the openldap project
<luca_> so i ask if there is some
<luca_> well definied task
<luca_> or subsystem from which i may start
<luca_> 20 working days..
<purl> it has been said that 20 working days is usually the minimum in the UK. I'm on 25
<NotFound> Good question. From my recent experience, find a good delimited place where to start is difficult. Even tickets described as an easy task can have unexpected ramifications.
<luca_> you see...
<purl> No. I don't see.
<diakopter> luca_: you can disregard the bot purl
<dalek> r28011 | jkeenan++ | trunk:
<dalek> : Two files listed in FLUID_FILES_2 should have been in
<dalek> : CONFIGURE_GENERATED_FILES so that they will be cleaned up (only) by 'make
<dalek> : realclean'.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28011
<kid51> purl:  Be polite to our new guests.
<purl> kid51: huh?
<kid51> purl, forget you see...
<purl> kid51: I forgot you see
<luca_> and who is here from europe?
<luca_> mi idea
<NotFound> I'm from Spain.
<luca_> ah nice
<tetragon> kid51: Just tested #54986
<luca_> i was looking for some parrot hacker to meet in europe
<luca_> also
<luca_> on septmber there will be the yapc italian chapter
<luca_> and it would be very nice
<tetragon> kid51: And got an Internal Server Error when I submitted my reply to the ticket
<NotFound> I live in Santiago de Compostela, you can make a pilgrimage ;)
<luca_> to present some kind of parrot dev tutorial for the conference
<luca_> mmm
<luca_> i am chatolic
<luca_> and catholic
<purl> catholic is little boobies, short skirts and pot.
<kid51> jonathan travels all over Europe.  Other active contributors in Europe (not necessarily on channel now) are barney, fperrad, kjs, et. al.
<luca_> and i preach but not practice
<kid51> tetragon:  Can you try that again?
<luca_> jonathan is the guy that is after the perl6 compiler?
<NotFound> purl: forget catholic
<purl> NotFound: I forgot catholic
<tetragon> kid51: I just reloaded, it appears that the 500 happened after applying the update
<kid51> tetragon:  Which RT was that?
<tetragon> kid51: 54986
<kid51> I see your post there:  http://rt.perl.org/rt3/Ticket/Display.html?id=54986
% gmansi has left gmansi!~gmansi@190.55.35.246
% gmansi has joined #parrot
<kid51> I'm hoping to get *this* one of your patches tested on Intel OS X 10.5:  http://rt.perl.org/rt3/Ticket/Display.html?id=52214
<kid51> Then apply that to trunk; then try the next patch.
* tetragon looks at her family pack of OS X 10.5
<dalek> r28012 | rgrjr++ | trunk:
<dalek> : [CORE]
<dalek> : * src/ops/experimental.ops:
<dalek> :    + (find_sub_not_null):  Say "Could not find" instead of "Could not
<dalek> :      invoke" for the error message.  Since this op is also used for
<dalek> :      ".const .Sub", the word "invoke" could be confusing if the code is
<dalek> :      only going to make a closure out of it.
<dalek> :    + (substr_r):  Doc tweaks.
<dalek> : * t/pmc/namespace.t:
<dalek> :    + Update message in two test cases.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28012
<NotFound> A very nice addition the code_tests target, boys.
<tetragon> But for me to switch the Intel over to 10.5, it would take some time (and additional portable storage as its network link would not handle 0.5GB of patches all that well)
<NotFound> Now let's hope all people use it ;)
<dalek> r28013 | rgrjr++ | trunk:
<dalek> : [IMCC]
<dalek> : * compilers/imcc/pbc.c: 
<dalek> :    + (add_const_pmc_sub):  Show flags in hex in DEBUG_PBC_CONST output.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28013
<NotFound> code_tests++
% tetragon has left tetragon!~seneca@gw-312-705.somanetworks.com
<DietCoke> kid51++
% Zaba_ has joined #parrot
% luca_ has left luca_!~lucax@host-84-221-220-144.cust-adsl.tiscali.it
% tetragon has joined #parrot
% Zaba has left Zaba!~zaba@ip102.148.adsl.wplus.ru
<dalek> r28014 | jkeenan++ | trunk:
<dalek> : Eliminate import of Data::Dumper.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28014
<NotFound> Someone can take a look at #55154, readline detection in c++?
% AndyA has left AndyA!~andy@82.152.157.85
% AndyA has joined #parrot
<dalek> r28015 | jkeenan++ | trunk:
<dalek> : Implement simpler way of averting uninitialized value warnings in two tests.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28015
% kid51 has left kid51!~jkeen@pool-68-237-19-3.ny325.east.verizon.net
% nnunley has joined #parrot
% Theory has joined #parrot
% guru has joined #parrot
% guru has left guru!~guru@dsl-67-204-16-202.acanac.net
% guru has joined #parrot
<particle>         ..\..\parrot.exe ..\..\runtime\parrot\library\PGE\Perl6Grammar.pbc --output=TGE/Parser.pir TGE/Parser.pg
<particle> Null PMC access in invoke()
<particle> current instr.: 'parrot;PGE::Perl6Grammar::Compiler;compile' pc 200 (runtime\parrot\library\PGE\Perl6Grammar.pir:157)
<particle> same with -G
<particle> this is after realclean
<particle> afk &
<Tene> spinclad: I'm afraid of too much chance of confusion with squaak.
<dalek> r28016 | allison++ | pdd25cx:
<dalek> : [pdd25cx] Add a continuation-based branching operation, that can be used safely
<dalek> : with CPS and exceptions.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28016
% gmansi has left gmansi!~gmansi@190.55.35.246
% iblechbot has joined #parrot
% Theory has left Theory!~Theory@c-24-21-175-208.hsd1.mn.comcast.net
% jhorwitz has joined #parrot
<dalek> r28017 | jkeenan++ | trunk:
<dalek> : Per request from pmichaud, change 'make code_tests' to 'make codetest'.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28017
% guru has left #parrot
% ilbot2 has joined #parrot
% moritz has joined #parrot
% avar has left avar!avar@u.nix.is
% avar has joined #parrot
% gmansi has joined #parrot
% bacek has left bacek!~bacek@123-243-38-218.tpgi.com.au
<NotFound> Given that PIO_printf and PIO_eprintf don't use PIOs in his interface and that are general purpose, will not be better to rename them as Parrot_printf and move his declarations to embed.h ?
% bacek has joined #parrot
% nemesys has left nemesys!~nemesys@HSI-KBW-085-216-121-022.hsi.kabelbw.de
% Theory has joined #parrot
<moritz> uhm, I get a segfault in rakudo while running t/spec/S29-array/push.rakudo
% nemesys has joined #parrot
<NotFound> moritz: can't reproduce here.
% Ivatar has joined #parrot
<NotFound> Works both with parrot perl6.pbc and with ./perl6
<moritz> valgrind reports some noise
<NotFound> Will try with a c++ build.
<moritz> ==21724== ERROR SUMMARY: 182558 errors from 73 contexts (suppressed: 61 from 1)
<NotFound> Same result winth c++ build.
<NotFound> Linux i386
<NotFound> Uh, I was not updated, retesting.
% Theory has left Theory!~Theory@c-24-21-175-208.hsd1.or.comcast.net
% masak has joined #parrot
<NotFound> Same result, no segfault.
<moritz> I tried to reduce the test file, which of course made the bug go away
<NotFound> Will try with gcdebug core, but may take ages to finish ;)
<NotFound> Adding a --runcore option to perl6 executable will be nice.
<moritz> NotFound: I have a reduced test cases that fails with runcode=gcdebug
<moritz> NotFound: http://nopaste.snit.ch/13136 produces an assertions failure for me
<moritz> called as ../../parrot --runcore=gcdebug perl6.pbc push.rakudo
<NotFound> Trying...
<purl> succeeding.
<masak> purl: you sure talk a lot
<purl> masak: what?
<moritz> purl: forget trying
<purl> moritz: I forgot trying
<masak> 'the little bot who forgot to try'
% braceta has joined #parrot
<NotFound> Maybe is a Jedi knight.
<masak> Maybe?
<masak> how do I run the rakudo test suite?
<masak> `prove t/` doesn't seem to cut it :)
<Zaba_> make test
<purl> More like "make fail"
<masak> Zaba_: thx
<moritz> make test spectest_regression
<masak> purl: forget make test
<purl> masak: I forgot make test
<moritz> the segfault happens in the latter
<masak> moritz: ok
<NotFound> Many people forget that.
<masak> :)
% cognominal has joined #parrot
* moritz tries to reduce the test even more
<masak> whoa!? did it just stop by the pugs repo and got me a few more tests?
<moritz> masak: it fetches all test from t/spec/* and runs some of them
<NotFound> moritz: it's still running.
<NotFound> Well, stepping, better say.
<moritz> NotFound: takes a few minutes ;)
<masak> it's nice to see so many tests actually passing
<moritz> some of them are actually fudged ;)
<masak> though my adventures with rakudo often quickly run up against things that ought to be a failing test
<masak> moritz: meaning that they are hidden?
<masak> moritz: all tests successful here
<moritz> masak: yes, they aren't executed, mostly because they are parse failures
<masak> ah
<NotFound> It's a good cpu test executing something non-minimal with the gcdebug core.
<moritz> I reduced the test to:
<moritz> use v6;
<moritz> my @push = ();
<moritz> say +@push;
<NotFound> Failed!
<NotFound> src/string.c:514: failed assertion '!PObj_on_free_list_TEST(a)'
<moritz> same here
<moritz> bug report sent
<NotFound> moritz: mine is a g++ build.
<masak> hm, when I run it in parrot perl6.pbc, I get "Scope  not found for PAST::Var '@push'"
<masak> is that the same error?
<moritz> masak: the interactive thing doesn't really work
<Tene> pmichaud: ping
<NotFound> Will try now the re-reduced version.
<masak> ok :)
<moritz> masak: you have to put the statements on one single line in the interactive blah
<masak> yuck
<masak> though when I do that, everything's fine
<masak> and directly from cmd line too
<masak> did you say the bug is compiler-specific?
<NotFound> Not very fast, anyway.
<moritz> masak: it seems to be a garbage collector bug
<masak> in parrot?
<purl> hmmm... in parrot is false.
<moritz> yes
<NotFound> masak: looks like not, given that fails same way building with C and with C++
<masak> maybe I need to remake
<moritz> masak: the reduced test only fails for me when I run parrot with --runcore=gcdebug
<masak> oh
<moritz> masak: ..after 3 minutes execution time ;)
<NotFound> masak: it does not fail for unless I use the gcdbeug core.
<masak> pray tell, what is this gcdebug core?
<masak> seems it is neither fast nor safe to use :)
<NotFound> masak: the gcdebug core makes a full gc after each opcode.
<masak> haha
<NotFound> So it's no exactly very fast.
<NotFound> But helps reproduce bugs.
<NotFound> There's a chromatic's article in perl.com about it.
<masak> you should use some Monte Carlo method to randomly add and remove GCs between each opcode, keeping the ones that go wrong
<masak> but favoring minimal sets
<moritz> or just an option to run GC after each $n opcodes
<NotFound> A Las Vegas method will be more enlighten X-)
<masak> :)
<moritz> bug report is [perl #55164]
<NotFound> Same stacktrace with the reduced case.
<moritz> aye
<moritz> you can also remove the 'use v6;'
<moritz> which seems to be a no-op anyway
<masak> does it produce opcodes? otherwise, it doesn't matter
<masak> yup, it spat out the failed assertion here too
* moritz wonders if there is a way to execute the compilation with normal runcore, and only the execution wih the gcdebug runcore
<masak> moritz: oh. I'm starting to realize what took time now...
<NotFound> moritz: probably to add a --runcore option to the executable will be simpler.
% paco has left paco!~chatzilla@139.Red-80-36-122.staticIP.rima-tde.net
<japhb> moritz: output to .pir, and then run that manually?
<moritz> japhb: the generated pir doesn't run
<moritz> Class 'Perl6Array' not found
<japhb> moritz: eww
<moritz> so it seems to need some libraries, and I don't know how to include them
<japhb> Actually, I think that may be a valid RT -- when PCT is told to output to .pir, either that .pir should be directly runnable, or there should at least be an option to output the .pir with runtime setup included so that it *will* directly run.
<moritz> japhb: there's already a ticket for that
<moritz> japhb: don't know its status, though
<jhorwitz> moritz: try adding this to your generated PIR: load_bytecode '/path/to/parrot/languages/perl6/perl6.pbc'
<moritz> jhorwitz: error:imcc:syntax error, unexpected PARROT_OP, expecting $end ('load_bytecode')
<jhorwitz> paste the PIR?
<moritz> load_bytecode '/home/moritz/src/parrot/languages/perl6/perl6.pbc'
<moritz> as the first line in the file
<jhorwitz> ah
% paco has joined #parrot
<moritz> does it need to be inside a namespace, sub or whatever?
<jhorwitz> throw it in a sub w/ :init :load
<moritz> jhorwitz: thanks, that works
<moritz> ... and doesn't produce ane error with gcdebug anymore
<jhorwitz> excellent
<masak> but strange
<moritz> it's not excellent, because it means debugging is much harder
<NotFound> moritz: agree.
<jhorwitz> oh i see
<jhorwitz> kinda like when buggy code works in gdb.  :-P
<NotFound> As some tickets say, it must be the old gc problem manifesting again.
<masak> is there a way to compile but not run using gcdebug?
<moritz> masak: try ../../parrot --runcore=debug perl6.pbc -c $file
<masak> let's hope that variant breaks in the same way
<jhorwitz> afk &
<masak> i.e. that the error is in the compilation phase, and not an interaction between compilation and runtime
% bacek has left bacek!~bacek@123-243-38-218.tpgi.com.au
<moritz> $ ../../parrot --runcore=gcdebug perl6.pbc -c push.rakudo 
<moritz> syntax OK
% masak has left masak!~user@nl119-202-176.student.uu.se
% paco has left paco!~chatzilla@139.Red-80-36-122.staticIP.rima-tde.net
<pmichaud> kid51++ jkeenan++   # changing code_tests to codetest
% masak` has joined #parrot
<masak`> it compiles ok :/
<masak`> moritz: so the GC problem requires both compilation and running to manifest
% bacek has joined #parrot
% paco has joined #parrot
<moritz> masak`: seems like, yes
<masak`> (whoever finally solves it)++
<pmichaud> (gc problem)   switching runcores in the middle probably doesn't help find a gc bug, because gc bugs are very dependent on the order of execution
<pmichaud> i.e., even changing a single instruction is often enough to make the evidence of the bug disappear
<NotFound> That's why gcdebug is so useful.
<pmichaud> correct.
<pmichaud> slow, but useful.
<pmichaud> actually, "slow" doesn't describe it well.  Excruciatingly slow.  Feels like working on geologic timescales :-)
<moritz> pmichaud: do you think an option to executed GC only each $N opcodes would be helpful?
% Zaba has joined #parrot
<pmichaud> moritz: I don't know enough about parrot internals to know for sure.
<NotFound> Yes, but the solution is to have a more robust gc, in order to not have to use it frequently.
<moritz> NotFound: ah right. Go ahead and implement it... (/me ducks very deep)
<NotFound> moritz: I think an option to not use it until n opcodes executed will be useful.
<masak`> plus a script that does binary search on the exact point of error
<NotFound> That way one can try several n values in each case.
<NotFound> masak`: binary search if fast enough to do it by hand.
<masak`> NotFound: not sure I parsed that.
% Zaba_ has left Zaba_!~zaba@ip102.148.adsl.wplus.ru
<NotFound> masak`: I think there is no need for the script in the majority of usages.
<tetragon> Ick, t/pmc/exception.t started failing recently on my box (no segfault/sigbus)
<moritz> if you want good binary search, used gi + git-svn
<masak`> NotFound: you may be right
<masak`> time to go home and plug in the laptop
% masak` has left masak`!~user@nl119-202-176.student.uu.se
<pmichaud> (standalone PIR)  I worked on getting pir-output modules to run standalone last night, but got too tired to finish.
<pmichaud> however, I don't think they'll be useful for gc debugging, since it's another case of the sequence of operations changing to hide the bug
<Tene> pmichaud: I have a question about interactive execution.
<NotFound> pmichaud: I think most gc bugs are hidden because after a few opcodes the offending vars are reused, in that case will be useful. But I don't have any evidence, just a thought.
<Tene> pmichaud: actions.pm currently uses @?BLOCK to keep track of symbol tables and such, and a few other similar tricks.  That environment isn't available after compilation ends.  Is there a plan to fix this such that interactive mode can actually work?
<pmichaud> Tene: That's not really the problem.
<pmichaud> by the time a program is running, @?BLOCK may be long gone
<pmichaud> (e.g., imagine the case where we compile to .pir and then run the .pir
<pmichaud> )
<pmichaud> in interactive mode, we need to be able to execute code that modifies the current lexical environment
<pmichaud> but Parrot isn't really set up for that, since lexical scopes are tied to .sub, and the smallest unit of execution we have is a .sub
<Tene> pmichaud: So the actions.pm stage needs to be able to look things up in the current lexical environment, then?
<pmichaud> partially.  But consider this:
<pmichaud> if I do:    my @past = ();
<pmichaud> when compiled, it acts as if I had done
<pmichaud> { my @past = () }
<Tene> Right.
<pmichaud> (because the statement has to be in a sub)
<pmichaud> which means that the @past variable was local to that one interactive line
<pmichaud> so we have to have a way for interactive mode to be able to propagate the lexicals it defined into its outer environment
<Tene> Yes, I understand that problem.
<pmichaud> to me that's the fundamental blocker at the moment.
<Tene> I'm more referring to the fact that "say $a" doesn't do a lookup in the lexical environment, it looks at $?BLOCK, which won't exist anymore.
<pmichaud> uh, that's not correct.
<pmichaud> say $a does do a lookup in the lexical environment
<pmichaud>     find_lex $P17, "$a"
<pmichaud>     $P18 = "say"($P17)
<pmichaud> oh, if you mean the process of compiling say $a doesn't have $a defined.... you're correct.  But that's a very fixable problem.
<Tene> In the compilation stage.  actions.pm:1572
<pmichaud> I'm not worried about that.
<Tene> Okay.
<pmichaud> we'll just pass the current lexical environment to the compiler, and the compiler can check for lexical variables when one isn't defined in $?BLOCK
<Tene> Clever.
<pmichaud> so that's not the hard issue in my eyes :-)
<Tene> Okay, thanks.
<Tene> I didn't think it was the most significant issue, just wondering if you had a solution planned yet.
<pmichaud> I do now.  :-P
<pmichaud> but I knew I could solve that problem so hadn't thought about it.  The "propagate lexicals to another scope" is trickier.
% mire has left mire!~Frodo@195-169-222-85.adsl.verat.net
<pmichaud> it may be that we'll have a special flag that we pass to the compiler that tells it to generate code to move the lexicals to the outer scope when finished.
<pmichaud> Although I don't know if Parrot yet allows us to modify another sub's lexical table.
<pmichaud> at any rate, it hasn't been high on my list of things to solve just yet.  Interactive mode is still a "nice to have" feature -- not a "gotta have".  :-)
<NotFound> I've made a proof of concept implementation and it catches the bug.
<NotFound> paste?
<clunker3> http://pasta.test-smoke.org/ or http://paste.husk.org/ or http://nopaste.snit.ch:8001/ or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/
<purl> it has been said that paste is (see: nopaste) or like glue but a little safer to sniff. or http://nopaste.snit.ch:8001/ or http://scsys.co.uk:8001/ anywhere shadowpaste is or mmm, sticky paste or You there! Eating the paste. or <see> 2 girls, 1 paste
<nopaste> "NotFound" at 213.96.228.50 pasted "gcdebug N test" (35 lines) at http://nopaste.snit.ch/13138
<NotFound> PARROT_GCDEBUG_SKIP=85000 ../../parrot --runcore=gcdebug perl6.pbc testprog.p6
% mire has joined #parrot
<moritz> NotFound: testing now...
<NotFound> With the 85000 value, just a few seconds before fail.
<NotFound> 85080 is almost exact.
<moritz> NotFound: your example (with 85000) didn't reproduce the error for me
<moritz> NotFound: I'll try with different values
<NotFound> moritz: try a lower value.
<moritz> the run time seems strongly non-linear in PARROT_GCDEBUG_SKIP
<NotFound> Will build with C and retest.
<NotFound> Fails with 85082
<NotFound> moritz: the algorithm is simple, if it does not fail or finishes in a short time, break and try other value ;)
% bacek has left bacek!~bacek@123-243-38-218.tpgi.com.au
<NotFound> 85084 is the exact point in my build.
<moritz> that seems build dependent
<NotFound> Strange.
<purl> But true.
<moritz> so it's a a nice way to identify GC bugs
<moritz> but not good for really tracing them
<moritz> anyway, nice patch
<moritz> NotFound++
<NotFound> For tracing, aborting at that point may help.
<NotFound> Or run under debugger and setting a breakpoint when counter exhausts.
<moritz> that's actually a nice idea
% bacek has joined #parrot
% Andy has left Andy!~AndyL@host3130.follett.com
<NotFound> Patch sended to the mailing list.
% scrottie has joined #parrot
<scrottie> Hi pmichaud >=)
<scrottie> pmichaud, thought I'd try to catch you active, re: the Larry Wall doll, but my clunky old irc-II client doesn't save messages and I'm heading out in a bit, so I'll catch you on email.  cheers!  I'm scott@slowass.net, by the way.
% scrottie has left #parrot
<moritz> btw the newly fudged S04-statements/try.t could be added to /spectest_regression.data
<pmichaud> moritz++
<moritz> S29-list/map.t also passes, but I wouldn't add that before clarifying if the last two tests are actually correct
<moritz> if they aren't, they should test for the exact opposite
% Theory has joined #parrot
% jan has joined #parrot
<moritz> it's a bit frustrating that p6l seems to discuss various stuff that nobody implements atm, but simple stuff doesn't get answered
<moritz> (ok, I sent this one to p6c, maybe I am to blame here)
% donaldh has joined #parrot
<pmichaud> what isn't being answered?  I'll try to push it
<pmichaud> I might have just overlooked it.  Likely, given how code-focused I've been for the past week.
<moritz> the question if $_ is rw in a map block
<pmichaud> I think yes.  Just a sec --
<pmichaud> S02:  You may add a trait argument of <rw> to allow called routines to modify your value. $_, $!, and $/ are context<rw> by default.
<pmichaud> so, assuming that @array is rw, I'd presume that $_ in a map block would be rw as well  (assuming the map block doesn't have a signature)
<moritz> so the 'rw' doesn't have to be in map's signature?
<pmichaud> you mean on the block argument?  That would seem to make the block itself rw
<moritz> ah right
<pmichaud> map doesn't assign to the block (I don't think) -- it simply invokes the block with arguments
<moritz> pmichaud: thanks for the clarification, I'll correct the test
% donaldh has left donaldh!~chatzilla@host213-123-171-12.in-addr.btopenworld.com
* moritz changed 'svn up' to 'svn up || true' in his local makefile to have 'make spectest' working offline
<pmichaud> smart.
% donaldh has joined #parrot
<pmichaud> isn't there a character that can be put in front of a makefile line that says "ignore the result code?"
<pmichaud> I think it's @ or + or something?
<moritz> dunno, my Makefile-fu is rather bad
<moritz> pmichaud: http://sial.org/pbot/31176
% Zaba_ has joined #parrot
<NotFound> pmichaud: @
<NotFound> Ops, no, this is to not echo the command.
<pmichaud> hyphen
<pmichaud> spectest_regression:
<pmichaud>         -svn up
<pmichaud> or, more precisely
<pmichaud>         -cd test && svn up
<NotFound> Yes, that is.
% Zaba has left Zaba!~zaba@ip102.148.adsl.wplus.ru
<pmichaud> ($_ in map block)   Also, S04 says this about the 'for' statement:  
<pmichaud> If you rely on $_ as the implicit parameter to a block, then $_ is considered read/write by default. That is, the construct:
<pmichaud> for @foo {...}
<pmichaud> is actually short for:
<pmichaud> for @foo <-> $_ {...}
<pmichaud> so you can modify the current list element in that case.
<pmichaud> so, I'd take it that bare blocks tend to treat $_ as rw
<particle> pmichaud:  the makefile prefix char for nmake is ! iirc
<moritz> so it's incompatible? d'oh
<particle> @ is so it won't print
<particle> @echo 'foo' # prints 'foo', not "echo 'foo'\nfoo"
<particle> oh, i see, -
<particle> no, i don't recognize that for nmake
<pmichaud> –[[number ]]command
<pmichaud> Turns off error checking for the command. Spaces and tabs can appear
<pmichaud> before the command. By default, NMAKE halts when any command returns
<pmichaud> an error in the form of a nonzero exit code. This modifier tells NMAKE to
% arbingersys has joined #parrot
<pmichaud> ignore errors from the specified command.
<particle> that's microsoft nmake, and not at&t or bell or whatever?
<pmichaud> I'm pretty sure it is.
<particle> well, go ahead and try it
<particle> 'course, win32 build is broken atm
<pmichaud> the PDF I took the above from was describing the Microsoft Program Maintenance Utility (NMAKE)
<pmichaud> version 1.20.
<particle> ooh, that's damned old
<particle> but probably still accurate
<pmichaud> so it's likely available :-)
<dalek> r28018 | pmichaud++ | trunk:
<dalek> : [rakudo]:
<dalek> : * Add S04-statements/try.t to spectest_regression (moritz++)
<particle> Microsoft (R) Program Maintenance Utility Version 8.00.50727.42
<particle> Copyright (C) Microsoft Corporation.  All rights reserved.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28018
<pmichaud> rakudo spectest_regression:  55 test files, 1012 tests
<pmichaud> is "1012 tests" the number of passing tests, or the number of tests including todo and skip ?
<particle> pmichaud: did you see that allison++ has added branchcc and return_branchcc for you to pdd25cx ?
<pmichaud> I did, yes!
<pmichaud> that should be... interesting.
<particle> i figured it would make you happy :)
<pmichaud> well, the devil is in the details -- I have to actually try using it first.  :-)
<pmichaud> but as I reviewed my various exception handling code, the thought of having to create a separate sub to handle each exception was looking really nasty.
<moritz> pmichaud: I think fudged tests emit something like "ok $count - skip $reason", so they'll be counted as passing
% Ivatar has left Ivatar!~graham@tu055.demon.co.uk
<particle> i figured we'd write a library file with an exception handler generator or something
<pmichaud> that still seemed pretty icky.
<particle> well, anyway, there may not be a need
<pmichaud> that's what I'm hoping.
<particle> i've gotta go epoxy some bolts in cement &
<japhb> kinky
<pmichaud> moritz: you're correct, skipped tests count as passing tests.  Wish we could fix that somehow.
<moritz> pmichaud: maybe just emit 'not ok # TODO' stuff?
<pmichaud> that might be nicer.
<moritz> it sounds like fudge could easily do that, but the only time I actually looked at its source I got scared off by too many local()s
<moritz> maybe I just need to ignore that ;)
<moritz> and hope that larry's code is so good that all my changes DWIM
<nopaste> "pmichaud" at 76.183.97.54 pasted "What's the purpose of eval here?" (13 lines) at http://nopaste.snit.ch/13140
<pmichaud> (from S04-statements/try.t)
<moritz> pmichaud: probably that the test won't exit if the implementation is faulty
<pmichaud> ....but we use #?rakudo skip   for that now
<pmichaud> or #?pugs skip  or   #?<impl>  skip
<pmichaud> and the whole point of a "try" block is to catch exceptions, including exits.  :-)
<moritz> I have to admit that I fudged that file rather brainlessly
<pmichaud> well, I agree that the test needs fudging for Rakudo
<pmichaud> I'm just thinking that there may be a lot of 'eval's that were done from before we had the ability to  #?impl skip
<pmichaud> and I'd like to get rid of those.  :-)
<moritz> I didn't even check if CATCH blocks should work inside or outside the try { ... }, because I hoped that the one who moved it to t/spec/ checked it ;)
<pmichaud> inside.
<pmichaud> that way the CATCH block has access to the try block's lexicals.
<moritz> seems logical ;)
<moritz> how well does eval() work in rakudo?
<pmichaud> okay, as long as you aren't trying to access outer-scoped lexicals.
<pmichaud> my $a;  eval "$a = 4";    # won't work.
<moritz> t/spec/S29-num/pi.t looks like much fun
<moritz> #?rakudo 6 skip 'eval not implemented'
<moritz> is_approx((eval("Num::pi "), $PI),  "Num::pi");
<pmichaud> unless Num::pi is being declared locally in the test, I don't think we need eval there.
<moritz> aye
<pmichaud> i.e.:   is_approx(Num::pi, $PI, "Num::pi")     should be sufficient.
<moritz> that's one of the cases you mentioned
<moritz> > say Num::pi
<moritz> [oops; continuation 0xb64b97a0 of type 22 is trying to jump from runloop 274 to runloop 50]
<moritz> Null PMC access in clone()
<pmichaud> oh, that's interesting.
<pmichaud> aha, old artifact.
<pmichaud> fixing.
<moritz> should type names propate from eval to the outside?
<pmichaud> example?
<moritz> mx $x = eval 'class A { }; A.new()"; say $x.WHAT
% Zaba has joined #parrot
<pmichaud> no, eval has its own lexical scope
<pmichaud> eval 'my $a = 3;';    does not give the outer scope a lexical $a
<pmichaud> at least, I don't think it does.
<pmichaud> (checking)
<moritz> so, what should the example print? "A\n"?
<moritz> and is the output useful?
<pmichaud> oh, the output would still be 'A', yes
<pmichaud> but the classname 'A' would not be in scope.
<moritz> that's kinda... ugly
<pmichaud> S04:  Blocks are delimited by curlies, or by the beginning and end of the current compilation unit (either the current file or the current eval string).
% Zaba_ has left Zaba_!~zaba@ip102.148.adsl.wplus.ru
% arbingersys has left #parrot
% braceta has left braceta!~Braceta@bl10-34-59.dsl.telepac.pt
<pmichaud> (Null PMC access) -- not sure where that NULL is coming from.  :-|
<pmichaud> oh, yes I am.  :-)
<moritz> speaking of "sure" - I should surely go te bed right now ;-)
<moritz> good localtime() everybody, see you tomorrow
<pmichaud> a demain
<bacek> moritz, see you... It's 8AM of tomorrow here :)
% iblechbot has left iblechbot!~iblechbot@ppp-62-216-205-141.dynamic.mnet-online.de
<dalek> r28019 | pmichaud++ | trunk:
<dalek> : [rakudo]:
<dalek> : * Clean up slurpy arguments to 'print'
<dalek> : * Fix lookups of non-existent package items
<dalek> : * moritz++
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28019
% donaldh has left donaldh!~chatzilla@host213-123-171-12.in-addr.btopenworld.com
% Whiteknight has joined #parrot
<pmichaud> particle: ping
% teknomunk has joined #parrot
% nemesys has left nemesys!~nemesys@HSI-KBW-085-216-121-022.hsi.kabelbw.de
% jan has left jan!~chatzilla@bdv75-4-82-226-115-243.fbx.proxad.net
<Whiteknight> chromatic++ has been on a rampage this weekend! My inbox is filled with tales of his travels
<Whiteknight> with s/travels/commits/
<jonathan> hi all
<jonathan> Back from France.
<jonathan> Three words. I hate flying.
<jonathan> On the upside, I did some Rakudo hacking at Orly airport. And fixed a GC bug in Parrot.
<jonathan> ci's coming in a bit
<NotFound> A tax-free patch?
<jonathan> It's all EU.
<jonathan> If I want tax-free stuff, I have to nip over to the Ukraine
<NotFound> I hate eastern europeans, they don't voted the "Chiqui Chiqui" in Eurovision X-)
<jonathan> By that logic, as a Brit, I hate almost all of Europe. ;-)
<NotFound> Yes, but the easterns the worse.
% toft has left #parrot
* jonathan points out that he moved to east europe from west europe. :-)
<NotFound> Anyway, the "Chiqui Chiqui" is destined to conquer the world, like the Macarena X-)
<NotFound> I'm wondering why ldd parrot shows libstdc++.so in non-c++ build.
<NotFound> And the same with ldd blib/lib/libparrot.so
<tetragon> I see the same on OS X (with otool)
<tetragon> Some of the build steps are done (on OS X) with g++ for some reason I haven't looked at
<tetragon> The OS X build uses "c++" instead of "ld" as the linker
<NotFound> Looks strange, given all "Why C" stuff in parrot faqs.
<jonathan> NotFound: I believe it's for ICU linking to work.
<jonathan> Since ICU is written in C++.
<tetragon> It looks like the Linux config defaults to using g++ when $conf->data->get('link') isn't set
<NotFound> jonathan: Stranger again.
<tetragon> jonathan: I've worked with ICU.  It has a Java, C++, and a pure C API
% Limbic_Region has joined #parrot
<Whiteknight> an optimized build fails "make perl6". Is this a known issue or should I create a ticket?
<jonathan> Whiteknight: provided there's no obvious link with the other currently open ticket on something not working on an optimized build yes. (I forget what the other currently being discussed issue is...)
<NotFound> By the way, I think some document must say that parrot is not written in C, but in the common subset of C and C++.
<Whiteknight> chromatic opened one about the optimized build failing a few of the threading tests
<jonathan> Ah, OK.
<jonathan> That doesn't feel like an obvious link to me.
<jonathan> So please do submit a ticket.
<Whiteknight> done and done
<Whiteknight> i need to submit some follow-up info (i didn't copy the error messages the first time, and need to rebuild)
<NotFound> Specifying --link=gcc , same result.
<Eevee> so.  what needs doing that will get my..  ankles wet
<dalek> r28020 | coke++ | trunk:
<dalek> : RT #53602 (remove or convert tools/docs/search-ops.py)
<dalek> : implementation by kid51++
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28020
% Limbic_Region has left Limbic_Region!~Limbic_Re@c-68-49-236-220.hsd1.md.comcast.net
% tetragon has left tetragon!~seneca@69-196-141-26.dsl.teksavvy.com
% kid51 has joined #parrot
* kid51 logs on and for the first time in days is not deluged with > 20 Viagra/Cialis spams
<kid51> More emails from Parroteers than spam -- perhaps there's hope for humanity yet!
% Ademan has left Ademan!~dan@h-69-3-232-86.snfccasy.dynamic.covad.net
<jonathan> OK, need sleep...will review and throw in my airport patches tomorrow.
<Eevee> night
% tetragon has joined #parrot
% ank has joined #parrot
% AndyA has left AndyA!~andy@82.152.157.85
% bacek_ has joined #parrot
% mire has left mire!~Frodo@111-173-222-85.adsl.verat.net
% AndyA has joined #parrot
% bacek_ has left bacek_!~bacek@mcas-151.usr.optusnet.com.au
% bacek_ has joined #parrot
% radhios has joined #parrot
% Whiteknight has left Whiteknight!~nobody@c-71-230-33-251.hsd1.pa.comcast.net
<japhb> Eevee: (what needs doing to get acquainted) Documentation consistency patches.  Several places there are two documents that discuss overlapping topics (typically because they are related but in some sense distinct).  Often, the overlap is inconsistent between the 2+ documents.  It should be factored out, made consistent, or one location should point to the other.  This would get you looking through lots of docs (and having to understand th
<japhb> em), and comparing against code to check which of two conflicting things is "right".  That would really get you ready for code patches then.
<japhb> For example, someone pointed out a couple days ago that the PDD doc for NCI has an incomplete list of NCI types, compared to the list in config/gen/call_list/core.in and the actual NCI code.  This should be fixed up.
<japhb> But that's just my personal hot button.  Any RT for "cage cleaner" work is a good foot-wetting thing to work on as well.
% Whiteknight has joined #parrot
<GeJ> japhb: isn't NCI worked on by tewk  for gsoc?
<kid51> Eevee:  This search string will show you all RT tickets with the [CAGE] designation:   http://tinyurl.com/4wwnct .
% grim_fandango has joined #parrot
<kid51> We can provide similar strings for [PATCH], [BUG], [TODO], etc.  If you let us know what your skill sets are, we can steer you in a good direction.
% DietCoke has left DietCoke!~coke@cpe-72-228-52-192.nycap.res.rr.com
<japhb> GeJ: Yes.  But that A) doesn't help us now, and B) shouldn't prevent us from fixing documentation (especially since good documentation would help tewk as well)
% radhios has left radhios!~roberto@190.19.126.84
% jhorwitz has left jhorwitz!~chatzilla@96.245.16.45
% rdice has left rdice!~richarddi@CPE001ff33cb98b-CM00159a01d44c.cpe.net.cable.rogers.com
* Eevee return
% ank has left ank!~ank@ppp121-44-210-24.lns1.hba1.internode.on.net
<Eevee> kid51: thanks, will have a look through these.  re skill sets, uh, I have a large assortment of hats I am comfortable wearing
<Eevee> incidentally
% Andy has joined #parrot
<Eevee> is there any outsider-friendly documentation for p6?  there's a spec, but besides that just seems to be the exegeses, which are full of corrections to themselves and of questionable recency
<kid51> Eevee:  C?  Perl 5?  Other dynamic languages?  Compilers?
<kid51> Specific OSes?
<Eevee> aha
<kid51> Parrot:  "I am multitudes."
<Eevee> *familiar* with C, but haven't really touched it in years
<Eevee> pretty solid with perl 5
<Eevee> know a decent bit of python; had a glance at ruby but nothing much
% davidfetter has joined #parrot
<Eevee> used gcc a bit, mostly on others' code when their makefiles suck, and msvc a handful of times
% grim_fandango has left grim_fandango!~matt@bas2-kingston08-1167934907.dsl.bell.ca
<kid51> See http://rt.perl.org/rt3/Ticket/Display.html?id=48260.  This is one of those tickets that always needs work, because there are always more C functions needing documentation.
% bobwhoops has joined #parrot
* kid51 must sleep
<purl> $kid51->sleep(8 * 3600);
<Eevee> aha, thanks
% cotto_work has joined #parrot
% cotto-work has left cotto-work!~cotto@tide531.microsoft.com
% grim_fandango has joined #parrot
% kid51 has left kid51!~jkeen@pool-68-237-3-221.ny325.east.verizon.net
% davidfetter has left davidfetter!~chatzilla@start.fetter.org
% Theory has left Theory!~Theory@c-24-21-175-208.hsd1.or.comcast.net
% Zaba_ has joined #parrot
% Zaba has left Zaba!~zaba@ip102.148.adsl.wplus.ru
% Zaba_ has left Zaba_!~zaba@ip102.148.adsl.wplus.ru
% Zaba has joined #parrot
<dalek> r28021 | pmichaud++ | trunk:
<dalek> : [rakudo]:
<dalek> : * Fix Test.pm so that it reports skipped tests correctly.
<dalek> diff: http://www.parrotvm.org/svn/parrot/revision?rev=28021
% Theory has joined #parrot
% jan has joined #parrot
% Whiteknight has left Whiteknight!~nobody@c-71-230-33-251.hsd1.pa.comcast.net
% Andy has left Andy!~Andy@64.81.227.163
% Theory has left Theory!~Theory@c-24-21-175-208.hsd1.mn.comcast.net
% Auzon has left Auzon!~ak9@24-171-76-148.dhcp.mtvr.il.charter.com
% Zaba_ has joined #parrot
% Zaba has left Zaba!~zaba@ip102.148.adsl.wplus.ru
<bacek_> summon pmichaud 
<bacek_> you broke List.map... MMD cant add Failure to Int...
<bacek_>  is ~(1,2,3,4).map({ $^a + $^b + $^c + $^d }), "10",  "map() works with 4-ary functions";
% ank has joined #parrot
<bacek_> http://www.parrotvm.org/svn/parrot/revision?rev=28021
% uniejo has joined #parrot
% Zaba has joined #parrot
<bacek_> very strange changes in lines 81-82...
% Ademan has joined #parrot
% Zaba_ has left Zaba_!~zaba@ip102.148.adsl.wplus.ru
<bacek_> t/spec/S29-num/roots.............................. All 10 subtests passed 
<bacek_> 	(less 11 skipped subtests: -1 okay)
<moritz> pmichaud++ # correct reporting of TODO tests
<bacek_> moritz: but incorrect for skip...
<moritz> bacek_: it works for me
<bacek_> moritz: shit... I have to take a brake at work...
<bacek_> it's really works...
% Zaba_ has joined #parrot
<bacek_> with current 't/fudge'.
<bacek_> but lines 81-82 should be 'skip(1, ..', not 'proclaim(1,..)'
% Zaba has left Zaba!~zaba@ip102.148.adsl.wplus.ru
% bacek_ has left bacek_!~bacek@mcas-151.usr.optusnet.com.au
% Zaba_ has left Zaba_!~zaba@ip102.148.adsl.wplus.ru
% iblechbot has joined #parrot
% Zaba has joined #parrot
% cosimo has joined #parrot