% PerlJam has left PerlJam!duff@feather.perl6.nl
% jonathan has left jonathan!jonathan@feather.perl6.nl
% chip has left chip!chip@feather.perl6.nl
% leo_ has left leo_!lt@feather.perl6.nl
% chip has joined #parrotsketch
% jonathan has joined #parrotsketch
% leo_ has joined #parrotsketch
% PerlPilot has joined #parrotsketch
% pmichaud has joined #parrotsketch
<pmichaud> just a note that I'll be late again this week
<pmichaud> I expect to be here at hh:15
% PerlJam has joined #parrotsketch
<chip> likewise
% smash has joined #parrotsketch
<obra> (allison also)
<obra> we'll start 10 minutes late today
<particle> ok
% chromatic has joined #parrotsketch
<chromatic> Happy 18:00 all.
<obra> chromatic: we're starting 10 minutes late.
<obra> (noted a moment before you joined)
<obra> allison, chip and pmichaud are all running late
<leo_> hi folks
<obra> hey leo
% allison has joined #parrotsketch
<obra> hey allison
<obra> still no chip and no patrick, I think
<allison> hi obra
<obra> but let's get going anyway. they'll be here soon
<obra> so. particle, leo, allison, then other people who pipe up
<particle> ~ hackathon was a smashing success (it can't be said enough times)
<particle> ~ 337 commits last *week*, 264 commits last *month*
<particle> ~ applied many parrot improvements and bugfixes, courtesy of hackathon attendees
<particle> ~ led parrot hackers in a pmc/object design meeting and provided summary with recommendations for allison (docs/dev/pmc_object_design_notes.pod)
<particle> ~ turned jim keenan (of phalanx fame) loose on pmc2c.pl -- expect a refactoring and proper test coverage in the next few weeks
<particle> ~ release prep, including doc updates, test triage, following release instructions and providing chip a patch... all tests pass on windows
<particle> ~ ported languages/forth/ to windows
<particle> ~ added two new configure.pl options to allow config w/o gdbm or gmp, even if they're installed (workaround for #40803)
<particle> ~ added new perl6 ops: infix:<min> and infix:<max> to parser
<particle> ~ playing with the new past and languages/abc/
<particle> ~ unfortunately, there was not enough time in chicago to spend time with everyone on my schedule, so some planned tasks remain unfinished
<particle> .end
<leo_> * bits of hacking, testing, bugfixing during hackathon
<leo_> * explained various corner cases and the deeper reason
<leo_> still doubleplusunplenty parrot time
<leo_> ^D
<obra> allison?
<allison> - I've been reading through particle's summary of the object discussion. It's  great, thanks to all who contributed and especially particle.
<allison> - I followed parts of the hackathon via IRC, but it really wasn't the same as being there. :(
<allison> EOR
<obra> chromatic: you're up
<obra> (then chip)
<chromatic> I fixed a lot of failing tests on multiple platforms.
<chromatic> That includes a lot of tests I thought I could never fix.
<chromatic> I also helped a few other people do a few other things.
<chromatic> Parrot::Embed is very close to working.
<chromatic> Unfortunately, some change in PGE broke Pheme and I'm debugging that with little avail.
<chromatic> ^D
<chip> hi all.  a non-report note: leo++ for referencing a classic work of English lit, when I wouldn't be able even to name any Austrian literary works, let alone quote them
<chip> Past:
<chip> * Hackathon rocked.  andy++ sachmet++
<chip> * Thanks to Andrew N. Johnson, "interp" is now the standard interpreter variable name.  He sent me a 2nd patch but I'm holding off applying it until Jonathan merges his packfile rework.
<chip> * We (Jerry, chromatic, I, and probably others who I didn't pay the deserved attention in all the activity) have been knocking down the test failures, and we're as close to zero as we're going to get.  Hence:
<chip> Present:
<chip> * The next release will be today, barring circumstances beyond my control. (And there was much rejoicing.)
<chip> Future:
<chip> * Squeaky wheels include Patrick's object difficulties and Jerry's need for an IPC::Open3 analog for ParrotInterpreter objects.
<chip> * I'm getting close to understanding the object system and its problems. I'm at the point where I'm coming up with increasingly large changes that either don't help or have serious probles.  Soon I'll be coming up with ideas I can't shoot down myself, and I'll write them up for the list and compare them with others' proposals.
<chip> * I'd like to help Jerry with Interpreter::Open3 but I haven't looked into it yet.
<chip> ^D
<obra> who else is here?
<obra> pmichaud: ping. coke. tewk jonathan 
<particle> chip++ for unmentioned sprintf fixes on windows
<obra> failing other folks: a brief aside into question time.
<obra> chromatic: go for it
<chromatic> I talked to Jonathan.  He, Matt, and I need to meet to discuss something (I forget what, but it was in the object discussion).  After #parrotsketch next week works, else I have a conference call available.
<particle> oh, jonathan can't make it today, as he's vacationing in chicagoland, but he created a branch and began implementation of the newly-approved packfile pdd
<obra> excellent
<particle> also, he worked closely with kelly ireland, who did parrot C work for the first time
<particle> s/kelly/kelli/
<obra> anyone else got anything?
<allison> three cheers for new parrot hackers :)
<chip> sprintf("%e") is a problem: Win32 is different from all others (it defaults to three-digit exponents instead of two).  We should either let it be platform-dependent or we have a significant mini-project making it consistent.
<chromatic> Reminds me of that dynload problem I found on Darwin.  Ugh.
<chip> I hacked enough to make our current tests pass but this will come up again.
* particle casts a vote for platform-independence
<obra> allison: I was supposedf to harass you about where in the PDDS the platform list got canonized
* allison agrees
<chip> good
<allison> obra: not sure where to put it
<allison> obra: but I'll add it to my list for the week
<obra> A section "Portability" in PDD 1?
<chromatic> Any thoughts on objects so far, allison?
<allison> obra: ok
<allison> chromatic: I'm typing response comments into particle's summary
<allison> c: will check it in tonight
<chip> PDD07 has a section on portability
* pmichaud is here, finally
<chip> but it's more about how rather than a list of specific targets
<obra> yeah. 1 feels more like "design goals" than "howto"
* chip nods
<allison> chip/obra: I'll check both to see if they need any specific comments added
<chip> kthnx
<particle> where does perl5 keep platform info?
<obra> allison: well, 1 has no section on portability yet. it was just an idea ;)
<obra> particle: it's more of an ad-hocracy on p5. there is no canonical list
<allison> obra: indeed
<obra> but we're well past Perl 5 1.0 ;)
<allison> we're probably going to end up needing a portability PDD
% coke has joined #parrotsketch
<obra> ok. anything else before I let patrick report?
<chip> perldoc perlport has info
<chip> inclyuding a list of platforms that work
<chip> (that's all)
<obra> ok. patrick. how's things?
<pmichaud> good.  unfortunately I don't have a pre-typed report (spent my morning writing code instead of report) so I'll type it in manually.  Apologies to the group
<pmichaud> Last week I added some more tests to the Capture PMC to demonstrate issues
<pmichaud> * updated compilers to use new :init pragma (tewk++)
<pmichaud> * Changed the compile sequence somewhat
<pmichaud> * refactored PGE to use the new compile sequence
<pmichaud> * Added an HLLCompiler class
<pmichaud> * Moved pgc to be a P6Grammar class
<pmichaud> * Changed Glob/P6Grammar so that the .pbc's can be invoked as command-line compilers
<pmichaud> (useful for debugging)
<pmichaud> * switched perl6 to use the new HLLCompiler
<pmichaud> * created a new PAST/POST implementation, in compilers/past-pm
<pmichaud> * so far the new implementation looks really sweet -- much easier to deal with
<pmichaud> * started converting perl6 to the new past-pm implementation
<pmichaud> * expanded particle++ 's "bc" grammar into a bc implementation based on past-pm
<pmichaud> * the bc implementation is languages/abc/ , and hopefully it'll be the "tutorial" for past/complers
<pmichaud> * reported various bugs and tried to be available for hackathon questions
<pmichaud> * discussed having a Dallas/Texas hackathon with Richard Dice
<pmichaud> plans:  continue with new past implemention, implement variables, loops
<pmichaud> update perl6 compiler to use new past and continue attacking the test suite
<pmichaud> <end>
<obra> excellent
<obra> ok. more reports? questions?
<chromatic> patrick: it looks like a ws rule change somewhere broke the Pheme parser.
<chromatic> Can you take a quick look and see what i need to change in the grammar?
<pmichaud> it's likely anything where whitespace appears in a character class
<pmichaud> S05 changed the interpretation on me
<pmichaud> so  <[xyz abc]>  is now <[xyzabc]>  (ws is ignored)
<chromatic> Hm.
<pmichaud> to match a space, backwhack it
<pmichaud> (e.g.,  <[xyz\ abc]>)
<pmichaud> we had the sam eproblem with tcl
<chromatic> I'll try that; it might be a quick fix.
<pmichaud> I'll look at it, too
<chromatic> That looks a lot better already.
<chromatic> I think that may have done it.
<chromatic> Running the test suite now....
<chromatic> Yes, thank you.
<chromatic> I'll put Pheme in the languages Makefile so that it runs in the tests.
<pmichaud> that would be great, thanks
<coke> [report] Some minor work on tcl this past week, inspired by Paul's bug-opening frenzy. will continue in this vein for a while. also got a new snapshot of how the tcl tests (not partcl's) are faring: ISTR we're back up to 13% with our current less hacky approach.
<coke> I will continue, apparently, to be late to PS each week. =-)
* pmichaud has at least one Q.
<obra> pmichaud: go for it
<pmichaud> can we change the #parrotsketch time to be 18:30 UTC instead of 18:00?
<coke> fine with me, obviously. =-)
<pmichaud> it is unlikely that I'll be able to make 18:00 until next spring
<obra> how does that work for everyone else?
<chromatic> It's fine here.
<allison> is 18:30 better than 19:00?
* leo_ acks
<particle> worksforme
<pmichaud> anytime after about 18:20 works for me
<pmichaud> (so yes, 19:00 would work too, but I think I like 18:30 better)
<allison> ok
<obra> allison: is 18:30 bad for you?
<allison> just difficult to schedule other meetings around it, but let's try it
<obra> ok. let's try it next week
<pmichaud> thank you
<obra> I'll need to get robrt to op me to fix the msg
<pmichaud> next q:  any thoughts (chip, allison) about the compiler question I posted to p2p last week?
<pmichaud> i.e., is a compiler an object or is it a sub?
<leo_> or an object that can be invoked?
<pmichaud> the problem with invokable object is that I haven't found how to pass parameters to it
<pmichaud> (specially slurpy named)
<leo_> yep that's missing I presume
<pmichaud> I thought I could do invokable object until I ran across that wall
<pmichaud> /across/into/
<chromatic> I think I ran into that a few weeks ago.
<allison> I say it's an object
<pmichaud> I'll update PDDs to match, then
<leo_> slurpy named could use a '%' sig char
<leo_> skurpy array is already done as '@' iirc
<pmichaud> I'll also draft a "compilers pdd" (I think I had said I would do this at oscon, but wasn't ready for it until now)
<allison> note that the namespaces PDD expects to have a Compiler PMC already
<pmichaud> yes, that's what prompted me to more formally ask the question
<pmichaud> so, the items listed in pdd21 will be incorporated into the compilers pdd
<pmichaud> (okay to duplicate, probably)
<pmichaud> (at least for now)
<allison> excellent, thanks! (yeah, clone and expand is a good strategy)
<pmichaud> okay, compiler as object is what I wanted to hear
<pmichaud> I'll have to figure out how to migrate the PGE compilers to that strategy without lots of breakage, but we'll get there :-)
<chromatic> Agreed.
<pmichaud> (since PGE::P6Regex and PGE::P5Regex are currently compiler-as-sub)
<pmichaud> I think that's all the questions I have for today
<obra> Anyone have anything else?
<chromatic> How do we encourage more people to 'make smoke'?
<chromatic> Also, why don't the smokes show the expected/received output?
<allison> send them t-shirts ;)
<pmichaud> more people or more platforms?
<pmichaud> (or both)
<allison> pmichaud: both
<chromatic> I know a guy at OSDL, but they only differentiate in hardware platforms, not OS type.
<allison> statistically, you need broad coverage to figure out if a problem is specific to a platform or just a particular install
<particle> smokeclient and server need an overhaul, i think
<leo_> is everything to 'make smoke' in Bundle::Parrot on CPAN (or was that Parrot::Bundle) ?
<particle> client doesn't show expected/received
<chromatic> I'd like to have some way of tying a particular patch to breakage on one or more platforms, so we don't have to do this triage and debugging again, ever.
<particle> server doesn't make it easy to view status of bug across smokes
<pmichaud> leo++ -- the biggest barrier to "make smoke" for me was the number of cpan modules involved
<particle> it's Bundle::Parrot, and it's still dev 0.4.6_02
<leo_> Es konnten keine Ergebnisse für "Bundle::Parrot" in der Rubrik module. gefunden werden. Versuchen Sie bitte einen anderen Suchbegriff.
* obra blinks
<particle> grr, search.cpan is saying _01... where's my new upload?
<obra> Anyone else have anything parrotsketch specific?
<allison> chromatic: well, multiple hardware platforms on some version of Linux would be a good boost
<particle> any solaris box would be nice, too
<chromatic> I'll talk to my contact and see what we can do.
<pmichaud> I wonder if any of my connections here in Dallas could get me access to more platforms
<particle> oh, and a shared windows dev box would be fabulous
<leo_> there is a solaris box - please contact robrt
<pmichaud> hmmmm
<allison> do we have a good clean set of instructions for people to set up automated regular smokes?
<allison> A landing page for the "Parrot Smokers"?
<chromatic> 1) Stick the filter end in your mouth
<chromatic> 2) Light the other end
<PerlJam> Does "make smoke" also smoke languages?
<particle> i know ^conner set a smokehouse up before, but i didn't pay attention to any documentation
<particle> pj: no, make languages-smoke does
<obra>  ok. I think we're back into #parrot territory here
<leo_> PerlJam: languages-smoke
<obra> I'll catch everybody next week at our NEW TIME. 18:30 GMT
<PerlJam> Do you guys think languages-smoke is as important as smoke?
* particle waves
<allison> thanks jesse
<chromatic> Yes.
<leo_> thanks jesse
<pmichaud> thanks obra, all
<allison> perljam: yes, so instructions should help people set up both
% allison has left allison!~chatzilla@adsl-75-7-0-115.dsl.pltn13.sbcglobal.net
% coke has left #parrotsketch
% allison has joined #parrotsketch
% mdiep has joined #parrotsketch
% allison has left allison!~chatzilla@207.47.10.130.static.nextweb.net
% PerlPilot has left PerlPilot!duff@feather.perl6.nl
% jonathan has left jonathan!jonathan@feather.perl6.nl
% pmichaud has left pmichaud!pmichaud@feather.perl6.nl
% leo_ has left leo_!lt@feather.perl6.nl
% chip has left chip!chip@feather.perl6.nl
% chip has joined #parrotsketch
% pmichaud has joined #parrotsketch
% PerlPilot has joined #parrotsketch
% leo_ has joined #parrotsketch
% jonathan has joined #parrotsketch
% allison has joined #parrotsketch
% chromatic has left chromatic!~chromatic@sub17-30.member.dsl-only.net
% particle has left particle!~particle@144.81.84.129
% particle has joined #parrotsketch
% allison has left allison!~chatzilla@208.106.28.34
% allison has joined #parrotsketch