Perl 6 - the future is here, just unevenly distributed

IRC log for #parrotsketch, 2009-10-06

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

All times shown according to UTC.

Time Nick Message
00:33 cotto joined #parrotsketch
02:52 cottoo joined #parrotsketch
02:54 cotto joined #parrotsketch
07:09 kurahaupo joined #parrotsketch
09:09 mikehh joined #parrotsketch
09:20 particle joined #parrotsketch
09:55 particle joined #parrotsketch
11:41 kid51 joined #parrotsketch
13:08 particle joined #parrotsketch
13:54 PacoLinux joined #parrotsketch
14:30 PacoLinux joined #parrotsketch
16:08 davidfetter joined #parrotsketch
16:50 cotto_work joined #parrotsketch
17:05 whiteknight joined #parrotsketch
17:13 pmichaud What I did:
17:13 pmichaud * Mainly worked on regex refactors
17:13 pmichaud * Started work in branches/pct-rx in parrot svn, but have since decided  to move all of the work to nqp-rx repository on github
17:13 pmichaud * End result will be a new implementation of NQP that is also able to compile Perl 6 regular expressions (including protoregexes and longest-token-matching semantics).
17:14 pmichaud * Have a working implementation of protoregexes, now just need to tie all of the new pieces together
17:14 pmichaud * Have a draft implementation of a new extensible operator precedence parser, with NQP source equivalent
17:14 pmichaud What I'm doing this week:
17:14 pmichaud * More regex engine work
17:14 pmichaud * Fixing up the build system of the nqp-rx project
17:14 pmichaud * Trying to figure out how to get all of the components bootstrapped
17:14 pmichaud What I'm blocking on:
17:14 pmichaud * Time
17:14 pmichaud EOR
17:18 NotFound joined #parrotsketch
17:19 Util joined #parrotsketch
17:24 cotto_work no report this week
17:30 whiteknight WHAT I DID:
17:31 whiteknight * Helped organize and advertise the weekend's hackathon. Was a big success!
17:31 whiteknight * Worked on PCC system, fixing tests mostly.
17:31 whiteknight * Wrote up some documentation about how argument processing happens
17:31 whiteknight * Updated some *Tasklist pages on the wiki
17:31 whiteknight * Answered questions on IRC
17:31 whiteknight WHAT I PLAN TO DO:
17:31 whiteknight * PCC: strike while the iron is hot. Would LOVE to get it merged before 1.7.0
17:31 whiteknight * Trunk seems to have some problems, especially with --optimize. Need to dig into that
17:31 whiteknight * Miscellaneous JIT preparations
17:31 whiteknight * Documentation
17:31 whiteknight WHAT I AM BLOCKING ON:
17:31 whiteknight * Not enough time
17:31 whiteknight EOR
17:35 japhb DONE:
17:35 japhb * http://use.perl.org/~geoffrey/journal/39697
17:35 japhb * http://use.perl.org/~geoffrey/journal/39713
17:35 japhb * Plumage now able to install Rakudo
17:35 japhb WILL DO:
17:35 japhb EOR
17:37 japhb Hmmm, that's odd, my report is missing lines in http://irclog.perlgeek.de/parrotsketch/today ...
17:40 jonathan joined #parrotsketch
17:45 japhb Oh interesting, I think I got throttled but my local client showed them as sent
17:45 japhb sending report again in pieces ....
17:45 japhb DONE:
17:45 japhb * http://use.perl.org/~geoffrey/journal/39697
17:45 japhb * http://use.perl.org/~geoffrey/journal/39713
17:45 japhb * Plumage now able to install Rakudo
17:45 japhb WILL DO:
17:45 japhb * Hack, hack, hack
17:45 japhb POTENTIAL BLOCKERS:
17:45 japhb * Need to find a place to host metadata repo.  Suggestions?
17:45 japhb EOR
17:46 japhb OK, that looks better
17:49 mikehh What I did in the last week:
17:49 mikehh * only managed 11 builds/fulltests on trunk only one of which was on i386 - the rest on amd64
17:49 mikehh * (there weren't that many commits to trunk)
17:49 mikehh * most of my time was spent building/testing on pcc_reapply branch
17:49 mikehh * make/make world does not build - used make corevm (with normal config parameters on amd64)
17:49 mikehh * make codetest/make manifest_tests did run and managed to fix (eventually) all failures there
17:49 mikehh * make headerizer did not work in the branch
17:49 mikehh * spent nearly all my waking time on Friday/Saturday/Sunday on the branch - Monday was a complete write-off
17:49 mikehh * make smolder_coretest got only 39 failures at r41710/#28557 but this increased to 59 again at r51727/#28571
17:49 mikehh * and was the same today r51733/#28614 - kid51 had a lot more problems on i386
17:49 mikehh * meant to install Ubuntu 9.10 beta - never got around to it
17:49 mikehh What I intend to do in the next week:
17:49 mikehh # continue testing in trunk and branches
17:49 mikehh # install Ubuntu 9.10 betas and test parrot on it
17:49 mikehh .eor
17:51 Tene * Work on PCC.  Going to make another try to finish fill_params tonight.
17:51 Tene That's probably all for this week.
17:52 Tene Unless someone assigns me a specific task.
17:52 Tene KTHXBAI
17:57 pmichaud I should also report that I'm extremely pleased to see all the progress on the PCC refactors.
18:02 jonathan * Back from my vacation in far east Asia...probably my last serious vacation before Rakudo *.
18:03 jonathan * Spoke at YAPC::Asia and for Seoul.pm too - plenty of interest from both :-)
18:03 jonathan * Got my grant for working on signature refactors approved
18:03 jonathan * "Eased" into it today by doing some refactors that make actions.pm cleaner and will let me change the way we build signatures more easily
18:03 jonathan * Expecting to make loads of progress on signature changes this week
18:03 jonathan * Also need to do some blogging - Blizkost also made progress on my Rakudo day before I went away as well as while-I-was-away hacking. One big bug to fix, then I think it may start to be quite useful, at least on Win32 (will need help on other platforms).
18:03 jonathan * Am liable to block somewhat on the PCC branch, so *very* happy it's a many person effort now - thanks everyone!
18:03 jonathan .end
18:04 NotFound 2009-oct-06
18:04 NotFound What I did:
18:04 NotFound * Add deprecation notice for TT #918.
18:04 NotFound * Some bug haunting on pcc_reapply.
18:04 NotFound What I will do:
18:04 NotFound * No plan.
18:04 NotFound EOR
18:07 kurahaupo joined #parrotsketch
18:09 allison joined #parrotsketch
18:10 allison - Created a new branch for the PCC refactor, merging the existing changes with a fresh branch of trunk. Fixed up merge problems to ready the branch for group effort.
18:10 allison - The hackathon this weekend was very productive. Spent as much time as possible on IRC, answering questions so others could work on the branch too.
18:10 allison - Changed generated VTABLE stubs in Object to use the new PCC.
18:10 allison - Broke arg/param/return/result handling in to a separate file for sanity.
18:10 allison - Worked out a new arg/param handling algorithm with Whiteknight on IRC to be more maintainable, implemented it, fixed several bugs/found missing features in the process.
18:10 allison EOR
18:15 Util # Done
18:15 Util * Poked at broken std_hilite/STD_syntax_highlight
18:15 Util = inspired by moritz++ : http://perlgeek.de/blog-en/perl-6/parse-tree.html
18:15 Util = working kludge achieved, but no peer review requested yet, so no patch yet.
18:15 Util * Removed bogus properties from files and root dir in Parrot repo. (r41736,r41737)
18:15 Util * TT#600 (bytecode testing framework) - Stalled
18:15 Util = have not replied to comment from dukeleto++ , nor led discussion yet.
18:15 Util * TT#994 (git-svn auto-props) - Testing in branches/autoprops
18:15 Util * TT#389 (keeping subs with :method out of the namespace) - Caught up with Allison's thinking.
18:15 Util # Plan for next week:
18:15 Util * Continue work on TT#994 (to completion) and TT#389 (time allowing), and prod #parrot for discussion of TT#600.
18:15 Util # Blockers:
18:15 Util * None.
18:15 Util .end
18:15 kurahaupo Converted ResizableIntegerArray tests to PIR, discovered & fixed some missing test coverage in the process <EOR>
18:21 Coke joined #parrotsketch
18:21 barney joined #parrotsketch
18:21 darbelo joined #parrotsketch
18:22 darbelo THE PAST:
18:22 darbelo (parrot)
18:22 darbelo * Obliterated the old and unused GC_IS_MALLOC code.
18:22 darbelo * Ripped out the old pmc2c tests
18:22 darbelo * Removed strstart uses in the debugger with the help of dukeleto++
18:22 darbelo * Some more strstart removal is pending on win32 testing. (http://nopaste.snit.ch/18236)
18:22 darbelo (partcl)
18:22 darbelo * Updated copyright and licensing information to PaFo and Artistic 2.0
18:22 darbelo * Error: partcl install failed with ENOTUITS
18:22 darbelo (plumage)
18:22 darbelo * Generated partcl metadata file. Runs 'make test'
18:22 darbelo THE FUTURE:
18:22 darbelo (parrot)
18:22 darbelo * Take a shot at fixing the frame builder on the pcc_reapply branch.
18:22 darbelo (partcl)
18:23 darbelo * keep working on the 'install' target.
18:23 darbelo THE IMPEDIMENTS:
18:23 darbelo * Tuit shortage.
18:23 darbelo .EOR
18:27 dukeleto Past: My write up about Perl and Parrot in GSOC2009 got published on the Google Open Source Blog: http://google-opensource.blogspot.com/
18:27 Coke (All partcl)
18:27 Coke Done: - switching over to github. (dukeleto++) - worked on switching to use parrot's call chain. Looks like we can't yet
18:27 Coke - running spec test occasionally.
18:27 Coke Coming up: - finalize switchover to github (dukeleto++)
18:27 Coke .
18:27 Coke - ping pmichaud on pct cutover. - combine 3 makefiles into a single one, improve deps (See http://code.google.com/p/pa​rtcl/issues/detail?id=111)
18:27 Coke Blocking on: - tuits
18:27 dukeleto Future: Continue to convert tests to PIR and benchmark released versions of parrot against trunk with euler_bench
18:28 dukeleto Blocking On: Too many cool things to hack on
18:30 japhb Coke: q1q
18:31 allison looks like no chromatic this wek
18:32 allison Let's start with the roadmap review.
18:32 allison https://trac.parrot.org/parrot/report/14
18:32 allison hll export?
18:33 whiteknight I had expressed interest but have done no work on it
18:33 allison okay, interested parties absorbed by other priority (pcc branch)
18:33 allison documentation for parrot_debugger?
18:33 chromatic joined #parrotsketch
18:34 allison iirc, that was dukeleto
18:34 allison I'm taking that as no progress this week
18:35 allison seed libraries?
18:35 allison looks like japhb has been making good progress
18:35 japhb allison, yeah, Plumage is going well
18:35 allison at least, he mentioned hosting for the metadata
18:35 allison excellent
18:35 allison what do you need hosted?
18:36 allison mysql + web interface?
18:36 japhb I need a place that's outside my firewall to have a mod_perl + mysql instance, where I can install arbitrary modules
18:36 allison (i.e. html and json)
18:36 japhb yeah
18:36 NotFound Mysql is now able to be installed from plumage in unix-like platforms
18:36 allison mod_perl?
18:36 japhb It's not time critical, I'm mentioning it now to keep it out of my critical path later
18:37 allison you're using mod_perl for serving the web pages and data?
18:37 japhb allison, I suppose CGI or FastCGI is possible,
18:37 allison japhb: mod_parrot eventually? :)
18:37 whiteknight any reason it can't be mod_parrot :)
18:37 japhb I'm just a mod_perl guy, so it's my default way to host something like this.
18:37 japhb allison, heck yeah.  And in fact,
18:37 japhb I'd be happy to do that if it was ready by the time I was.
18:37 allison okay
18:37 japhb (And the performance wasn't abysmal)
18:38 allison I'm thinking this should be hostable on parrot.org
18:38 allison will check with our admins
18:38 japhb allison, cool.
18:38 japhb thank you.
18:39 allison bytecode testing framework?
18:39 * japhb smiles at the thought of writing mod_parrot code.
18:39 Util TT#600 (bytecode testing framework) - only progress is a vote from dukeleto++ for running both .pir and .pbc versions of the testing as the default. I need to carry it into #parrot; among other issues, running both will more than double our current `make test` time.
18:39 Util I will make time tomorrow for the discussion, at the latest.
18:39 allison Util: aye, it's something that should be an option for fulltest, but not part of the default tests
18:40 allison sounds good
18:40 allison prune c data structures?
18:40 chromatic I have notes.  I haven't put them on the wiki or implemented them yet.
18:40 chromatic I'm a bit leery of making large trunk changes pending a PCC merge.
18:40 allison chromatic: yes, that's sensible
18:41 allison chromatic, would you like to take over for the weekly priority review
18:41 allison (so far we've done roadmap review)
18:41 chromatic Is the roadmap review over?
18:41 allison yes, that's the last item
18:41 whiteknight yes
18:42 chromatic Last week's priority: the hackathon.
18:42 chromatic Thoughts?
18:42 allison a huge success
18:42 whiteknight huge
18:42 pmichaud agreed
18:42 pmichaud let's do it again.  now.  :)
18:42 whiteknight and we're coming out of it with a lot of momentum and a much higher bus number on that system
18:42 allison it had a big impact on motivation and collaboration
18:42 darbelo We should have one every week.
18:43 chromatic Is every week too often?
18:43 whiteknight every week might be pushing it, if we don't have clear goals
18:43 allison yes, probably
18:43 whiteknight busywork is no good
18:43 allison it's definitely a push
18:43 whiteknight I say about once per month, assuming we have good candidate tasks to work on
18:43 allison with a corresponding dip in productivity right after
18:43 whiteknight maybe more, maybe less
18:43 allison one a month is probably about right
18:44 darbelo one or two a month is a bit more realistic.
18:44 allison depending on the tasks at hand
18:44 chromatic How about the size of the hackathon tasks?
18:44 allison and, towards the early side of the month for development merges
18:44 allison this one was a good size
18:44 allison a branch that's nearly finished
18:45 allison we dropped it from 900 failing tests to 58-ish
18:45 whiteknight the "saturday" for the hackathon started friday evening and ended sunday evening
18:45 whiteknight so that might have been too big a task
18:45 whiteknight (but very necessary)
18:45 chromatic Other thoughts?
18:45 allison whiteknight: even if we'd only cut the failing tests in half it would have been a good hackathon
18:46 whiteknight yes, and we knocked "half" out of the water!
18:47 allison I'd like to revive the announcements of "testing sprints" before releases
18:47 whiteknight agreed
18:47 chromatic Testing sprint?
18:47 allison sort of related to hackathons, but not quite the same
18:47 particle previously known as "bug day"
18:48 * particle arrives, late
18:48 whiteknight development sprints towards the beginning of each cycle, and testing sprints towards the end
18:48 mikehh we still need to to get the full build to work in pcc_reapply - and I would like to test the other cores
18:48 allison that reminds me: q1q
18:48 whiteknight yes, lots of work still on pcc_reapply, and then I'm sure the merge will be eventful
18:48 allison (chromatic: japhb had a q1q at the start)
18:48 whiteknight plus HLL testing
18:49 chromatic Shall we tentatively plan a testing weekend on the 17th and 18th?
18:49 darbelo +1
18:49 mikehh +1
18:49 allison +1
18:49 chromatic Okay.  Let's go to questions.
18:50 chromatic japhb?
18:50 japhb It sounds like partcl will be switching repos this week, correct Coke?
18:50 particle q1q
18:51 japhb If so, can you send me the new repo info, or a patch against the metadata in Plumage, please?
18:51 dukeleto japhb: i am working on partcl->github, so I can help with that
18:51 japhb That was it, but we seem to have lost Coke.
18:51 dukeleto 'ello, by the way
18:51 chromatic allison?
18:51 japhb dukeleto, ah, cool, thank you.
18:51 darbelo japhb: I can do that when the move's done.
18:51 japhb excellent
18:51 Util q1q
18:52 allison my question is about src/frame_builder.c
18:52 allison it's the last remaining bit of the old JIT
18:52 allison as I understand, kept for speed gains on x86?
18:53 NotFound allison: IMO his own advantage is that it doesn't need a reconfigure when you need a new signature.
18:53 allison it was bunged up a good bit when the old JIT was removed, so a bit shaky
18:53 NotFound s/own/only
18:53 allison and, it completely doesn't survive the PCC transition
18:53 NotFound Let's kill it!
18:54 allison so, the question is, is it worth delaying the PCC branch merge to get it working again?
18:54 chromatic It's the only useful part of the old JIT and it does get used.
18:54 allison or, can we disable it temporarily, and get it working again later?
18:54 whiteknight but alternatives are already popping up
18:54 whiteknight we have a libJIT-based framebuilder on github for starters
18:54 mikehh it was causing major problems with i386 in the branch
18:54 whiteknight an LLVM-based one would be a good kickoff to the rest of our JIT work
18:55 chromatic There are three different questions here. (more)
18:55 mikehh kid51 had to disable it to get tests to pass
18:55 chromatic 1) Should it block the branch merge?
18:55 allison mikehh: at the moment the code doesn't work with the new PCC, guaranteed
18:55 chromatic 2) Is it useful as-is?
18:55 chromatic 3) What are likely replacements?
18:55 mikehh it doesn't bother me on amd64 though :-}
18:55 whiteknight ditto :)
18:56 allison mikehh: as far as I understand, it's always disabled on amd64
18:56 NotFound To avoid problems with testers we just need to make --buildframes=0 the default.
18:56 allison mikehh: that is, it only ever works on i386
18:56 mikehh didn't do the old jit
18:57 whiteknight If we have a firm plan to have a replacement in trunk by 1.8, can we remove it now?
18:57 chromatic Let's take the questions one by one.
18:57 chromatic Should it block the branch merge?
18:57 particle i think as long as we have a replacement by 2.0 we can remove it
18:57 particle and the branch merge should not block.
18:57 whiteknight no. Branch has been too far delayed as-is
18:57 allison to one, I'd say no, it shouldn't block the merge
18:57 whiteknight and HLLs are hurting for it as is
18:57 chromatic Objections?
18:58 particle we must have a replacement soon.
18:58 chromatic <crickets />
18:58 chromatic #2 -- is it useful as-is?
18:58 mikehh no
18:58 particle and as a side, it's sad to see the last bit of gsoc code removed from parrot.  again.
18:58 whiteknight barely, only on x86 I think
18:58 chromatic It improves startup time on x86 32-bit by ~15%.
18:58 chromatic It reduces libparrot.so size measurably.
18:59 particle *as an aside
18:59 chromatic On the con side, it doesn't handle all signatures appropriately.
18:59 NotFound It fails to run opengl examples
18:59 particle that might be opengl's fault
18:59 chromatic I think that's NCI in general.
18:59 japhb NotFound, thanks for not making me be the one to point that out again.
18:59 whiteknight I would say it's benefit is overall neutral
18:59 whiteknight and only on x86
18:59 chromatic ... which is the majority of our platforms right now.
19:00 particle it's benefit is unknown without asking our users
19:00 whiteknight not among developers
19:00 allison is its limited usefulness on x86 blocking progress on other platforms?
19:00 particle our users are hll devs. ask them.
19:00 whiteknight allison: nobody is working on a frame builder for any other system. Don't know why
19:00 allison sounds like it's largely neutral, then
19:00 whiteknight I would suspect more interest in developing a "proper" portable one
19:01 chromatic Summary then: it's somewhat useful as is, but not compellingly so.
19:01 chromatic Third question: are there likely alternatives in the near future?
19:01 whiteknight yes
19:01 allison If we get it working with the new PCC, it's probably worth keeping it until something better comes along.
19:01 particle can they be completed along with our other goals?
19:01 whiteknight there's a drop-in libJIT solution on github now, and an LLVM-based one could be added in a month
19:01 mikehh is it worth the effort needed to get it working on one platform ?
19:02 Coke I'm back, sorry.
19:02 particle whiteknight, why on earth would we want to make parrot rely on both libjit and llvm?
19:02 allison mikehh: depends on the amount of effort, and the readiness of alternatives.
19:02 whiteknight particle: either/or
19:02 allison particle: they would both be optional
19:02 whiteknight configure finds what we have, and uses anything it finds
19:02 chromatic Did I hear someone mention hand-coded assembly?
19:03 whiteknight TiMBuS and I are idly working on something like that
19:03 particle so, a llvm gc and libjit frame builder is possible?
19:03 whiteknight particle: no idea about the GC, but an LLVM-based frame builder is, yes
19:03 allison particle: yes, it doesn't matter how you compile the code down to machine code
19:03 whiteknight and a compelling first stepping stone in our larger JIT rewrite
19:03 allison particle: it's machine code in the end
19:03 particle eh, whatever.  just ask the users about removing functionality we currently provide.  they don't seem to be here.
19:04 * particle <--- user advocate today
19:04 chromatic It's mostly not user visible.
19:04 allison only as speed
19:04 allison which is important for 2.0
19:04 particle and recompiling
19:04 chromatic There's one spot where it's visible.
19:04 whiteknight in the same way the old JIT system was "mostly not user visible"
19:04 chromatic If they try to use an NCI signature we haven't compiled in, it's possible the frame builder can make the appropriate thunk.
19:04 particle hey, i need a new sig. i have to recompile parrot.  oh, wait, i have a debian package.
19:05 mikehh speed like in premature optimisation...
19:05 NotFound Is visible for people that have problems with selinux config blocking the frame creation.
19:05 Coke I'm sorry,did someone leave the snark on?
19:06 whiteknight ah yes, I forgot about selinux
19:06 chromatic We've answered the important first question though.
19:06 kurahaupo Doesn't JVM have same problem on SElinux?
19:06 chromatic We've agreed that it shouldn't block merging the PCC branch.
19:07 allison chromatic: the answers to 2 and 3 are a process
19:07 particle agreed.
19:07 whiteknight we can have a replacement by next release, which coincidentally would be a great topic for a hackathon
19:07 chromatic Any other comments for now?
19:08 chromatic particle, you had a question.
19:08 particle we've had multiple releases since 1.0, our target for 'stable platform for hll developers'.  how have we  failed to meet hll devs needs, and how can we address that?  we can use the knowledge gained from a discussion on this topic to address the 2.0 'business-ready' target.
19:09 chromatic HLL developers?  Coke?
19:09 whiteknight <pmichaud>Parrot hasn't really been stable for HLL developers since 1.0 though</pmichaud>
19:09 Coke 1.0 was really not stable enough at all.
19:09 particle we need a forum where these issues can be discussed candidly, with hll devs.
19:09 Coke I am still forced to update every week or so, not tied to releases.
19:10 particle can someone set up an internet chat, here or on #parrot, soliciting hll devs for input and discussion?
19:10 whiteknight like a #parrot-hll?
19:10 chromatic or #parrotowthedeephurting
19:10 allison #parrotlang
19:10 whiteknight or a "chatathon"?
19:11 particle they're all fine names.
19:11 Coke given that you have two hll folks on the board, you could potentially solicit feedabck there pretty easily.
19:11 jonathan To pick a specific example of one disappointing thing, while I'm very happy to see so many people working hard on PCC stuff now, it's a disappointment that it's taken so long to get to this point. Especially given it was flagged up at PDS.
19:11 dukeleto BTW: I did make incremental progress on the parrot debugger docs, just wasn't here when it was mentioned.
19:11 allison particle: did you have a particular time in mind?
19:11 whiteknight particle: are you asking about a dedicated chat place, or dedicated chat time?
19:11 allison particle: or more of an ongoing thing?
19:11 dukeleto q1q
19:11 particle allison: whatever works for them.  once, for now.  but we need to make this a regular thing
19:11 chromatic jonathan, I think that's a process failure more than anything else.
19:12 particle there are critical (imo) issues to address, and scheduled meetings is one of them
19:12 chromatic We've been modifying our processes to improve and avoid such things.
19:12 particle we need someone to act as a user advocate
19:12 allison would the language devs here like to meet for 20 minutes right after #parrotsketch?
19:12 Coke particle: what's driving this?
19:13 particle the devs are noisy, but the users are quiet
19:13 allison particle: the users are the devs
19:13 Coke so, are pmichaud and I not users, from that standpoint?
19:13 whiteknight Yeah? Coke is as noisy as they come!
19:13 whiteknight :)
19:13 allison particle: though, I have to say, it would be useful for me to talk from the perspective of Pynie development
19:13 particle no, patrick is an infrequent core contributor to parrot
19:13 allison that is, it's more about topic focus than about who participates
19:13 particle his focus is rakudo.
19:13 particle allison: correct
19:14 allison #parrotsketch is for core development
19:14 kurahaupo I have question/comment about ResizableXxxxArrays and their lack of initialization guarantees. Seems to me that that makes them "not fit for purpose" for any purpose I can think of which isn't fulfilled by a FixedXxxxArray. If you have the language devs together, should bring it up then?
19:14 allison kurahaupo: that's a coredev question
19:15 allison but, best on the mailing list
19:15 allison or #parrot
19:15 Coke kurahaupo: I find the Fixed ones completely unfit for purpose. so your MMV. =-)
19:15 allison to wrap up this question, is there general interest in a regular meeting for language devs?
19:16 Coke No.
19:16 kurahaupo A one-off meeting?
19:16 allison Coke: you're satisfied with #parrotsketch/#parrot/parrot-dev to meet language dev needs?
19:16 particle i'm happy to talk to hll dev teams separately
19:16 particle they can always talk to each other on #parrot or otherwise
19:16 mikehh are you going to be the user advocate?
19:17 whiteknight one-on-one meetings with individual dev teams might be nice
19:17 Coke allison: if I have a need, I know where to go, yes.
19:17 particle i seem to be the only one making noise, mikehh, so yes :)
19:17 allison Coke: makse sense
19:17 Coke particle: and why are you making noise?
19:17 whiteknight Coke: users are something we should be concerned with more then we are
19:17 particle coke: for one, rakudo devs are disappointed in parrot's lack of delivering pcc
19:18 particle and many more are disappointed in the lack of parrot api stability
19:18 particle what have we done to address that? nothing.
19:18 allison the stability we promised is "you can update only every 6 months"
19:18 whiteknight to be fair, rakudo have pushed for changes in the past that broke stability
19:18 allison I'm doing that with Pynie, and it works well
19:19 particle pynie is a slow moving target
19:19 allison partly, yes, but more I keep it very intentionally decoupled from Parrot development
19:19 dukeleto +1 to the idea of having a channel specifically for HLL devs
19:19 particle what happened to cardinal development? it's trailed off.  why? does anybody know?
19:20 particle these are questions we want to have answered.
19:20 whiteknight I haven't seen Treed around in a while
19:20 jonathan allison: Rakudo isn't coupling more tightly just for the fun of it, you know.
19:20 particle right, out of steam, out of time, or out of patience?
19:20 chromatic Let's summarize before this becomes a free for all.
19:20 particle thanks, c.
19:20 whiteknight chromatic: agreed
19:20 chromatic particle, will you talk to HLL devs and get some data for us to discuss?
19:20 mikehh treed has started a new $job, hasn't been around that much
19:20 allison jonathan: I know, it's because the Rakudo developers are adding Parrot features they need, which is good for both
19:20 particle c: i will.
19:21 chromatic jonathan, Coke, allison, tene, whomever, will you help particle?
19:21 jonathan Sure.
19:21 chromatic Great, thank you.
19:21 chromatic Util, you had a question.
19:21 whiteknight particle: create a place, such as on the wiki, where this info can be put
19:21 Util In his report, pmichaud said that the new (branched) NQP will support protoregexes and LTM.
19:21 Util Is that the last major obstacle to STD.pm working on Rakudo? If not, what else are blockers?
19:21 Tene chromatic: help particle with what exactly
19:21 Tene chromatic, particle: I'll do whatever you ask me to do.
19:22 particle make me a sandwich
19:22 whiteknight sudo make me a sandwich
19:22 chromatic Those are the two big ones I know of, Util.  jonathan and pmichaud will know better.
19:22 Tene particle: it was mostly just treed working on cardinal while he wa sunemployed/job hunting.  He has a job now, and is busy with life.
19:23 barney same here, with Pipp
19:23 jonathan Util: It will go a long way towards that.
19:23 allison Tene: I think that was "help particle summarize hll developer's needs"
19:23 jonathan Util: However, there's a few other bits (embedded closures, context variables in regexes, etc)
19:23 particle ok, that's great news. i only mentioned it because we don't, as a group, know the answers.  more regular contact with our users will give us better info.
19:23 jonathan That are also important.
19:24 particle pcc is important, too
19:24 pmichaud sorry, I was afk for a bit
19:24 whiteknight lots of important things
19:24 pmichaud 19:18 <allison> the stability we promised is "you can update only every 6 months"
19:24 Util I am trying to keep an eye out for things that can be worked on in parallel to protore&LTM, since Rakudo+STD will be such a big step forward.
19:24 pmichaud both Coke and I have commented that this form of stability is basically useless for HLL devs
19:24 Tene q1q
19:25 pmichaud I'll take it to #parrot
19:25 Util Thanks. EOQ.
19:25 chromatic Tene?
19:25 Coke pmichaud: +1 - it's not /harmful/ to me, per se, but it doesn't help.
19:25 allison pmichaud: we were running a bit off topic there, and decided we needed longer discussion elsewhere,
19:25 pmichaud allison: yes, I just caught up with scrollback
19:25 allison #parrot is good or whatever
19:25 pmichaud sorry for the untimely interjection
19:26 Tene Last weekend's hackathon was a big success.  I've heard a few people express interest in doing it again.  Do we want to schedule another hacking event, possibly with a different focus?  pirc?  any suggestions?
19:26 allison pmichaud: it's good confirmation that particle is right and we do need to talk more
19:26 dukeleto chromatic: i had a question queued as well
19:26 Tene EOQ
19:26 chromatic pirc is a good candidate.  We roughly agreed to do this the weekend after next to focus on testing before the release.
19:26 Tene a testing hackathon would be great.
19:26 allison Whiteknight suggested jits, which might be a good one after pirc
19:27 Tene we could make a wiki page with hackathon targets.
19:27 particle what about a frame builder?
19:27 whiteknight so long as JIT happens before 2.0 I am fine
19:27 whiteknight particle: frame builder is part of JIT
19:27 particle yes, it is.
19:27 chromatic Does someone volunteer to make that page?  We can put further discussion there.
19:27 allison if we do a testing one the weekend before the release, and a development one soon after that should work well
19:27 particle i think that's more important than pirc atm
19:28 mikehh whiteknight: do you think we will have something workable by then?
19:28 whiteknight I'll create the page
19:28 whiteknight mikehh: by when?
19:28 pmichaud I'm fine with another focus as long as the pcc development doesn't lose pace again
19:28 mikehh 2.0
19:28 chromatic dukeleto, you had a question?
19:29 allison pmichaud: this would be a hackathon after 1.7, so not distracting from finishing off pcc
19:29 pmichaud okay
19:29 pmichaud good, and thanks.
19:29 mikehh imcc is really causing a lot of problems - I think pirc need to be emphasised
19:29 dukeleto chromatic: yes
19:30 dukeleto is pcc_reapply going into 1.7 ?
19:30 particle we have a roadmap to help us select code sprint topics
19:30 Tene dukeleto: several people really want it to happen.
19:30 allison dukeleto: can't say for sure yet, but we'll try
19:30 chromatic If it's mergeable in the next week, definitely.
19:30 chromatic After that... depends on how mergeable it is.
19:30 allison yes, no merges in the week before the release
19:31 mikehh I'll keep testing as soon as there are any commits
19:31 chromatic Other questions?
19:32 dukeleto allison: since I am release manager, i would like to know where to focus my energies. i would like to get it in 1.7, if possible
19:32 dukeleto i must run, i will catch up on the backlog in a few minutes
19:32 allison the first focus is on return handling
19:32 allison that's the last big gap in functionality
19:33 chromatic Final thought: weekly focus.
19:33 chromatic Anything besides PCC?
19:33 particle pcc and pcc.
19:33 mikehh let's get pcc out of the way asap
19:33 particle how about coverage on some of those files?
19:33 particle is there a heavily used pmc there?
19:34 chromatic CallSignature.
19:34 allison particle: CallSignature PMC
19:34 particle amen.
19:34 allison which has a single test
19:34 allison (can I be instantiated)
19:34 allison plenty to test there
19:35 mikehh I will try for some more
19:35 chromatic Anything else before we adjourn?
19:36 chromatic Until next week then.  Thank you all.
19:36 cotto_work If CallSignature is extensively used in that branch, we should make sure we're not testing code that's already well-exercised.
19:36 mikehh cheers
19:36 cotto_work vim
19:36 cotto_work ww
19:36 Topic for #parrotsketchis now Vision for 2.0: Production Users | Priority for 1.6: Merge Branches | Priority for this week: get PCC branch mergable, increase test coverage on CallSignature PMC | https://trac.parrot.org/parrot/w​iki/ProposedParrotsketchProtocol | Note: This channel is only for our Tuesday status meetings; you probably want #parrot instead.
19:36 PacoLinux left #parrotsketch
19:37 Util left #parrotsketch
19:38 chromatic left #parrotsketch
19:40 darbelo left #parrotsketch
19:41 NotFound left #parrotsketch
20:57 davidfetter left #parrotsketch
20:58 Whiteknight joined #parrotsketch
21:09 Coke left #parrotsketch
21:42 Whiteknight joined #parrotsketch
22:57 jonathan left #parrotsketch
23:08 cotto joined #parrotsketch

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