Camelia, the Perl 6 bug

IRC log for #parrot, 2010-10-14

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:17 davidfetter magnush++ for making that git transition actually happen :)
00:22 cotto is he the osuosl guy?
00:22 cotto also, isn't that a bit premature?
00:23 davidfetter um, for pg, the git transition is done
00:23 davidfetter for parrot, not yet. :)
00:23 dmalcolm left #parrot
00:23 davidfetter magnush == magnus hagander. he's worked at getting the pg transition done for years.
00:24 cotto I see.  props to him then.
00:25 kid51 The article was a fine documentation of the challenges which OS projects face in making infrastructural transitions.
00:26 * davidfetter heaps praise on the people who do all the non-sexy stuff that makes the world run smoothly
00:26 davidfetter like maintaining and upgrading infrastructure, for example
00:27 cotto It really depends on your definition of "sexy".
00:30 dalek parrot: r49530 | jkeenan++ | branches/tt1824_ipv6_configure:
00:30 dalek parrot: Creating branch to work on �http://trac.parrot.org/parrot/ticket/1824.
00:30 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49530/
00:30 davidfetter o/` i'm // too sexy for my source code management system // too sexy for my source code management system // no way i'm disco dancing o/`
00:31 davidfetter sorry. right said fred moment.
00:31 cotto np.  It happens to everybody.
00:31 davidfetter really?!?
00:32 cotto Sure.  I thought it was common knowledge.
00:45 dalek parrot: r49531 | jkeenan++ | branches/tt1824_ipv6_configure (2 files):
00:45 dalek parrot: Create directory and files needed for adding configuration step auto::ipv6.  (No real content yet.)
00:45 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49531/
00:50 davidfetter left #parrot
00:53 kid51 left #parrot
01:01 dalek parrot: r49532 | jkeenan++ | branches/tt1824_ipv6_configure (3 files):
01:01 dalek parrot: Begin to translate patch from kurahaupo++ into configuration step probe.  Not actually set up yet.
01:01 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49532/
01:28 theory joined #parrot
01:45 theory left #parrot
01:46 cotto What's a good way to create an inner runloop in pure pir?
01:47 cotto nm. FIA.sort looks like it'll do nicely.
01:49 simcop2387 joined #parrot
01:51 cotto Hmmm.  Custom sort subs aren't tested.  I'm taking bets on whether they work.
01:51 cotto (for fia)
01:52 cotto segfault
02:02 dalek parrot: r49533 | cotto++ | trunk/t/pmc/fixedpmcarray.t:
02:02 dalek parrot: [t] remove comment about an obsolete ticket
02:02 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49533/
02:04 davidfetter joined #parrot
02:15 mikehh joined #parrot
02:20 cotto At least FPA works
02:22 cotto And after all that, no inner runloops.
02:29 theory joined #parrot
02:30 davidfetter left #parrot
02:35 janus left #parrot
02:44 theory left #parrot
03:03 Andy joined #parrot
03:33 dngor_ joined #parrot
03:37 dngor left #parrot
03:39 dngor_ is now known as dngor
03:43 ledrug joined #parrot
03:44 ledrug left #parrot
03:44 janus joined #parrot
04:04 jsut joined #parrot
04:09 jsut_ left #parrot
04:16 jsut_ joined #parrot
04:21 jsut left #parrot
05:21 cotto What's the magic invocation to TODO a test in nqp?
05:23 cotto also, why is "ok" available when "todo" isn't?
05:25 cotto Oh.  It's explicitly built-in.
05:49 bacek_at_work cotto, I use parrot's Test::More in PIRATE.
06:00 cotto I'll steal that.
06:03 cotto Q:PIR looks much easier than messing around with pure nqp
06:38 fperrad joined #parrot
06:39 Andy left #parrot
07:12 dalek parrot: r49534 | cotto++ | trunk/src/runcore/profiling.c:
07:12 dalek parrot: [profiling] factor out op time recording into a separate function
07:12 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49534/
08:07 bacek ~~
08:08 PerlPilot joined #parrot
08:12 PerlJam left #parrot
08:12 cotto Parrot keeps track of runloop levels.  Who knew.
08:14 jjore left #parrot
08:14 jjore joined #parrot
08:15 Arktosis left #parrot
08:16 Kulag joined #parrot
08:29 dalek parrot: r49535 | cotto++ | trunk (3 files):
08:29 dalek parrot: [profiling] use parrot's internal runloop tracking and add a test to make sure
08:29 dalek parrot: that exceptional control flow doesn't mess up pcore output
08:29 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49535/
08:33 cotto An additional benefit is that the user will now have an extra sizeof(INTVAL) bytes free while profiling.  It should help out on machines with low memory.
08:35 contingencyplan left #parrot
08:36 moritz you mean, per process? or per PMC?
08:40 cotto per instance of parrot
08:40 cotto no need to thank me
08:40 mikehh left #parrot
08:44 moritz if you run parrot often enough, you'll see the difference :-)
08:58 fperrad_ joined #parrot
09:01 fperrad left #parrot
09:01 fperrad_ is now known as fperrad
10:29 cotto seen chromatic
10:29 aloha chromatic was last seen in #parrot 1 days 11 hours ago saying "Most of it is doable by a C capable novice.".
10:37 dngor left #parrot
10:46 cotto left #parrot
10:46 cottoo joined #parrot
10:47 cottoo is now known as cotto
10:50 fperrad_ joined #parrot
10:52 dip joined #parrot
10:53 dngor joined #parrot
10:53 fperrad left #parrot
10:53 fperrad_ is now known as fperrad
11:04 dngor_ joined #parrot
11:04 dngor left #parrot
11:11 ruoso left #parrot
11:16 fperrad left #parrot
11:48 tadzik joined #parrot
11:49 tadzik left #parrot
11:51 tadzik joined #parrot
12:06 tadzik left #parrot
12:08 tadzik joined #parrot
12:10 tadzik left #parrot
12:45 jan left #parrot
12:45 dngor_ is now known as dngor
12:48 bluescreen joined #parrot
12:55 whiteknight joined #parrot
13:31 ruoso joined #parrot
14:03 kurahaupo joined #parrot
14:09 kurahaupo left #parrot
14:11 Andy joined #parrot
14:22 kurahaupo joined #parrot
14:27 particle left #parrot
14:29 particle joined #parrot
14:42 brianwisti joined #parrot
14:57 dmalcolm joined #parrot
15:10 contingencyplan joined #parrot
15:19 theory joined #parrot
15:23 kid51 joined #parrot
15:27 rurban joined #parrot
15:46 particle left #parrot
15:48 particle joined #parrot
15:51 particle left #parrot
15:53 particle joined #parrot
15:55 ruoso left #parrot
15:58 particle1 joined #parrot
16:01 bluescreen left #parrot
16:01 bluescreen joined #parrot
16:02 particle left #parrot
16:16 theory left #parrot
16:48 particle joined #parrot
16:50 particle1 left #parrot
16:55 particle left #parrot
16:56 particle joined #parrot
17:19 particle left #parrot
17:25 particle joined #parrot
17:26 kid51 left #parrot
17:29 brianwisti Oh hey neat. Ubuntu 10.10 has parrot 2.6.0, with parrot-nqp available in parrot-dev.
17:45 theory joined #parrot
17:48 moritz \o/
17:48 moritz brianwisti: is there also a rakudo package?
17:48 brianwisti Sadly, no.
17:50 moritz ah well, I hope it's only a matter of time :-)
17:51 brianwisti There is a Rakudo package in Debian Experimental, though: http://packages.debian.org/experimental/rakudo
17:54 brianwisti Looks like Ubuntu grabbed from Debian unstable for their parrot deb files. So maybe if/when Rakudo gets into unstable?
17:58 allison brianwisti/moritz: the guy who was doing experimental packaging for Rakudo on Debian kind of disappeared. I'd be happy to help anyone who wants to start it up again.
17:59 allison brianwisti/moritz: I do the packaging for Parrot on Debian/Ubuntu, and have to explicitly request the update for each new supported release
18:00 brianwisti allison: Ah, I should have known. Thanks for that! I've never done any sort of distro packaging, so I won't pretend I can help with Rakudo.
18:01 allison brianwisti: no worries, glad to know people are using the 2.6 packages in 10.10 :)
18:03 particle left #parrot
18:07 particle joined #parrot
18:37 davidfetter joined #parrot
18:47 kurahaupo left #parrot
18:49 simcop2387 left #parrot
18:50 simcop2387 joined #parrot
18:53 brianwisti left #parrot
19:14 cotto dukeleto, are the git workflow, branching/merging and release docs done?
19:16 whiteknight do we have any kind of an ETA for git-switch yet?
19:17 cotto after the next release
19:17 cotto preferably soon after
19:18 whiteknight nice. I'm very much looking forward to it
19:18 whiteknight hopefully I'll be able to get back into development after that
19:18 cotto The big blocker is a strategy for letting HLLs use a specific non-release version of Parrot or something later.
19:24 dukeleto I talked with allison, kid51 and chromatic about that stuff last night at the PDX Hackathon
19:25 dukeleto Consensus is that I was making life too difficult, and the simplest possible thing that can work will do for now.
19:25 fperrad joined #parrot
19:26 dukeleto Releases of HLL's will only be able to depend on a release of Parrot, but arbitrary version of an HLL can depend on a sha1, as long as the parrot repo is present.
19:26 dukeleto This will be much easier to implement. I just need some time to do it.
19:28 cotto I'm glad that's (temporarily) out of the way.
19:28 dukeleto I was letting rare edge cases complicate things, which I am sometimes very good at.
19:29 cotto Is it too early to send out an announcement with a firm timeline for the migration?
19:29 dukeleto I think so, since I can't promise any dates currently. Once we have the last two tools converted over, and have talked to OSUOSL a bit more about what we need from them, we can do that.
19:30 dukeleto I will be visiting the Portland office of OSUOSL tomorrow, so maybe I can talk to them about it
19:30 allison dukeleto: are we planning to host with OSUOSL, or with github?
19:30 cotto Are you familiar with now to configure the navadd plugin for trac?
19:30 cotto allison, github
19:31 cotto http://github.com/parrot/parrot
19:31 allison cotto: good, I'm not even sure if OSU OSL does git hosting :)
19:32 dukeleto allison: github will be our main mirror, but we can decide to mirror in as many places as we want. We just need to tell people "for now, X is canonical"
19:32 dukeleto if we decide we don't like github, we can always move. But github makes managing teams and commit access very easy.
19:33 dukeleto I would like to set up mercurial and bzr mirrors as well.
19:33 dukeleto I don't want anybody left in the cold.
19:33 dukeleto Actually, I would like there to be other mirrors, but *I* don't want to set them up :)
19:33 allison dukeleto: I'll use the bzr mirror, if it works
19:34 cotto dukeleto, what'd be the right github API key for trac.parrot.org to use?
19:34 dukeleto allison: ok.
19:34 dukeleto cotto: good question. not sure at the moment
19:34 * dukeleto looks to see if the parrot org has it's own api key
19:34 cotto that'd be the logical one
19:34 cotto if it exists
19:35 dukeleto cotto: what does it need the api key for?
19:35 cotto the post-receive url
19:36 cotto so that nobody except github can inject commits into the db
19:37 cotto Let me make sure I have that right.
19:39 dukeleto cotto: i can't find an api key for an organization, but I only looked quickly. Please add notes to the wiki page about this.
19:39 cotto verified; api token is necessary
19:40 cotto dukeleto, do you know how to configure the navadd plugin for trac or should I include instructions on it with the github plugin instructions?
19:43 dukeleto cotto: i have never done that before, so please add instructions
19:43 dukeleto cotto: i would imagine that we need to create a custom github account with an API key that has access to the repo, which will only be used for trac
19:44 dukeleto cotto: since I don't want my API key in there.
19:44 cotto ok.  Also, it looks like an api token isn't needed, just an arbitrary string.
19:44 dukeleto cotto: ah, cool.
19:44 cotto I'll use "qwerty".  That should be secure.
19:44 cotto ;)
19:53 whiteknight left #parrot
19:53 fperrad left #parrot
20:15 cotto dukeleto, are all the docs (workflow, branching/merging, release) updated for git?
20:17 dukeleto cotto: i believe they are sufficient to migrate to git, but I would like to add more docs for how to optimally setup git and such. But the answer is "Mostly Yes"
20:17 dukeleto cotto: i even created a git terminology document, that will hopefully grow
20:17 cotto I'll cross them off GitMigration then.
20:19 dukeleto cotto++
20:20 cotto Will we keep svn.parrot.org around read-only, and for how long?
20:33 allison cotto: for the sake of avoiding confusion, I'd rather drop svn.parrot.org entirely
20:33 kurahaupo joined #parrot
20:33 theory left #parrot
20:35 cotto I'd prefer to keep it read-only, but I can see it being a source of confusion.
20:38 cotto dukeleto++
20:38 * cotto goes off to draft some messages to parrot-dev
20:41 dukeleto cotto: github has a svn read-write mirror of all git repos, so we can just turn it off and tell people to use the github svn interface
20:42 dukeleto cotto: we should probably keep the old svn repo around for maybe a few days or a week, while we tell people how to migrate
20:43 cotto wfm
20:46 dukeleto i just added it to the migration plan
20:50 cotto dukeleto, for the actual migration it looks like we need a time when you, me and an osuosl admin will all be around to coordinate.  Can you think of anyone else?
20:50 * davidfetter waves to dukeleto
20:53 dukeleto davidfetter: howdy
20:53 dukeleto cotto: hmmmm
20:54 davidfetter dukeleto, so about PL/Parrot
20:55 davidfetter who's the first end user, and why?
20:56 dukeleto cotto: we should have someone on the rakudo side that we notify when we do things, but they don't need to be available
20:57 dukeleto davidfetter: Rakudo Perl 6 is our first end user, PL/Perl6 is the first useful application of PL/Parrot and Perl 6 grammars are the first useful feature of that application
20:57 dukeleto davidfetter: who will use grammars to do what is still up in the air
20:57 brianwisti joined #parrot
20:57 cotto Logical.  I'd think pmichaud, but I haven't seen him around here much.
20:57 dukeleto davidfetter: and currently pl/perl6 grammars are broke due to changes in Rakudo
21:00 dukeleto These are the kind of links that I want to provide for the web interface to mk_language_shell: http://matt.might.net/articles/imp​lementing-a-programming-language/ . A mixture of theory and parrot docs.
21:05 davidfetter dukeleto, interesting. what domains would you expect DSLs to be helpful in?
21:05 davidfetter dukeleto, frex, are they something that would come in handy in biology?
21:07 dukeleto davidfetter: could be, but biologists mostly don't use DSL's, or at least don't call them that. But that doesn't mean they couldn't benefit from them.
21:08 davidfetter well, few people use DSLs because constructing a language is a giant PITA
21:08 davidfetter ...or was until now :)
21:08 * dukeleto added the PDX Parrot hackathon to the google calendar
21:09 dukeleto davidfetter: yep, there are many future possibilities due to changing conditions
21:09 dukeleto davidfetter: but i think getting people in the door with PL/Perl6 will be easier, then they will have time to see the light ;)
21:09 brianwisti dukeleto: I'm still not sure how much I can contribute to the hacking part of the hackathon, but I'm looking forward to it just the same.
21:10 dukeleto brianwisti: yes, it will be very nice to meet people and just talk about stuff face-to-face
21:10 dukeleto brianwisti: we can talk about parrot babysteps posts and things like that
21:10 davidfetter brianwisti, there's a persistent and silly idea that the only skill needed at a hackathon is Deep Programming Knowledge
21:11 davidfetter the persistence is almost as silly as the idea itself
21:11 dukeleto brianwisti: also, parrot wants to solicit feedback from users, like what stuff needs improvement and things like that
21:11 cotto brianwisti, do you live in WA?
21:11 dukeleto cotto: both of you are in Seattle, methinks
21:11 brianwisti Right. And I have a doc site project on the backburner that I could pull out and work on. Along the  lines of perldoc.perl.org
21:11 brianwisti yup
21:12 brianwisti cotto: I live near Greenlake, work in Belltown.
21:12 cotto not quite Seattle, but close
21:12 brianwisti cotto: Where are you at?
21:12 cotto Redmond
21:12 brianwisti Pretty close indeed.
21:15 brianwisti Hm. I seem to have found encoding weirdness in my laptop setup.
21:18 lucian joined #parrot
21:23 theory joined #parrot
21:25 cognominal left #parrot
21:26 cognominal joined #parrot
21:26 brianwisti Wait, this might actually be a bug in Parrot 2.6 on my Ubuntu 10.10 install. Checking out via SVN to see if http://gist.github.com/626686 still causes me problems.
21:32 bluescreen left #parrot
21:33 cotto dukeleto, how long does it take to sync the github mirror with svn?
21:37 cotto Why did dalek stop following wiki updates?
21:38 whiteknight joined #parrot
21:42 dukeleto cotto: less than a minute
21:42 dukeleto cotto: i am still thinking about doing a full repo translation with a proper author map. it is more work now, but will make searching for things in the future easier
21:43 dukeleto cotto: that could take a few hours, but i can do a test run to see exactly how long it would take
21:44 cotto ok
21:45 dukeleto that is an optional step right now.
21:45 cotto That sounds like it'd mean that people should avoid committing to local clones of the github repo during the migration.
21:45 dukeleto cotto: it would change every sha1 in parrot.git
21:46 brianwisti Okay, what I found *was* a bug, but it's been fixed since 2.6.0. Looking forward to 2.9 release :)
21:46 cotto that's a yes
21:51 cotto what's the leto/testing branch?
21:53 cotto looks like nothing important
21:56 mikehh joined #parrot
21:58 cotto All messages about the migration are drafted.  I'm itching to send them out.
22:11 dukeleto cotto: leto/testing can be deleted. just a "does this work" branch
22:14 dngor_ joined #parrot
22:15 dngor left #parrot
22:15 brianwisti left #parrot
22:19 dngor joined #parrot
22:20 dngor_ left #parrot
22:31 whiteknight dukeleto: copy the svn repo locally before doing the import into git, if you can
22:31 whiteknight saves a hell of a lot of time in network transactions
22:32 sorear How does one copy a svn repo?
22:34 dukeleto whiteknight: i requested someone to give me a local copy of the svn repo, but i haven't gotten one and i forgot who I asked
22:34 dukeleto sorear: a tarball of the actual svn repo on disk
22:34 whiteknight sorear: ah, I can't remember the command off the top of my head. svnadmin something
22:39 rurban_ joined #parrot
22:42 rurban left #parrot
22:42 rurban_ is now known as rurban
22:45 bacek_at_work aloha, humans
22:46 davidfetter 'sup bacek_at_work
22:46 bacek_at_work davidfetter, "_at_work" :)
22:47 * davidfetter too
22:47 davidfetter fortunately, work is a pretty cool place to be, atm
22:47 bacek_at_work yeah
22:47 davidfetter i get to work on Neat Stuff™
22:47 bacek_at_work Playing with Plack/PSGI. It's really cool
22:48 * davidfetter doing things with postgresql
22:49 whiteknight bacek++ # The generational gc work
22:49 bacek_at_work whiteknight, hey! Do you have time to look at it? I'm slightly stuck...
22:50 dmalcolm left #parrot
22:50 whiteknight I can't tonight. probably tomorrow
22:50 whiteknight I don't have my dev laptop back yet
22:50 bacek_at_work fair enough
22:56 whiteknight I will tomorrow. What do you need looking at?
23:00 bacek_at_work whiteknight, overall logic. Does it make sense, etc.
23:00 whiteknight ok
23:00 bacek_at_work To increase Bus Number :)
23:18 whiteknight yeah
23:18 silug left #parrot
23:21 kurahaupo left #parrot
23:53 theory left #parrot

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

Parrot | source cross referenced