Camelia, the Perl 6 bug

IRC log for #parrot, 2010-11-01

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:03 whiteknight joined #parrot
00:13 dalek parrot: r49760 | plobsing++ | branches/gsoc_nci/src/pmc/nci.pmc:
00:13 dalek parrot: fix calculation of number of elements in values
00:13 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49760/
00:24 dngor left #parrot
00:24 dngor joined #parrot
00:50 dngor left #parrot
00:53 dngor joined #parrot
01:04 jsut_ joined #parrot
01:09 jsut left #parrot
02:12 whiteknight left #parrot
02:17 AzureSto_ joined #parrot
02:18 GeJ_ joined #parrot
02:18 particle1 joined #parrot
02:20 theory left #parrot
02:20 jan left #parrot
02:20 x3nU left #parrot
02:20 confound left #parrot
02:20 he left #parrot
02:20 bacek_at_work left #parrot
02:20 aloha left #parrot
02:20 pmichaud left #parrot
02:20 sECuRE left #parrot
02:20 autark left #parrot
02:20 moritz left #parrot
02:20 dukeleto left #parrot
02:20 jnthn_ left #parrot
02:20 Tene left #parrot
02:20 ascent left #parrot
02:20 Maddingue left #parrot
02:20 szabgab left #parrot
02:20 Infinoid left #parrot
02:20 KatrinaTheLamia left #parrot
02:20 nopaste left #parrot
02:20 AzureStone left #parrot
02:20 PacoLinux left #parrot
02:20 contingencyplan left #parrot
02:20 chromatic left #parrot
02:20 Kulag left #parrot
02:20 cosimo left #parrot
02:20 PerlPilot left #parrot
02:20 hudnix left #parrot
02:20 dzoe left #parrot
02:20 atrodo left #parrot
02:20 particle left #parrot
02:20 cotto left #parrot
02:20 GeJ left #parrot
02:20 mikegrb left #parrot
02:20 dip left #parrot
02:20 simcop2387 left #parrot
02:20 slavorgn left #parrot
02:20 frodwith left #parrot
02:20 ingy left #parrot
02:20 Util left #parrot
02:21 frodwith joined #parrot
02:22 Kovensky left #parrot
02:22 athomaso1 left #parrot
02:22 Util joined #parrot
02:22 PerlJam joined #parrot
02:25 dip joined #parrot
02:27 cosimo joined #parrot
02:28 athomason joined #parrot
02:29 mikegrb joined #parrot
02:30 ingy joined #parrot
02:32 cotto joined #parrot
02:32 contingencyplan joined #parrot
02:32 PacoLinux joined #parrot
02:33 s1n joined #parrot
02:33 atrodo joined #parrot
02:33 chromatic joined #parrot
02:33 slavorgn joined #parrot
02:33 dzoe_ joined #parrot
02:33 [hudnix] joined #parrot
02:33 simcop2387 joined #parrot
02:33 Drossel joined #parrot
02:33 theory joined #parrot
02:33 jan joined #parrot
02:33 x3nU joined #parrot
02:33 ascent joined #parrot
02:33 nopaste joined #parrot
02:33 szabgab joined #parrot
02:33 he joined #parrot
02:33 jnthn_ joined #parrot
02:33 bacek_at_work joined #parrot
02:33 aloha joined #parrot
02:33 Maddingue joined #parrot
02:33 dukeleto joined #parrot
02:33 pmichaud joined #parrot
02:33 Tene joined #parrot
02:33 KatrinaTheLamia joined #parrot
02:33 moritz joined #parrot
02:33 sECuRE joined #parrot
02:33 autark joined #parrot
02:33 Infinoid joined #parrot
02:33 Kovensky joined #parrot
02:33 davidfetter left #parrot
02:48 cotto ~~
02:53 atrodo Evening cotto.  looked like there was quite the good lorito chat
02:53 sorear hollo atrodo
02:53 sorear hello cotto
02:53 cotto orly?
02:54 cotto atrodo, what are you referring to?
02:55 atrodo The one that you and sorear had yesterday-ish
02:56 cotto ok, so it's nothing I missed
02:57 atrodo nope.  I was commenting on it since I missed it
02:58 atrodo one thing that stuck out to me was the idea of JIT'ing lorito on a stack based vm
02:58 cotto I threw a batch of questions on the wiki at the bottom of LoritoDesignQuestions based on my own questions and what I got from whiteknight and atrodo.
02:59 cotto Yes.  That was an interesting question.
03:00 sorear I can write a very simple deregisterizer
03:01 sorear a value can stay on the stack if 1. its lifetime is contained in one basic block 2. its lifetime is properly nested wrt all other such values
03:01 atrodo it also brings an interesting idea that with lorito, if we have specs and test suite, it doesn't necessarily have to be only written in C
03:02 cotto atrodo, not at all.  We could do a prototype in Perl to make it easier to hack on.
03:03 atrodo or, be really crazy and write a js engine so parrot code, sans external code, ran in a web browser.  I can't imagine that'd be efficient tho
03:03 cotto or nqp, but that's convenient and slow instead of convenient and fast
03:04 cotto I'm not going to stop anyone from doing that, but I will bring popcorn.
03:04 cotto that said, Google's done a lot with gwt, which compiles Java to js
03:05 sorear one deeper thing I need to do is to implement real continuations
03:05 sorear I don't know how much rippling this will cause
03:05 sorear my loritoid only does coroutines so far
03:08 sorear which is fine for p6 but no good for lorito
03:10 cotto sorear, continuations will be on top of Lorito, not built in
03:14 sorear Eh?
03:15 sorear You mean, like, in C code?
03:15 cotto http://irclog.perlgeek.de/p​arrot/2010-07-23#i_2600954
03:29 confound joined #parrot
03:56 ash_ joined #parrot
04:09 jsut joined #parrot
04:13 jsut_ left #parrot
04:19 s1n left #parrot
04:21 dalek parrot: r49761 | plobsing++ | branches/gsoc_nci (8 files):
04:21 dalek parrot: eliminate use of libffi-specific nci.pmc
04:21 dalek parrot: libffi framebuilder now invoked using IGLOBALS_NCI_FB_{CB,UD} hooks
04:21 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49761/
04:25 Limbic_Region left #parrot
04:27 jhelwig joined #parrot
04:37 dalek parrot: r49762 | plobsing++ | branches/gsoc_nci (4 files):
04:37 dalek parrot: eliminate gen::libffi
04:37 dalek parrot: we're no longer copying custom versions of nci.pmc around
04:37 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49762/
04:40 cxreg plobsing++ indeed
04:40 cxreg plobsing: any idea how much works remains on that branch?
04:42 plobsing cxreg: it now passes make test and no longer does any tricky copying around of near-identical pieces of code
04:42 cxreg sweet.  looks like you put in a good 20 commits on it the last couple days
04:43 plobsing there are a few areas that need fixing up before this gets merged
04:43 plobsing there's some todos (mostly in src/nci/libffi.c) and codetest is hurting pretty bad
04:44 plobsing in terms of the subsystem though, there is much work left to do. But I'll leave that for another branch.
04:45 plobsing I'd like to have this branch in a testable/mergable state by end of week
04:45 dukeleto plobsing: are we talking about gsoc_nci ?
04:45 plobsing yes
04:45 dukeleto plobsing: awesome!
04:45 plobsing only 2 months to late!
04:45 plobsing s/to/too/
04:46 ash_ :-(
04:46 plobsing if there's one thing I do well, it's procrastinate
04:46 dukeleto plobsing: a half a loaf is better than none
04:46 ash_ thats my fault, i have dropped the ball on getting gsoc_nci merged
04:46 dukeleto ash_: we will give you the appropriate amount of lashings
04:47 ash_ grad schools has been eating into my spare time (ie. i don't really have any right now)
04:47 cxreg ash_: were it not for you, this would not have gotten done at all :)  ash_++
04:47 ash_ plobsing: i can try to fix up any of my code that doesn't pass the coding standard
04:48 dukeleto ash_: writing a blurb about what new things people can do after the gsoc_nci branch lands, that would be useful. Something for NEWS
04:49 ash_ okay, i can do that right now
04:49 ash_ does test 42 (nci_vP) fail still or has someone tracked that down? (completely out of the loop on the state of gsoc_nci since i did my last few commits)
04:50 plobsing unless it was skipped, it is working now. fulltest gets all the way to codetest before bailing
04:52 dukeleto plobsing: so all tests pass in "make test" ?
04:52 plobsing yes. I managed to get the last two somehow. Push enough code around and problems go away apparently.
04:54 dukeleto plobsing++ # awesomesauce
05:03 ash_ did it go away by just re-organizing my code? (means i probably had a scoped variable freed in the wrong scope, thats my guess anyway)
05:05 plobsing I'm not entirely sure myself
05:05 plobsing I did rearrange a lot. I also had to kill some parts for the greater good.
05:07 ash_ thats fine, i think i did some 'Bad Things ™' in my code in some places
05:10 ash_ hope it wasn't to unreadable and made sense to someone other than me
05:11 plobsing that's what I'm hoping now.
05:12 ash_ my use of void* and malloc'ed chunks of data felt wrong, but i didn't know a better way of doing it, but i see you thought of doing  a much more organized strcut with a union (pmc_holder_t), thats a lot better than me mallocing random memory and just pointing to it later on
05:14 plobsing oh we still do a little of that elsewhere. I'm not terribly fond of it, but I have yet to think of a better way.
05:21 plobsing an important point for testing: make sure to configure --without-core-nci-thunks and --without-extra-nci-thunks. otherwise, the framebuilder will be shaddowed by a pre-filled nci thunk cache
05:21 sorear I wonder how much of our branch merge yak shaving could be handed off to GCI
05:24 ash_ plobsing: did the signature parse get changed?
05:24 plobsing sorear: the pain of svn merges is not the first (or one of the first) impressions of open source I'd like people to have
05:24 plobsing ash_: yes. it was kinda tangled around generating the cif and the pcc signature (which was coming out wrong)
05:25 ash_ so, the nci signatures will not be update yet? i guess that will be held off till a more logical/simple system can be specified?
05:25 ash_ (signature parsing, i mean)
05:25 dukeleto sorear: yak-shaving doesn't make a good task
05:26 dukeleto plobsing: i don't quite understand what you mean about an "important point for testing"
05:26 plobsing ash_: yes. that's the biggest hit the branch had to take
05:26 plobsing dukeleto: if you don't configure special like that, the builtin thunks (which cover all tested functions) will be built in to parrot and the framebuilder will never get called
05:27 ash_ dukeleto: if you built it with the default nci-thunks and it detected one, it would use those instead of using libffi
05:27 plobsing frame builder only gets called on an NCI frame cache miss
05:29 ash_ plobsing: should i move that into a library maybe? it was a basically the same syntax as perl's pack with a few modifications to accommodate the old nci system
05:31 plobsing ash_: I kinda like the array of int-flags that I'm using internally. no parsing required. I'm considering exposing that as an interface (it is probably easier to use programatically, eg: zavolaj)
05:33 ash_ what about complex data structures?
05:34 plobsing what about them? they can't be used in calls. the most complex thing, a struct (without nesting), might make that tricky to do; but I'm sure we'll figure something out.
05:36 ash_ well, that was sorta part of the whole parse thing i wrote, it could parse a structure (written like "{ii}" which means struct { int; int }; ), maybe there is a different way we could go about that?
05:37 plobsing not sure. haven't really thought about it. I just was sick of dealing with raw chars. flags are explicit about what they mean.
05:40 ash_ thats fine
05:52 fperrad joined #parrot
05:59 fperrad_ joined #parrot
06:00 theory left #parrot
06:03 fperrad left #parrot
06:03 fperrad_ is now known as fperrad
06:53 chromatic left #parrot
06:57 ash_ left #parrot
07:08 uniejo joined #parrot
07:26 cognominal left #parrot
08:34 lucian_ joined #parrot
08:35 lucian_ left #parrot
08:38 lucian left #parrot
08:50 jsut left #parrot
08:50 confound left #parrot
08:50 Kovensky left #parrot
08:50 Infinoid left #parrot
08:50 autark left #parrot
08:50 sECuRE left #parrot
08:50 pmichaud left #parrot
08:50 aloha left #parrot
08:50 bacek_at_work left #parrot
08:50 jnthn_ left #parrot
08:50 he left #parrot
08:50 x3nU left #parrot
08:50 jan left #parrot
08:50 Drossel left #parrot
08:50 moritz left #parrot
08:50 ascent left #parrot
08:50 simcop2387 left #parrot
08:50 atrodo left #parrot
08:50 Maddingue left #parrot
08:50 [hudnix] left #parrot
08:50 szabgab left #parrot
08:50 KatrinaTheLamia left #parrot
08:50 dzoe_ left #parrot
08:50 slavorgn left #parrot
08:50 nopaste left #parrot
08:50 dukeleto left #parrot
08:50 Tene left #parrot
08:50 PacoLinux left #parrot
08:50 contingencyplan left #parrot
08:50 cotto left #parrot
08:52 jsut joined #parrot
08:52 confound joined #parrot
08:52 Kovensky joined #parrot
08:52 Infinoid joined #parrot
08:52 autark joined #parrot
08:52 sECuRE joined #parrot
08:52 moritz joined #parrot
08:52 KatrinaTheLamia joined #parrot
08:52 Tene joined #parrot
08:52 pmichaud joined #parrot
08:52 dukeleto joined #parrot
08:52 Maddingue joined #parrot
08:52 aloha joined #parrot
08:52 bacek_at_work joined #parrot
08:52 jnthn_ joined #parrot
08:52 he joined #parrot
08:52 szabgab joined #parrot
08:52 nopaste joined #parrot
08:52 ascent joined #parrot
08:52 x3nU joined #parrot
08:52 jan joined #parrot
08:52 Drossel joined #parrot
08:52 simcop2387 joined #parrot
08:52 [hudnix] joined #parrot
08:52 dzoe_ joined #parrot
08:52 slavorgn joined #parrot
08:52 atrodo joined #parrot
08:52 cotto joined #parrot
08:52 contingencyplan joined #parrot
08:52 PacoLinux joined #parrot
08:53 uniejo left #parrot
08:55 uniejo joined #parrot
09:30 bacek left #parrot
09:52 muixirt joined #parrot
09:52 muixirt good morning
09:54 muixirt the line numbers parrot emits in the case of an error don't look to be reliable, any advice?
10:06 moritz muixirt: file tickets for those cases where they are off
10:08 bacek joined #parrot
10:09 muixirt moritz: so line numbers aren't generally off?
10:40 tadzik joined #parrot
10:41 tadzik o/
10:48 tadzik left #parrot
10:50 tadzik joined #parrot
10:55 moritz muixirt: no, often they are right
10:55 moritz unless you produce a weird syntax error
10:57 muixirt I meant in the case of an error like 'Null PMC access in type()'
10:57 muixirt or is that related to Ticket #1652 too?
10:58 moritz that sounds like a rakudo error, not a parrot error
11:00 moritz (the "null pmc access in type", that is)
11:00 moritz there the line number has a meaning, but it doesn't point into the user's source file
11:00 moritz which is entirely rakudo's fault, afaict
11:01 muixirt moritz: fyi I'm fiddling with the ecmascript corpse
11:31 plobsing left #parrot
11:32 PacoLinux left #parrot
11:34 dalek parrot: r49763 | bacek++ | branches/generational_gc (4 files):
11:34 dalek parrot: Get rid of PObj_is_special_FLAG. They are all special now.
11:34 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49763/
11:34 dalek parrot: r49764 | bacek++ | branches/generational_gc/src/pmc.c:
11:34 dalek parrot: Rewrite code to make gcc happy
11:34 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49764/
12:29 he_ Hm, can someone please do cp config/gen/platform/freebsd/sysmem.c config/gen/platform/netbsd/ and commit the result?  Smoke test succeeds on r49680 with that change in place.
12:33 tadzik left #parrot
12:33 whiteknight joined #parrot
12:35 whiteknight good morning, #parrot
12:52 whiteknight good morning, #parrot
12:52 whiteknight ...?
12:54 moritz oh hai whiteknight
13:08 whiteknight hello moritz. How are you doing today?
13:09 moritz whiteknight: great. Last week I was on Mallorca, today is a public holiday in parts of Germany (including where I live)
13:19 muixirt moritz: why didn't you say that earlier, wanted to shop but stood in front of closed doors :-)
13:19 moritz muixirt: I had no idea where you live :-)
13:20 moritz muixirt: I thought somewhere in any <de at ch> maybe, but the holliday rules are not very uniform over that area
13:21 muixirt forgot that I live in a catholic area
13:24 muixirt so no dancing today: Tanzverbot!
13:27 moritz at least not in public venues
13:27 moritz http://www.perlmonks.org/?node_id=868728
13:28 moritz sorry, wrong channel, wanted to post to #perl6
13:28 estrabd joined #parrot
13:36 lidden joined #parrot
13:51 particle1 is now known as particle
14:02 PacoLinux joined #parrot
14:36 whiteknight left #parrot
14:42 tadzik joined #parrot
14:43 Andy left #parrot
14:48 tadzik hello
14:48 cognominal joined #parrot
14:53 PacoLinux left #parrot
15:01 ash_ joined #parrot
15:06 atrodo left #parrot
15:06 atrodo joined #parrot
15:11 bluescreen joined #parrot
15:12 atrodo left #parrot
15:13 atrodo joined #parrot
15:13 uniejo left #parrot
15:31 muixirt left #parrot
15:41 mjl69 joined #parrot
15:49 PacoLinux joined #parrot
15:50 cotto tilde tilde
15:50 tadzik oh slash
15:51 theory joined #parrot
15:52 atrodo backslash oh
15:58 allison joined #parrot
16:17 silug left #parrot
16:19 whiteknight joined #parrot
16:24 plobsing joined #parrot
16:39 patspam left #parrot
16:44 cotto whiteknight, have you tried this? http://today.msnbc.msn.com/id​/39831216/ns/today-foodwine/
16:45 atrodo hurray!  That's the only meal I need all day!
16:48 whiteknight cotto: I am angry at you for introducing me to this concept
16:52 cotto If you look like a balloon next time I see you, I'll buy you some medifast or something.
16:53 cotto ;)
16:55 cotto seen dukeleto?
16:55 aloha dukeleto was last seen in #perl6 1 hours 44 mins ago joining the channel.
17:11 dukeleto 'ello
17:11 cotto 'i
17:11 cotto So, when's git o'clock?
17:12 whiteknight is the big migration happening sometime soon?
17:12 dukeleto cotto: how much do you want to help?
17:12 dukeleto whiteknight: how much time do you have to donate to the migration?
17:12 whiteknight dukeleto: besides well-wishing, basically 0 until I get home
17:12 * dukeleto also has this odd thing called a $job
17:13 whiteknight in about 5 hours, I can focus on it pretty well, I think
17:13 dukeleto i still don't have a local dump of the subversion database.
17:13 dukeleto which i need to regenerate the repo with a proper authormap in a reasonable amount of time.
17:15 * cotto updates his local dump of the svn repo
17:19 cotto dukeleto, I'll put a tarball of a svnsync'd db (as of r49764) up as soon as it's finished being tarballed.
17:21 cotto Will that help?
17:21 cotto Updating it is pretty simple, if a bit slow.
17:22 dukeleto cotto: yes, that will help.
17:23 dukeleto cotto: if people commit to the svn repo after your tarball is created, i will probably have to do this again. or perhaps i can get away with incremental fixes
17:25 cotto Why couldn't you do incremental updates like you've been doing, only with authmap in place?
17:26 dukeleto cotto: authmap happens at conversion time. authmap requires a new conversion (new sha1's) or git filter-branch (new sha1's as well)
17:26 dukeleto cotto: perhaps there is an option to git-svn to use the authmap on every rebase, but i haven't used it before
17:37 cotto dukeleto, http://mksig.org/parrot/parrot_svn_repo.tar.bz2
17:38 cotto that'll expand into a dir named "repo"
17:38 cotto update with svnsync sync file:///path/to/repo
17:42 dukeleto cotto: thanks
17:43 cotto I'm not sure how to solve the authmap problem, but at least it'll be easy to keep up-to-date
17:44 cotto If we have to say "don't commit to svn or git during the switch", I guess we can do that.
17:44 fperrad_ joined #parrot
17:44 moritz sounds sane.
17:44 whiteknight we can put out an email asking people not to commit during the dump/migration process, or we can got nuclear and lock the repo completely for the duration
17:45 moritz or you can just undo any dangling commits in the resulting git repo
17:45 cotto or both
17:47 fperrad left #parrot
17:47 fperrad_ is now known as fperrad
17:48 cotto If that'll be the plan, you should definitely do a dry run first.
17:51 tadzik dukeleto: I can donate some time to the migration (today at least)
17:56 dukeleto tadzik: that is awesome to hear
17:56 cotto tadzik++
17:57 dukeleto tadzik: have you looked at the GitMigration wiki page?
17:57 tadzik dukeleto: I visit it from time to time
17:57 dukeleto tadzik: awesome to hear!
17:57 * tadzik feels motivated to take a walk to the wiki
17:57 cotto dukeleto, is it too early to send out a migration schedule to parrot-*?
17:58 cotto tadzik, it's disorganized but there's some good stuff there.
17:58 cotto Feel free to organize it to whatever degree you have the tuits for.
17:58 dukeleto cotto: if you want to cut the cord, go for it.
17:58 patspam joined #parrot
17:59 dukeleto cotto: if you all promise me beer and fine cheese, perhaps i can cut over to git tonight
17:59 cotto 1) That would be amazing.
17:59 tadzik cotto: what can I do? I don't feel like messing with the wikipage as I don't know what is it expected to show, besides stuff that it alredy does
18:00 cotto 2) I don't want it to be too sudden for people who might not be aware that it's coming.
18:00 cotto tadzik, if one page seems like it should be linked from another, add a link
18:01 cotto dukeleto, also, it depends on how fine a cheese.  They can get pretty spendy.
18:01 tadzik cotto: is the help needed with the git migration, or with /wiki/GitMigration?
18:01 cotto The wiki page outlines the process and is mostly current.
18:02 dukeleto cotto: i like Goat Gouda
18:02 tadzik so what needs to be done, from a practical POV?
18:02 * cotto googles
18:03 * cotto is a fan of Jarlsberg
18:04 dmalcolm joined #parrot
18:05 cotto dukeleto, if it didn't happen tonight, how soon could it happen?  Tonight is an option, but I also want to give people more notice if possible.
18:06 dukeleto I need to regen a new git repo with an authormap. Then push that to github, and migrate old github branches to the new repo with patches
18:06 whiteknight We definitely should send a warning email out to parrot-dev and maybe elsewhere ASAP if we are planning an imminent migration
18:07 dukeleto cotto: i will be away on business for about a week soon, so i would like to do this sooner rather than later
18:07 cotto whiteknight, I have the message drafted already so I just need to fill in the blanks.
18:07 whiteknight don't patch branches. That should save some time
18:07 cotto I've had it ready since Oct 14th.
18:07 dukeleto whiteknight: wanna send the email? If you guys start ringing the alarm bells, i will feel more motivated to get something done.
18:08 cotto I'll nopaste the draft.
18:08 dukeleto whiteknight: creating patches for the branches is trivial, there are only 2
18:08 whiteknight I'm happy to send it, sure
18:08 nopaste "cotto" at 192.168.1.3 pasted "DRAFT git migration email to parrot-dev and parrot-users" (18 lines) at http://nopaste.snit.ch/25114
18:08 * dukeleto thinks somebody should send out an email to parrot-dev, and warn them that git migration is imminent
18:09 dukeleto cotto: don't forget to send it to the rakudo peeps
18:09 whiteknight dukeleto: if you're doing the majority of the heavy-lifting, you dictate what time works best for you
18:09 cotto perl6-internals?
18:09 dukeleto cotto: perl6-language probably, or perl6-announce
18:09 whiteknight I'll send out the email to the correct recipients with the correct times
18:09 tadzik #perl6 maybe
18:10 cotto whiteknight, deal
18:10 dukeleto i need people to interact with the OSUOSL people for me, to coordinate some of that stuff
18:10 tadzik Perl 6 folks rely on svn parrot only in Rakudo's Configure.pl script
18:10 whiteknight dukeleto: what do you  need from OSUOSL?
18:10 dukeleto whiteknight: look on the GitMigration wiki, there is a bunch of stuff only they can do? like trac stuff
18:14 cotto dukeleto, how long will the migration take with the authmap?
18:17 dukeleto cotto: shouldn't take more than a few hours with a local svn dump. I haven't done it yet, tho.
18:17 dukeleto cotto: it could take as little as half an hour. Not sure.
18:18 cotto is a dry run practical?
18:22 silug joined #parrot
18:32 dukeleto cotto: sure. i will try to do a dry run today, to time it
18:33 cotto ok.  Once you know that it works and how long it can take, whiteknight can send out the announcement.
18:33 * cotto is excited
18:33 lucian joined #parrot
18:33 moritz so git migration is pretty immediate?
18:33 * moritz gits all excited
18:33 dukeleto moritz: with well wishes and promises of beer, yes :)
18:33 cotto barring difficulties, yes
18:33 allison left #parrot
18:34 * cotto buys pbr and some goat gouda
18:34 * cotto carefully notes that dukeleto never specified what beer
18:34 moritz dukeleto: when we meet in real life, we have to go to a pub and talk about parrot, Perl 6, solitons and other geek topics... yes, you'll get a beer :-)
18:35 dukeleto moritz: awesome! Will you be at YAPC::NA next year? I will do my best to be there.
18:36 moritz dukeleto: very unlikely. I'm going to become a father soon, so priorities shift
18:37 dukeleto moritz: ah, i see. I will have to make it to a conference on your side of the pond.
18:37 dukeleto moritz: congrats on your new fatherhood :)
18:37 moritz ETA is 2010-02-01
18:37 moritz thanks
18:37 moritz erm, 2011-02-01
18:38 moritz offby1error hit again!
18:55 cotto ENOBABY
19:01 whiteknight dukeleto: if you ever make it out out PA, I've got beer for you
19:01 whiteknight I'm a gouda fan myself, so I'll pick up some of the good stuff
19:02 dukeleto whiteknight: sweet. I won't forget that.
19:02 dukeleto whiteknight: i will pay you a visit the next time i visit family in the Northeast
19:06 whiteknight where in the northeast?
19:07 Tene Where is YAPC::NA next year?  Things are lining up to be much more reasonable for me to go finally.
19:08 Tene ah, http://www.yapc2011.us/yn2011/
19:09 jsut_ joined #parrot
19:14 jsut left #parrot
19:16 lucian_ joined #parrot
19:20 lucian left #parrot
19:24 jsut joined #parrot
19:28 allison joined #parrot
19:29 jsut_ left #parrot
19:36 whiteknight I'm probably going to delete my embedding API branch and reapply the changes to a new git branch after the migration
19:47 cotto whiteknight, are you not happy with it?
19:54 ash_ left #parrot
20:00 whiteknight I am happy with it. I haven't done any work on it in two days, and would rather start fresh in the new system
20:00 whiteknight anyway, leaving now
20:00 whiteknight left #parrot
20:04 patspam1 joined #parrot
20:07 patspam left #parrot
20:17 fperrad left #parrot
20:27 arnsholt left #parrot
20:27 jsut_ joined #parrot
20:32 jsut left #parrot
20:47 arnsholt joined #parrot
20:51 mjl69 left #parrot
20:53 ash_ joined #parrot
21:21 whiteknight joined #parrot
21:32 theory left #parrot
21:34 patspam1 left #parrot
21:49 dngor_ joined #parrot
21:49 dngor left #parrot
22:01 bacek left #parrot
22:07 dngor_ is now known as dngor
22:11 patspam joined #parrot
22:14 cotto dukeleto, have you tried a dry run of the git import with an authmap?
22:27 dngor left #parrot
22:33 kid51 joined #parrot
22:43 dukeleto cotto: not yet. Had a some errands to run. I will try it nowish.
22:48 tadzik left #parrot
22:49 theory joined #parrot
22:56 * dukeleto is untarring the svnball now
23:03 bluescreen left #parrot
23:15 dngor joined #parrot
23:16 bacek_at_work ~~
23:17 dukeleto I am missing an email in CREDITS for GeJ
23:17 dukeleto GeJ_: ping
23:17 dukeleto cxreg: ping
23:18 dukeleto cxreg: i am missing info in CREDITS for you
23:18 cxreg cxreg@pobox.com
23:20 dukeleto cxreg++
23:20 dukeleto cxreg: do you want a realname in CREDITS?
23:20 GeJ_ is now known as GeJ
23:20 GeJ dukeleto: pong
23:21 cxreg sure, Dave Olszewski
23:21 dukeleto GeJ: which email would you like to use for CREDITS
23:21 * dukeleto is generating an authormap for Git, which reads CREDITS and creates the proper committer name + email for git
23:25 dukeleto cotto: ping
23:25 cotto pong
23:25 * cotto got back from a haircut at just the right time, apparently
23:26 dukeleto cotto: so if i want to get the latest svn rev into my svn repo dump, i do "svnsync sync file://repo" ?
23:26 cotto yes
23:26 cotto Nothing will happen now because nobody's committed anything.
23:26 cotto file:///path/to/repo
23:27 dukeleto cotto: I have :)
23:27 cotto It has to be an absolute path because svnsync is nvb
23:27 dukeleto cotto: yes, i just got it to work. Sweet.
23:28 cotto dukeleto, are you thinking tonight or later?
23:28 cotto If tonight, we should get the notification out asap.
23:28 dukeleto cotto: i am going to do a dry run pretty soon
23:28 dukeleto cotto: to see how long regenerating the repo with a new authormap takes
23:29 cotto dukeleto, makes sens
23:29 cotto e
23:38 dalek parrot: r49765 | dukeleto++ | trunk/CREDITS:
23:38 dalek parrot: Update CREDITS in preparation for git conversion
23:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49765/
23:41 theory left #parrot
23:44 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#879) fulltest) at r49764 - Ubuntu 10.10 i386 (g++-4.5)
23:46 mikehh dukeleto: what is happening re tests that work on svn
23:48 mikehh or use info from svn eg smoke/smolder
23:48 cotto mikehh, we have a git branch that rips them out
23:48 dukeleto mikehh: "make smoke" knows how to grab info from git repos already. that was added long ago
23:49 dukeleto mikehh: and any test that looks at svn props will be deleted in the git repo
23:49 mikehh cotto, dukeleto: ok, waiting to try all that :-}
23:51 * kid51 is overjoyed; he just recruited a new smoketester;  http://smolder.parrot.org/app​/projects/report_details/880
23:54 mikehh ok another question, I usually run config with --maintainer, and if it regens  compilers/imcc/imclexer.c or h I usually just rm the files and svn up, unless I want to commit, will this work in git?
23:54 kid51 is now known as kid51_at_dinner
23:54 dukeleto mikehh: not sure. try it out.
23:55 bacek_at_work mikehh, git checkout compilers/imcc
23:55 dalek parrot: r49766 | dukeleto++ | trunk (3 files):
23:55 dalek parrot: [tools] Make the author map code an actual file and remove it from CREDITS
23:55 dalek parrot: review: http://trac.parrot.org/parrot/changeset/49766/
23:55 cotto mikehh, why do you do that?
23:56 mikehh dukeleto: I will, but I think I have had problems with this type of thing before (in git that is), need to work it out
23:57 cotto Is there any benefit?
23:57 cotto (if you're not hacking on imcc, that is)
23:58 mikehh cotto: usually because the change it trivial (usually differences with bison version or something)
23:59 ash_ left #parrot
23:59 mikehh cotto: adds some additional testing that I incorporate and check on

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

Parrot | source cross referenced