Parrot 2.9.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | merge gsoc_nci |
Set by moderator on 2 November 2010.
kid51 whiteknight: Yes. By not doing it today, we inevitably run into conflicts as we get closer to US Thanksgiving 00:00
dalek rrot: r49795 | jkeenan++ | trunk/config/init/hints/freebsd.pm:
Correct syntax error spotted by doughera++; TT #1841.
 00:07
bacek_at_work ~~ 00:16
Infinoid sorear: Yeah, looks interesting. Does it work? :) 00:21
00:24 bacek left
sorear Infinoid: It's live; I hope it works! 00:24
00:32 kid51 is now known as kid51_at_dinner
Infinoid Awesome 00:47
sorear: It looked like you had a single ongoing queue file, does that keep getting larger over time or is it cleaned out somehow?
I saw the poll script but didn't see the botnix plugin in your repo
sorear It just grows over time 00:49
Cleaning would be a manual process, if it ever happens; I don't expect the file to grow more than about 1MB/year
GeJ kid51_at_dinner: FYI make fulltest PASSes on FreeBSD 7 amd64. 00:50
Infinoid Maybe we should have a #dalek-plugins channel to discuss its design :)
GeJ kid51_at_dinner: and as for the __APPLE_CC__ warning (in TT #1841) it has more to do with gettext and FreeBSD than Parrot. I have it on FreeBSD 7, and I'm almost certain that I don't have it on my 8.x box at home. I will confirm it tonight. 00:53
01:07 kid51_at_dinner is now known as kid51
kid51 GeJ: Thanks for those reports. The taptinder box is reporting successful build as well. And discussion with mikehh earlier led us to the same conclusion re __APPLE_CC__ 01:08
GeJ Do you submit smolder reports on that FreeBSD 7 amd64 box?
GeJ I may need to install some missing deps here. But I definitely do it on my home box. 01:10
kid51 They would be helpful. 01:11
dalek TT #1841 closed by jkeenan++: Add 'osvers' key-value pair to Configure.pl 01:15
TT #1841: trac.parrot.org/parrot/ticket/1841
01:15 whiteknight left 01:31 arnsholt joined
dalek rrot: r49796 | jkeenan++ | trunk/lib/Parrot/Headerizer.pm:
Correct spelling error. Improve some POD formatting.
 01:41
01:56 lucian left
dalek rrot: r49797 | jkeenan++ | branches/tt532_headerizer_refactor:
Create branch to work on ļæ½trac.parrot.org/parrot/ticket/532.
 01:56
01:57 dngor_ joined 01:59 dngor left 02:06 dngor_ is now known as dngor
mikehh haven't got any emails from parrot.org for a few hours 02:12
dalek rrot: r49798 | jkeenan++ | trunk/config/init/hints/darwin.pm:
Correct erroneous argument to /usr/sbin/sysctl.
rrot: r49799 | jkeenan++ | branches/tt532_headerizer_refactor (4 files):
Begin to move functions out of tools/dev/headerizer.pl and into new module lib/Parrot/Headerizer/Functions.pm.
dukeleto waves hello from NC 02:15
sorear aloha, seen bacek_at_work 02:16
aloha sorear: bacek_at_work was last seen in #parrot 2 hours ago saying "~~".
bacek_at_work sorear, pong
bacek_at_work is have to run in less than 10 minutes
sorear bacek_at_work: Is it just me, or did aloha handle messages at one point? 02:17
bacek_at_work sorear, erm.
msg sorear It was from early ages of aloha
aloha OK. I'll deliver the message.
sorear Did you remove that b/c of phenny and purl?
bacek_at_work "b/c"?
dukeleto mikehh: i haven't seen any email to parrot-dev either. I never saw cotto's email about PDS either
sorear because
kid51 filed Trac ticket with OSU OSL re lack of mail to parrot-dev 02:18
sorear oh, it's #parrot only
that makes sense
bacek_at_work I should have aloha on #perl6 as well
kid51 Mail is being received and is being listed on our pipermail archive site, but it's not being distributed.
dukeleto that sucks.
bacek_at_work sorear, aloha should cross-deliver messages I think. 02:19
Ah. I didn't load Msg plugin in "#perl6" version. 02:20
sorear, do you need it?
sorear no
plobsing tried to use it, failed, and I was wondering why
I showed him phenny
bacek_at_work fair enough 02:21
afk # meetings...
02:27 stilgar joined
dalek rrot: r49800 | jkeenan++ | branches/tt532_headerizer_refactor/tools/dev/headerizer.pl:
Move declaration of main() into first position among functions.
 02:28
rrot: r49801 | jkeenan++ | branches/tt532_headerizer_refactor (2 files):
Encapsulate some code within main() into a function. Place that function in Parrot::Headerizer::Functions.
sorear dukeleto: Infinoid just asked me in pm if parrot will be moving to github
dalek rrot: r49802 | jkeenan++ | branches/tt532_headerizer_refactor (2 files):
Create a file to hold tests of Parrot::Headerizer::Functions.
 02:43
rrot: r49803 | jkeenan++ | branches/tt532_headerizer_refactor (2 files):
Write basic tests of read_file() and write_file().
 02:59
03:12 kid51 left
dalek rrot: r49804 | jkeenan++ | branches/tt532_headerizer_refactor/t/tools/dev/headerizer/01_functions.t:
Add tests of print_headerizer_warnings().
 03:15
rrot: r49805 | petdance++ | trunk/src/exceptions.c:
consting
 03:31
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#968) fulltest) at r49803 - Ubuntu 10.10 i386 (g++-4.5) 03:34
03:42 patspam joined
cotto pongs sorear 03:44
sorear cotto: Why do you want to keep the wiki scraper?
cotto I don't care as long as the work that it does gets done in some other way
sorear ok. 03:45
cotto Something more elegant would be great. Wiki scraping is a hack. 03:46
did you have something in mind? 03:47
sorear the replacement is misc/dalek-conf.json in the mu repository 03:48
not everything has been migrated 03:49
cotto How many people have a bit to update that? 03:50
sorear over a dozen, and they're handed out no questions asked 03:54
mu is the spiritual successor to pugssvn
although the larger projects have been moved out of it
cotto wfm 03:55
dukeleto, poing 03:57
(It's like ping, but it bounces.) 03:58
04:03 patspam left
cotto sorear, who do I talk to to get a mu commit bit? 04:16
sorear What's your github id?
cotto cotto
sorear moritz can definitely do it 04:18
I thought I could
cotto moritz, ping
sorear but I can't find the "Admin" button
while logged in as 'perl6'
cotto curious 04:25
since the repo is "perl6/mu"
04:34 lidden left
dukeleto cotto: pong 05:27
05:31 rkazak joined
sorear hello rkazak 05:37
cotto dukeleto, what time tomorrow do you want to do the migration? 05:40
dukeleto cotto: good question. What part of the migration do you mean? 05:42
cotto: the repo switch?
cotto dukeleto, primarily that, yes 05:43
rkazak hi 05:44
cotto also, can you modify the script to mark both autrijus and audrety as audreyt? 05:45
rkazak, hi
rkazak I built parrot on mac osx 10.6, the say command seems to fail? 05:50
cotto rkazak, what exactly are you trying to do that fails? 05:52
rkazak run a simple .pir script 05:53
cotto can you nopaste the script, how you're running it and what the output is?
aloha, nopaste?
aloha cotto: nopaste is is nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl)
plobsing rkazak: can you nopaste the script? (perl tools/dev/nopaste.pl)
rkazak sure 05:54
05:56 bluescreen left
dukeleto cotto: yes, i will fix autrijus -> audreyt 05:57
nopaste "rkazak" at 192.168.1.3 pasted "say failing?" (9 lines) at nopaste.snit.ch/25389 05:58
plobsing rkazak: your string delimiters don't match 05:59
(yes the error message is LTA)
rkazak ouch - sorry... 06:00
that fixed it.
LTA ?
plobsing less than awesome 06:01
dukeleto rkazak: Less Than Awesome
cotto rkazak, no worries. It's hard to know when something's your fault and when it's Parrot's when you're just learning.
rkazak :)
dukeleto rkazak: that error message is pretty bad. I would consider that a bug.
bacek_at_work Existence of IMCC is a bug 06:02
cotto dukeleto, sure. All we have to do is fix imcc.
dukeleto bacek_at_work++
cotto smop
dukeleto cotto: i'll get right on that, after I rewrite the Perl 5 regex engine
cotto 30 minutes, tops
bacek_at_work throwing Dragon Book into cotto 06:03
cotto om nom nom
dukeleto cotto: if by "fix" you mean "burn with a flamethrower", i can do that.
cotto dukeleto, patience. PIRATE's day will come and there will be much rejoicing as imcc goes down in flames. 06:04
dukeleto so, are we not getting emails to parrot-dev? 06:05
cotto dukeleto, apparently not
dukeleto seems like bad timing, with the git migration and all...
cotto thoroughly 06:06
sorear dukeleto: you need a regex engine? I've only written about five, I think I'm the go-to person for those 06:07
06:07 bluescreen joined
rkazak ok guys, thanks - I need to hit the pit as it's late for me. More for me later. 06:08
dukeleto sorear: i will let you know when I need one, thanks :)
dukeleto fiddles with the authormap for the final git conversion 06:09
cotto rkazak, g'night
06:09 rkazak left
cotto weak. The parrot-dev archive shows my reply to the pds scheduling thread, even though nothing went out. 06:10
dukeleto cotto: yeah, kid51 mentioned that the archives are getting stuff, but not the list 06:14
cotto If anyone here wants to attend the online developer's summit, please add your availability to www.doodle.com/q4nwdt9i5k929n2s
06:15 d4l3k_ joined, dalek left 06:16 d4l3k_ is now known as dalek
dukeleto ug. looks like there is only one svn username with spaces in it: "Paul C. Anagnostopoulos". Hopefully my stuff can deal with it. 06:18
i have no CREDITS info for "barney" 06:19
does anybody know anything about "barney" ? 06:20
06:20 dalek left
dukeleto prepares to get trolled 06:20
cotto dukeleto, what about him? 06:21
06:21 dalek joined
dukeleto cotto: he is in svn history, but no CREDITS info 06:21
so i don't have a valid email or name for him/her
06:21 davidfetter left
cotto him 06:21
just a sec 06:22
sorear check what they committed, cross-reference it against project history
?
cotto For some reason, his username is listed as "bernhard"
06:22 rurban joined
cotto he and barney are the same person 06:23
dukeleto barney ~~ bernhard, got it
dukeleto thinks he has a spiffy new author map script 06:24
cotto: when are you planning on telling people to not commit to svn anymore?
cotto as late as possible 06:25
dukeleto cotto: i am going to regenerate parrot with my new authormap in a few mins, are you going to be around for a bit?
cotto sure
06:28 rurban left
dukeleto cotto: the doodle poll says "Virutal Summit". Is that a new kind of summit? ;) 06:31
cotto: i filled out my info on the doodle poll
dukeleto goes back to migrationz
cotto So far there's exactly one slot that'll work for everyone. 06:32
06:33 rurban joined
dukeleto cotto: dec 5? i thought i saw one with a bold "6" at the bottom 06:34
cotto yes, at 1700 left coast time 06:35
dukeleto kicks off a new filter-branch of parrot.git and cranks up some NIN 06:37
it is more than 25% done already. 06:42
moritz cotto: what's your github ID? 06:48
cotto: added github ID 'cotto' to the perl6 team - if that's not your ID, please scream asap :-) 06:49
cotto moritz, cotto 06:50
cotto doesn't scream
sorear moritz: How do I add people to mu? 06:51
moritz sorear: do you have access to github.com/organizations/perl6/teams/14366 ? 06:52
dukeleto now learns about the --all option to git filter-branch 06:53
sorear I can view the list 06:54
but there's no obvious button
moritz for me there's an input field and an 'add' button at the bottom
dukeleto moritz: you are an admin on the perl6 org, most others are not 06:55
cotto: what is our policy for current github branches? 06:56
moritz sorear: I've now added you to perl6 org "owners" team, so you should be able now
sorear: when somebody wants commit access to the mu/std/specs/roast/whatever repo, please add them to 'perl6' team 06:57
cotto dukeleto, anything not containing a slash is liable to be nuked
dukeleto cotto: i think i can rewrite cotto/git_aware_docs and the few others with filter-branch, and replace them
cotto great
dukeleto cotto: one thing to keep in mind is that if we keep old branches with the new repo, they will be disjoint histories, and hence take up lots of space. We don't want that.
cotto: i think i will run filter-branch one more time, with --all, which will filter-branch all branches, and then I can force push everything and hopefully unicorns come out the other side 06:59
cotto dukeleto, what would we lose by removing those branches? Would commits to them still appear in the repo? 07:00
dukeleto cotto: i don't quite know what you mean by "removing those branches"
cotto: it isn't a big deal, i have them locally, and I can apply them as patches, but i think filter-branch can take care of it 07:01
cotto wfm
dukeleto cotto: they will get merged in very soon after the new parrot.git goes live
filter-branch just finished, after about ~28m 07:04
cotto That was way faster than last time.
dukeleto Indeed. I just started it again with --all 07:06
so in ~28 minutes, we may have a new parrot.git 07:07
cotto shiny
dukeleto i may have to hack my mirror script a bit to force push everything for the first push of the new parrot.git, but that should be easy
dukeleto reminds parrot devs to put their tray tables and chairs in the upright position 07:08
cotto In the event of a water landing, purl may be used as a flotation device. 07:09
sorear moritz: did you just create a bunch of teams, or were "book" et al only visible to owners? 07:10
cotto dukeleto, what time tomorrow do you want to flip the switch? Do you know if the osl folks will be around in the evening? 07:14
moritz sorear: github automatically creates one team per repo 07:15
sorear: book is one of the few repos where we have separate comitters list atm 07:16
dukeleto cotto: not sure. I am on east coast time this week. 07:23
cotto that complicates things
to get this migration done I'm fine shifting my schedule around as necessary, but I'm not sure about osuosl. 07:24
dukeleto cotto: have you talked to them at all? 07:26
cotto no 07:27
do you think it'd be better for me to sync with them?
07:34 fperrad joined 07:38 dalek left 07:39 dalek joined 07:42 bacek joined
bacek aloha, humans 07:48
cotto parrot-dev is down at a most inopportune time, so here's the current plan: dukeleto and I will try to get svn set to read-only as soon as we can. 07:53
After that, dukeleto will sync any final changes to the github clone and I'll work with osuosl to install the github trac plugin.
Once that's done, we'll be good to go. It'll be safe to clone and push to the github repo as soon as dukeleto's done syncing.
He'll post when that's ready to go. 07:54
dukeleto I am doing a hopefully final filter-branch, and I have emailed support@osuosl.org asking to make SVN read-only 07:55
We are very close to being Git-enabled. 07:56
cotto at this point, svn commits are discouraged 08:00
dukeleto Every time you commit to svn between now and when we convert to git, a fairy loses her wings. 08:01
cotto opbots, names 08:02
moderator Parrot 2.9.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | GIT MIGRATION IMMANENT: svn commits are discouraged | parrot-dev is down 08:02
08:03 theory left
sorear Does OSUOSL know we're moving to git? 08:04
dukeleto sorear: yes
sorear: they are #osuosl on freenode, and i sent in support tickets 08:05
sorear wonders how well they approve of the move to github
cotto 'night
dukeleto sorear: i assume they will like it, since it means less work for them
cotto sorear, I don't think they care too much. It's a free service that they provide which we're relying on less post-migration.
cotto is out
08:06 theory joined, theory left 08:36 dalek left 08:37 dalek joined
cxreg dukeleto: still around? 10:00
i'm wondering what filter-branch --all does to merge commits with some mild paranoia 10:01
10:28 masonkramer_ joined, masonkramer___ left 10:42 AzureStone left 10:45 AzureStone joined 10:56 rurban_ joined 10:59 rurban left, rurban_ is now known as rurban 11:27 contingencyplan left
dalek rrot: r49806 | gerd++ | trunk (2 files):
add a sha256 message digest function from the SHA-2 family
 11:39
11:57 jsut joined 12:02 jsut_ left
dalek rrot: r49807 | gerd++ | trunk/config/gen/makefiles/root.in:
add the generation of the according pbc file from sha256.pir
 12:11
12:26 patspam joined, patspam left
dalek rrot: r49808 | gerd++ | trunk/MANIFEST.generated:
install sha256.pbc
 12:27
12:31 darbelo joined 13:16 mtk joined 13:17 lidden joined 13:31 whiteknight joined
whiteknight good morning, #parrot 13:39
moritz good mornin' whiteknight 13:42
darbelo aloha: clock? 13:43
aloha darbelo: darbelo: LAX: Mon, 05:43 PST / CHI: Mon, 07:43 CST / NYC: Mon, 08:43 EST / UTC: Mon, 13:43 UTC / LON: Mon, 13:43 GMT / BER: Mon, 14:43 CET / TOK: Mon, 22:43 JST / SYD: Tue, 00:43 EST
dukeleto waves hello 13:48
whiteknight good morning duke 13:49
particularly bad timing for a mailinglist outage, eh?
darbelo I'm guessing the outage is going to delay the migration. Right? 13:50
whiteknight I posted a notice on my blog this morning. I don't know how the readship of that compares to the readership of parrot-dev 13:51
recent list traffic suggests a decision in my favor though :)
darbelo You could also post to the parrot.org home page. 13:52
Prejudiced people might argue that is more authoritative than some dude's blog.
dukeleto I am not going to let the mailing list change my migration plans.
git filter-branch may be effective in driving me insane this morning, though.
whiteknight darbelo: Even if it's the same dude writing the stuff? 13:53
darbelo You can't beat prejudice, man.
whiteknight actually, I don't think my blog is on planet.parrotcode yet, still
darbelo ;)
13:56 bluescreen left
dalek rrot: r49809 | gerd++ | trunk/runtime/parrot/library/Digest/sha256.pir:
remove a trailing whitespace
 14:01
dukeleto THE HORROR 14:03
cotto: ping! 14:04
cotto: did you spam everyone individually to tell them not to commit?
cotto: evidently gerd++ didn't get the message
14:04 mtk left 14:06 bluescreen joined
whiteknight getting the message is for losers 14:07
I know, because none of us got the message :)
moritz well, I got it (from the /topic) 14:11
dukeleto moritz++
dalek rrot: r49810 | gerd++ | trunk/runtime/parrot/library/Digest/sha256.pir:
remove trailing whitespace
 14:17
dukeleto another fairy loses their wings 14:20
dukeleto just emailed gerd++ to stop
darbelo Can't we just make svn read-only for the migration? 14:21
dukeleto OSUOSL peeps are not awake yet.
darbelo: I would love to. I can't.
If anyone can make the svn repo read-only, that would be amazing. But I think the OSUOSL peeps need to do it.
darbelo Ouch.
whiteknight I can go in and remove all committers 14:28
not quite the same thing
moritz let's DDOS it!
*SCNR*
dukeleto is talking to OSUOSL peeps now 14:30
whiteknight dukeleto: where at? 14:31
dukeleto svn: Commit failed (details follow): 14:32
svn: Server sent unexpected return value (403 Forbidden) in response to MKACTIVITY request for '/parrot/!svn/act/8d506c6d-7c62-41f6-88fb-71d6c1761c04'
SVN is read-only, and I am going to ignore those last few whitespace commits.
whiteknight we can re-add them manually 14:33
dukeleto cotto: ping! 14:37
14:41 gerd joined
dukeleto made a mistake and is starting over, so I will pull in those whitespace commits 14:45
whiteknight dukeleto does not make mistakes. He simply finds a higher level of perfection 14:47
14:49 gerd left, PacoLinux joined
dukeleto Wheels within wheels... 14:49
UGGGGG 15:04
someone emailed me 3 minutes ago about changing their name. 15:05
Should I restart my filter-branch?
plobsing can't names be changed afterwards? 15:08
(by someone with admin privs of course)
dukeleto plobsing: it changes the sha1's all all history after it
15:09 Andy left
plobsing yes rewriting history has that property, but it seems like something we might want to do from time to time 15:09
15:09 Andy joined
dukeleto plobsing: no, we won't do that from time to time 15:11
plobsing: we do it once. Now.
plobsing ah. ok.
dukeleto plobsing: changing history would make everyone's clone become divergent
plobsing can't rebase fix that? 15:13
dukeleto plobsing: if i didn't know better, I would think that you are trolling me. 15:14
plobsing: rebasing public history is a big NO NO NO. If you rebase public history, that means that everyone else that has that history in their repo has a divergent (different) history
PROTIP: Never ever rebase public history. 15:15
plobsing: read the "recovering from an upstream rebase" at www.kernel.org/pub/software/scm/git...ebase.html 15:16
Excerpt: Rebasing (or any other form of rewriting) a branch that others have based work on is a bad idea: anyone downstream of it is forced to manually fix their history. This section explains how to do the fix from the downstreamā€™s point of view. The real fix, however, would be to avoid rebasing the upstream in the first place.
darbelo "History has changed. The world is no longer as you remember it." 15:17
dukeleto Upstream rebasing = making everyone's life really complicated.
dukeleto starts a filter-branch ONE MORE TIME 15:24
plobsing "never ever" just seems artificial. it seems to me that if it was done sufficiently infrequently and with enough advanced warning, it wouldn't be that bad.
"never ever" does work though. 15:25
just seems suboptimal in my eyes
darbelo If it makes you feel better "Never ever" can be revised later :)
Unless we decide to never ever reconsider our never ever policy... 15:26
plobsing my head asplode
dukeleto plobsing: i don't understand why you want to make life more complicated 15:27
we can put a pre-commit hook to disallow it, and we most probably will. 15:29
plobsing I agree that normal commiters shouldn't be allowed to do such things. I look it as a form of repo maintenance. 15:31
dukeleto plobsing: i hear what you are saying. But as the number of clone that exist increases, it becomes increasingly impossibilish to change history without causing horror and catastrophe. 15:32
when we are on github, there could be hundreds to thousands of clones of parrot.git 15:33
whiteknight hopefully there will be thousands of clones 15:37
I know I'm planning to fork as soon as it's live 15:49
15:55 theory joined 15:56 darbelo left 16:09 gerd joined 16:10 gerd left 16:11 darbelo joined, gerd joined
cotto dukeleto, pong 16:16
dukeleto cotto: welcome to the fun.
cotto wheee
what's the current status of the migration? 16:17
trac time?
dukeleto cotto: i am about to force push our new git repo to github. Shall I do it?
cotto of course
dukeleto cotto: yes, we need to bug people in #osuosl
cotto do it now
www.youtube.com/watch?v=u6ALySsPXt0 16:18
moritz once the github repo is our official standard, we should have a policy against force pushes
cotto dukeleto, let me know when the repo is updated so I can make sure the revision map generated for the trac plugin is still ok 16:20
dukeleto presses a large red button 16:22
cotto dukeleto, how long will the push take?
videosift.com/video/Dont-Push-the-H...ser-Button 16:23
moritz IWBN if github had an option that disabled forced pushes for a repo
dukeleto is force-pushing the new repo as we type
moritz: that can be done in a pre-commit hook 16:24
master is updated. The other branches are rolling in now. 16:25
don't reclone, yet.
16:27 brianwisti joined
dukeleto i think everything is pushed. But i think github is caching stuff, because I am not seeing the new stuff on the github web interface 16:29
i am cloning a fresh github parrot.git to see what it looks like
whiteknight I'm seeing your pushes on my dashboard
dukeleto whiteknight: does the most recent commit show just "gerd" or the whole name?
whiteknight just "gerd" 16:31
dukeleto Ok, it looks like github is caching stuff, because when you clone, the name is correct
whiteknight: go look at the gsoc_nci branch
cotto dukeleto, so cloning is safe now?
moderator Parrot 2.9.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | GIT MIGRATION IN PROGRESS: svn commits are disabled | parrot-dev is down 16:32
dukeleto cotto: please clone it yourself and take a look 16:32
cotto: i want to do some sanity checking for a few minutes before the hordes are notified
cotto dukeleto, doing so now 16:33
dukeleto cotto: i am going to commit something to see if that triggers the github cache to DTRT 16:34
atrodo fetching as well
whiteknight it looks to me like the github cache is correct 16:35
gerd has a github account with username gerd
dukeleto whiteknight: i just realized that
whiteknight if you scroll way down, you see cotto is listed by his full name
dukeleto If whiteknight and cotto say that the repo doesn't look like it is total garbage, then I think we can notify others.
whiteknight and you also notice that cotto needs to get a damned github account!
dukeleto I just pushed a new commit to github, can y'all test to see if you can pull it down with "git pull --rebase" ?
whiteknight I see it in github 16:37
cotto whiteknight, I probably just need to add whatever email address the svn import uses. 16:38
dukeleto whiteknight: cool
cotto: are you insinuating that you want me to redo stuff? 16:39
whiteknight cloning now
dukeleto cotto: you can create a github account and add multiple email addresses for it
whiteknight++
cotto dukeleto, not even close
I'm perfectly happy adding another email address to my github account.
whiteknight ...actually not taking super long either
dukeleto cotto: good :) 16:40
whiteknight looks good to me 16:41
dukeleto So, i will take a break for a few mins. If y'all tell me that the repo looks good, I can tell @parrotvm to announce to the hordes
16:42 contingencyplan joined
cotto dukeleto, the plugin appears to be happy 16:43
dukeleto cotto: you are working on the plugin stuff already? sweet.
moderator Parrot 2.9.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | svn is dead. Long live git! | parrot-dev is down 16:44
moderator Parrot 2.9.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | svn is dead. Long live git! |
cotto dukeleto, of course. 16:44
dukeleto darn.
cotto darn?
moderator Parrot 2.9.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | svn is dead, long live git! github.com/parrot/parrot | parrot-dev is down 16:44
atrodo git++ 16:44
dukeleto cotto: i messed the /topic up for second :) 16:45
cotto I don't like "darn"
dukeleto, don't scare me like that
I need to head off to $dayjob.
dukeleto cotto: no worries, just "darning" my /topic mishap ;)
cotto: have fun!
so shall we say that github is live?
and let the hordes know?
cotto the instructions for the trac plugin should work without a hitch.
dukeleto, very yes 16:46
brianwisti Stupid $dayjob. Getting in the way of perfectly good Parrot hacking.
whiteknight :)
cotto goes to w3rk
dukeleto OSUOSL tells me the emails should be flowing again 16:49
let me know if y'all get them
it might have a big backlog, so it could take a few mins 16:50
whiteknight awesome
dukeleto whiteknight: we need to do the language repo 16:51
whiteknight: but it is complicated, because you split some out to git repos 16:52
whiteknight we're not going to do the language repo as a whole. We're leaving them as the subrepos 16:53
I don't think any of those have seen any action. I'm comfortable with rewriting history in-place for those repos to get author info right 16:54
dukeleto whiteknight: yep, now that i have all the scripts written, it should be trivial
whiteknight I'll gladly do the work myself if you send me the scripts and some gentle instructions 16:56
17:12 darbelo left, darbelo joined 17:17 cotto_work joined
dukeleto whiteknight: ok 17:19
17:25 rurban left
cotto_work dukeleto: anything interesting happening atm? 17:26
dukeleto cotto_work: i just merged git_aware_tools, kill_svn_tests and git_docs 17:28
and now I am going to lunch
cotto_work woot 17:29
17:31 rurban joined
whiteknight I'm going to merge stdhandle_meths soon, probably. I would also like to start working towards a merge of gsoc_instrument soon if possible 17:32
I will probably hold off on the later until after 2.10. I want to get *something* embed-related in before the release if I can 17:34
17:43 gerd left 17:46 rurban left 17:49 stilgar left
cotto_work whiteknight: I should be able to help with gsoc_instrument now that I have a decent working knowledge of the plobsing++'s dynop_mapping code 17:50
I'd love to see that in trunk again.
17:51 rurban joined
whiteknight I think we should be close. Last time I was looking at it the upconversion looked straightforward but I ran out of steam and got distracted by other things 18:09
I suspect we can get it merged in with an afternoon of decently-focused hacking
dalek TT #1843 created by doughera++: NetBSD fails compiling sysmem.c 18:35
TT #1843: trac.parrot.org/parrot/ticket/1843
cotto_work dukeleto: is parrot-dev back in action? 18:38
dukeleto cotto_work: seems to be 18:39
dukeleto just got back from lunch
18:45 dmalcolm joined
moderator Parrot 2.9.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | svn is dead, long live git! github.com/parrot/parrot 18:54
Infinoid ooo, parrot git 18:55
That's been a long time coming. :) 18:56
18:57 rurban_ joined
darbelo An announcement on parrot.org would be nice. 18:58
19:00 rurban left, rurban_ is now known as rurban
cotto_work darbelo: thanks for volunteering ;) 19:00
darbelo I think I have the right privileges to post it but I have no idea about what to say about it. 19:01
whiteknight "ZOMG IZ T3H GITZ, LOL"
cotto_work "mai git. let me show you it"
whiteknight "source control: ur doin it wrong"
"GIT PWNS J0" 19:02
cotto_work I actually have a message drafted at home. If nobody beats me to it, I'll send it out.
moritz fwiw, rakudo builds on parrot-from-git, and the first few spectests pass 19:09
whiteknight awesome 19:12
whoever does it, make sure the announcement goes out to parrot-dev and parrot-users
sorear fetch_url: failure fetching trac.parrot.org/parrot/timeline?ti...ormat=rss: 500 Internal Server Error
dalek is broken
well, not dalek in general 19:13
just the wiki scraper + TTs
whiteknight blah 19:14
darbelo sets git-svn on fire. 19:18
cotto_work trac is down for upgrades 19:22
darbelo I'll do my best not to finds bugs in the meantime, then. 19:24
:)
19:25 cotto_work is now known as cotto_work2 19:32 cotto_work joined
cotto_work2 goodbye, mibbit 19:32
19:32 cotto_work2 left
atrodo have a real irc client now? 19:34
cotto_work yes 19:35
sorear ooh 19:45
hey darbelo
dukeleto adds gas to the git-svn fire 19:47
cotto_work ā™© We didn't^H^H^H start the fire. ā™© 19:49
darbelo hey sorear 19:57
cxreg dukeleto: did you see my question last night? 19:59
dukeleto cxreg: yes 20:00
cxreg: my filter-branch command changes author names and emails. That's it.
cxreg except that those break sha1's 20:04
which means that parents of merges are possibly wrong
ie, your old commit's parent could be pointing to a now-replaced other commit
filter-branch is smart enough to fix that in /one/ branch, but i'm not sure across multiple branches 20:05
might Just Work though
sorear well, if you change the author names and emails in the same way in two branches, the new SHAs will be the same 20:06
SHA1 maps identical inputs to identical otuputs
darbelo Such a clever little hash, he is. 20:07
dukeleto cxreg: i read the man page for filter-branch, and I think that i read that it deals with that, but I am not sure. Is there a way to you can take a look at the repo and see if it looks sane? 20:09
cxreg sure, wheres it at?
dukeleto cxreg: it is in the /topic :)
cxreg orite 20:10
sorear: i know, assuming everything was properly updated. thats why i said paranoia :) 20:12
darbelo You mean it'd be dependent on the order in which artifacts are precessed. 20:13
Ugh. Now I'm paranoid too. 20:14
cxreg i'm not actually sure that matters. and the repo does look ok.
i think if there were such problems, "git shortlog --all -ns" would show them 20:15
and that looks correct
or does it?
darbelo I think the tool should be smart enough to properly walk the DAG, but I have never investigated enough to convince myself that a git-svn DAG is sane in all the ways it should be. 20:16
cxreg 3255 Will "Coke" Coleda
3216 coke
hrm
ah crap. maybe it /is/ broken 20:17
b95c77cd6d is the same commit as 96873f4965, for example.
darbelo Well, *that* sucks.
cxreg it's a matter of going back and doing parent filters 20:18
but i dont know /how/ to do that in an efficient manner :( 20:19
and without introducing the same horrible cascading issue. i guess you'd need to do it topologically from the top
darbelo dukeleto: ping 20:21
cotto_work dukeleto: can you update www.parrot.org/download to add any configuration issues for git users?
20:21 rurban left
dukeleto darbelo: pong 20:22
darbelo dukeleto: irclog.perlgeek.de/parrot/2010-11-08#i_2976751
There seems to be trouble with the git repo. 20:23
dukeleto what is the problem?
cxreg the merges are pooched, as i had hoped they weren't
the unrewritten history is still there 20:24
ah!
i think it's /tags/ 20:25
not branches
yes, confirmed
for example, RELEASE_2_9_1
dukeleto what is the problem, exactly?
cxreg so its not merges, its tags
the tags were not filtered, and are in the alternate history with the unfixed authors
dukeleto cxreg: hmmmmm 20:26
cxreg which is far from the worst case scenario, it means that the branches are fine
moritz and since the alternate history is still referenced, it can't be GC'ed
cxreg whew
dukeleto moritz: indeed.
cxreg: do you have any recommendations?
darbelo Would re-tagging the releases and deleting the 'alternate history' tags DTRT? 20:27
cxreg darbelo: yeah probably 20:28
how many tags are there?
hm, 74's not bad
dukeleto a few dozen
cxreg: would you like to fix this? I will buy you a beer when we finally meet up :)
cxreg heh :)
dukeleto cxreg: or your beverage of choice
cxreg i can help sure 20:29
the good news is, peoples clones should be safe
they might want to "git fetch --tags" when this is done
dukeleto cxreg: tags are pulled by default 20:30
cxreg: but not pushed by default
cxreg no
changed tags are not
dukeleto cxreg: hmmm. If you say so, I thought that was the case, tho.
cxreg its a security feature
nope
dukeleto cxreg: oh, yes "changed tags"
cxreg: what commands will you actually need to do to fix this? 20:31
cxreg i'm not entirely sure. i'm going to skip filter-branch and just use maybe git patch-id to find the right "real" commit
20:32 rurban joined
moritz youc an even look at the svn-commit to git-commit map, and use the tags from svn 20:34
cxreg oh right, forgot about those
dukeleto OH! 20:38
wait
i forgot to push the new tags
cxreg: ^^^
cxreg: just fixed it :)
everyone needs to do a "git fetch --tags" if you already have a clone 20:39
darbelo fetches. 20:41
darbelo gc's 20:42
$ git gc
Counting objects: 375003, done.
Delta compression using up to 2 threads.
Compressing objects: 100% (112973/112973), done.
Writing objects: 100% (375003/375003), done.
Total 375003 (delta 252658), reused 374994 (delta 252658)
whiteknight december 5th at 8PM is looking like the winning time for a PDS
cxreg dukeleto: LOL 20:43
dukeleto LOLLERSKATES
moritz running git gc --prune increased the size of my .git/ repo 20:44
s/repo/dir/
darbelo dukeleto: shouldn't a git gc collect something after the tag update?
cxreg dukeleto: still some relics. leto/testing ?
that did it 20:45
dukeleto twitter.com/#!/parrotvm/status/1736627248435200 <-- directions for getting the new tags
cxreg: i believe that was just a "testing branch" to test my mirror script with. It can be sent to /dev/null
cxreg yeah 20:46
kill it
dukeleto cxreg: you want me to?
ah, that is keeping the gc from happening
just rm'ed it 20:47
dukeleto does a gc
20:47 bluescreen left
dukeleto darbelo: pull and try again 20:51
darbelo $ git pull 20:53
Already up-to-date.
I'm not seeing any changes.
dukeleto darbelo: you may need to do a "git branch -d leto/testing" 21:00
wait, you probably don't have it checked out 21:01
hmmmmm
darbelo I only have *master in my repo.
atrodo what does "git branch -a" show?
darbelo remotes/origin/leto/testing amongst others. 21:02
mikehh dukeleto: kI was logged in to github and it says I should use something like mikehh@github.com/parrot/parrot.git is that right?
dalek tracwiki: v1 | cotto++ | GithubPluginTest 21:03
tracwiki: trac.parrot.org/parrot/wiki/GithubP...ction=diff
mikehh s/kI/I/
dukeleto darbelo: easiest thing is to clone again
darbelo: leto/testing won't be there 21:04
sorear \\o/
who fixed trac
darbelo dukeleto: Not a big deal, it's not like I need the diskpace back.
dukeleto mikehh: i think so 21:05
darbelo mikehh: I did a "git clone git@github.com:parrot/parrot.git" which goes over ssh, but you need to set up your ssh keys on github for that to work.
dukeleto mikehh: cloning over http/https is very slow, and will take a long time
mikehh: cloning over ssh is best
allison where is the wiki page that lists our procedures for git? 21:06
as in, the standard way to clone, setup our branches, submit patches, etc 21:07
that is, "the parrot project standard", since git has dozens of incompatible ways to do each 21:08
cotto_work allison: I believe dukeleto update docs/project/branching_guide.pod 21:09
or not
moritz dozens incomapatible ways? that's news to me
allison moritz: not going there :) 21:10
moritz remains unenlightened
cotto_work Trac is back up and should be back to 100% with the github plugin as soon as the kind folks at osuosl and I get some final issues ironed out.
allison consensus and documentation on our procedures was a requirement for the git switch
so we just switched with no procedures in place? 21:11
Tene remains tened.
cotto_work I know there are updated docs.
dukeleto: what happened to those?
whiteknight dukeleto did merge a few branches today mostly concerning docs 21:12
darbelo docs/project/git_workflow.pod strikes me as relevant 21:13
but the committer guide looks unupdated.
whiteknight no matter what is or is not written, we are going to have some learning to do as a community 21:14
allison the branching guide is certainly not updated
cotto_work allison: I think it's just an issue of things getting temporarily lost in the chaos rather than them not existing. We're still getting everything updated and making sure people know where what is.
aloha: clock?
aloha cotto_work: cotto_work: LAX: Mon, 13:14 PST / CHI: Mon, 15:14 CST / NYC: Mon, 16:14 EST / UTC: Mon, 21:14 UTC / LON: Mon, 21:14 GMT / BER: Mon, 22:14 CET / TOK: Tue, 06:14 JST / SYD: Tue, 08:14 EST
darbelo There's a 'Creating and Switching Branches' section in the git_workflow file
allison whiteknight: true, but we need to start with at least a strawman procedure to improve on, otherwise it's "everyman for himself" 21:15
whiteknight allison: yes. As cotto said, it's probably a case of things still being early in the migration. We do have docs, it's a matter of making sure they are in the right places, and that they are comprehensive 21:16
the committer guide should be updated. That was probably an oversight
and the metacommitter guide too, I venture to guess
although that one is probably in more flux since Github has changed some things vis organizations lately 21:17
allison I'm not blaming anyone, it's past. But we need to fix it quickly, as in today or tomorrow.
a nice prominent page like github.com/rakudo/rakudo/wiki/step...te-a-patch 21:18
whiteknight I very much agree with that sentiment. I'll dig through it when I get home. I suggest several other people put eyes on it too to identify areas that are lacking
21:20 darbelo left
whiteknight there were docs on trac too. I don't know how many of them made it into the repo 21:21
tangentially, I wonder what the long-term role of the docs/project folder in the repo is, if we have other places for project documentation
another topic for another day
dukeleto is back 21:22
cotto_work wb dukeleto 21:23
dukeleto I did update the branching guide. 21:25
whiteknight I'll take a look at the committer and metacommitter guides when I get home. I do want other opinions on them too. The conversion should be straight-forward I think
dukeleto hmmm. something didn't go through, because I *did* update that branching guide 21:27
ah, i remember. I started a new document.
allison: github.com/parrot/parrot/blob/mast...rkflow.pod 21:29
whiteknight: github.com/parrot/parrot/blob/mast...rkflow.pod
the old branching guide was very svn specific, so i didn't want to modify it
whiteknight we should remove it from the repo, if it's completely useless 21:30
dukeleto whiteknight: you just volunteered
cotto_work There's a line. 21:31
sorear Aaaand the wiki scraper is DEAD!
allison dukeleto: it needs "how to submit a patch"
whiteknight is heading home. Will look at things later. Send me a list of all the things that I've apparently been volunteered for
allison dukeleto: commiters are the minority 21:32
dukeleto: but otherwise looks good
21:32 whiteknight left
dukeleto allison: hokey dokey 21:33
cotto_work Can someone push a commit? I need to test the post-receive callback. 21:34
dukeleto cotto_work: sure 21:35
cotto_work: just pushed
Infinoid sorear++ # hacking dalek
sorear btw, with the new design, we can do instant notifications 21:36
cotto_work dukeleto: it works 21:37
trac.parrot.org/parrot/wiki/GithubPluginTest
that last one is what you just pushed
mikehh t/configure/018-revision_to_cache.t - Failed test: 5 - 'Got numeric value for revision number' in perl Configure.pl --test etc 21:39
cotto_work I can feel my stomach unknotting.
21:40 cotto_work left, cotto_work joined
cotto_work trac.parrot.org should be 100% functional. If you find anything surprising, let me or dukeleto know. 21:40
note the shiny tooltips on trac.parrot.org/parrot/wiki/GithubPluginTest 21:43
atrodo Shiny
dukeleto mikehh: some of those tests will need to change to know about sha1's. Sorry I didn't catch those before the transition
dalek tracwiki: v2 | cotto++ | GithubPluginTest 21:50
tracwiki: trac.parrot.org/parrot/wiki/GithubP...ction=diff
tracwiki: v3 | cotto++ | GithubPluginTest
tracwiki: trac.parrot.org/parrot/wiki/GithubP...ction=diff
tracwiki: v4 | cotto++ | GithubPluginTest
tracwiki: trac.parrot.org/parrot/wiki/GithubP...ction=diff
21:52 jan joined
mikehh dukeleto: just trying to figgure it out - cd02859ff353ea8e15ea64668e19189fdcdd43e4 - does not pass like($Parrot::Revision::current, qr/^\\d+$/, "Got numeric value for revision number"); 21:53
dukeleto: what would the rggex be? 21:55
regex
test for 40 hex digits?
moritz ^ <xdigit>**30 $ 21:56
mikehh let me try 21:57
dukeleto mikehh: [a-f0-9]{30} perhaps 22:00
cotto_work shouldn't that be 40?
dukeleto cotto_work: yes 22:02
i just added the new parrot.git to the languages wiki page, so maybe we will get updates about git commits from dalek soon 22:06
a better solution is needed
dalek tracwiki: v147 | dukeleto++ | Languages
tracwiki: trac.parrot.org/parrot/wiki/Languag...ction=diff
cotto_work dukeleto: The current (or soon to be implemented) plan is to use a json config file in the mu repo for dalek. 22:07
misc/dalek-conf.json or something similar
sorear yes 22:10
I've already done the final scraper run, to regenerate dalek-conf.json 22:11
cotto_work sorear: is that how it works now?
sorear only on localhost
cotto_work you mean that the json file is used but has to be copied to the local machine? 22:12
sorear no
I mean that I'm running a test instance of dalek here
with newer code
cotto_work ah
sorear actually I think it's working well enough now 22:13
12 files changed, 246 insertions(+), 914 deletions(-)
dukeleto sorear: how often does dalek look at the languages wiki page?
sorear dukeleto: every 5 minutes 22:14
well, never, as of the live dalek update just now :)
you need to commit changes to the mu repo
if you don't have a commit bit, ask me, moritz, or TimToady 22:15
cotto_work notices that he has a mu bit
dukeleto sorear: i have a bit
sorear if using github.com/perl6/mu.git is politically unacceptable, create a second json file of the same format and I'll add it to the poll list
dukeleto sorear: does your conf have parrot.git in the conf?
sorear not yet 22:16
dukeleto sorear: i added it a few minutes ago to the languages wiki page
sorear: just rescrape and you will get it
22:17 fperrad left
sorear I reran the scraper and it reported everything twice, because it's using both the old scrape stored in json and the wiki 22:19
so I'll just add this manually
dukeleto sorear: sounds fine to me. i just want to see git commits in here :) Thanks! 22:20
22:20 cotto_work left
sorear also, that link is broken. 22:20
22:21 cotto_work joined
sorear at least, for tools 22:21
22:21 dngor left, dngor joined
dukeleto tools link? 22:21
sorear What branches do you want reported? They have to be enumerated explicitly 22:22
yeah, the http: -> https: redirection only works for browsers
programmatic access to github.com is dead
that's what broke dalek last week - I had to fix all the links
dukeleto sorear: thank you for working on this 22:23
sorear just master for now?
dukeleto sorear: well, we at least want to see master. I am not sure which other branches we want to see
sorear: sure, wfm
22:23 brianwisti left
sorear git is not very good at diffing json 22:23
dukeleto sorear: you can write custom diff-handlers 22:24
22:24 cotto_work left
sorear Infinoid: Is there any reason to keep around modules/local/parrotlog.pm? 22:25
22:29 masonkramer___ joined, masonkramer_ left, cotto_work joined
sorear Infinoid: would you take a look at my latest commit? 22:46
this is what I was talking about as "stage 2"
Infinoid sorear: No reason to keep parrotlog around unless you want the basis for a generic trac commit parser. 22:51
It's in source control if we need to dig it out, anyway
mikehh dukeleto: did you change tools/dev/mk_manifest_and_skip.pl or t/manifest/* 22:58
22:58 bluescreen joined 23:07 dngor_ joined 23:08 dngor left 23:24 gerd joined 23:33 gerd left 23:35 gerd joined 23:44 stilgar joined 23:45 gerd left 23:46 dngor joined 23:50 dngor_ left, particle1 left

Powered by Cro and the RakuĀ® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!