www.parrot.org | Parrot 1.6.0 "half-pie" released: feel free to tear out the JIT! | Testing priorities: Exception and MultiSub
Set by moderator on 17 September 2009.
nopaste "kid51" at 71.247.53.57 pasted "Is nopaste.snit.ch working?" (1 line) at nopaste.snit.ch/17986 00:00
kid51 Here's a smolder on that branch from linux/i386: smolder.plusthree.com/app/public_pr...ails/27558 00:01
darbelo not for me.
kid51 Running smolder on kill_jit branch on darwin/ppc. Back after food. 00:02
darbelo kid51++
dalek tracwiki: v3 | Util++ | TracSpammers 00:07
tracwiki: trac.parrot.org/parrot/wiki/TracSp...ction=diff
tracwiki: v4 | Util++ | TracSpammers
tracwiki: trac.parrot.org/parrot/wiki/TracSp...ction=diff
rrot: r41329 | darbelo++ | branches/kill_jit/config/auto/frames.pm:
Add an else to the conditional chain in the interest of test-coverability. kid51++ for pointing it out.
 00:08
cotto_work Util, what's the point of TracSpammers? 00:13
dalek tracwiki: v5 | Util++ | TracSpammers 00:17
tracwiki: trac.parrot.org/parrot/wiki/TracSp...ction=diff
Util cotto_work: It is Coke's creation; I suppose it is to coordinate and keep history of the work done (and needing to be done) to eliminate spam. Without it, I would not have known how to help when I saw spam on a ticket. 00:21
cotto_work works for me 00:22
darbelo I would expect that he's also harvesting the addresses for a blacklist. 00:23
Util Trac admins: be advised that the TracSpammers wiki page lists the work that I was unable to do, that I need one of you to do. Thanks! 00:24
00:28 whoppix joined 00:30 klaus_ joined 00:36 jrtaylor joined
cotto_work It'd be good if someone could start an ObjectionsToGit wiki page with the points allison metnioned. I'd do it but for work priorities atm. 00:37
dalek kudo: a7ac444 | pmichaud++ | src/parser/grammar.pg:
Fix handling of twigil in param_var for anonymous invocant (RT #69206, masak++)
 00:38
00:42 athomason joined
darbelo cotto_work: Don't worry. We'll wait :) 00:44
cotto_work um, thanks
could be a while
dalek rrot: r41330 | darbelo++ | branches/kill_jit/t/steps/auto/jit-01.t:
Shorten t/steps/auto/jit-01.t to match the changes in the auto::jit configure step.
 00:45
00:45 jrtayloriv joined 00:46 darbelo left
pmichaud cotto_work: trac.parrot.org/parrot/wiki/GitObjections 00:47
cotto_work pmichaud++
dalek tracwiki: v1 | pmichaud++ | GitObjections 00:48
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
dukelet0 does anyone know which version of Trac parrot.org is running? 00:49
cotto_work it's at the bottom of each page 00:52
00:53 payload joined
dukelet0 cotto_work++ # i don't read instruction manuals either ; 00:57
;) even
dalek tracwiki: v2 | dukeleto++ | GitObjections 00:58
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
kid51 Smolder report on kill_jit branch, performed on darwin/ppc: smolder.plusthree.com/app/public_pr...ails/27560
dukelet0 i am tempted to start an SVNObjections page
dalek rrot: r41331 | jkeenan++ | branches/kill_jit/t/compilers/imcc/syn/macro.t:
Removing a TODO condition on a test. (Was able to do same in trunk yesterday.)
 00:59
dukelet0 i just don't get how people thing merging and branching in svn is easier/less work/less error-prone
chromatic If you've never successfully used Git, Git seems like a lot more work than all of the manual SVN stuff. 01:01
dukelet0 chromatic: yeah. it just seems like allison had a really bad git experience a while ago because she was coming at it without any previous experience and no one helping her. I can understand that. Git is very intimidating at first. 01:02
i need to come up with some examples that blow svn out of the water in ease of use, during a merge 01:03
and I think a lot of the complexities of git-svn are getting confused with pure-git-goodness
mattp i was surprised when i found out parrot was still in svn 01:04
dukelet0 the objection about developer training is real, but it seems that many parrot devs are using git-svn these days
mattp: better than rcs ;)
chromatic I had the same experience with Git when starting. 01:06
It took you and Schwern saying "Look, it's this simple!" to me and Darian to get over that. 01:07
PerlJam git is like perl. There's a bit of a learning curve, but once you get over it, there's loads of power available to you. 01:08
dalek tracwiki: v3 | pmichaud++ | GitObjections 01:15
tracwiki: refactor to make objections and answers a bit clearer
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
pmichaud I think we have a significant number of parrot devs who have said that they git merging much easier than svn. I don't know why we need more examples than that. 01:16
Either that or we simply believe our developers have no clue about what they're talking about. 01:17
s/they git/they find git/
afaik, none of the people who say "git merging is easier" are doing so based on speculation. It's all from hard experience in doing merges in both svn and git. 01:18
02:07 darbelo joined, ilbot2 joined
moderator www.parrot.org | Parrot 1.6.0 "half-pie" released: feel free to tear out the JIT! | Testing priorities: Exception and MultiSub
dalek tracwiki: v7 | dukeleto++ | GitObjections 02:13
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
tracwiki: v8 | dukeleto++ | GitObjections 02:20
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
tracwiki: v9 | dukeleto++ | GitObjections 06:15
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
06:15 mokurai joined, ilbot2 joined
moderator www.parrot.org | Parrot 1.6.0 "half-pie" released: feel free to tear out the JIT! | Testing priorities: Exception and MultiSub
06:24 fperrad joined 06:27 theory joined 06:34 HG` joined 06:57 iblechbot joined
dalek TT #1020 created by jessevdam++: subclassing pmc from pir + lot more 07:17
07:18 einstein joined, notostraca joined 07:22 donaldh joined 07:27 notostraca_ joined 07:38 dukeleto joined
dalek TT #1021 created by vkon++: typo in inspect description 07:44
07:48 desertm4x joined
einstein bye monday I will be back 07:50
dalek tracwiki: v14 | dukeleto++ | GitObjections 07:52
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
07:52 notostraca joined
dalek rrot: r41332 | chromatic++ | trunk/docs/book/pir/ch06_subroutines.pod:
[book] Clarified inspect op on Sub PMC (TT #1021, Vadim Konovalov).
 07:55
tracwiki: v15 | dukeleto++ | GitObjections
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
07:56 pdcawley_ joined
dukeleto github.com/notbenh/euler_bench/blob...004/01.pir is 2.3 times *slower* in trunk than Parrot 1.0.0 :( 07:58
dalek TT #1021 closed by chromatic++: typo in inspect description
chromatic_afk Time to pull out callgrind. 08:00
Though that means Parrot 1.0 was a lot faster than Perl 5. 08:01
dukeleto chromatic_afk: i've got euler_bench timing various builds of parrot for me now. I am going to hook it up to some pretty graphing code soon 08:04
chromatic_afk: if you want me to benchmark some code across versions, let me know. both builds are default with no optimization flags 08:05
chromatic_afk I'd like to see the callgrind.out files from both 1.0 and trunk for that slowdown.
chromatic Though without optimization, we're probably using the slow register access in trunk. 08:06
dukeleto chromatic: i can give whatever flags to the builds if you want. and which flags to callgrind?
chromatic alias cg='time valgrind --tool=callgrind --dump-instr=yes --trace-jump=yes'
the --optimize flag to Configure.PL should suffice. 08:07
That should make a big difference too. 08:09
dukeleto chromatic: ok, recompiling both with --optimize now
nopaste "dukeleto" at 69.64.235.54 pasted "Example lib/config.yaml file for euler_bench to benchmark different parrot builds" (32 lines) at nopaste.snit.ch/17989 08:15
"dukeleto" at 69.64.235.54 pasted "Example output of euler_bench of Parrot 1.0.0 (git-svn) vs Parrot r41331 (svn) for Euler Problem #004" (5 lines) at nopaste.snit.ch/17990 08:17
dukeleto chromatic: that is the output for unoptimized builds. the optimized builds are benchmarking now 08:18
mikehh All tests PASS (pre/post-config, smoke, nqp_test, fulltest) at r41331 - Ubuntu 9.04 i386
nopaste "dukeleto" at 69.64.235.54 pasted "Example output of euler_bench of *optimized* Parrot 1.0.0 (git-svn) vs Parrot r41331 (svn) for Euler Problem #004" (4 lines) at nopaste.snit.ch/17991 08:19
chromatic Trunk is still slower then. 08:22
Optimization helps trunk more though.
dukeleto warning: no debug symbols in executable (-arch i386) <-- is that normal? 08:23
chromatic Oh, my configure script adds -g back to the executable with the optimized builds. 08:24
nopaste "dukeleto" at 69.64.235.54 pasted "callgrind on parrot trunk for /Users/leto/work/euler_bench/parrot/004/01.pir" (22 lines) at nopaste.snit.ch/17992 08:25
"dukeleto" at 69.64.235.54 pasted "callgrind on parrot 1.0.0 for /Users/leto/work/euler_bench/parrot/004/01.pir" (21 lines) at nopaste.snit.ch/17993 08:26
08:26 NotFound joined
dukeleto chromatic: my output files are really small. should I recompile with -g ? 08:26
chromatic Yes.
dukeleto chromatic: what is your alias again?
chromatic You'll probably have to bzip them and upload them somewhere.
alias cg='time valgrind --tool=callgrind --dump-instr=yes --trace-jump=yes' 08:27
dukeleto chromatic: the alias for hacking on the -g flag
chromatic: you should be able to do this with euler_bench trunk. the bin/bench script takes a yaml file, where you can specify a list of interpreters for each language. notbenh just showed me tonight. it is pretty cool. 08:28
chromatic perl Configure.pl --maintainer --optimize --prefix=/home/chromatic/dev/tmp/parrot && perl -pi -e 's/-DDISABLE_GC_DEBUG=1 -DNDEBUG -O2 /-O3 -DNDEBUG /' Makefile && perl /home/chromatic/bin/rewrite_config_lib
dukeleto chromatic: sweet
what is rewrire_config_lib ?
chromatic You can drop the --prefix part if you never install.
dukeleto rewrite_config_lib, i mean 08:29
chromatic That fixes config_lib.pasm to have a SVN revision. It's only useful if you use something other than SVN and need the parrot_config executable (to build Rakudo, for example).
dukeleto chromatic: ok
mikehh partcl r742 builds on parrot r41331 - make test PASS - Ubuntu 9.04 i386 08:32
rakudo (a969c9e) builds on parrot r41331 - make test / make spectest (up to 28280) PASS - Ubuntu 9.04 i386 08:33
BTW - that is my major objection to git - wtf does a969c9e... tell you - you can use revision numbers in bzr and hg afaik 08:35
dukeleto chromatic: looks like I should have actually put a -g in the regex that you gave me. time to recompile. 08:37
chromatic: i am having issues with callgrind right now. i will try to figure this out tomorrow. 08:52
chromatic Fine by me. 08:54
09:08 cotto joined 09:22 mokurai left
szbalint mikehh: what is a revision number anyway? I think git deemphasizes it because non-linear commit tree 09:29
*because of the
mikehh: you can tag revisions in git based on an arbitrary rule, so you could have "revisionnumbers" 09:32
09:33 Whiteknight joined
mikehh szblint: sure, I just find it inconvenient - I look at r41331 and I know where I am at but looking at a969c9e... doesn't tell me much 09:45
moritz how can I participate in the straw poll? 09:46
do I have to edit the poll, and bump up the number I want by one?
mikehh I know you can get that info fairly easily, for recent commits anyway - but it is still an inconvenience 09:47
moritz ah well, if git-describe gives you 2009-09-2-ga969c9e you know you're 2 revisions after the 2009-09 tag 09:48
that's not much worse than r41331 09:49
mikehh argh - sorry szbalint
09:50 pdcawley_ joined
Whiteknight moritz: you should just be able to click the "Vote" button 09:51
moritz ENOVOTEBOTTON 09:52
Whiteknight www.parrot.org/content/straw-poll-w...parrot-use
and you need to be logged in
moritz I am
still no vote button 09:53
Whiteknight hmmm...somebody must have monkeyed with permissions since yesterday then 09:55
mikehh also there are something like 200+ commands in git - steep learnibg curve - I have seen tutorials like Everyday GIT With 20 Commands Or So 09:57
learning
purl well, learning is not wasting time
Whiteknight moritz: try now 09:59
mikehh Whiteknight - me too
moritz Whiteknight: works now, Whiteknight++ 10:00
Whiteknight mikehh: try now 10:01
I changed it so all "editors" can vote. That should enable everybody
10:01 payload joined
Whiteknight not everybody, but actual members 10:01
moritz Whiteknight++ 10:03
mikehh Whiteknight - not for me - I thought getting a commit bit sort of gave you membership but maybe not 10:08
Whiteknight mikehh: let me look at it 10:09
mikehh: try now 10:10
mikehh Whiteknight - ok it's there - thanks 10:12
maybe I should vote for bzr - it's a bit lonely - but I do find it easy to use and it has revision numbers :-} 10:13
I once made the mistake of checking out the parrot with branches etc - it took ages and was something like 1.3GB 10:21
JimmyZ irc? 10:22
mikehh perl which has a lot more history is only 58.4MB
purl irc is serious fucking business
JimmyZ irclog?
purl hmmm... irclog is irclog.perlgeek.de/parrot/today
JimmyZ 1 + 1
purl 2
10:38 pdcawley__ joined 10:42 MoC joined 10:45 sjn joined 11:18 pdcawley_ joined
dalek lscript: 1c63389 | fperrad++ | (4 files):
remove prefix library in path for load_bytecode
 11:31
11:32 payload joined 11:40 pdcawley__ joined 12:04 quek joined 12:06 whiteknight joined 12:13 ruoso joined
whiteknight good morning #parrot 12:15
12:21 quek left, quek joined, masak joined
Util good morning, whiteknight 12:22
12:23 pdcawley_ joined, bacek joined
whiteknight hello Util 12:25
12:32 Austin joined 12:37 HG` joined
dalek rrot: r41333 | coke++ | trunk/runtime/parrot/library (2 files):
Add some clarifying documentation to Tcl::Glob, and have each Glob style
 12:58
TT #1022 created by coke++: PGE::Glob needs docs. 13:04
13:12 bluescreen joined 13:17 sri joined 13:33 pdcawley__ joined 13:40 pdcawley_ joined
dalek kudo: 4b141a8 | mberends++ | tools/test_summary.pl:
[tools/test_summary.pl] implement some simple relative benchmarking
 13:40
13:54 AndyA joined 13:59 Austin joined
Coke ./tclsh -e ':::::::::::::::::::::::::::::::puts hi' 14:11
davidfetter happy new year, Coke 14:13
Coke !? 14:15
davidfetter well, it's new year at sunset, wherever you are
Coke !?
davidfetter 5770
hey, i celebrate the CE new year. might as well share the joy :) 14:16
Coke ah. that gives me enough google clue. =-)
14:16 whiteknight joined
Coke shana tova. 14:16
davidfetter hatima tova :) 14:17
davidfetter never sure how to spell that, as the hebrew alphabet isn't isomorphic to latin 14:18
NotFound Spell in hebrew, we fully support Unicode
Coke maybe parrot does, but ssh/screen/irssi over putty makes for some difficulty. 14:31
NotFound Write a parrot irc client, then ;) 14:32
jonathan Coke: I managed to get it working...just about. 14:37
Coke: Though this week I had to switch to a non-standard PuTTY in order to get Korean char support.
14:40 Psyche^ joined 14:44 riffraff joined
Coke NotFound: just need to get ircbot.tcl working again. 14:44
(the io cleanup months ago killed it)
Coke wonders if he can trick jonathan into getting rid of tcl's home grown call chain! 14:45
14:59 sri_ joined
davidfetter שנה טובה 15:00
davidfetter wonders how parrot, etc. handle right-to-left languages
Coke davidfetter: no better than ICU does. 15:01
davidfetter heh
Coke (and probably not that well atm.)
davidfetter well, there are two important ones i can think of, the big one being arabic
15:02 theory joined 15:05 Austin joined, sri_ joined
Coke wonders again how he can make [uplevel] work with parrot's call chain. (upvar seems doable, if a pita.) 15:33
how can I run code in the context of something up the chain?
with the homegrown version, I pop the end off the chain, run the code, and then restore the chain manually. 15:36
anyone? =-) 15:40
15:46 desertm4x_ joined 15:52 quek left
particle 1/3.7 15:52
purl 0.27027027027027
Util double-takes while reading the scrollback backwards, *certain* that Coke was making a bicycle reference. Doh! 15:53
16:04 bacek joined
cotto_work hi bacek 16:14
clock?
purl cotto_work: LAX: Fri 9:14am PDT / CHI: Fri 11:14am CDT / NYC: Fri 12:14pm EDT / LON: Fri 5:14pm BST / BER: Fri 6:14pm CEST / IND: Fri 9:44pm IST / TOK: Sat 1:14am JST / SYD: Sat 2:14am EST /
cotto_work too much coffee?
bacek cotto_work: just arrived at home after Friday partying :) 16:16
16:37 jan joined 16:39 particle left
Austin pmichaud ping 16:41
16:45 mokurai joined
Austin Coke: I think the interpreter maintains a stack of continuations, no? 16:46
16:46 darbelo joined
Coke Austin: yes. 16:47
so I can, say, do $P1 = interp ['continuation'; 3] ... then what? 16:48
Austin Can you then use, or clone, that Continuation to do your uplevel? 16:49
Coke "how"?
purl i heard "how" was different for different languages
Austin (Keep in mind that I don't know what I'm talking about...) 16:50
Coke I think you're on the right track, fwiw. =-)
Austin In core.ops there's an "invokecc" which is a "normal" call, and there's "invoke" which takes a continuation as second arg. 16:51
It seems like there ought to be a "call-the-continuation" which doesn't take a sub. 16:52
But in your case, I think that "invoke" is what you want.
Coke looks plausible. if only those ops docs were more explicit. =-) 16:53
Austin That's what $PARROT/src is for... :-|
Coke if I switch to invoke(PMC), I'm going to need to manage my own args, yes? 16:56
dukeleto Coke: ping! re: trac.parrot.org/parrot/ticket/1012
Austin purl, msg pmichaud In close, I have a block containing two subs. The second sub is mal-formed, and so the parser reports a syntax error on the outer block. How can I get better error location?
purl Message for pmichaud stored.
Coke dukeleto: yes?
dukeleto Coke: do you have the powers to close said ticket? 16:57
Austin Coke: Sorry, I don't know. I'm just RTFMing trying to answer your question?
Coke dukeleto: ah. will do sometime today.
dukeleto Coke: sweet! that would be awesome
dalek rrot: r41334 | dukeleto++ | trunk (3 files):
[TT #1019] Make ok and nok use a PMC argument in test_more.pir and add tests for nok, flh++
 16:58
Austin Coke: That being said, can you parse the uplevel call into an eval with no args? 17:00
dalek TT #1019 closed by dukeleto++: Make Test;More;ok and nok use a PMC argument 17:01
Austin That is, treat "uplevel 1 foo(a, b)" as "uplevel 1 sub anon { foo(a,b) }"
Coke You're conflating tcl and pir there. 17:02
Austin I don't know tcl.
Coke I already invoke the PIR compiler too much; but I suppose that would be doable.
I'd rather do it it the right way.
Austin I remember uplevel and upvar, because they were so cool, but not much beyond that, syntax-wise.
I don't think there is a right way. The "right" way puts the arg info out early, so the PIR/IMCC/whatever can mangle it into a set_args/get_results byte stream. 17:03
17:15 particle joined 17:16 MoC joined, chromatic joined 17:17 DrForr joined
dalek a: 8c33dc2 | fperrad++ | config/makefiles/root.in:
add platform.longsize
 17:31
17:40 fperrad_ joined 17:44 joeri joined 17:53 fperrad_ joined
darbelo Aha! 17:59
18:01 fperrad_ joined
dalek rrot: r41335 | darbelo++ | branches/kill_jit/src (3 files):
With the JIT dead, nobody has exec capability anymore.
 18:05
chromatic Hm, one more runcore to remove. 18:06
dalek rrot: r41336 | darbelo++ | branches/kill_jit/src/frame_builder.c:
Bring back Parrot_jit_clone_buffer and Parrot_jit_free_buffer, platforms with PARROT_HAS_EXEC_PROTECT need them.
 18:12
darbelo Right now it's disabled by the configure probe, so there's no hurry to remove code.
cotto_work it'll be interesting to see how many lines get clobbered when that branch merges 18:17
darbelo Merge will probably be non-straighforward, I'm pretty sure. 18:18
There's also the 'moved files' thing 18:19
The git camp will probably get some ammunition out this branch merge. 18:21
cotto_work the more the merrier 18:33
whiteknight the poll is more lopsided then I expected 18:36
15:5
and CVS, the control group, is getting no votes 18:37
so that's a good sanity check 18:38
darbelo I would probably vote for it i could :) 18:39
whiteknight darbelo: can't vote?
darbelo is not an editor. 18:40
cotto_work that's probably a good thing ;)
whiteknight darbelo: try it now 18:41
darbelo Yup. whiteknight++ 18:42
Austin WhiteKnight: You have privs on the website?
18:43 notostraca joined
darbelo goes off to deface the website. 18:43
whiteknight Austin: yeah, I'm bt ebomb-diggity
and I have stupid fingers
Austin Can you see what's up with "Austin_Hastings" ? I did the set-a-new-password thing, but it denies me still.
And FWIW, switching to a new VC system is always a fashion show. You should expect mad enthusiasm for whatever new thing comes along after git, too. 18:44
chromatic If it's an improvement over Git as much as Git is an improvement over SVN, I'm willing to consider it. 18:46
darbelo The Parrot development workflow is branch-intensive, I think it would benefit from a tool with better branch management.
whiteknight I don't worry about change, so long as each change brings improvement 18:47
Austin: I don't see a user "Austin_Hastings" 18:48
18:49 AndyA joined
Austin :( 18:49
whiteknight Austin: which username do you want, "Austin_Hastings" or "Austin"? 18:51
Austin Machts nichts, whiteknight - as long as Firefox remembers it for me, it doesn't matter.
whiteknight Okay, "Austin" it is
Austin Wow. People can be on a first name basis with me. Woo-hoo. 18:52
whiteknight look at your nick here on IRC: We're already on a first-name basis
Austin What's my password?
whiteknight I don't know, haven't gotten that far
Austin Ah, okay. 18:53
whiteknight okay, the usernames "Austin" and "Austin_Hastings" don't exist, but your email address is listed as already registered
so that's lousy
dukelet0 'ello 18:54
Austin Now you can be as frustrated as me.
Hey, duke. 18:55
dukelet0 Austin: howdy 18:56
purl hi, dukelet0.
dukelet0 i still can't vote in the poll thingy
whiteknight duke: what's your username there?
Austin You haven't paid the poll tax.
whiteknight maybe I'll have more success with your account 18:57
dukelet0 whiteknight: dukeleto
purl dukeleto is Duke Leto Atreides I (10,140-10,191 A.G.) is a fictional character in the Dune universe created by Frank Herbert, or twitter.com/dukeleto
Austin You have to be a property owner in Philadelphia, a college graduate, and subscribed to the git-users mailing list. Then you can vote.
dukelet0 Austin: lulz 18:58
whiteknight Austin: Try username "Austin Hastings" with no underscore
Coke: ping 18:59
or chromatic, I think either could help
18:59 notostraca joined
Austin You rock! 19:00
purl Dis is the drum
whiteknight yay!
dukeleto: try the poll now
Austin How do I vote? 19:01
whiteknight you don't see a vote button?
dukelet0 now I see a button
Austin I see Preview Comment
whiteknight okay, try again 19:02
dukelet0 Austin: maybe you didn't pay the taxes on your land in Philly
Austin Boy, don't get me started on taxes.
I paid off my mortgage this year, so now I have to send those checks in by myself.
("You spend HOW MUCH on IT support for an Elementary School?!?!?!?!?") 19:03
Aha. Whiteknight, I got it after coming in from the top again. 19:04
Now I can log out and never touch it for another year.
whiteknight Austin: at least you have a mortgage.
Austin whiteknight++
Actually, I don't. I paid it off. 19:05
Now I can say "It's my house." and mean it.
dukelet0 whiteknight++ for being a poll ninja
whiteknight the bank keeps telling me "the economy is down blah blah", and "no credit" and "you'd be worth more ground up as dogfood", etc
Austin lol
What kind of down payment did you offer? 19:06
19:06 sri_ joined
whiteknight Whatever was in my wallet and I had a pack of skittles in the car 19:06
"lightly used"
Austin I think if you go above 20%, their minds turn to much and they agree to anything.
s/much/mush/
chromatic Copies of the PIR book don't work. Trust me. 19:07
whiteknight chromatic: on the website what's the difference between "editor", "poster", and "member"?
which groups should I apply to which disenfranchised souls?
Austin A "poster" is a big picture left over from college, probably involving a naked woman and a big snake. 19:08
chromatic Every committer should be a member.
Austin An "editor" is ed, or an imposter.
chromatic I don't know the rest of that.
whiteknight chromatic: okay, that's all I needed to hear
19:10 sri joined 19:11 integral joined
dukelet0 Austin: i'll have one of what you are having 19:13
19:13 flh joined
Austin serves up another shot of straight unix. 19:14
19:24 iblechbot joined
Austin Is there a document specifying how multi dispatch works at the parrot level? 19:25
dalek tracwiki: v16 | dukeleto++ | GitObjections
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
19:25 notostraca joined
darbelo Last time I checked it said "Slowly". 19:26
But that was before the magical coding robot optimized it.
dalek rrot: r41337 | darbelo++ | branches/kill_jit (5 files):
Remove some more exec-related dead code.
 19:27
Austin Not much of a document, but I'll take what I can get. Where is it? 19:28
darbelo Austin: docs/multidispatch.pod but its old. 19:32
Coke whiteknight: pong. 19:35
whiteknight Coke: I'm trying to figure out what the differnet roles mean on the website 19:36
"poster", "editor", "member", etc
chromatic says "members" are committers, so I think I know that muc
chromatic Anyone who's made two or more contributions to Parrot in the past calendar year is a member. 19:37
darbelo contributions? 19:38
purl contributions are very welcome
chromatic Patch, commit, bug report, documentation, et cetera. 19:39
Austin Hmm. Is ":invocant" still used? 19:40
darbelo Austin: I think so, but the PCC branch was going to change something about it, IIRC 19:41
Tene considers investing tonight in pcc. 19:42
darbelo pdd27 still mentions it FWIW
Tene++ 19:43
Tene oh, I do have the entire weekend mostly unscheduled...
and pcc does need to happen...
darbelo Tene++
Coke whiteknight: they're just roles. each role can have various perms, but in general, and editor can edit stories. a poster can submit stories. a member is a member of the pafo (i.e. a committer) 19:44
Austin Pdd27 is what I'm reading now.
whiteknight Austin: the :invocant flag was planned but neveri mplemented 19:46
the PCC refactors were supposed to either add that, or improve the landscape so it could be added
Austin In favor of the :multi(_,_) syntax, I guess.
Probably better - I was dreading trying to parse that.
Tene OH NO SVN IS CATCHING UP 19:50
17:7 now
Austin Quick, Tene, create some fake accounts!
Tene Oh, good idea!
purl Tene: Good Idea: Taking up a new hobby like bird calling. Bad Idea: Taking up a new hobby like buffalo calling.
Tene I just need to send in fake CLAs for them so they can become committers... 19:51
Austin Naw, just get WhiteKnight to help you. 19:52
He needs a down payment for a mortgage, so ...
Maybe $1,000 each user? 19:53
whiteknight if I could afford to pay 1000$ for each user, I'd already have a house 20:00
")
:)
(stupid fingers today) 20:01
Austin The plan was that you'd get paid for each fake user.
20:10 notostraca_ joined 20:37 bluescreen joined 20:42 AndyA_ joined 20:44 zerhash joined
dalek rrot: r41338 | darbelo++ | branches/kill_jit (3 files):
Recycle the old auto::jit exec_protect probes for auto::frames. Shuffle some files around.
 20:49
darbelo Tests of kill_jit r41338 on x86 appreciated!
20:53 cotto_w0rk joined
chromatic auto::jit - Determine JIT capability.............................no. 20:55
auto::frames - Determine call frame building capability............yes.
dalek rrot: r41339 | darbelo++ | branches/kill_jit/include/parrot/runcore_api.h:
Oops. Forgot to headerize.
 20:56
chromatic Building and testing now.
darbelo I got a passig fulltest on linux before restoring the exec_protect fucntionallity, I expect few surprises there. 20:57
kid51++ submitted an all-green smolder for pcc too. 20:58
I developed that last patch on x86 OpenBSD, and all test pass there, with fulltest still running. 21:00
It's starting to look good.
21:18 quek joined
chromatic All tests pass here. 21:19
darbelo I guess that's enough configure-poking then. 21:23
Time to make frame_builder.c less ugly.
dalek rrot: r41340 | darbelo++ | branches/kill_jit/t/steps/auto/frames-01.t:
Add tests for the new Configure step.
 21:24
chromatic Hm, startup time is faster too.
darbelo Lots of code removed.
21:24 notostraca joined
darbelo I think a struct or two might have lost members as well. 21:26
chromatic Not a huge amount, but a little bit.
21:27 Austin joined 21:43 jrtayloriv joined 21:52 theory joined
dalek rrot: r41341 | mikehh++ | branches/kill_jit/src/frame_builder.c:
codetest failure - pod syntax
 21:52
rrot: r41342 | darbelo++ | branches/kill_jit/MANIFEST:
Regenerate MANIFEST.
21:53 zerhash joined 21:54 Whiteknight joined
mikehh darbello: I get manifest_tests failures - do you want to run tools/dev/mk_manifestand_skip.pl or shall I 21:54
ok you did it :-} 21:55
dalek rrot: r41343 | darbelo++ | branches/kill_jit (2 files):
Remove jit2c.pl
 22:00
jdv79 Coke: yeah, i've talked to mpeters about it. he says the next version is more stable. 22:01
Whiteknight mattp: ping 22:03
mikehh darbelo: looks good 22:21
darbelo mikehh: What platform are you on? 22:22
mikehh oops - forgot to apply the svn properties patch 22:23
Ubuntu 9.04 i386 at the moment
make test / make fulltest PASS
dalek rrot: r41344 | mikehh++ | branches/kill_jit/t/steps/auto/frames-01.t:
codetest - add svn properties
 22:24
darbelo Ah, my svn-props commit just failed.
mikehh codetest - TODO pass -
darbelo TODO pass aren't a worry in branches, we can get them in trunk after the merge. 22:25
Hmm I'll need somebody to test this on win32 before we can merge it. 22:26
mikehh t/codingstd/c_function_docs.t - TODO passed: 48 - ok 48 - compilers/pirc/src/bcgen.c # TODO Missing function docs
I'll check later to see how it fares on Ubuntu 9.04 amd64 22:42
Whiteknight I'll do that now
darbelo amd64 should see no changes from the point we branched, the frame vuilder is a x86 thing 22:43
Whiteknight I can still test it 22:44
Whiteknight likes feeling productive
darbelo (Just and informative comment, not meant in any way to discourage testing)
Whiteknight: actually, give 2 minutes and I'll have a commit for you to test there. 22:45
Whiteknight ok 22:47
I can coretest almost 4 times in 2 minutes
darbelo ok, commit is done. test away! 22:48
Whiteknight ok
dalek rrot: r41345 | darbelo++ | branches/kill_jit/src/frame_builder.c:
Kill some leftover JIT functions, they are not used in the frame builder.
 22:49
darbelo mikehh: if you are still on x86 a make test would be appreciated. 22:51
Whiteknight tests are good 22:53
darbelo *-amd64, OpenBSD-x86 and Linux-x86 cover all the #ifdefs in this code AFAICT.
mikehh darbelo - make test PASSes
darbelo ok, let's see what else I can rip out of there. 22:54
23:05 tetragon joined
mikehh need some sleep 23:30
bbl
23:32 sri joined 23:42 MoC joined 23:45 bacek joined 23:46 desertm4x__ joined
cotto_w0rk good morning bacek 23:49
bacek good morning cotto_w0rk 23:50
dalek rrot: r41346 | darbelo++ | branches/kill_jit (9 files):
Obliterate exec. Kill it dead.
 23:52
cotto_work darbelo, how long until you merge that sucker? 23:58

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!