#parrot Parrot 2.2.0 "Like Clockwork" Released! | parrot.org/ | Channel log: irclog.perlgeek.de/parrot/today | Priority: documentation sprint and pre-release testing for 2.3, fix line number annotations | Review and vote GSoC applications
Set by moderator on 13 April 2010.
sorear Invalid charset number '1537' specified 00:03
parrot trunk does not build for me
darbelo Did you try a 'make realclean' ? 00:04
sorear known?
purl somebody said known was good
sorear no
darbelo You probably need to. 00:05
sorear :(
darbelo We changed the packfile format a bit.
sorear How is the deprecation cycle stuff going to work? 00:06
darbelo Deprecation cycle stuff?
sorear with immutable strings and changed packfile format
darbelo Any parts of the API that change in a backwards-incompatible way get listed in DEPRECATED.pod, with a date after which you can't rely on them anymore. 00:08
The packfile/bytecode format has no such explicit guarantee. 00:09
purl: msg plobsing Do your changes to packfiles need a PBC_COMPAT bump? 00:12
purl Message for plobsing stored.
cotto I don't think immutable strings will be merged prior to the release. It's only necessary to be sure to get the proper deprecation notices in place. 00:17
darbelo AFAICT immutable strings *can't* merge before 2.3 ships. 00:18
sorear Is the entire string API going to be deprecated in 2.3? 00:20
darbelo Not really. Most of the functions are fine the way they are now. 00:21
Coke I would tend to doubt that.
(to sorear)
darbelo The ones that modyfy strings in place have to go, obviously.
And bacek took an axe to a few ops IIRC. 00:22
But there shouldn't be too many changes to most things. 00:23
bacek_at_work I did put deprecation notice for strings, afair. 00:28
darbelo =item STRING COW [eligible in 2.4]
Yep.
But not the inplace ops. 00:29
bacek_at_work I updated ticket but not DEPRECATED.pod... 00:30
darbelo, can you update it please?
darbelo Sure, TT# ?
bacek_at_work 1540 00:31
darbelo bacek_at_work: done. 00:36
It's r45678 if you want to check that I didn't miss anything.
bacek_at_work darbelo, thanks 00:38
00:38 tcurtis joined
Whiteknight allison++ #email to the list about immutable strings 00:44
darbelo Just read it, it looks like we'll need to add some more stuff to the deprecated pile. 00:45
Whiteknight oh, darn 00:46
I hate deprecating things </lies>
darbelo Just make sure to deprecate them before 2.3 ships out.
dalek rrot: r45678 | darbelo++ | trunk/DEPRECATED.pod:
Update the list of deprecated items to reflect the altest changes to TT #1540.
 00:51
darbelo Someone needs to update my updates in light of allison's mail to the list.
darbelo has to go now.
00:55 abqar joined
davidfetter did anybody ever get started on a p5 implementation atop parrot? 01:16
cotto davidfetter, blitzkost is the closest to that effort 01:18
davidfetter 600613s
cotto though it embeds Perl 5 in Parrot rather than trying to parse it correctly
01:18 jsut joined
cotto msg bacek Make sure to add a list of deprecated items related to the immutable strings branch. 01:21
purl Message for bacek stored.
01:40 plobsing joined
dalek rtcl-nqp: 1891de6 | Coke++ | (2 files):
run this test modulo todos.
 01:43
rtcl-nqp: f75d6b1 | Coke++ | (2 files):
implement as much of [vwait] as we had in partcl-original.
rtcl-nqp: 59cc1c2 | Coke++ | (2 files):
implement [format] in the style of partcl-original.
01:49 davidfetter joined 02:07 chromatic joined
sorear davidfetter: I'm working on Blizkost, which uses the P5 guts in order to be compatible with promiscous CPAN XS modules 02:57
Larry is working on STD_P5, which is probably going to turn into Ponie 2 at some point
davidfetter i take it ponie2 is *not* perl2/parrot 02:59
sorear STD_P5 is a parser for Perl 5 03:01
do not confuse this with Punie
or Pynie, for that matter
davidfetter heh 03:02
Coke panie. 03:06
03:09 patspam joined 03:10 brooksbp joined 03:16 abqar joined
dalek parrot: e4fe808 | dukeleto++ | (2 files):
Wrap PIR subroutines

users do not have the ability to override other functions		 03:18
parrot: cf6540a | dukeleto++ | (2 files):
Merge branch 'subwrap'
03:29 patspam joined 03:32 janus joined 04:08 Austin joined 04:26 brooksbp_ joined 05:10 brooksbp joined 05:17 allison joined 05:18 rt7 joined 05:34 brooksbp_ joined 05:55 chromatic joined, eternaleye joined 06:06 jhelwig joined 06:07 uniejo joined 06:32 aukjan joined 06:35 gerd joined 06:39 aukjan joined 07:16 aukjan joined 07:25 aukjan joined 07:28 iblechbot joined
tcurtis Does make cover really need to take several hours? 07:34
sorear if it's covering Perl code, yes
the DB:: interface is hideously slow 07:35
if it's mostly running over C, it shouldn't take more than 2-3x as long as make fulltest
07:42 Gerd joined
dalek parrot: 3cd468d | dukeleto++ | (2 files):
Add naive support for timestamp-related datatypes
 07:48
07:49 gerd joined 08:22 tcurtis joined 08:31 fperrad joined 08:32 fperrad_ joined 09:20 allison joined 09:21 bacek joined
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#33202), fulltest) at r45678 - Ubuntu 10.04 beta amd64 (g++ with --optimize) 09:45
bacek o hai 09:46
mikehh hi bacek - need any branch testing done?
moritz is the immutable strings branch far enough to get HLLs running on it? 09:47
bacek moritz, yes. But rakudo's binder fails epically.
moritz does NQP-rx run on it? 09:48
bacek mikehh, immutable_strings testing would be nice.
moritz, yes, afaiu
mikehh backe - ok will work on it as soon as I reboot (updateing Ubuntu 10.04 beta - kernel changes) 09:49
09:50 snarkyboojum joined
bacek mikehh, ok 09:50
mikehh bbiab
09:58 allison joined 10:05 mikehh joined
dalek rrot: r45679 | bacek++ | branches/immutable_strings_part1/src/string/api.c:
Remove useless string_capacity function
 10:10
rrot: r45680 | bacek++ | branches/immutable_strings_part1/src/string/api.c:
Don't "copy" old string in Parrot_str_clone. Just create new fresh string header.
10:24 allison joined
mikehh bacek: immutable_strings_part1 won't build with g++, gcc - make corevm builds, lots of failures with make coretest, make fails 11:06
bacek mikehh, ah... sorry, I forgot to dcommit last one. 11:07
done, r45681
mikehh bacek: immutable_strings_part1 - with gcc make corevm/make coretest, make world/make test PASS 11:14
dalek rrot: r45681 | bacek++ | branches/immutable_strings_part1/src/string/api.c:
Revert previous commit. It causes some problem with building PGE.
 11:15
mikehh bacek: immutable_strings_part1 - g++ fails to build - src/string/charset/unicode.c 11:17
bacek mikehh, thanks 11:31
11:34 clinton joined 11:41 payload joined 11:42 payload1 joined 11:55 khairul joined
dalek kapo: 51facfb | austin++ | (2 files):
Updated IsNull matcher
 12:01
kapo: 5a3b504 | austin++ | (2 files):
Got is-a-type matcher working
kapo: 78210be | austin++ | (2 files):
Got AllOf Matchers working, tested
kapo: b60568c | austin++ | (2 files):
Got AnyOne Matchers working, tested
kapo: 7fa5a4c | austin++ | (2 files):
Got Defined matcher up, tested
kapo: a7ed2e5 | austin++ | (2 files):
Got IsString matcher tested
12:01 snarkyboojum joined
dalek kapo: d2170da | austin++ | (2 files):
Got IsInteger matcher tested
 12:01
kapo: a9bff64 | austin++ | (2 files):
Got IsNumber matcher tested
rrot: r45682 | bacek++ | branches/immutable_strings_part1/src/string/api.c:
Actually avoid useless str_copy in str_clone and str_append
 12:04
kapo: dd2db16 | austin++ | (2 files):
Got PassFail matcher tested
 12:06
12:06 whiteknight joined
bacek purl, Good morning, #parrot is <reply>Good morning, Mr. Whitworth! 12:08
purl OK, bacek.
whiteknight ... 12:09
good morning, #parrot
purl Good morning, Mr. Whitworth!
whiteknight :)
bacek :)
whiteknight how are you doing this morning, bacek? 12:13
bacek whiteknight, pretty ok. Broke branch during syncing with trunk... 12:14
12:17 tetragon joined
whiteknight ouch. hate it when that happens 12:18
bacek whiteknight, nm. I fixed it already :)
whiteknight of course. it's easy for a magical coding robot 12:19
dalek rrot: r45683 | bacek++ | branches/immutable_strings_part1 (109 files):
Sync branch with trunk

  \tsrc/pmc/namespace.pmc
  \tsrc/pmc/string.pmc
  \tsrc/string/api.c
  \tsrc/string/encoding/fixed_8.c
  \tt/pmc/string.t		 12:20
rrot: r45684 | bacek++ | branches/immutable_strings_part1/t/pmc/string.t:
Fix test plan after merge.
Austin good morning, #parrot
purl Good morning, Mr. Whitworth!
Austin Hmmm
Needs work. 12:22
bacek purl, forget good morning, #parrot
purl bacek: I forgot good morning, #parrot
Austin For reasons which are unclear to me, I've encountered a whole bunch of references to REBOL recently. 12:23
Bafflingly, it's a closed/unfree language which a bunch of people are pushing as "the way to do it". 12:24
moritz looking at rebol examples, I'm hardpressed to actually see a structure in this thing
looks like a huge pile of built-in libraries 12:25
Austin Yeah, it's like the Cobol version of PHP.
Chromatic linked to a language rant that I would have sworn came from like 1993, but didn't. 12:26
"You shouldn't have to declare your variables!"
moritz I've read that too... 12:27
Austin "You shouldn't have to import anything to do I/O!"
moritz the author does have some points, but some of it is pretty much nonesense
Austin Heh. The 1990's called, they want their shoes back.
moritz "one single aggregate type"
what a bad idea 12:28
kurahaupo Javascript anyone?
12:33 ruoso joined
dalek rrot: r45685 | bacek++ | branches/immutable_strings_part1/src/string/api.c:
Throw exception on attempt to replace in NULL strings.
 12:37
Coke javascript is a nice language. =-) 12:40
I see that austin's been busy! 12:41
oooh, Mono on MacOS X 12:42
12:43 JimmyZ joined 12:47 iblechbot joined
dalek rrot: r45686 | mikehh++ | branches/immutable_strings_part1/MANIFEST.SKIP:
regenerate MANIFEST.SKIP
 12:53
rrot: r45687 | mikehh++ | branches/immutable_strings_part1 (10 files):
set svn properties
12:55 lucian joined 12:56 lucian joined 12:58 JimmyZ_ joined 13:06 bubaflub joined
dalek p-rx: 210c0b7 | pmichaud++ | src/HLL/Actions.pm:
Avoid global-namespace lookup of quotemod_check (TT #389).
 13:19
p-rx: c7127ad | pmichaud++ | src/stage0/ (3 files):
Update bootstrap.
pmichaud good morning, #parrot 13:25
mikehh hi pmichaud 13:26
dalek rrot: r45688 | bacek++ | branches/immutable_strings_part1/include/parrot/string_funcs.h:
Rerun headerizer.
 13:27
bacek hi pmichaud
mikehh bacek: you still have a c++ comment in src/string/encoding/fixed_8.c
rrot: r45689 | bacek++ | branches/immutable_strings_part1 (5 files):
Constify to_encoding
rrot: r45690 | bacek++ | branches/immutable_strings_part1/src/string/api.c:
Remove commented out write_COW. They are actually not required.
rrot: r45691 | pmichaud++ | trunk/ext/nqp-rx/src/stage0 (3 files):
[nqp-rx]: Eliminate use of namespace-lookup for quotemod_check. (TT #389)
bacek mikehh, it's gone. I'm just checking locally
13:31 atrodo joined 13:42 dalek joined
bacek killed dalek... Again... 13:47
Austin You bastard!
bacek just don't like competition with other robots. 13:49
13:53 patspam joined
dalek rrot: r45693 | bacek++ | branches/immutable_strings_part1/src/string/encoding (4 files):
Actually remove now unused set_bytes and set_codepoints functions
 14:01
bacek mikehh, ok. Branch should be ready for tests. 14:08
mikehh bacek: looks good - starting fulltest now 14:10
bacek mikehh, thanks
dalek tracwiki: v36 | bacek++ | BranchDescriptions 14:18
tracwiki: remove pcc_mergecells branch
tracwiki: trac.parrot.org/parrot/wiki/BranchD...ction=diff
kudo: 0f1a5cd | pmichaud++ | src/Perl6/Actions.pm:
Avoid direct-namespace lookup of HLL::Grammar::quotemod_check.

a different HLL-root.		 14:22
particle pmichaud: related to TT#1542 perhaps? 14:23
14:23 snarkyboojum_ joined 14:24 JimmyZ joined
pmichaud particle: looking 14:28
Coke schwern++ 14:29
pmichaud particle: yes, TT#1542 looks *very* suspicious.
particle pmichaud: i thought so, based on your excellent commit message 14:30
mikehh immutable_strings_part1: 14:32
All tests PASS (pre/post-config, make corevm/make coretest, smoke (#33211), fulltest) at r45693 - Ubuntu 10.04 beta amd64 (g++ with --optimize)
14:33 theory joined
mikehh bacek: everything ok and g++ too 14:33
bacek mikehh, I did one more commit. I promise, it was last one! :) 14:34
dalek rrot: r45694 | bacek++ | branches/immutable_strings_part1 (26 files):
Remove Parrot_str_append function which was actually synonym for str_concat
rrot: r45695 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
[distutils] use TAP/Harness library for step 'test' (instead of P5 prove)
rrot: r45696 | fperrad++ | trunk/runtime/parrot/library/TAP/Parser.pir:
[TAP] small fix
rrot: r45697 | fperrad++ | trunk/examples (2 files):
[examples] update setup.pir
mikehh bacek: that blew up a bunch of tests 14:42
bacek mikehh, at r45698? 14:43
14:43 khairul joined
mikehh no r45697- I'll svn up 14:44
14:45 JimmyZ_ joined
bacek mikehh, i forgot to grep outside of src/. My bad. 14:45
dalek rrot: r45698 | bacek++ | branches/immutable_strings_part1 (5 files):
Remove last references to Parrot_str_append
 14:51
mikehh immutable_strings_part1 branch: 15:01
All tests PASS (pre/post-config, make corevm/make coretest, test, fulltest) at r45698 - Ubuntu 10.04 beta amd64 (gcc with --optimize)
it builds and passes corevm/coretest, test for g++ as well 15:07
bacek: looks good - part2? 15:13
bacek mikehh, after merge.
more cleanups, remove deprecated inplace ops, etc. 15:14
consting of strings passed
mikehh bacek: excellent
15:15 jsut_ joined
mikehh bacek: when are you going to merge - after 2.3? 15:15
bacek yes
There is a lot of incompatible changes.
(And Rakudo still failing epically...) 15:16
mikehh 'k - back to testin' and fixin' trunk 15:17
dalek a: 360858c | fperrad++ | setup.pir:
use tapir instead of prove
15:22 chromatic joined
dalek rrot: r45699 | fperrad++ | trunk/tools/dev/tapir.pir:
[tapir] fix exit status
 15:24
rrot: r45700 | fperrad++ | trunk:
add tapir in svn:ignore
he 86 failed tests on NetBSD/macppc 4.0, ref. smolder.plusthree.com/app/projects/...ails/33212 - that's a lot worse than it used to be. 15:44
darbelo It's all packfiles, ouch. 15:46
15:51 brooksbp joined
darbelo he: Did you make realclean, start from a fresh checkout, or just "svn up; make" ? 15:52
15:53 muixirt joined
muixirt hi 15:55
purl what's up, muixirt.
dalek lscript: 6d84106 | fperrad++ | setup.pir:
use tapir instead of prove
muixirt for what is 'nci_thunk_gen' good for
Austin thunks? 15:56
purl thunks are implemented in the singleton POE::Session. There's one thunk per POE call signature. You include a bit of data with the outbound event that maps the inbound event back to you.
Austin What do I win?
darbelo muixirt: It generates nci thunks ;)
Austin: A parrot bug that breaks your code on the next release. 15:57
Austin darbelo: Cool. I'm trying to collect the entire set. 15:58
muixirt darbelo: and what have i done wrong because invoking it i get an error msg: "load_bytecode" couldn't find file 'data_json.pbc'
darbelo There was something bootstrapy about that code. But I can never remember what. 16:00
muixirt so basically one have to do bindings for system libraries manually? 16:01
16:01 plobsing joined
darbelo muixirt: plobsing wrote that, IIRC. He'll know better. 16:02
plobsing convenient timing. what did I miss?
darbelo muixirt is having troubles with nci_thunk_gen. 16:03
muixirt so how would i start for getting some bindings to system libraries? 16:04
plobsing darbelo: wrt PBC_COMPAT - last time I bumped it, I got reverted, so I'm not sure what the rules are for bumping. But I notice that all stable releases get a bump, so why bother?
muixirt plobsing: invoking nci_thunk_gen gives the error msg: "load_bytecode" couldn't find file 'data_json.pbc'
plobsing hmmm... you don't have data_json.pbc in parrots library search path. that's a fail right there. I thought we fixed that. 16:05
16:06 brooksbp_ joined
plobsing muixirt: to bind to systems libraries call 'loadlib' to get a shared library object and then call 'dlfunc' to get native function objects out of that library object 16:06
muixirt: you only need to use nci_thunk_gen if parrot does not currently support the signatures of the functions you need to bind. 16:07
you'll know that because it will give you an exception stating as much when you try and call that function (thunk finding is lazy right now, I'm considering whether we want to make it eager at some point) 16:08
muixirt once i tried nci bindings for libcurl but it was a tedious task, and a few months later it didn't worked anymore but i didn't bother to look into it 16:10
whiteknight plobsing: all stable revisions don't automatically get a bump
plobsing: release policy is that if we've had several small bumps between releases we aggregate them all into a single large bump
plobsing oic, that's why my bump got rolled back 16:11
whiteknight who did the rollback? you should have at least gotten a friendly explanation
Coke will give whiteknight a dollar the next time we have a stable release that doesn't have a bump,though. =-)
Austin Yeah. The only revisions that gets bumps are the ones with dots in them. 16:12
plobsing muixirt: if you're looking for something to automatically generate bindings by inspection, nci_thunk_gen is not it
whiteknight Coke: no, I'm not denying that one often brings the other
but it's not a rule that it must
plobsing muixirt: I don't think we have that yet
muixirt well it's a hard problem 16:13
plobsing projects/patches welcome
darbelo I always thought 'You need to run mk_native_pbc' => 'You need to bump PBC_COMPAT' 16:14
Since, hey, you just made PBC incompatible.
plobsing I'll look into who rolled back my bump and ask them.
darbelo: I'll admit to that. And at a pretty fundamental level too.
muixirt: if you don't want to do low level stuff to get at NCI, check out github.com/jnthn/zavolaj/ 16:16
muixirt plobsing: thanks for the pointer 16:17
16:30 davidfetter joined 16:32 tcurtis joined 16:44 theory joined 16:45 theory_ joined 17:10 chromatic joined
cotto_work morning, humans 17:39
darbelo fsvo humans. 17:41
cotto_work starts to wonder 17:42
davidfetter is "morning humans" anything like "morning coffee?" 17:43
cotto_work nom
darbelo Slightly less tasty, but definitely more crunchy.
cotto_work I don't really like coffee. 17:44
darbelo I'm the opposite. With few exceptions, I can't stand humans. 17:47
17:47 whiteknight joined
cotto_work speaking of humans, here's whiteknight 17:48
whiteknight TADA!
moritz I thought you were speaking of humans :-)
darbelo suspects everyone here is a bot. 17:49
whiteknight especially bacek
dalek rrot: r45701 | mikehh++ | trunk/MANIFEST.SKIP:
regenerate MANIFEST.SKIP
 17:50
rrot: r45702 | mikehh++ | trunk/runtime/parrot/library/distutils.pir:
fix codetest failure - pod syntax
18:02 lucian_ joined
dukeleto 'ello 18:04
cotto_work hi dukeleto 18:05
dukeleto cotto_work: good $localtime 18:06
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#33213), fulltest) at r45702 - Ubuntu 10.04 beta amd64 (g++ with --optimize) 18:13
18:28 jan joined 18:30 jsut joined 18:32 iblechbot joined
dalek rrot: r45703 | chromatic++ | branches/immutable_strings_part1/src/pmc/eventhandler.pmc:
[PMC] Made EventHandler's get_string VTABLE return a compile-time constant,
 18:39
darbelo dukeleto: ping 18:47
18:53 davidfetter joined
chromatic wants -flto from GCC 4.5 18:55
dukeleto darbelo: pong 19:00
moritz just sent chromatic a bug report :-)
darbelo dukeleto: Now tapir is bundled in parrot, is development going to happen in svn or will it be 'pulled from upstream' like nqp-rx? 19:02
arnsholt The precedence limit in NQP-rx's grammars, that means "only allow operators binding tighter than this level" right?
moritz chromatic: is there public source repository for Pod::PseudoPod::LaTeX anywhere?
chromatic I have my own SVN repo I haven't converted to Git yet. 19:03
Coke darbelo: is that even the same tapir, or is it a reimpl?
moritz chromatic: ping me if you do, I might get around to attack some of the bugs that affect the Rakudo * bug 19:04
darbelo Good question. I just saw a tapir executable and assumed it was the same one.
chromatic moritz, I'm checking to see if there are any changes in that repo since the last release to the CPAN. I don't *think* so.
Confirmed. 19:05
Let me fork the gitpan repo.
19:06 integral joined
chromatic github.com/gitpan/Pod-PseudoPod-LaTeX 19:06
dukeleto darbelo: I didn't know tapir was bundled with parrot
darbelo: i thought fperrad reimplemented something, but I am not sure 19:07
darbelo Oh... ok.
dukeleto darbelo: he never told me about it either, i just started seeing TAP/* stuff in the parrot repo 19:15
darbelo: i think he has a commit bit on Tapir, so I don't quite know what is up
19:16 hercynium joined 19:39 joeri joined
Coke wishes for a vi flag that said "edit this file and start me on the first occurence of "some pattern". 19:44
particle hrmm... 19:46
i can get you to a tag with -t
chromatic vi -c /pattern file
particle coke: from cmdline, or inside the tool?
chromatic Or at least vim -c /pattern 19:47
Coke use case - I run a test, I get a failure, I have the test name, but not the line number it occurred on.
particle well, ctags can fix that
Coke chromatic: woot. that's it.
particle but vim -c will work
chromatic++
cotto_work line numbers for test failures would be nice too
Coke I was trying --cmd, which does the command BEFORE the file was loaded. *facepalm* 19:48
cotto_work www.youtube.com/watch?v=ZOq2y5fTaHc 19:50
very shiny jvm inspector 19:52
20:07 ircleuser joined
cotto_work is building gcc 4.5. This might take a bit. 20:13
moritz msg chromatic please review (and if ok, apply) the patches from <github.com/moritz/Pod-PseudoPod-LaT...mits/>, they improve the rendering of the Perl 6 book significantly 20:17
purl Message for chromatic stored.
chromatic Will do. 20:18
moritz I've bumped the version number a bit, but you can revert that if you don't like it 20:20
that was just for locally building Debian packages from it
moritz.faui2k3.org/tmp/book.pdf for the PDF 20:21
the only remaining problem is rendering of » 20:22
Austin What do you call an instantiation of a role? 20:24
class Foo does R ; my $f := Foo.new ;
what's $f?
(in respect to R)
20:25 kjeldahl joined
moritz an implementation of R, or so 20:25
or $f does Foo
Austin I was thinking that Foo was an implementation of R. 20:26
moritz right 20:27
particle an object
purl an object is pretty simple, there are no lazy defaults
Austin So is $f an "instance" of R, just as it would be an instance of some parent class of Foo? 20:28
particle roles are composition-time only
er, an instance, not an object
Austin particle: I get that.
But there's still a relationship between $f and R, and I'm trying to give it a name.
particle if you try to instantiate a role, a class is punned
moritz $f does R
20:28 cotto_work joined
particle $f isa <anon-Foo> does Foo 20:29
Austin $f is a doer of R
chromatic moritz, if you send me a pull request I'll apply those.
moritz chromatic: sent 20:31
chromatic msg schwern Can you give me committer access on github.com/gitpan/Pod-PseudoPod-LaTeX ? 20:43
purl Message for schwern stored.
chromatic Hm, that wasn't what I intended.
particle the future is angry at you 20:44
chromatic Yes, because I beat it up.
moritz chromatic: you can just fork it yourself, and consider your fork the "official" version 20:45
chromatic Sure, and I may do that. I want to know Schwern's intention with gitpan first though.
20:50 theory joined
chromatic moritz, I just uploaded 1.101050 to the CPAN. 20:50
moritz chromatic++
20:59 brooksbp joined
dalek rrot: r45704 | fperrad++ | trunk (3 files):
[TAP] improve formatter
 21:06
cotto_work cc1: error: LTO support has not been enabled in this configuration 21:19
grrr
NotFound winxed pass all tests with the immutable_strings_part1 branch 21:21
cotto_work it'd be nice to know that I need to pass --enable-lto to gcc's configure before I spent a bunch of time buiding it
This is a time when it'd be nice if Parrot had a separate build dirs a la gcc and Firefox. 21:23
bacek ~~ 21:25
Good morning, biological beings. 21:26
cotto_work ~~~
cotto_work wonders if that makes him 50% smarter
Coke is reminded to test partcl. 21:28
tcurtis How do locals and lexicals differ? 21:29
moritz tcurtis: what do you mean by "locals"?
.local pmc foo?
PerlJam almost answered with the standard perl response 21:30
tcurtis moritz, yes.
moritz tcurtis: .local is always tied to the current .sub
tcurtis: wheras lexicals (via findlex etc.) present HLL lexicals 21:31
tcurtis: and a HLL can compile an inner block to a different subroutine
so HLL lexicals can spread around multiple PIR subs
21:31 Whiteknight joined
moritz (as far as I understand) 21:32
tcurtis Thanks. moritz++ 21:33
particle .local attaches a human-readable name to a register inside a sub 21:35
21:45 kid51 joined 21:49 cotto_work joined
dalek kudo: 3a1eb98 | jonathan++ | src/Perl6/Actions.pm:
We should only go looking for outer lexical candidates when introducing a lexical multi, not an our-scoped one. Fixes a regression.
 21:56
21:57 |Ryan52 joined
dalek kudo: 5fd0669 | jonathan++ | t/spectest.data:
Add a # icu marker for a test that seems to need it.
 22:02
cotto_work looks like there's a small improvement between gcc 4.4.1 and gcc 4.5.0 with -flto for an optimized parrot 22:23
22:26 Mokurai joined
cotto_work 2005807033/1999880110 22:27
purl 1.00296363915535
cotto_work valgrind shows a decrease, though wallclock time looks slightly faster 22:28
darbelo Did you run it at least 50 times and do a statistical comparison? 22:29
;)
cotto_work That's why I say "looks".
I'm too lazy to get something where I can say "is".
darbelo I wonder if valgrin could be convinced to output something a bit more like walltime... 22:31
cotto_work I thought it was smart about emulating the cost of various instruction but I didn't find anything in the docs about it. 22:32
darbelo My problem is that instruction counts, branch prediction and cache miss analysis are separate, so I can't really trust 'optimizations' that help one but hur another. 22:35
*hurt
That's exactly what happened when I ripped out next_for_gc. instruction counts went down, but branch mispredicitons and cache misses went up. 22:36
Technically, we could develop a 'wheight criteria' of 'A gain of W% in measure X is only justified if it hurts Y by les than Z%' but it'd be easier if it came with the tool. 22:42
22:44 pjcj joined
darbelo OTOH, I know enoguh about hardware to know that whatever we do It'll suck on at least one arch ;) 22:45
dalek kudo: 78faa0c | jonathan++ | src/Perl6/ (2 files):
Bring our parsing of array and hash subscripting in line with STD. This also unbreaks zen slices.
 22:48
22:52 cotto_work joined 23:05 TiMBuS joined 23:09 ruoso joined 23:10 davidfetter joined 23:14 tetragon joined 23:51 lucian joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!