Parrot 2.1.1 Released! | parrot.org/ | Tasks: PCC deprecations branch, HLL subclassing and MMD branch
Set by moderator on 19 February 2010.
00:03 snarkyboojum joined
bacek_at_work Heh :) 00:05
ttbot disagree with fperrad
NotFound bacek_at_work: Do you trust a bot better than a buddy? ;) 00:08
00:20 Myhrlin joined 00:33 AndyA_ joined 00:38 GeJ joined 00:39 NotFound joined, nbrown_ joined, fperrad_ joined, dngor joined 00:42 payload joined 00:43 kthakore joined, preflex joined 01:00 abqar joined
Whiteknight I'm forming an idea in my mind where we can create extremely cheap subclasses when the subclass has all the same VTABLEs as the parent 01:00
and because of the way :vtables and :methods are stored in classes, not the VTABLE struct, this applies to all user-defined classes 01:01
we take things like _class and whoami out of the vtable struct, and create a second, smaller "rider" struct with this info. 01:04
alternatively, no rider struct, every PMC contains a pointer directly to it's class PMC or Proxy PMC
a cheap subclass then can be created by pointing to an existing vtable, but pointing to a new Class object and maybe new Namespace object 01:05
This would allow us to cheaply subclass builtin types, like how PDD23 wants lots of subclasses of exception 01:06
each type has the same vtables, attrs, and methods, but a different class name and maybe a different initialization routine (which can be stored in the Class) 01:08
NotFound Whiteknight: Do you mean to avoid creating the proxy when there is only a parent and that parent is a pmc?
Whiteknight NotFound: Lower than that 01:09
avoid creating a new vtable if it's the same as the parent
this system would probably involve creating more proxies earlier 01:12
but, we're moving in that direction anyway 01:13
Austin What exactly is the point of a pmc proxy?
Whiteknight introspection of c-based types that don't have a class 01:15
two different introspection mechanisms, so two different types
terrible name, however
Austin Huh?
Whiteknight ? 01:16
Austin A class is "blah blah blah all this stuff". Why wouldn't a C based type, compiled into the system, have that stuff?
Whiteknight it does, but a diferent form
because one is written in C, the other in PIR 01:17
Austin Okay. Why a different form? If the point of all these things is to serve as backing mechanisms for objects, shouldn't they be doing that?
Whiteknight I dont claim to know the motivation
Austin Perhaps I should ask, what *is* a pmc proxy?
Whiteknight I would like to move towards unification 01:18
Austin I mean, a class is just a big array of sub pmc's, right?
Whiteknight as an extreme simplification, yes
Austin Here's your foo method, here's your bar method, etc.
So is the proxy building pir-to-c wrappers around the compiled functions? 01:19
NotFound Austin: and support for multiple ineritance and for vtable overrides.
Whiteknight example: an Object stores attributes in a hash. A C-based type stores them in a c struct
Austin, yes 01:20
Austin Sure, but we don't need to know how anything is stored, because we access it by calling one of the forty-eleven different attribute subsystem methods..
getproperty, getattribute, inspect, edit, freeze, absquatulate, defenestrate, whatever.
Whiteknight I've always hated VTABLE_absquatulate 01:22
Austin NotFound: Granted classes participate in MI and Vtable overrides, but PMCs by definition are cast in stone, no?
Whiteknight O
NO
Austin Or are we wanting to insert "parents" behind the pmc types?
Whiteknight that is possible, yes 01:23
Austin Whiteknight: More specific?
Whiteknight actually, I'm wrong 01:24
Austin gasp! unpossible!
Whiteknight You can't change the MRO after you've created an object of the class
Austin Okay, but ... 01:25
NotFound There are confusing statements about that in some docs and comments, but the current way is that, yes adding or removing parents is forbidden after first instantiation.
Whiteknight you can do some duck-typing and method injection
Austin So, since nobody creates very many ResizableBooleanArray objects, I can probably insert a parent class behind it? 01:26
Whiteknight Austin: sure, I guess
NotFound Austin: you can't insert a parent class because isn't a class.
Austin Is that the point of the proxy? 01:27
Whiteknight NotFound, but it has a proxy, with an MRO
at least, it should work
i think
try it
Austin Because I'm pretty sure we could do without *that* feature.
Whiteknight I doubt adding a new parent class on a c-based type ould change the vtables 01:28
NotFound Whiteknight: I'm not sure about what is the current behaviour, either always instantiating a proxy instead of the pmc or not. 01:29
Whiteknight what? 01:31
NotFound new 'Integer' instantiates the Integer PMC, or his proxy?
its 01:32
Whiteknight the Integer 01:33
purl the integer is a big purple monster that will slowly melt my hardware if you find it in my code. If you do find it, delete all references quickly!!!
Whiteknight there is only one proxy per type
Proxy is like a class
again, bad name
NotFound That's the point. The instance is an Integer PMC, or an Object instantiated from the proxy? 01:34
Whiteknight The instance is an Integer PMC
only PIR-based types use Object
NotFound That is the intention, but I'm not sure about the code. 01:35
Coke is there a way to tell svn merge to just postpone everything?
Austin Put a colon in front of the command line? 01:36
Coke Austin: I cannot tell if that a joke or a serious suggestion.
aha. --accept p
Austin Coke: In shell parlance, a colon as a command just eats everything behind it. 01:37
Coke so, a joke then.
nopaste "Austin" at 68.37.46.53 pasted "P6metaclass.register problem? Or Proxy trouble?" (12 lines) at nopaste.snit.ch/19729
Coke svn merge --accept postpone ^/trunk
NotFound There wasn't a plan to rename p6object to some more language agnostic? 01:41
Austin Sure.
It used to be called Perl6Object. 01:42
Now the p stands for pmichaud.
NotFound And the 6? ;)
Austin ta-da!
Probably a counting error - like i19n
except it should have been P7 01:43
atrick ?
That's 6
NotFound I feel much better now. 01:48
Austin NotFound, Whiteknight: When I do get_class 'FileHandle', I get a PMCProxy. If I inspect the proxy, I get a hash. If I try to dump the hash, I get an error.
"Handle cannot be instantiated" etc.
But if I do a get_class 'Handle' there's no problem - I get a proxy, everything's fine, etc.
Any ideas what the difference is?
NotFound No 01:49
Whiteknight Austin: post your code, and I'll check it out locally 01:53
Whiteknight just pulled a large amount of what looks like white pea-gravel from his kitchen faucet 01:54
and now I have like 10x more water pressure
Whiteknight vows to never, ever drink the tap water here
davidfetter where's "here?" 01:57
Whiteknight davidfetter: north-east of philadelphia
not that drinking the water in this area was ever recommended to begin with 01:58
but now apparently there is white gravel in the water pipes
davidfetter white gravel is better than a lot of other things that could be in there 01:59
things that could actually hurt you, e.g.
Whiteknight could be goddamn anthrax for all I know 02:04
or worse yet, some kind of salt, which implies my water is coming from street runoff
dalek kudo/master: 14f803a | (Solomon Foster)++ | src/core/Iterator.pm:
Get rid of GrabAndSay, not spec and it is easily done now with batch($n) and say.
 02:06
bacek_at_work donsbot.wordpress.com/2010/02/21/sm...m-codegen/ 02:10
Looks interesting. 02:11
NotFound whishtles 'In the streets of Philadelphia...'
02:19 JimmyZ joined
dalek tracwiki: v47 | bacek++ | ParrotQuotes 02:24
tracwiki: p6object demystified by Austin and NotFound
tracwiki: trac.parrot.org/parrot/wiki/ParrotQ...ction=diff
02:25 kid51 joined
Coke treads more water in his branch, trying to keep up with the trunk merging. 02:33
dalek rrot: r44285 | coke++ | failed to fetch changeset:
merge latest changes from trunk
 02:55
02:57 hudnix joined
Coke ugh. building rakudo sloooow. 03:01
dalek TT #785 closed by coke++: "hello world" nqp fails when run with installed parrot 03:05
TT #687 closed by coke++: compilers/imcc missing dependencies 03:22
03:24 cotto joined 03:25 jimk joined
dalek rrot: r44286 | coke++ | trunk (2 files):
actually complain about TT #633
 03:28
03:28 theory joined
ttbot Parrot trunk/ r44286 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/207557.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 03:34
Coke does a once-through on trac. 03:39
03:46 TonyC joined 03:58 mikehh joined
Tene Austin_away: rethrow continues walking the original iterator of exception handlers that the exception gathered when it was first thrown. 04:00
Austin_away: *PLEASE* feel free to update pdd23 however you like. I would love to get updates to the exceptions pdd from someone who is confident about opinions about an exceptions system. 04:01
Austin_away: I plan to implement class-based exception filtering as soon as I get an evening free. If you are confident about what types and severities or whatever should exist, I would be very happy to see the PDD reflect that. If you don't want to edit the PDD, just email me whatever recommendations you have, and I'll update the pdd to reflect that. 04:04
Austin_away: I agree with every complaint you listed. Exceptions rework is at the top of my priority list right now.
Coke msg austin be sure to backsroll on austin_away. 04:11
purl Message for austin stored.
04:19 janus joined
Coke if I need to store stuff in a config file format readable by the config system, do we have any already-in-use? json/yml/INI, etc? 04:29
looks like "in perl" is the current favorite. 04:31
dalek rrot: r44287 | coke++ | branches/rm_cflags/config/auto/warnings.pm:
Split out compiler warnings for gcc/g++; should help to add a third compiler...
 04:34
Coke I am leaning towards json, since we'll probably keep json core, since it's so small and easy. Then I just need to see about pulling in a JSON parser for config. (adding more perl5 modules--, but it'd only be ``short term'') 04:40
04:45 patspam joined
ttbot Parrot trunk/ r44288 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/207622.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 04:58
Parrot trunk/ r44289 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/207640.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 05:02
dalek rrot: r44288 | mikehh++ | trunk/runtime/parrot/library/distutils.pir:
fix missing = on back for podchecker
 05:07
rrot: r44289 | mikehh++ | trunk/src/pmc/oplib.pmc:
fix codetest failure - cuddled else
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32322), fulltest) at r44289 - Ubuntu 9.10 amd64 (g++ with --optimize) 05:19
06:18 JimmyZ joined
dalek rrot: r44290 | coke++ | branches/rm_cflags/config/auto/warnings.pm:
move gcc/g++ under 'warnings'
 06:28
dukeleto 'ello
cotto hi 06:30
purl hi, cotto.
JimmyZ hello 06:55
NotFound Coke: my vote is for json 07:47
fperrad ttbot ? 07:48
purl it has been said that ttbot is TapTinder build bot owned by mj41 and reporting tt.ro.vutbr.cz/buildstatus/pr-Parrot/rp-trunk build errors. or a master of timing.
07:51 iblechbot joined 08:33 particle joined 08:54 bacek joined
bacek o hai 09:12
did I missed anything interesting?
ttbot Parrot trunk/ r44291 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/207801.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 09:35
dukeleto bacek: o hai 09:42
bacek: not much
purl same here, dude
dukeleto has his parrot benchmark environment setup on a new server
bacek dukeleto, can it benchmark branches? And have some kind of web interface? 09:44
dalek rrot: r44291 | bacek++ | failed to fetch changeset:
Merge branch 'sys_mem' into trunk
 09:46
dukeleto bacek: that is exactly what I am working on 09:55
bacek++ 09:56
bacek: i am going to start small and make it do one thing well and then branch out
bacek dukeleto, no way. I did nothing apart from crappy suggestions.
Oookey. My dinner is ready. 09:57
dukeleto bacek: the ++ was for merging in the branch :)
bacek :)
dukeleto bacek: deciding which branches are "active" is interesting
i guess looking at the date of the last commit is easy enough
bacek "last commit"
yeah, it is :)
ttbot Parrot trunk/ r44292 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/207877.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 10:13
dalek rrot: r44292 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
[distutils] minor
 10:18
rrot: r44293 | fperrad++ | branches/sys_mem_reduce:
branch has been merged into trunk
bacek fperrad, thanks! (Hope I didn't broke Lua badly) 10:20
10:43 payload joined 10:50 gaz joined 10:55 contingencyplan joined
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32331), fulltest) at r44293 - Ubuntu 9.10 amd64 (gcc with --optimize) 11:42
11:46 cghene joined
mikehh rakudo - make test PASS, make spectest_smolder #32234 - 2 tests FAIL - t/spec/S32-num/rat.rakudo and t/spec/S32-str/uc.rakudo - No subtests run 12:11
rakudo - on parrot at r44293 - Ubuntu 9.10 amd64 (gcc with --optimize)
12:16 payload joined
dalek kudo/master: 7fe9e3e | moritz++ | t/spectest.data:
re-enable test for statement-modifying for-loop, and add a new test file which we have never passed before
 12:17
TT #1473 created by Austin_Hastings++: Handle PMC type is broken 12:26
kapo: 8ffad9d | austin++ | (12 files):
Various patches inspired by working on close.
 12:32
12:34 fperrad joined 12:43 whiteknight joined
whiteknight good morning #parrot 12:44
moritz bacek: I hope you know that you can delete svn branches without a svn check 12:45
dalek rrot: r44294 | bacek++ | branches/boehm_gc_2 (300 files):
Merge branch 'master' into boehm2

  \tcompilers/imcc/imcparser.c
  \tcompilers/imcc/imcparser.h
  \tsrc/gc/api.c
  \tsrc/gc/gc_inf.c
  \tsrc/main.c
  \tt/op/string_mem.t
moritz bacek: svn rm $branch_url works
purl rumour has it src/gc/gc_inf.c is causing codetest and manifest_tests failures
whiteknight purl forget src/gc/gc_inf.c 12:47
purl whiteknight: I forgot src/gc/gc_inf.c
whiteknight that file is PERFECT. you hear me? PERFECT
mikehh whiteknight: check out donsbot.wordpress.com/2010/02/21/sm...m-codegen/ 12:48
whiteknight mikehh: Yeah, I read it last night 12:49
very cool 12:50
he shows a lot of cases where LLVM creates big speedups, but doesn't show any of the cases where it doesn't
so it's hard to really get a comparison because he's using toy examples and showing that they are fast, but then briefly mentions that these numbers aren't typical
mikehh it also seems early days, without using llvm optimize features 12:51
kthakore whiteknight: hi ya 13:03
Austin sings, "I love tornadoes, I love arach-anids! I love hot magma, I love the giant squids..." 13:06
whiteknight boomdiada 13:07
13:13 ruoso joined 13:22 AndyA joined
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32337), fulltest) at r44294 - Ubuntu 9.10 amd64 (g++ with --optimize) 13:34
Coke (deciding which branches are active) see the branchstatus tool. and the wiki page. 13:35
13:38 ruoso joined
Coke will regen the wiki page. 13:40
13:45 tetragon joined
mikehh rm_cflags branch - make corevm/make coretest, make test PASS, couple of codetest failures (fixed in trunk) - at r44294 - Ubuntu 9.10 amd64 (g++ with --optimize) 13:46
lots of extra warnings though
Coke trac-- 13:53
pmichaud 01:41 <NotFound> There wasn't a plan to rename p6object to some more language agnostic? 14:00
No. "P6object" specifically means that we're adding Perl 6 object metamodel properties to Parrot classes.
And it's "P6" instead of "Perl6" because it's not the complete Perl 6 metamodel. 14:01
dalek tracwiki: v29 | coke++ | BranchDescriptions 14:03
tracwiki: rerun script.
tracwiki: trac.parrot.org/parrot/wiki/BranchD...ction=diff
Coke Pe6! 14:04
Coke finds himself missing ClearCase. 14:07
Austin didn't think he'd *ever* hear that... 14:08
dalek rrot: r44295 | coke++ | failed to fetch changeset:
merge latest changes from trunk
 14:09
Coke Austin: I am probably misremembering how much I hate other bits of Clear*, but the merging policy actually made some sense. 14:12
unlike any other tool I use. (git, svn, cvs)
Austin :) 14:13
It wasn't the industry leader for 20 years without some reason...
Coke bah. I very much hesitate to include YA perl5 module in parrot core.
but it's still got to be better than rolling my own JSON.
Austin :) 14:14
What's the purpose, here?
Coke putting configure information into an actual config file instead into raw perl.
(specifically for the CFLAGS replacement)
Austin Okay. What's the purpose of that?
Coke maintainability. readability. 14:15
Austin IMO, readability and json will never be in the same sentence.
And isn't the config info generated by a script?
Coke really? what's your cross-language config file format of choice, and why isn't there a parrot module for it?
Austin: this IS the script.
Austin ?
If you want something readable by parrot, write it in pir. 14:16
Coke *ARGH*
hang on.
purl hmmm... hang on. is this actually "session is still there but user has been deleted" ?
Coke when configure is done, we have a makefile and a few generated files. the makefile will have information about what warnings to use.
Austin Right. 14:17
Coke this information comes from the perl script config/auto/warnings.pm
Austin Okay.
Coke (and on trunk, CFLAGS, but that's going away)
ok. config/auto/warnings.pm has a lot of configuration information, like "this compiler might support these warnings. this other compiler supports these warnings. Use these extra warnings for compiler foo if --cage was given to configure. (it will shortly allow you to say) which files should I not use warning foo on ever? 14:18
that big chunk of information is more maintainable in "not a perl datastructure in a .pm file". that's the config data I'm trying to put somewhere more maintainable. 14:19
Austin But this is definitely *pre* parrot, right? 14:20
Coke today, absolutely.
(I hear of a future I cannot comprehend, where parrot is used to build parrot.) 14:21
Austin (I'm assuming that 'Makefile' implies "parrot not built yet")
Coke which is why I'm talking about having to include a p5 module to process JSON, which I don't want to do.
Austin Yaml?
purl i guess Yaml is not going to work because you need to assign thse two keys to that particular model or comitte or commie or the new SGML or as bad as python when it comes to *exact* indentation or horrific and amazing at the same time or more human readable than xml for some kinds of data or Yet Another Multicolumn Layout or not json for some reason
Austin You're going to have to include a p5 module to do anything, right?
Unless you want to code it by hand in p5, which is just wrong.
Coke Austin: yaml is also not included in core, and at least json has the advantage of having parrot support.
(which yes, does not help us today during the configure process.) 14:22
Austin The "I don't want another p5 module" claim seems moot.
Everything you do leads to another module, except doing the wrongest thing. 14:23
Coke exactly.
Austin So it's down to maintainable.
And YAML is way up there on the maintainable scale.
Coke Austin: I disagree, having followed with some interest ingy's plot to finish yaml on perl. =-)
Austin The module, or the format? 14:24
Coke also, there is no yaml support in parrot, should that ever matter.
Austin: the module.
purl the module is literally what i pasted above, perhaps with a coercion from Num -> BigNum or somewhat useful as a worked example, of course
Coke module(s)
Austin Non-argument. The benefit of modules is they are someone else's responsibility.
And they're done!
(Obviously a prerequisite.) 14:25
Coke my argument-to-the-future is, if yaml is so good, why do why not support it yet in parrot. =-)
(because as the configure process sheds perl, we're going to want to do more and more with a miniparrot during the build... 14:26
Austin You know, I understand the "not available on parrot" argument, but there's a hell of a lot to be said for "this solves today's problem today." The parrot project seems to suffer a lot from not prioritizing that highly.
Coke if I want to solve today's problem today, I just leave the config script in perl. 14:27
done.
Austin (!)
Amen.
Next! 14:28
Coke ... I think you'd be more sympathetic if you spent more time in Configure. ;P
Austin I wanted to, a while back, but I couldn't figure out where it was.
Seriously, the control flow is impossible to grasp.
Coke ah, but it super testable. :P 14:29
*it's
Austin Woohoo
Coke is beaten down, and will just cram all this information into config/auto/warnings.pm
Austin I still have this symlink problem, which I submitted a ticket for. But I'm not smart enough to figure out how to find out where the test is located.
(I'm not even sure if this is a config problem or not.) 14:30
14:36 lucian joined
mikehh rm_cflags branch - post-config test FAIL - t/steps/auto/warnings-01.t - Failed tests: 17, 22, 24 14:36
# Failed test 'runstep() returned true value' 14:37
# at t/steps/auto/warnings-01.t line 72.
# Failed test 'runstep() returned true value'
# at t/steps/auto/warnings-01.t line 88.
# Failed test 'Got expected verbose output'
# at t/steps/auto/warnings-01.t line 90.
14:38 ruoso joined
Coke mikehh: (*^&4~#$ configure tests. 14:39
I'll deal with them before the mergeback. =-)
Austin: yaml's conceit of posting their website IN YAML is cute and horrific.
Austin Coke: Kind of like posting the POD synopsis in POD?
mikehh Coke: I will look in a bit 14:40
Coke mikehh: I'm in the middle of editing that config step anyway.
mikehh 'k
dalek p-rx: 216bacc | pmichaud++ | build/PARROT_REVISION:
Bump PARROT_REVISION.
 14:43
p-rx: 3a51f6c | pmichaud++ | (2 files):
Fix statement_modifier on line after close curly bug.
mikehh rm_cflags branch - manifest_tests, codetest, corevm/coretest, test PASS 14:44
14:51 hudnix joined
nopaste "coke" at 65.91.151.194 pasted "potential data structure for config/auto/warnings.pm" (32 lines) at nopaste.snit.ch/19734 14:54
NotFound Coke: I don't see a big problem: let's use json now to store the info, provide some functions (or objects, or whatever) to handle it, and later we can change the storage format if wanted.
Coke NotFound: where do the functions come from?
(also, feedback on that nopaste welcome.) 14:55
14:55 iblechbot joined
NotFound Coke: we write them, using as implementation detail the modules needed for the underlying format. 14:56
Coke NotFound: meh. more code I have to write. 14:57
14:57 lucian joined
NotFound Coke: that's the bad part. 14:57
nopaste "coke" at 65.91.151.194 pasted "potential data structure for config/auto/warnings.pm" (25 lines) at nopaste.snit.ch/19735 14:58
dalek kudo/master: eacccf9 | masak++ | docs/metamodel.pod:
[docs/metamodel.pod] fixed small typo
Coke NotFound: having the code contain the config is the least painful option for now, especially if I want the config to contain conditionals for andy.
NotFound Coke: anyway, that can be done step by step, start by using directly the module wanted, encapsulate it in some abstract interface later. 14:59
(or never) 15:00
Coke: the intention in that proposal is to handle g++ separately from gcc? 15:01
Coke NotFound: no, we're already doing that. more to allow for specifying exclusions or modifications based on filename.
15:02 whiteknight_ joined
NotFound Coke: ok. Looks good. 15:02
Coke recent versions have split out the gcc vs. g++ handling that was kind of intertwined before. 15:08
(should be possible to add in another compiler now with probes without too much effort.
ttbot Parrot trunk/ r44296 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/208038.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 15:14
dalek rrot: r44296 | whiteknight++ | trunk/src/pmc/socket.pmc:
[pmc] Socket PMC should not override VTABLE_does unnecessarily. Put the role in the pmclass definition
 15:15
15:17 AndyA joined
nopaste "mikehh" at 94.10.35.176 pasted "rakudo patch to get it to build on parrot r44294 with g++ - Ubuntu 9.10 amd64 (g++ with --optimize)" (13 lines) at nopaste.snit.ch/19736 15:19
Austin Why isn't ++ implemented for Undef? Shouldn't that autoconvert to integer? 15:21
Coke mikehh++
Austin: I could see some languages doing that.
dalek kudo/master: 5afabfa | pmichaud++ | src/Perl6/Grammar.pm:
Fix bug with close-curly not terminating statement before apparent
kudo/master: dcc84e6 | pmichaud++ | docs/metamodel.pod:
Merge branch 'master' of git@github.com:rakudo/rakudo
15:22 davidfetter joined 15:24 PacoLinux joined 15:35 ashish joined 15:38 patspam joined 15:41 bubaflub joined 15:44 darbelo joined 15:47 iblechbot joined 15:49 Psyche^ joined 15:53 theory joined 16:00 lucian joined
dalek rrot: r44297 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
[distutils] add a variant of setup (helper for nqp)
 16:04
16:06 fperrad joined
ttbot Parrot trunk/ r44297 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/208093.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 16:07
16:21 payload joined 16:26 Myhrlin joined 16:35 lucian_ joined
dalek tracwiki: v11 | darbelo++ | Modules 16:36
tracwiki: Change the digest-dynpmcs url to point to fperrad's repo.
purl dalek: that doesn't look right
dalek tracwiki: trac.parrot.org/parrot/wiki/Modules...ction=diff
Coke is there a bug where t/op/gc-leaky.t is removed on realclean? 16:47
dalek rrot-plumage: 1ac1c87 | darbelo++ | metadata/digest-dynpmcs.json:
Update digest-dynpmcs metadata to point to fperrad's github repo.
16:51 cotto_work joined
kthakore whiteknight: thanks man! 17:02
whiteknight kthakore: no problem!
I like to put the spotlight on cool new projects
kthakore I like how you think SDL is cool
whiteknight it opens entirely new classes of application to Parrot developers 17:03
Plus, you write the bindings once, and all the HLLs magically get them 17:04
kthakore yup
cotto_work whiteknight, typo in your blog post: s/live on Github/live on Gitorious/ 17:12
whiteknight ...?
purl Yada yada yada hasn't been implemented yet! (unless you run bleadperl)
cotto_work the ParrotProjects post 17:13
whiteknight right, what's the typo?
cotto_work The last word in the disgest-dynpmcs section should be Gitorious, not Github. 17:14
(unless those PMCs are hosted in both places)
whiteknight ah, okay. Fixed. cotto++ 17:15
17:16 lucian joined 17:24 cotto_w0rk joined
whiteknight austin's mini-rant yesterday really got me thinking about Exceptions 17:25
it's funny how you don't think about something when nobody's complaining about it 17:26
17:29 ashish left 17:35 nbrown joined
dukeleto Coke: t/op/gc-leaky.t stays around after I do a realclean 17:39
'ello, by the way 17:41
bubaflub mornin' dukeleto
dukeleto bubaflub: howdy fine sir 17:42
17:42 payload joined 17:50 lucian joined 17:59 plobsing joined 18:09 cotto_w0rk joined
bubaflub hola #parrot - i'm converting some old perl5 tests into PIR, but some tests can't be converted and so the old file still needs to stick around. should i just rename the old test file to whatever_old.t and move the pir tests to whatever.t? 18:23
(and by perl5 tests i mean those under t/*) 18:24
whiteknight bubaflub: Yes
that's exactly perfect 18:25
dalek kudo/master: a3a650e | pmichaud++ | src/ (2 files):
Change lazy signature setup so that it uses block references instead
purl dalek: that doesn't look right
kudo/master: d98b492 | pmichaud++ | src/Perl6/Actions.pm:
Refactor method creation a bit, making sure that

to the underlying Parrot sub, so that the PAST::Block for the sub itself can be nested within the class/package/role main block. This resolves the problem with methods being unable to access lexical and package-scoped variables in class definitions.
bubaflub whiteknight: do i need to worry about svn-properties on the new file?
whiteknight bubaflub: PRobably not. Make sure you add it to MANIFEST though
bubaflub whiteknight: ok. will a make fulltest catch anything i forget to do? 18:26
whiteknight yes
it should, anyway
or mikehh will start yelling at you about it :)
18:27 hercynium joined
bubaflub hahaha. alright. i've got a few patches moving some tests over to PIR. 18:27
besides PASM tests, are there tests i should *not* convert? 18:28
whiteknight bubaflub: if you can convert it and still maintain the integrity of the test, do it
some of the PASM tests test for very specific low-level things that the PIR tests cannot do 18:29
bubaflub whiteknight: ok. i was more thinking along the lines of tests that we are eventually going to throw away.
whiteknight or tests specifically for PASM syntax
bubaflub whiteknight: ok. 18:30
whiteknight no tests really get thrown away. They just get updated to match new syntax and capabilities
ttbot Parrot trunk/ r44298 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/208161.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 18:39
Coke dukeleto: ok. probably a result of merging wierdness. 18:41
bubaflub: yes, unfortunately, you always must worry about the svn properties. 18:42
(or face the wrath of the whoever it is that runs those tests. =-)
18:43 payload joined
bubaflub Coke: doh! I've got a copy of the repo checked out with git svn, should i have a regular ole svn checkout? 18:43
Coke at the end of the day, the repo is svn, and you'll have to deal with it or find someone who is willing to trade your awesome patches for their twiddling of the repository. This should not be a hard bargain to make. =-) 18:45
basically just send mikehh a pint.
bubaflub Coke: i do have some homebrew i could send his way...
dalek rrot: r44298 | plobsing++ | trunk (3 files):
refactor loader generation to predeclare loader function (avoids warnings)
 18:47
dukeleto bubaflub: what are you hacking on these days? 18:48
bubaflub: are you still an eligible student this year? thinking of submitting a parrot project?
bubaflub dukeleto: i've been busy for a bit, but have been working through t/* converting some tests to PIR 18:49
dukeleto: i'm still an eligible student, so i may submit a project
i was looking at whiteknight's suggestion about the string subsystem
and reading through pdd28
whiteknight bubaflub++
dukeleto bubaflub: awesome! there will be quite a few different parrot projects to choose from 18:50
bubaflub i can handle myself with PIR (and know where to look when I can't) but it's been a while since i've done C and i haven't really touched parrot internals
dukeleto whiteknight++ for all the good blog posts recently
bubaflub dukeleto: i was also interested in getting cardinal up and running. the ruby lang people have released an official spec for 1.9 18:51
and at least that would give us something tangible to aim for
but then again, everyone has a Ruby VM these days
dukeleto bubaflub: sounds like a good idea
treed bubaflub: News to me.
Link?
purl or "Link is ... like ... this pointy eared goblin that walks around in midi-music land with a letter opener attacking circles and things and wooing princesses but not bannon, you know?" or preaction is Error.
bubaflub treed: rubyspec.org/
i believe it's a work in progress, but somewhat mature
treed I didn't realize that was official. 18:52
dukeleto bubaflub: treed hacks on cardinal
treed Our goal is official recognition and support from all the organizations currently involved in Ruby implementations"
Also, Cardinal is nowhere near being able to even bootstrap rubyspec. :-(
bubaflub treed: ah. so it's unofficial. that's strange.
dukeleto treed: can you point bubaflub in the right direction? he wants to help hack on cardinal
bubaflub treed: summer of code or not, i'd help out
treed I've been trying to work on the object model.
bubaflub: Cool! 18:53
I think once I get the object model working, cardinal can accelerate very quickly.
But it's tricky right now because you can't subclass parrot;Class.
So I have this funky facade misdirection thing going on. 18:54
bubaflub: The standard "Work on this for now" thing is adding methods to the stdlib classes like Array.
Class methods don't work on master, though, AFAIK.
(And I can't verify whether they work on iss41 until I fix the current bug in my way.) 18:55
whiteknight treed: why can't you subclass clas?
class*
treed whiteknight: Crashes, IIRC.
Ask Tene, he actually tested it.
I just assumed from the start that it wouldn't work and went my way.
Because I've had so many problems with subclassing in the past.
Tene asked why I wasn't subclassing parrot;Class and then tested it.
(Proving my point.) 18:56
dukeleto treed: do we have tests in parrot core that show what the subclassing bug is?
treed I think Tene finally fixed the Exception issue recently.
whiteknight treed: well, I feel like that's something that should work or else Parrot doesn't have a suitable object model
treed dukeleto: Doubt it.
whiteknight: I agree. 18:57
whiteknight treed: so we need to make a ticket and fix it, if a ticket doesn't exist already
I would be very interested to see a failing test case
treed I'll ask Tene to see if he filed a ticket.
And if not to either do it or dig up his PIR test to file one.
18:57 payload joined
treed It was a pretty simple PIR case. 18:57
bubaflub treed: either way, i can look at some array methods to add. do you want me to just send a pull request via github when i have something?
treed bubaflub: Sounds good. 18:58
purl Sounds good. is there a good way for me to find out when branches are merged, other than read every svn commit?
dukeleto treed: i can help with writing the test, if needed 19:01
19:01 AndyA joined
Tene treed: I haven't had time to work on subclassable exceptions yet. I have some stuff locally that hasn' tbeen committed, and should be pretty simple to commit, but I haven't had a free evening in a while. 19:02
treed: Also, I'm looking forward to feedback on the Exceptions API from Austin. 19:03
dalek rrot: r44299 | whiteknight++ | trunk (2 files):
[TT #968] remove last remnants of PObj_active_destroy_* which were supposed to be removed as part of this ticket
 19:04
ttbot Parrot trunk/ r44299 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/208209.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 19:05
treed Tene: Was the parrot;Class subclassing thing ever filed? 19:07
Tene treed: I have no idea. 19:08
19:08 theory joined
treed Tene: If you didn't do it, then I doubt it. 19:08
Tene: I'm looking back over your gists, and not finding the test case. 19:10
I see one kinda looks like it's related.
gist.github.com/251807
Tene treed: I've occasionally posted gists when logged in as cardinal
miiiiiiiiiiiight check there
that's subclassing Integer, not Class 19:11
treed Yeah.
19:11 fperrad joined
treed Checked cardinal, don't see anything there 19:11
Might have been a private gist? 19:12
Tene no, but it might have been when I wasn't logged in.
ttbot Parrot trunk/ r44300 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/208255.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 19:15
whiteknight Tene: if we can get a ticket opened and a test case, I'll explore it 19:16
treed checked the #cardinal logs for gist.github.com
didn't find it
remember where we talked about it?
19:18 AndyA joined
Tene just write a test case yourself. subclass Class, instantiate it, then instantiate an instance of your instance. 19:20
dalek rrot: r44300 | plobsing++ | trunk/tools/dev/nci_thunk_gen.pir:
ignore CR as well as LF as whitespace in nci signature files. This should make windows explode less.
Tene dawg.
bubaflub i heard you like instances 19:21
treed If we got that fixed, I think my object model work would be much smoother.
It takes a large amount of concentration to be able to hold the multiple levels of indirection in my head. 19:22
I end up rereading things a lot while I work on it.
19:23 AndyA joined
plobsing do we have a good way of knowing what the file separator is from PIR yet? 19:24
darbelo whiteknight: ping
whiteknight darbelo: pong 19:25
Coke plobsing: yes. if you absolutely must use native slash and not just '/', you can peek in Config's 'slash'.
perldoc runtime/parrot/library/config.pir
darbelo whiteknight: digest-dynpmcs isn't available on gitorious, fperrad maintains them on github. 19:26
plobsing Coke: I need it. Make on windows is converting '/'s into '\\'s which is giving "interesting" failures
Coke plobsing: ... I turned that off in the rm_cflags branch.
it's useless.
whiteknight darbelo: yeah, I saw that changed. I wrote that post a few days ago
Coke bubaflub++ #xhibit reference. 19:27
darbelo whiteknight: Just saw a mention on the irclog going the other way around. Wanted to let you know.
I *do* have dbm-dynpmcs (actually only gdbmhash for now) on gitorious, if you want to pimp that one ;) 19:28
whiteknight darbelo: sure! Link?
darbelo gitorious.org/dbm-dynpmcs/ 19:29
Just copied over the pmc, haven't translated the tests yet. It's in not-yet-embryo state, but it'll work as well a trunk does. 19:30
That reminds me... 19:31
Coke plobsing: so, don't go crazy trying to work around it. 19:32
darbelo purl: msg bubbaflub I saw in the irclog you were looking to convert tests to PIR. You'll be my personal hero if you convert t/dynpmc/gdbmhash.t ;) 19:33
purl Sorry, I've never seen bubbaflub before.
Coke it's bubaflub.
darbelo purl: msg bubaflub I saw in the irclog you were looking to convert tests to PIR. You'll be my personal hero if you convert t/dynpmc/gdbmhash.t ;)
purl Message for bubaflub stored.
bubaflub darbelo: i have a patch that converts it to PIR, but it gets some strange results. 19:34
whiteknight treed: ping
bubaflub darbelo: i can send you the patch and see what you can make of it
darbelo bubaflub++
Dude, that's some *impressive* response time. 19:35
bubaflub darbelo: lemme verify that real quick.
darbelo: yeah, i tried converting this a while back but i couldn't figure out why the tests were failing. i'll gist the patch for ya to try out. 19:36
darbelo: perhaps you'll have to patch my patch
yo dawg. heard you like patches.
dalek TT #1474 created by coke++: runtime/parrot/library/config.pir refers to long dead "None" PMC.
darbelo can patch patches and diff diffs.
19:37 cotto_w0rk joined
dalek rrot: r44301 | fperrad++ | trunk/tools/dev/nci_thunk_gen.pir:
[build] fix nci_thunk_gen.pir on Windows : pathname with \\
 19:37
rrot: r44302 | fperrad++ | trunk/config/gen/makefiles/root.in:
partial revert of r44291
darbelo I can probably diff patche and patch diffs too, if I put my mind into it ;)
bubaflub darbelo: i probably did something boneheaded, so all bonus points go to you if you can fix this 19:38
darbelo: gist.github.com/311400
19:40 payload joined
darbelo bubaflub: Could you put up the finished file as well? 19:41
bubaflub oh, sure
dalek nie: r109 | allisonrandal++ | wiki/Roadmap.wiki:
Edited wiki page through web user interface.
 19:42
darbelo patch insists on rejecting all hunks.
bubaflub darbelo: gist.github.com/311403
20:01 cotto_work joined
dalek rrot: r44303 | plobsing++ | trunk/tools/dev/nci_thunk_gen.pir:
use configslash? to determine native file separator (in case it is neither '\\\\' or '/')
 20:11
nie: r110 | allisonrandal++ | trunk/Lib/test/bootstrap/ (2 files):
Adding tests for deleting items from lists and dictionaries. Modified
 20:17
rrot: r44304 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
[distutils] fix cc_run on Windows
 20:27
20:34 cotto_work joined
bubaflub darbelo: i've got class tonight till about 10 p.m. CST but feel free to msg or email me about gdbmhash stuff. 20:38
i'm also taking requests for converting other tests as well
20:40 fperrad_ joined 20:41 bacek joined 20:43 joeri joined
darbelo msg bubaflub You had extra newlines on a few strings. All tests pass now. 20:44
purl Message for bubaflub stored.
dalek m-dynpmcs: 242cc5d | darbelo++ | setup.pir:
Update setup.pir to take advange of new distutils features. Thanks to fperrad for adding them.
 20:46
m-dynpmcs: d9ab7bd | darbelo++ | t/gdbmash.t:
Add PIR tests for gdbmhash. Converted by bubaflub++ from the original perl tests.
m-dynpmcs: 1c1699a | darbelo++ | t/old/gdbmhash.t:
Remove old perl-based tests.
bubaflub darbelo: arg, that's frustrating. glad you found it, though. 20:49
Coke ... False ~~ True is correct? whoa.
I'm sure that won't be at all confusing. =-)
darbelo You were also not deleting a temp file after use. That was being done in perl, but it didn't get translated into PIR. 20:50
bubaflub darbelo: good catch. 20:51
20:52 bubaflub left
Coke note: just changing the vim: line is probably not sufficient. =-) 20:55
the mode: cperl needs to be updated also.
(I'm guessing there's a distro test that for.) 20:56
... for that.
darbelo Isn't mode cperl used for PIR too?
Coke ... really? I would highly doubt it.
but I could be wrong.
darbelo Checking... 20:57
Coke IANAEN.
darbelo Ah, there's a mode: pir
Coke++
Coke if that didn't fail t/codingstd/pir_code_coda.t, that's a bug. 20:58
dukeleto++ #rapid benchmark response. 21:00
dalek rrot: r44305 | darbelo++ | trunk/t/dynpmc/gdbmhash.t:
Convert gdbmhash tests to PIR.
21:00 ash_ joined
Austin Coke: If False ~~ True, Perl6 can replace Ada as the gov't standard programming language. 21:02
Coke ... someone proof this: 21:07
# test files (.t) files might need testing.
# ignore everything else.
return 0 unless $filename !~ /\\.t$/;
moritz unless !~ is double negated, no? 21:08
Coke so it's returning false IF it's a .t; yes? 21:09
moritz aye
we don't reject no double negations, nay!
Coke whee. wond3er how much will start reporting failed when I fix this. =-) 21:10
moritz :-) 21:11
treed return 0 if $filename ~~ /\\.t$/;
No?
purl Maybe.
moritz =~ instead of ~~ if you want to support perl 5.8
darbelo And you do.
treed I mean, it's the same as
remove both negations and it should do the same thing. 21:12
and be more readable
Coke treed: no, we're KEEPING .t files.
treed ?
bacek morning...
treed Is what I said the same thing as what you said earlier?
Coke treed: just see the latest commit.
treed: yes, both of which are wrong. =-)
treed Yeah. 21:13
I was saying "If you want to be sure about it, just remove the negation."
And then you know "Oh, yes, it's doing the wrong thing."
Coke mikehh: hey, yah, sorry about killing the codingstd tests. =-)
Coke wonders moving perlcritic tests into a completely separate make target. =-) 21:15
*ponders
darbelo make critic
Sounds good to me.
But I don't have perlcritic installed, so it doesn't really do much of a difference ;) 21:16
dalek rrot: r44306 | darbelo++ | trunk/t/dynpmc/gdbmhash.t:
Put the right mode in the PIR coda.
rrot: r44307 | coke++ | trunk/lib/Parrot/Distribution.pm:
Fix the sense of this check .

to start failing various codingstd tests.
Coke gets 5 failing codingstd tests. (some with 100s of files) 21:18
darbelo looks like mikehh is going to have lots of fun ;) 21:20
mikehh working on it now :-} 21:21
Coke hey, we screwed up the test the same way on perl files. 21:26
fixed.
mikehh a lot of fun fixing that lot :-} - gonna take a while though 21:27
Coke whoof. perlcritic is going to be a *(#$. =-) 21:28
mikehh wonders if he can automate some of it
Coke I'll mail the list with a heads up.
Austin sings, "The captain wired in: he had water coming in, and the good ship and crew were imperiled..." 21:29
Coke wondes how long that's been broken. 21:31
nopaste "coke" at 65.91.151.194 pasted "codetest failures." (29 lines) at nopaste.snit.ch/19738 21:32
dalek rrot: r44308 | coke++ | trunk/lib/Parrot/Distribution.pm:
This check was also broken for perl files.
 21:33
21:34 hercynium joined
darbelo Coke: Now that you mention broken, checkout nopaste.snit.ch/19739 21:34
Coke darbelo: that's on rm_cflags? 21:36
darbelo Yes, with sun cc.
Coke ... you shouldn't even be trying to compile anything in that step at this point. 21:37
oh.
moment.
darbelo I'm not compiling anything. "cc: illegal option -Wno-unused-parameter" 21:38
I just included the cc list of flags in case you needed it. 21:39
Coke darbelo: I fixed that, I think.
(and now the build will fail because some things specifically tack on a gcc-specific -W flag.) 21:40
darbelo Coke: yes, I did a svn up and Configure is fail-free, but the build is still borked. 21:42
Coke yup. 21:43
21:43 Andy joined
Coke darbelo: I don't see anything in that flags list like gcc's warnings options. 21:44
dalek rrot: r44309 | bacek++ | branches/boehm_gc_2/src/gc/api.c:
Remove finalize_gc_system from mandatory methods
 21:49
rrot: r44310 | bacek++ | branches/boehm_gc_2/src/gc/api.c:
Remove destroy_child_interp from mandatory GC methods.
rrot: r44311 | bacek++ | branches/boehm_gc_2/compilers/imcc/pbc.c:
Use zeroed allocations for constants
rrot: r44312 | bacek++ | branches/boehm_gc_2/src/gc/api.c:
Remove more methods from mandatory GC API
rrot: r44313 | bacek++ | branches/boehm_gc_2/src/gc (2 files):
Update GC Boehm to recent API
rrot: r44314 | bacek++ | branches/boehm_gc_2/src/packfile.c:
Use zeroed allocations in PackFiles
rrot: r44315 | coke++ | branches/rm_cflags/config/auto/warnings.pm:
must return success even though we didn't do anything.
rrot: r44316 | mikehh++ | trunk/t/library/test_class.t:
Copyright, Id and coda
rrot: r44317 | mikehh++ | trunk/t/oo/methods.t:
fix codetest failure - Copyright format
rrot: r44318 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
[distutils] cc_run : fix link with mingw
darbelo Coke: There's -v (warn more), -w (don't warn) and -erroff=list (enable or disable warnings acording to list) 21:51
Coke ah, danke. 21:52
darbelo You'll probably have to play games with -erroff=%tag,no%othertag,%yetanothertag and the like.
putting %tag on the list disables that warning, putting no%tag on the list enables. And for bonus torture points: Order matters. 21:54
Coke can you have multiple erroffs?
darbelo Coke: Yes. 21:58
But you have to watch out for ordering issues. 21:59
Coke that's fine.
darbelo The list gets traversed left-to-right and all options are applied in order. 22:00
So, no%sometag,%all disables everything. But %all,no%sometag disable everything except sometag. 22:01
Coke that's fine - the config step allows for that.
(keeping things in order).
(mostly. will get hairy for enable/disable on individual files.) 22:02
heading out. 22:03
darbelo msg Coke this might come in hady if you want to please Sun cc (docs.sun.com/source/806-3567/cc_options.html) 22:06
purl Message for coke stored.
dalek nxed: r424 | julian.notfound++ | trunk/winxedst1.winxed:
string indexing in stage 1
rrot: r44319 | bacek++ | trunk/src/gc/gc_ms.c:
Move definitions of mem_sys_foo functions inside GC MS
rrot: r44320 | bacek++ | trunk/t/library/test_class.t:
Fix stange line in t/library/test_class.t
darbelo msg Coke In particular, you might want to consider something like this: docs.sun.com/source/806-3567/sun.html#52600 22:07
purl Message for coke stored.
dalek nie: r111 | allisonrandal++ | trunk/Objects:
Add a directory for pynie objects.
 22:08
22:23 AndyA joined
dalek rrot: r44321 | mikehh++ | trunk/t (5 files):
fix codetest failures - cuddled else
 22:23
rrot: r44322 | mikehh++ | trunk/t/codingstd/cuddled_else.t:
fix codetest failure - cuddled else
rrot: r44323 | fperrad++ | trunk/runtime/parrot/library/Configure/genfile.pir:
[gengile] don't replace slash on source file (only on Makefile)
rrot: r44324 | mikehh++ | trunk/t/pmc (3 files):
fix codetest failures - missing or incorrect Id
kudo/master: b9b10eb | pmichaud++ | build/Makefile.in:
Add some dependencies to the t/*/*.t targets.
kapo: 6dcb666 | austin++ | (8 files):
Got a basic Mock generator working.
 22:26
nie: r112 | allisonrandal++ | trunk/ (9 files):
Move objects into a python-standard directory structure.
 22:27
22:29 Whiteknight joined
dalek rrot: r44325 | mikehh++ | trunk/t/library (2 files):
fix codetest failures - missing Id
 22:39
rrot: r44326 | mikehh++ | trunk/t/codingstd/c_macro_args.t:
fix codetest failures - incorrect Id
rrot: r44327 | mikehh++ | trunk/t/compilers/pge (2 files):
fix codetest failures - missing Id
tracwiki: v12 | fperrad++ | Modules 22:43
tracwiki: add parrot-MT19937
tracwiki: trac.parrot.org/parrot/wiki/Modules...ction=diff
22:44 fperrad joined 22:48 fperrad_ joined 22:53 samlh joined 22:54 cotto_work joined 23:03 fperrad_ joined, samlh left
japhb approves of the last three songs on his Pandora station: Iron Man, Pour Some Sugar On Me, and Highway to Hell 23:10
sorry, ww
dalek rrot: r44328 | mikehh++ | trunk/compilers/ncigen/t (5 files):
fix codetest failures - missing Id, plus Copyright and Coda
 23:12
tracwiki: v19 | allison++ | JITRewrite 23:16
tracwiki: trac.parrot.org/parrot/wiki/JITRewr...ction=diff
23:18 payload joined
cotto_work LIR? 23:19
LIR is LLVM intermediate representation
LIR?
purl i heard LIR was LLVM intermediate representation
plobsing can someone who knows about mmd comment on TT #1039? it appears to be fixed. 23:21
23:24 payload joined
cotto_work plobsing, is there a test for it? 23:25
plobsing there's a file in the ticket that dtrt (but was apparently failing before) 23:27
cotto_work Yeah. Once there's a test, the ticket can be closed.
cotto can take care of it if nobody beats him to it. ;) 23:28
plobsing I'll do it, I'm just fairly ignorant of multi-dispatch, so I wasn't sure if the whole issue was resolved
darbelo We should adopt 'add test before closing ticket' policy. 23:29
dalek rrot: r44329 | mikehh++ | trunk/t/codingstd (3 files):
fix codetest failures - fix Copyright
rrot: r44330 | mikehh++ | trunk/t/native_pbc (3 files):
fix codetest failures - fix Copyright
cotto_work That's from before the pcc refactor. It's reasonable that the refactors would have fixed it.
rrot: r44331 | mikehh++ | trunk/t (5 files):
fix codetest failures - fix Copyright plus some trailing whitespace
Whiteknight any tests need a'writin' tonight? I have a handful of free tuits 23:30
cotto_work Whiteknight, trac.parrot.org/parrot/ticket/1039 (though it's hardly more than a copy/paste)
unless plobsing was going to do that first
23:32 kid51 joined
Whiteknight TT #1039 passes now? There's no comment there 23:35
cotto_work plobsing only seems to have noticed recently
I guess he's the only one to have looked since the pcc refactors. 23:36
Whiteknight plobsing++
A good ticket triage could make for a nice hackathon
everybody take 20 tickets and see what the status is 23:37
plobsing Whiteknight: more like (whoever fixed that)++
cotto_work As long as it's well-coordinated, that wouldn't be a bad idea.
A PiratePad doc would work well for that. 23:38
Just get a list of tickets, break them into groups of 5-10 and let people claim them by editing the doc. 23:39
dalek rrot: r44332 | plobsing++ | trunk/t/pmc/fixedpmcarray.t:
add test script from TT #1039 (already passing)
 23:45
rrot: r44333 | mikehh++ | trunk/t/pmc/packfileannotation.t:
fix a test that got corrupted in editing
TT #1039 closed by plobsing++: MMD bug in FixedPMCArray.sort 23:46
TT #1475 created by plobsing++: [Pmc2c] make METHOD and MULTI consistent in terms of how PCC is done
plobsing wtf how did that get assigned to cotto right away? 23:47
cotto_work ?
beats me 23:48
kid51 hackathon? Did whiteknight say hackathon? 23:53
kthakore kid51: Yes at TPM for SDL 23:55
right Whiteknight ?
plobsing purl, msg bacek are you still interested in TT #903? I don't see anything in DEPRECATED.pod. 23:58
purl Message for bacek stored.

Powered by Cro and the Raku® Programming Language.

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

Thank you!