Parrot 3.7.0 "Wanda" | parrot.org | Log: irclog.perlgeek.de/parrot/today | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 25 August 2011.
dalek rrot: 4757aba | cotto++ | src/dynpmc/Defines.in:
don't build the Select dynpmc on window for now
 00:05
ttbot Parrot 4757abad i386-linux-thread-multi make error tt.taptinder.org/cmdinfo/45589 00:06
kid51 cotto_work: Please re-try tt_2185 branch on Win32, with some options to Configure.pl 00:09
cotto_work kid51: doing so now 00:18
benabik cotto_work: Your commit appears to break the build. 00:23
cotto_work kid51_at_dinner: win32 fix works for me 00:31
benabik: sad face
benabik cotto_work: If I move the comment to before the DYNPMC_TARGETS= line, it works. 00:32
cotto_work benabik: pushed 00:33
benabik++
00:33 soh_cah_toa joined
dalek rrot: 46f0466 | cotto++ | src/dynpmc/Defines.in:
fix build
 00:33
cotto_work kid51_at_dinner: any objections to merging that thing? 00:34
benabik Hm. select.t appears to leave README2 behind in the root. That seems poor. 00:35
cotto_work lta
droppings-- 00:36
I'm having second thoughts about letting the select branch into master. 00:40
not building on win32 is a significant black mark 00:44
00:48 Coke joined 01:07 woosley joined
kid51 Coke ping 01:52
cotto_work kid51: you mind if I merge the relcheck branch? 01:56
tt_2185
kid51 I don't care, but since Coke both created and took the ticket, I was trying to catch him 01:57
But I don't see that anything more needs to be done in branch 01:58
cotto_work good enough for me 01:59
dalek rrot: 448bcca | cotto++ | / (4 files):
Merge branch 'tt_2185'
 02:00
cotto_work pushed and deleted
coke++ and kid51++ for working on it 02:02
dalek TT #2185 closed by cotto++: "nmake release" fails on windows
TT #2185: trac.parrot.org/parrot/ticket/2185
cotto_work trac-- for not letting me modify that ticket, then taking me to another page from which I also couldn't modify the ticket.
02:23 PacoLinux joined
dalek rrot/all-hll-test: 6dc901f | cotto++ | tools/dev/all_hll_test.pl:
add a mostly hard-coded script to build and test Rakudo and nqp
 02:28
Coke cotto_work: if you tested that target, then that's fine with me, cloes away. 02:46
02:55 PacoLinux joined
cotto ~~ 03:09
cotto, I did, so good. 03:10
05:00 nbrown joined
cotto (our build)-- 05:23
cotto just spend 30 minutes chasing down the fact that he needed to run pmcrenumber after removing a PMC.
05:56 particle joined 06:08 cybergggirl joined 06:16 simcop2387 joined 06:21 he_ joined, preflex_ joined 06:26 jsut joined 06:51 dngor joined 06:52 mj41 joined
cotto dukeleto, ping 07:08
dukeleto, unping 07:13
07:24 dngor joined 07:38 jsut_ joined 07:49 nbrown joined 08:05 lucian joined
mikehh cotto: ping 08:11
08:50 fperrad joined 08:55 rfw joined
cotto mikehh, pong 09:00
09:08 woosley left
cotto mikehh, sleeping now. 09:29
'night
10:07 ambs joined 10:22 rfw joined 11:47 ligne joined
atrodo =~ 11:48
12:07 whiteknight joined
whiteknight good morning, #parrot 12:08
weird. For once I don't have any messages
tadzik good afternoon whiteknight
whiteknight hello tadzik
12:24 jsut joined 12:36 preflex joined 12:46 bluescreen joined
bubaflub ~ 12:49
12:53 woosley joined
whiteknight good morning, bubaflub 12:54
bubaflub morning whiteknight
any update on the housing situation? 12:55
whiteknight bubaflub: We put in an offer on a new house on monday. On tuesday morning we heard that the seller had too many emotional attachments to the house and was going to keep it as a rental property 12:56
bubaflub whiteknight: aghjkdhsagjasdhlgjhasdlgh
whiteknight so we're up to 7 houses that we've put bids in on, but lost
bubaflub that's frustrating
whiteknight yessir. "Frustrating" is the word I would use 12:57
13:05 coke joined
kthakore bubaflub: you pingd? 13:05
bubaflub kthakore: i did?
COKE_WORK seen dukeleto?
aloha dukeleto was last seen in #parrot 14 hours 14 mins ago saying "msg Coke I am getting more spam about your gsoc final eval not being done".
COKE_WORK seen whiteknight?
aloha whiteknight was last seen in #parrot 8 mins 27 seconds ago saying "yessir. "Frustrating" is the word I would use".
whiteknight I've seen him. Handsome fella
kthakore bubaflub: aghjkdhsagjasdhlgjha 'SDL'
gh 13:06
whiteknight Good morning COKE_WORK. how are you today?
COKE_WORK meh.
kthakore bubaflub: I am like a genie. You said the magic word. What do you want?
COKE_WORK still need to fill out the GSOC paperwork. I have ... just under 5 hours.
whiteknight COKE_WORK: at least you should be far enough from the hurricane that your weekend won't suck too badly
COKE_WORK yay. we have family drama that is hurricane related, but everyone should be fine. 13:09
er, "yah"
dalek nxed: 665f3a2 | NotFound++ | winxedst1.winxed:
tiny refactor of NewExpr classes
atrodo I've never known too many people to yell with their nick 13:10
whiteknight atrodo: Caps lock is cruise control for awesome
atrodo whiteknight++ that is so true
ligne the build currently seems to be failing with g++. is this a known problem? 13:19
src/pmc/filehandle.c:1200:37: error: ‘Parrot_io_async’ was not declared in this scope
13:20 nbrown joined
whiteknight do we have a Parrot_io_async? 13:28
13:40 jsut_ joined
ligne there's one in src/platform/generic/io.c 13:44
and a windows one in src/platform/win32/io.c
it's not in any headers though
whiteknight that might be a new addition from the Select PMC 13:54
14:32 logie joined
COKE_WORK needs a beer. 14:48
COKE_WORK will fix this when beer o'clock rolls around.
atrodo really, isn't always beer o'clock? 14:50
plus, is very disappointed that COKE_WORK doesn't yell all the time
whiteknight in real life, he isn't so aggressive 14:51
COKE_WORK be. aggressive. be. be. aggressive.
Hey, can anyone here draw tiny flags? I need about a dozen of them, and will beer you. 14:52
mikehh Hey parroteers: I had to install Windows 7 (64 bit) for a project I am working on, What do you need to test parrot on it? 14:59
and where do you get it?
cotto mikehh, trac.parrot.org/parrot/wiki/Platforms/Windows 15:00
was that what you pinged me about?
mikehh cotto: Yeah 15:01
cotto there you go
mikehh cotto: I couldn't find that - I looked 15:03
cotto: I have a TechNet install and Active State Perl - will that work? 15:04
cotto technet? 15:06
I don't know about activestate perl either way. Can you document what happens?
mikehh I got a licence from TechNet for Windows 7 Ultimate (64 bit) and Office 15:07
as part of the project I am working on 15:08
cotto mikehh, I'm sorry to hear that. I'm glad you're making the most of it. :)
mikehh Hopefully I can find some time to work on parrot in the next week or so (been seriously busy with the project) 15:09
Have had to run some Windows only programs for it - MathCad for instance (don't really like it but it is required on the project) 15:11
15:44 dmalcolm joined 15:57 nbrown joined 15:59 dodathome joined 16:04 COKE-WORK joined
COKE-WORK yawns. 16:44
16:47 contingencyplan joined 16:52 ambs joined
COKE-WORK MANAMANAH 16:53
DO DOOO DO DOO
MANAMANAH
COTTO OHAI
whiteknight why is everybody yelling? 16:56
atrodo CAPS LOCK IS CRUISE CONTROL FOR AWESOME 16:57
sorry, my bad
ATRODO CAPS LOCK IS CRUISE CONTROL FOR AWESOME
cotto THAT'S BETTER 16:58
COKE-WORK I'd like to get tt_2184 merged today. 17:03
I think NotFound++ fixed the last crow issue.
whiteknight what's that ticket about?
COKE-WORK killing ChangeL^WNEWS
whiteknight awesome 17:04
cotto +1
COKE-WORK one less thing to futz with.
cotto hands coke a match and some lighter fluid
use them wisely
17:05 AzureStone joined
whiteknight less things to futz with is more awesome 17:07
COKE-WORK hopes all the parotteers on the east coast will take care this weekend. 17:08
cotto removing thread stuff gets me a pretty segfaulty parrot
whiteknight yeah, it's going to take some work. There's a lot more to remove than you would expect 17:09
COKE-WORK: yeah, I've already gotten earthquaked. I'm not pumped about getting hurricaned 17:10
cotto yeah. I started a new branch to remove the gunk in a more piecemeal fashion, but even that's less easy than I'd expect.
COKE-WORK we have mentors slotted for the GSOC mentor summit, yes? 17:12
(that is to say: I ain't going?)
cotto dukeleto and I are on the list afaik 17:13
I've got my hotel room booked
COKE-WORK full up then, hokay. be awesome.
cotto prepares for awesome-being
dalek kudo/nom: 2742f6d | moritz++ | src/binder/multidispatch.c:
fix memory leaks, patch by mls++
 17:17
COKE-WORK ooh, that's nice.
dukeleto COKE-WORK: WHAT IS WITH ALL THE CAPS? 17:32
whiteknight they have to speak up over the ambient server noise 17:34
COKE-WORK this fan in my office is really loud. 17:38
atrodo no matter how many lowercases COKE-WORK uses, i think he's yelling
COKE-WORK you'll be rid of me soon enough. 17:39
moritz note that this isn't the big memory leak (I fear) 17:40
COKE-WORK every bit helps. 17:41
I'm happy that was a rakudo leak and not a parrot one! ;) 17:42
whiteknight don't worry, there will be plenty of new leaks when we merge 6model into parrot core 17:43
let's enjoy this calm before the storm 17:44
actually, if I'm barracaded in my house this weekend, maybe I'll start work on that project
rip out threads, prototype some sandboxing stuff, and start merging in 6model 17:45
moritz whiteknight: do you plan to do it like in nqp, where 6model coexists with the current parrot thingy?
atrodo write a new gc? maybe write a new threading model?
moritz and then gradually rewrite PMCs to be 6modelish? 17:46
whiteknight moritz: At first it will be in parallel, just like NQP
I want to do it the same way as NQP, so the transition to get NQP running on it instead of on its own version will be smoother 17:47
internally we will start rewriting Object, Class, and PMCProxy to start using 6model instead of the current garbage
then, the rest of the PMC types. We can do it piecemeal
cotto whiteknight, do you think there'd be any benefit in letting nine rip out the old threading code since he has some interest in making it shinier? 17:48
atrodo cotto> I would say yes. Ripping things out can really help in understanding where the little tendrils go
whiteknight cotto: In theory it's a good idea, but it's pretty well wedged in there. We have to make fixes to src/interp/*, src/exception.c, src/scheduler.c, etc 17:49
COKE-WORK in a branch, please. ;)
right, but if he has tuits, let him at it.
whiteknight okay, if nine wants to do it, I'll definitely be available to answer questions and double-check work
I don't feel like removing the current threading system is going to be too educational about what Parrot is all about or where we want to go 17:50
but, i'm nto going to tell others how to spend their tuits
cotto Sure. We'd help as much as he wants.
whiteknight I can't even adequately manage my own, with all the projects I keep volunteering for
cotto nine, ping
I'm not saying we should force him. Just give him the option. 17:51
not sure how we'd "force" him anyway. ;) 17:53
COKE-WORK ponders writing one time perl script to hand roll SQL for a CF app. 17:54
atrodo All you need is an address and a sack of oranges
COKE-WORK hopes that wasn't directed at him! ;)
cotto a sack of oranges sounds great, depending on how it's used 17:55
whiteknight but then the oranges get all smooshy and sticky 17:56
whiteknight is suddenly jonesing for an orange
atrodo mission accomplished
cotto aloha, atrodo?
aloha cotto: Search me, bub. 17:57
whiteknight aloha: whiteknight?
aloha whiteknight: whiteknight is The Count
cotto aloha, atrodo is a secret employee of the orange lobby
aloha cotto: Okay.
whiteknight ah ah ah
atrodo :)
whiteknight thats TWO factoids in aloha that are false! ah ah ah
I've been watching way too much sesame street 18:01
cotto happens to everyone
whiteknight my kid has this adorable shirt that says "Everything I know, I learned on the street", and there's a picture of Bert with sunglasses
my kid is totally gangster 18:02
18:03 whiteknight joined
dalek p: 0bfcef6 | moritz++ | src/6model/multi_dispatch.c:
fix a multi dispatch memory leak, patch by mls++
 18:06
whiteknight I would really really really like to fix Parrot's MMD system in the near future 18:08
seeing other projects with decent mmd always makes me jealous 18:09
moritz I guess parrot initially tried to provide enough semantics for Perl 6 MMD 18:14
and then failed, but still did too much, and did much at call time 18:15
whiteknight right 18:17
people kept jamming new functionality into places that couldn't handle it
now, we have the packfile loader building multis at packfile load time, and the MultiSub is just an array that does almost nothing related to being a sub or doing dispatch 18:18
All the sorting is done by the MMD system which doesn't keep state so there is no caching done. Every time we dispatch we need to re-sort the entire candidates list
And we have Subs which keep track of their own multisignature, so a Sub can't have more than one, and subclasses of MultiSub can't use different storage semantics 18:19
basically, every single part of Parrot MMD that can be done wrong, was done wrong
oh, and string-based class names in multi signatures, so we can't dispatch based on things like anonymous classes, or classes with names that have changed since the multi was created, or classes that don't stringify very well, etc 18:20
and we can't do some even very basic things, like dispatching by role, or dispatching based on data contents 18:21
atrodo wow 18:24
whiteknight Imagine adding some basic caching to multisub dispatches. Better yet, add in some PIC and do site caching with type inference 18:25
Oh, we also can't dispatch to subs based on the types of named parameters, based on optional parameters, with slurpy parameter lists, etc 18:26
so all those cool features that PCC has suddenly can't be intelligently used if we invoke the wrong kind of Sub 18:27
One thing the system does have going for it is a textbook implementation of a manhattan sort, even if some of the heuristics used are faulty 18:28
atrodo manhattan sort? 18:29
dukeleto atrodo: manhattan distance
atrodo dukeleto> thanks
18:33 soh_cah_toa joined 18:52 woosley left 18:59 soh_cah_toa joined 19:11 Eclesia joined
Eclesia hi 19:11
whiteknight hello Eclesia 19:14
Eclesia: The eria page on sourceforge can't be found 19:15
Eclesia whiteknight: you mean using a search engine ? 19:16
whiteknight sourceforge.net/projects/eria/
Eria shows up on the sourceforge search, but I go to the page and it doesn't work 19:17
Eclesia whiteknight: there : sourceforge.net/p/eria/
whiteknight oh, much better 19:18
Eclesia whiteknight: there are not much things useable in the project. I worked more on the module system then the code parser 19:19
19:20 jsut joined
Eclesia I was searching for a solution for fixed size number. and bubaflub give the solution ^^ s I'll be able to improve the code parser 19:20
whiteknight oh, nice 19:21
I'm playing with some of your reflection ideas in a new Rosella library
Eclesia whiteknight: ha? added some interesting thing ? 19:26
whiteknight a few things I'm playing with: github.com/Whiteknight/Rosella/tre...le/reflect 19:27
I'm going to start adding tools for building and modifying classes too 19:29
Eclesia whiteknight: ho, you mean if I create a new Rosella.Refect.Class, properly configure it, I might be able to compile it ? 19:31
whiteknight probably something else. I don't know what the interface will be
var builder = new Rosella.Reflect.ClassBuilder("ClassName"); 19:32
var class_obj = builder.build()
something like that
Eclesia looks like my project :p
whiteknight ah, I knew i was getting a good idea from somewhere
I won't try to do it if you're doing it already
19:33 rohit_nsit08 joined
rohit_nsit08 hello #parrot 19:34
Eclesia if you check the project, you'll see classes for all types, class, reference, declaration, instanciation, .... all objects to describe a class
that's the 'paradygm' module
whiteknight hello rohit_nsit08
I hadn't seen all this stuff before 19:36
COKE-WORK whiteknight: is there a documented plan of attack on the 6model stuff?
something I could start attacking in a branch?
rohit_nsit08 whiteknight: hello . I got an e-mail regarding failure of my project , so I wanted a feedback for carrying on my work further as I think I can finish it after some better understanding
19:37 preflex joined
whiteknight COKE-WORK: honestly, no. Certainly not what I am sure you would want to see 19:37
COKE-WORK whiteknight: hokay.
whiteknight COKE-WORK: the first stage of the plan is to get the source from the NQP repo moved into the parrot repo, and get it to build and work the same way
rohit_nsit08: okay. We do want the project still 19:38
19:38 dmalcolm_ joined
whiteknight rohit_nsit08: I've been playing with jison and cafe the last two days, and have some ideas about how to proceed 19:38
COKE-WORK rohit_nsit08: best of luck. Glad to see you in channel.
whiteknight COKE-WORK: once we have it working the same way as it does in NQP, we can start integrating it better into parrot core, and start changing other stuff like Object.pmc and Class.pmc to use it internally 19:39
COKE-WORK: The first step is pretty monolithic, because it compiles or it doesn't. I'm sure I'm going to need somebody to double-check my makefile work and the like 19:40
COKE-WORK I can do makefile scutwork. 19:41
moritz can occasionally compile with 'make -j' and jell if it's broken 19:42
that's a surprisingly good test at times :-)
rohit_nsit08 COKE-WORK: thanks. sorry to ask but was that the final review or something still can be done about that? because I really want to get this project finish it is only this recruitment process that is keeping me busy .
whiteknight rohit_nsit08: that was the final evaluation for the summer and the deadline is passed. We can still work on the code, but the grading is over 19:43
rohit_nsit08 whiteknight: hmm.. my bad :-(
19:44 dmalcolm__ joined 19:46 bluescreen joined
rohit_nsit08 COKE-WORK: can we have a discussion tomorrow regarding the points of failure so that I can know the areas I need to focus from now ? 19:46
whiteknight rohit_nsit08: I can have a meeting with you, if COKE-WORK is busy 19:48
rohit_nsit08 whiteknight: ya sure, that will also do .
than I think we can have it right now, if you are not busy 19:49
whiteknight I have 30 minutes, if that's enough 19:52
and do you want to talk in channel, or private message? 19:53
tcurtis ~~
whiteknight hello tcurtis
tcurtis hello, whiteknight 19:54
whiteknight tcurtis: I still need to get that walkthrough from you about lalrskate. You around tonight or this weekend?
19:58 preflex joined
Eclesia question : knowing the path "some.path.to.var" how can I get the pmc related ? 20:07
whiteknight Eclesia: what do you mean?
Eclesia whiteknight: I set a pmc at a given namespace position using : set_global 20:08
know I would like to acces it
whiteknight $P0 = get_global ns, "name"
Eclesia any winxed shortcut ? 20:09
whiteknight $P0 = get_global ["some"; "path"; "to"], "var"
I don't think so. How did you set it?
Eclesia in pir :D
tcurtis whiteknight: I may be able to talk tonight; I will definitely be able to do so at some point this weekend. 20:10
whiteknight tcurtis: okay. Now that final evaluations are out, I don't have any more leverage over you :) Whenever you're free I would appreciate it
I'll probably be trapped indoors for much of the weekend 20:11
COKE-WORK coworker just suggested getting cash, gas, and water. 20:12
20:13 mj41 joined
soh_cah_toa good. plenty of time to work on a new debug pmc ;) 20:13
whiteknight soh_cah_toa: I was worried for a bit that you weren't going to do that final evaluation
don't scare us like that! 20:14
soh_cah_toa :)
i got distracted
whiteknight no more distractions. You're almost a full-fledged parrot contributor now! All you have left to do is survive the trials 20:28
you have to pick up a construction cone with your legs, without using your hands, and walk it through an obstacle course of my own design
cotto I remember having to do that. 20:32
Eclesia hm .. I think I'll pass 20:33
soh_cah_toa :) 20:35
NotFound Eclesia: with the latest revision, you can use the namespace operator and keyed acces in it. 21:24
Soemthing like: (namespace ["some","path","to"])["var"] 21:25
Eclesia NotFound: and that return a class/function/pmc ? (depends what he find at this place) 21:26
NotFound Eclesia: that depends on the namespace PMC, not on winxed. 21:27
I think vars are prioritary,
Eclesia ok 21:28
NotFound: changing topic : did you added whiteknight commit on concatenate namespace declaration ? 21:29
to avoid : namespace n1 { namespace n2 { ...
NotFound Eclesia: yes, and you can do the same with class, too. 21:30
Eclesia Not the same with class ? 21:31
21:31 not_gerd joined
not_gerd hello, #parrot 21:31
NotFound Eclesia: class Foo.Bar is a shortcut for. namespace Foo { class Bar 21:37
Eclesia NotFound: hm... it might be a bit confusing. like for example : namespace NS1 { class NS2.User . what will be the class namespace ? NS1.NS2. User or NS2.User ? 21:39
NotFound: what is the result ? I suspect it's the first but not sure 21:45
NotFound The same as the long version, nested into the enclosing namespace. 21:46
I must add a test that verifies and exemplifies it. 21:47
21:56 tcurtis_ joined 21:59 Coke joined 22:21 rfw joined 22:29 Eclesia left, eternaleye joined 22:40 kid51 joined
kid51 ~~ 22:40
Are GSOC results publicly available? 22:42
We should publish a summary on www.parrot.org 22:43
23:34 Kulag joined 23:40 Drossel joined 23:55 Tene 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!