Topic for #parrot is: Parrot 3.3.0 released | parrot.org | Log: irclog.perlgeek.de/parrot/today
Set by moderator on 20 April 2011.
dalek p/ctmo: 4e2bed3 | jonathan++ | src/ (2 files):
Remove magical loading of the NQP meta-objects library, and just do it with a use statement in the setting.
 00:02
p/ctmo: 8147a0f | jonathan++ | src/stage0/ (6 files):
Update bootstrap, just to be sure the previous works out there too.
00:29 kid51_at_dinner is now known as kid51
cotto ~~ 01:12
01:12 dmalcolm left 01:25 theory left 01:33 whiteknight left 01:36 petdance joined 01:39 bubaflub left 01:57 kid51 left 02:02 mtk left 02:07 mtk joined, dngor left 02:12 dngor joined 02:33 ShaneC joined
dalek parrot: 52b996b | leto++ | META.json:
Add a skeleton pgxn meta.json file
 02:39
parrot: 5f4d360 | leto++ | META.json:
meta.json wants a semantic version number, so give it one
 02:50
parrot: 65e3360 | leto++ | META.json:
Add repo and bugtracker to meta.json
02:51 bubaflub joined
dalek parrot: c612728 | leto++ | META.json:
Add more details to meta.json
 03:05
03:11 S_Arms left, bubaflub left 03:26 Drossel joined 03:27 Kulag left 04:30 Eduardow left
dalek rrot: 84a6b3b | petdance++ | src/ops/var.ops:
consted a few variables
 04:50
rrot: 74e2932 | petdance++ | src/ops/set.ops:
consted a var
rrot: f99d63c | petdance++ | src/ops/object.ops:
removed an unnecessary setting of a var
rrot: 968fcad | petdance++ | src/ops/math.ops:
consted a var
rrot: 8f7ec49 | petdance++ | src/dynoplibs/ (2 files):
consted a var
rrot: 9f7969e | petdance++ | src/ops/core.ops:
consted a pointer. Undid some unnecessary var setting
05:06 birdwindupbird joined 05:28 petdance left 05:33 hudnix left 05:51 jrt4 joined 05:55 jrtayloriv left 06:16 mj41 joined 06:28 benabik joined
dalek parrot: 3c2246a | leto++ | Makefile:
Get rid of svn-- revision checking junk
 06:32
06:49 birdwindupbird left 07:09 birdwindupbird joined 07:47 contingencyplan left 08:07 ShaneC left 08:17 jrt4__ joined 08:21 jrt4 left 08:26 jrt4 joined 08:29 jrt4__ left 08:33 jrt4__ joined 08:37 jrt4 left 08:40 jrt4 joined 08:42 ShaneC1 joined 08:44 jrt4__ left 08:49 jrt4__ joined 08:53 jrt4 left 08:57 jrt4 joined 09:01 jrt4__ left 09:04 jrt4__ joined 09:08 jrt4 left 09:11 jrt4 joined 09:14 SHODAN joined 09:15 jrt4__ left 09:18 jrt4__ joined 09:22 jrt4 left, Drossel left 09:23 Kulag joined 09:27 Kulag left, Kulag joined 09:30 dodathome joined 09:34 Kulag left, Kulag joined 09:39 ShaneC1 left 09:44 ShaneC joined 09:46 ShaneC left 09:48 mtk left 09:51 Kulag left, Kulag joined 09:55 mtk joined 10:27 birdwindupbird left 10:33 birdwindupbird joined 10:34 Kulag left 10:35 Kulag joined 10:39 Drossel joined 10:41 Kulag left 10:42 alin joined 11:00 JimmyZ joined 11:02 lucian joined 11:06 lucian_ joined 11:10 lucian left
tadzik where does the select pmc live? 11:19
wasn't it merged? 11:22
ah, not 11:27
tadzik sad panda
cotto_work: ping 11:30
11:35 bacek left 11:41 JimmyZ left 11:45 lucian_ left 11:46 bacek joined 11:57 Patterner left, Psyche^ joined, Psyche^ is now known as Patterner 12:04 JimmyZ joined 12:34 birdwindupbird left 12:43 mj41 left 12:52 bubaflub joined 13:02 hudnix joined 13:20 Coke left, cosimo left, Coke joined 13:24 woosley joined 13:31 mrwall-e joined 13:33 contingencyplan joined 13:35 alin left, lucian joined 13:47 rohit_nsit08 joined
rohit_nsit08 hello #parrot 13:47
tadzik hello rohit_nsit08 13:49
rohit_nsit08 tadzik: hi, good morning
13:54 whiteknight joined
dalek sella: d8321a6 | Whiteknight++ | src/harness/TestFile.winxed:
swap_handles is a utility function, not a method
 13:54
sella: 2de0727 | Whiteknight++ | src/harness/ (3 files):
start adding debug output to test harness
whiteknight good morning, #parrot 13:55
tadzik good morning whiteknight
rohit_nsit08 whiteknight: hello 13:56
whiteknight hello tadzik, rohit_nsit08 13:57
13:59 hercynium joined
rohit_nsit08 whiteknight: Got surprised by how JavaScript get its dynamic nature :-) check the post rohitnsit08.blogspot.com/2011/04/ja...rnals.html 14:00
tadzik rohit_nsit08: "findout out more about it." Typo? 14:02
rohit_nsit08 tadzik: ya, the main was how execution contexts and scope chain works
the code execution can only run inside execution contexts and transfer of control with pushing earlier context on stack 14:03
whiteknight rubini.us/2011/02/23/introduction-to-fancy/ 14:07
rohit_nsit08: I'll read that post in a bit
rohit_nsit08 whitknight: okay 14:08
14:09 mrwall-e_ joined, mrwall-e left, mrwall-e_ is now known as mrwall-e 14:12 mrwall-e left, mrwall-e joined 14:14 jrt4 joined
arnsholt rohit_nsit08: Which font are you using for bodytext on your blog, OOC? 14:14
rohit_nsit08 arnsholt: hi, it's google font : Philosopher 14:15
14:18 jrt4__ left 14:19 mrwall-e left 14:20 mrwall-e joined, petdance joined 14:24 jrt4__ joined 14:28 jrt4 left
lucian rohit_nsit08: i think you may misunderstand how 'this' works in js 14:29
rohit_nsit08 lucian: 'this' simply points to the container object , isn't it? 14:30
lucian rohit_nsit08: usually, yes. but the containing object isn't always clear
closures are a case where 'this' often gets fucked up
event callbacks, etc 14:31
rohit_nsit08 lucian: ya that's true, closures happen when there are inner function declarations.
lucian: my guess is that in that case also , 'this' should point to the object containing that method only 14:32
lucian afaik, 'this' has an odd dynamic scoping, not lexical
due to its implicit nature
atrodo i thought in js, this was set to the object the function was called on (or with apply)
rohit_nsit08 atrodo: agree 14:33
lucian atrodo: there are corner cases where it seems that doesn't happen
bubaflub rohit_nsit08: as a side note, IE sometimes fails to handle "this" correctly
atrodo lucian: for instance?
lucian although it does, in fact
atrodo: let me see if i can find that article 14:34
rohit_nsit08 IE has its own Js specification, so can't comment on it.
'this' has some problem with closures, I'll have to test it more i guess 14:35
14:35 jrt4 joined
rohit_nsit08 I'm testing my js codes on nodejs 14:35
so they should work on chrome
lucian atrodo: hmm, jashkenas.github.com/coffee-script/ search for 'function binding'
not quite what i was looking for 14:36
rohit_nsit08: yeah, just ignore IE
closures in JS get lexically scoped, for the most part
except for this, which is dynamically scoped
atrodo lucian: right, a function call changes this
lucian rohit_nsit08: www.digital-web.com/articles/scope_...avascript/ might be helpful 14:37
atrodo: i think where it looks very odd is with nested closures
if you define a callback inside a closure that itself was a callback 14:38
14:38 jrt4__ left
atrodo lucian: right, a closure is a function. Calling that function changes this from what it was when you defined the function 14:38
lucian atrodo: yes, and you would like to get the 'this' at the beginning of the closure chain 14:40
you can do it explicitly with .apply, which is what CoffeeScript does with => 14:41
dalek rrot-linear-algebra: ab2c7b6 | Whiteknight++ | t/harness:
make sure to run tests inline
rrot-linear-algebra: 0340153 | Whiteknight++ | t/ (2 files):
remove debugging garbage from harness. fix typo in set_block.t
rrot-linear-algebra: a4e1696 | Whiteknight++ | t/methods/complexmatrix2d/convert_to_complex_matrix.t:
fix another typo. All tests run again (but most fail)
rrot-linear-algebra: b9a3697 | Whiteknight++ | t/harness:
fix typo in harness. Start reclaiming tests
rrot-linear-algebra: bf705ea | Whiteknight++ | t/ (10 files):
update several assertions and other calls to the new architecture. Reclaim a bunch more tests
rrot-linear-algebra: 32125af | Whiteknight++ | t/ (14 files):
fix references to test factory, a few other fixes. Reclaim several other tests
rrot-linear-algebra: 1a76233 | Whiteknight++ | t/testlib/methods/gemm.nqp:
fix gemm tests. methods are all tests, so utilities have to be subs
rrot-linear-algebra: aa5c604 | Whiteknight++ | t/testlib/methods/set_block.nqp:
fix setblock tests
rrot-linear-algebra: 8a6dca3 | Whiteknight++ | t/methods/ (45 files):
fix all remaing tests. no more failures
whiteknight note to self: Huge rewrites and refactors to your test library means huge rewrites to all test suites which use it 14:42
atrodo whiteknight: note to self, progress is work
atrodo hates progress some days
lucian atrodo: this is also ugly jibbering.com/faq/notes/closures/#clAc 14:43
whiteknight the changes are all necessary, and I'm much more happy now with Rosella.Test than I used to be
atrodo And those are the days i love progress. We have a love/hate relationship
whiteknight of course, I have several dozen test files in PLA which basically need identical updates, but which aren't amenable to a quick sedscript 14:44
14:44 jrt4__ joined
cotto_work ~~ 14:44
whiteknight good morning cotto_work
cotto_work hi whiteknight
14:45 petdance left
whiteknight cotto_work: did you get my email last night? 14:45
cotto_work whiteknight: I did. The part about making the concurrency proxies RO sounded a lot like the ro variant vtables we have. 14:46
whiteknight cotto_work: yes, very similar. The big draw of the proxy is that we can intercept calls and make intelligent decisions about them. Also, open them up for subclassing and hllmapping 14:47
14:47 jrt4 left
cotto_work Yeah. It's more fine-grained. 14:47
whiteknight A similar mechanism could be used to cut out ro_variant vtables entirely, which are a huge waste of space
an RO proxy object would be extremely easy to create (Rosella does it now in like 30 lines of winxed code) 14:48
cotto_work Overall it struck me as sensible. I'd like to take another hour and give it a second more thorough read.
whiteknight it's a first draft. I expect it to be more the start of the conversation than the end of it
14:50 mrwall-e left 14:55 rohitnsit08 joined
whiteknight what I want to avoid is getting stuck in a trap of providing too much or too little. We give people sane, safe defaults and let them get into trouble if they want 14:57
If somebody says "I want to be able to modify data across threads with no wait and no latency, but I'm going to provide my own implementation of locks" whatever, that's a personal decision
Parrot can warn you loudly, but we shouldn't insert a bunch of costly logic to prevent problems 14:58
14:58 rohit_nsit08 left
whiteknight one thing I like about the proxy idea is that the proxy can contain a reference to the owner thread of the target PMC, so every PMC in the system doesn't need to be saddled with ownership information 14:58
So long as you work through the proxy, everything is safe. If you insist on talking to the target directly, it's on you 14:59
15:03 rohitnsit08 left 15:07 rohitnsit08 joined 15:10 preflex joined
tadzik cotto_work: ping 15:11
cotto_work tadzik: pong 15:14
15:19 BloodyCoin joined
cotto_work tadzik: you pinged? 15:23
15:24 theory joined 15:26 woosley left
moritz \\o 15:27
I just noticed that rakudo's perl6.pbc is 11MB, and just 1.3 if gzip-compressed 15:28
what makes those PBC so huge?
11004336 / 1288994
aloha 8.53715067719477
moritz compressed by a factor 8.5
15:32 rdesfo joined
PerlJam run strings on it and see if there is a lot of textual data? 15:32
moritz $ strings perl6.pbc|wc -c 15:33
247949
247949 / 11004336
aloha 0.0225319365021206
moritz 2.3%
aloha 0.023
moritz not very significant 15:34
tadzik cotto_work: aye
whiteknight parrot de-duplicates strings internally, I think
tadzik cotto_work: #2034, you mentioned you want to review the Select PMC once again. Got some spare time to do that? 15:35
PerlJam moritz: I only get a factor of 4.8 compression on my system. Maybe you've got double wide alignment compared to me :-) 15:36
moritz PerlJam: do you have a 32bit system? 15:37
cotto_work tadzik: possibly. I might be slow to respond though. 15:38
PerlJam I don't think this system is 32bit 15:39
moritz PerlJam: if it's linuxoid, uname -a will tell you
15:47 ambs joined, rdesfo left 15:48 whiteknight left 15:52 simcop2387_ joined 15:54 simcop2387 left, simcop2387_ is now known as simcop2387
JimmyZ moritz: perl6.pbc is 5.34MB here, on windows 7 15:55
15:58 whiteknight joined
tadzik 11 MBs, 64bit linux 15:58
JimmyZ 32 bit windows 7 :) 16:00
16:01 petdance joined 16:03 lucian_ joined
moritz I just sent an email to parrot-dev, and it bounced with "Recipient address rejected: User unknown in virtual alias table 16:06
" 16:07
16:07 petdance left
moritz oh 16:07
I sent to @parrot instead of @lists.parrot
16:08 lucian left
whiteknight about the bytecode size issue? 16:08
dukeleto whiteknight: ping
whiteknight pong 16:09
moritz whiteknight: yes 16:10
16:30 petdance joined
cotto_work tadzik: will you be around tomorrow? 16:35
16:35 lucian_ is now known as lucian
dukeleto ~~ 16:38
whiteknight moritz: I haven't gotten that email yet, I don't think 16:41
16:44 petdance left
moritz whiteknight: I haven't resent it 16:46
whiteknight: I've said everything I have to say here on #parrot too
16:49 JimmyZ left
whiteknight ok 16:50
16:56 mrwall-e joined 16:58 ambs left
whiteknight now that the biggest portion of the IMCC refactors are done, we can start seriously looking at the serialization changes needed by rakudo 17:00
the imcc_compreg_pmc and then the packfile_wrap branches were prerequisites for that work
now we can pass packfiles around as PMCs throughout most of the system instead of structure pointers, manage them with GC, and link between them 17:01
rohitnsit08: I read your blog post. It all looks very good 17:04
tadzik cotto_work: yeah, probably 17:05
cotto_work: I also merged master into tewk/select
cotto_work ok
I'm afk for the rest of the day. 17:06
whiteknight++ for the release streamlining thoughts. If I weren't going out the door, I'd love to comment. 17:17
17:22 lucian_ joined
whiteknight there's always more time to talk! 17:25
17:26 lucian left 17:27 jrt4__ left 17:28 jrt4__ joined 17:35 mtk left, jrt4 joined
lucian_ whiteknight: i didn't realised you adhered to this so closely www.total-knowledge.com/~ilya/mips/ugt.html 17:37
17:37 rohitnsit08 left
lucian_ realises that whiteknight follows standards better 17:37
17:38 lucian_ is now known as lucian
whiteknight lucian: I like it 17:38
:)
17:39 jrt4__ left 17:41 rdesfo joined, mtk joined 17:42 mtk left 17:48 jrt4__ joined 17:50 rdesfo left 17:52 jrt4 left 17:53 rohitnsit08 joined 18:06 ShaneC joined
whiteknight msg soh_cah_toa Check out this link: blog.llvm.org/2011/04/regular-expre...mands.html 18:41
aloha OK. I'll deliver the message.
whiteknight msg soh_cah_toa nothing you need to worry about for your project, but still fun to think about :) 18:43
aloha OK. I'll deliver the message.
19:04 rohitnsit08 left 19:05 bbatha joined 19:06 rohitnsit08 joined 19:11 bbatha left 19:17 rohitnsit08 left 19:22 rohitnsit08 joined 19:31 whiteknight left 19:45 rdesfo joined
plobsing ~~ 19:45
19:51 rdesfo left 19:53 Coke left, Coke joined 20:00 ambs joined 20:03 Coke left, Coke joined 20:07 perlite_ joined 20:10 perlite left, perlite_ is now known as perlite 20:18 dodathome left 20:21 Coke left, Coke joined 20:27 Coke left 20:31 AzureSto_ left 20:33 Coke joined 20:34 rohitnsit08 left 20:36 AzureStone joined 20:43 rohitnsit08 joined 20:46 bbatha joined
dalek Heuristic branch merge: pushed 90 commits to parrot/tt1931-nci-parameters-deprecation by plobsing 20:48
20:50 Coke left, Coke joined 21:04 bbatha left 21:19 bbatha joined 21:22 jrt4 joined, bbatha left 21:25 jrt4__ left 21:30 jrt4__ joined, ambs left 21:34 jrt4 left 21:36 rohitnsit08 left 21:38 jrt4 joined 21:41 jrt4__ left 21:42 bacek left 21:55 soh_cah_toa joined 21:56 jrt4 left 21:57 jrtayloriv joined 22:07 hercynium left 22:17 bacek joined 23:16 lucian left 23:28 rdesfo joined, jrt4 joined 23:31 bbatha joined 23:32 jrtayloriv left 23:40 rdesfo left 23:41 jrt4__ joined 23:44 jrt4 left 23:54 mikehh left

Powered by Cro and the Raku® Programming Language.

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

Thank you!