Camelia, the Perl 6 bug

IRC log for #parrot, 2011-09-14

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:18 cotto whiteknight, better to fix parrot, but that's much harder
00:19 cotto or so I'm told.  It's one of those ideas I've inherited without testing.
00:21 slavorg joined #parrot
01:08 benabik joined #parrot
01:28 bluescreen joined #parrot
01:31 jsut_ joined #parrot
01:35 dalek Rosella/harness_test: 35f5d7e | Whiteknight++ | / (4 files):
01:35 dalek Rosella/harness_test: Add in some global accessors for Harness to get factories for TestExecutor, TapParser, and FileResult objects. The user can override these factories to provide different instances. Rearrange Harness a little bit to take advantage and make TestExecutor more testable
01:35 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/35f5d7e866
01:35 dalek Rosella/harness_test: 03dbe25 | Whiteknight++ | s (2 files):
01:35 dalek Rosella/harness_test: Fix setup.winxed to exit(1) when there is an error. +some comments and small cleanups in Harness
01:35 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/03dbe2527b
01:35 dalek Rosella/harness_test: c45a149 | Whiteknight++ | / (2 files):
01:35 dalek Rosella/harness_test: +tests for the new Harness accessor methods
01:35 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/c45a149912
01:35 dalek Rosella/harness_test: ec3b15d | Whiteknight++ | / (2 files):
01:35 dalek Rosella/harness_test: Stub tests for TestExecutor
01:35 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/ec3b15d569
01:35 dalek Rosella/harness_test: e2a8dc5 | Whiteknight++ | / (8 files):
01:35 dalek Rosella/harness_test: Several Harness test fixes
01:35 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/e2a8dc5309
01:35 dalek Rosella/harness_test: 09c731e | Whiteknight++ | src/harness/TapParser.winxed:
01:35 dalek Rosella/harness_test: Change TapParser to be more intelligent about whitespace, which fixes error reporting.
01:35 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/09c731e672
01:35 dalek Rosella/harness_test: 68c099b | Whiteknight++ | / (3 files):
01:35 dalek Rosella/harness_test: Trim out newlines from the tap stream. Fix parsing of TODO to be more robust.
01:35 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/68c099baa6
01:47 dalek Rosella/harness_test: ca77256 | Whiteknight++ | t/tap_harness/Harness/TestRun/Factory.t:
01:47 dalek Rosella/harness_test: +tests for TestRun.Factory
01:47 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/ca7725653b
01:56 Coke joined #parrot
01:57 bubaflub joined #parrot
01:59 woosley joined #parrot
02:11 Coke joined #parrot
02:26 Coke joined #parrot
03:15 Coke joined #parrot
05:07 dalek rakudo/nom: 9d6d7dd | moritz++ | src/Perl6/Metamodel/ (2 files):
05:07 dalek rakudo/nom: MRO optimizations by mls++
05:07 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/9d6d7dd582
05:11 dalek parrot/win64-workaround: ff18696 | cotto++ | config/auto/sizes.pm:
05:11 dalek parrot/win64-workaround: less incorrect fix (plus comment) to force long long intval on msvc/x64
05:11 dalek parrot/win64-workaround: review: https://github.com/parrot/parrot/commit/ff186967a9
05:11 dalek parrot/win64-workaround: 107f638 | cotto++ | / (2 files):
05:11 dalek parrot/win64-workaround: Revert "first try to force sizeof(void*) == sizeof(INTVAL) on 64-bit windows"
05:11 dalek parrot/win64-workaround:
05:11 dalek parrot/win64-workaround: This reverts commit 0950a0066a7ff1de911104c0553f840385314ee8.
05:11 dalek parrot/win64-workaround: review: https://github.com/parrot/parrot/commit/107f638b3b
05:11 cotto aloha, clock?
05:11 aloha cotto: LAX: Tue, 22:11 PDT / CHI: Wed, 00:11 CDT / NYC: Wed, 01:11 EDT / UTC: Wed, 05:11 UTC / LON: Wed, 06:11 BST / BER: Wed, 07:11 CEST / TOK: Wed, 14:11 JST / SYD: Wed, 15:11 EST
05:21 SHODAN joined #parrot
05:30 dalek misc: 7f22440 | cotto++ | timeline.md:
05:30 dalek misc: Merge remote-tracking branch 'wk-timeline-gist/master'
05:30 dalek misc: review: https://github.com/parrot/misc/commit/7f2244091b
05:31 dalek misc: 37bb054 | cotto++ | timeline.md:
05:31 dalek misc: Merge remote-tracking branch 'pm-timeline-gist/master'
05:32 dalek misc: review: https://github.com/parrot/misc/commit/37bb054f09
05:34 cotto git is fun!
05:42 rfw joined #parrot
06:15 davidfetter joined #parrot
07:01 UltraDM joined #parrot
07:26 mj41 joined #parrot
07:37 kthakore joined #parrot
07:43 jtpalmer joined #parrot
08:14 jsut joined #parrot
08:31 dalek rakudo/nom: f2857b7 | (Timothy Totten)++ | src/core/IO/Socket/INET.pm:
08:31 dalek rakudo/nom: IO::Socket::INET.accepts now returns the connected socket
08:31 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f2857b7fc9
09:11 TiMBuS joined #parrot
09:12 dodathome joined #parrot
09:17 tcurtis joined #parrot
09:27 Coke joined #parrot
09:36 dalek nqp: 8ba6001 | moritz++ | src/how/NQPClassHOW.pm:
09:36 dalek nqp: MRO calculation optimization by mls++
09:36 dalek nqp: review: https://github.com/perl6/nqp/commit/8ba60015ce
09:50 Tene dukeleto: I'm no longer on-call at work tomorrow, so I might be able to finally work on Parrot tomorrow night.
10:17 ambs joined #parrot
11:18 mj41 joined #parrot
11:20 felixdo joined #parrot
11:35 Psyche^ joined #parrot
12:22 whiteknight joined #parrot
12:23 whiteknight good morning, #parrot
12:39 atrodo =~
12:40 benabik joined #parrot
12:51 honza joined #parrot
12:51 benabik o/
12:52 alvis joined #parrot
12:53 alvis joined #parrot
13:11 mtk joined #parrot
13:14 honza Hi, I am going to make an interpretter for a Smalltalk-like language. Is there any possibility to create bytecode file with dynamically created objects in runtime with Parrot?
13:16 moritz you can freeze objects, store them in .pbc files, and then thaw them
13:16 moritz is that what you want to know?
13:17 honza can i use one pbc file as an image?
13:21 moritz it seems that the freeze opcode creates a string
13:21 moritz so you could embed the result of a freeze as a constant
13:22 moritz though note that parrot's freeze/thaw mechanism isn't all that general (at least not general enough to serialize Perl 6 classes :/ )
13:22 honza what's missing?
13:22 benabik moritz: I've been wondering about that?  Is that due to 6model or just freeze/thaw being insufficently clever?
13:23 moritz last I looked, we needed some way to freeze recursive structures, but stop at certain boundaries
13:24 whiteknight plobsing added backreferences a while back. I don't know if that covers all use-cases
13:24 honza I need it too, but whole tree of objects.
13:25 whiteknight honza: We can create bytecode and insert arbitrary constants into it. I don't think we have any users stress-testing the features so I'm sure there will be rough edges
13:25 benabik moritz: Rakudo needs to freeze _most_ of a tree?
13:25 benabik s/tree/graph/
13:25 whiteknight the bytecode system is one that we are planning a lot of work on, so feedback and feature requests would be very good
13:26 moritz benabik: correct
13:26 moritz benabik: suppose you write class Foo is Int { }
13:27 moritz benabik: class Foo needs to be serialized, and all its dependencies, except those which are already in the setting
13:27 benabik Oh.  Need to free Foo and connect it with the main Int when loaded.  Fun.
13:27 benabik *freeze
13:30 whiteknight I think that's the last part we are missing. We do have the backreferences, but we don't have the ability to partially freeze objects like that and reassemble
13:31 arnsholt IIRC jnthn will start working on that Real Soon Now for NQP
13:31 whiteknight I would really much rather have it done directly in Parrot, if possible
13:31 moritz http://6guts.wordpress.com/2011/09/13​/whats-coming-up-in-septemberoctober/
13:32 benabik_ joined #parrot
13:33 honza moritz: thank you for the information. Please, can you send me some link with example of freezeing object and thawing them?
13:34 moritz honza: t/pmc/freeze.t in the parrot repo
13:35 moritz https://github.com/parrot/parr​ot/blob/master/t/pmc/freeze.t
13:46 honza moritz: my beginner question - can be this code executed in a pir file?
13:47 moritz honza: that code is a perl script that contains lots of little runnable PIR scripts
13:48 moritz each of which can be run as a PIR file
13:49 honza moritz: thank you :) I will test it.
14:15 alvis_ joined #parrot
14:15 alvis left #parrot
14:29 contingencyplan joined #parrot
14:31 dalek Rosella/gh-pages: 9049485 | Whiteknight++ | libraries/ (2 files):
14:31 dalek Rosella/gh-pages: Updates to Harness and Query docs. Add new classes to Harness and some updated descriptions. Add information about Streams to the Query docs. Some of these later docs aren't 100% accurate because Streams need a refactor.
14:31 dalek Rosella/gh-pages: review: https://github.com/Whiteknig​ht/Rosella/commit/9049485677
14:33 benabik docs++
14:33 benabik whiteknight++ # documentation? that's crazy talk
14:33 benabik ;-)
14:46 ambs joined #parrot
15:08 dukeleto ~~
15:18 honza Where can I find assigment PMC -> library, for example File -> io_ops, please?
15:19 moritz honza: it's called 'grep' or 'ack' and is used on the parrot source tree
15:19 moritz I don't think there's a better way :(
15:21 whiteknight honza: what do you mean?
15:22 honza I don't know which library I have to load for e. g. 'OS' oject
15:22 honza object
15:23 whiteknight loadlib "os"
15:23 honza thank  you :) but it would be nice if there were helpful comments at the headers of the PMC
15:24 whiteknight We can add that
15:25 honza for newbies it would be excellent
15:32 dalek nqp: f3b8f73 | (Michael Schroeder)++ | src/how/RoleToClassApplier.pm:
15:32 dalek nqp: copy has_method implementation over from perl6
15:32 dalek nqp:
15:32 dalek nqp: The perl6 implementation uses the method_table instead
15:32 dalek nqp: of iterating over all methods. It ignores $local for now,
15:32 dalek nqp: this may need to be fixed in the future.
15:32 dalek nqp: review: https://github.com/perl6/nqp/commit/f3b8f73e40
16:04 SHODAN joined #parrot
16:05 ambs joined #parrot
16:21 honza Can someone explain to me the difference between an attribute and a property?
16:22 dmalcolm joined #parrot
16:24 sorear honza: the Parrot Way it to use multiple names for the same thing in some places and the same name for different things in others.  Can you be more specific about what you mean by each?
16:25 honza sorear: are there situations where properties should be used instead of attributes?
16:26 sorear honza: what do you mean by "properties" and "attributes"/
16:26 sorear both terms are used for multiple things in the source
16:30 honza I have found it - properties can be acessed as key in associative arrays and hold additional information about PMC
16:30 sorear that's not all they do
16:30 honza what's the rest?
16:32 sorear well, are you talking about getprop properties, or OO properties, or something else?
16:32 whiteknight attributes are things that objects in a class has
16:32 dukeleto honza: so you would like a document that describes all core libraries and which functions they contain?
16:32 whiteknight so a class has an attribute, and objects of that class also have those attributes. These are things you see in most languages
16:32 whiteknight properties are named values that can be set on an instance, and have nothing to do with the class
16:32 dukeleto honza: also, are you asking questions about writing PIR ?
16:32 sorear whiteknight: how can you be so sure that honza is talking about getprop and getattribute?
16:32 whiteknight honza: for most uses, you probably want attributes
16:33 sorear whiteknight: I wish people would help us help them
16:33 whiteknight sorear: in the context of Parrot, that's what those two words mean
16:33 dukeleto honza: or perhaps a searchable web interface?
16:34 dukeleto honza: we definitely need to make our documentation friendlier. Suggestions are very welcome.
16:37 honza thank you all very much.  It's clear.
16:37 dalek rakudo/nom: f8301da | Coke++ | t/spectest.data:
16:37 dalek rakudo/nom: track failures, run fudges.
16:37 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f8301da74b
16:41 Coke sorear: I'm not sure what was unclear about his question.
16:41 Coke I think you might be a little to close to the internals if that was confusing.
16:43 sorear ok.
16:59 HSlayer|2 joined #parrot
17:12 mj41 joined #parrot
17:18 dalek Rosella/gh-pages: 223991a | Whiteknight++ | libraries/ (3 files):
17:18 dalek Rosella/gh-pages: a few fixes/additions to Core docs
17:18 dalek Rosella/gh-pages: review: https://github.com/Whiteknig​ht/Rosella/commit/223991a375
17:19 dalek Rosella/query_streams: bd0b66e | Whiteknight++ | s (3 files):
17:19 dalek Rosella/query_streams: Query.Queryable.Stream -> Query.Stream. Make it not a Queryable anymore. Small cleanups to the source
17:19 dalek Rosella/query_streams: review: https://github.com/Whiteknig​ht/Rosella/commit/bd0b66e884
17:23 cotto I'm thinking about m0 again.  Is there any reason to support short (16-byte ints) as a type?
17:24 cotto I don't see one, but it seems like a good idea to ask before rejecting the idea
17:27 whiteknight cotto: lots of users have asked for various-sized integer types
17:27 whiteknight I think we can't go wrong if we offer, even if it's not optimized for performance, access to 8- 16- and 32-bit integer values
17:28 whiteknight or, if we have an easy buffer type that allows those types to be used pretty easily
17:28 whiteknight In either case, I know people do ask for them
17:32 dukeleto but what datatypes does m0 need to understand natively?
17:32 dukeleto whiteknight: i agree that we should have them, but i am not quite sure if m0 needs to know about all of them
17:37 whiteknight that's what I'm saying, we don't need to optimize for them, but we need a mechanism for working with them
17:40 whiteknight I know Rakudo wants them, and we don't want to make a short-sighted system that they are just going to have to work around
17:47 mls hi parrot folks!
17:48 mls what's the background of imcc not setting the line number for lines that don't start with whitespace?
17:49 mls (should have said: updating the line number)
17:52 cotto_work ~~
17:53 mls oh, there's ticket #1652, please ignore the noise ;)
17:56 fperrad joined #parrot
17:58 whiteknight mls: bad decisions were made. We're not proud of it
17:59 bubaflub whiteknight: don't know if you saw it, but http://amber-lang.net/ is smalltalk running on top of JS, maybe something fun for jaesop in the future
17:59 whiteknight bubaflub: yeah, maybe
17:59 whiteknight bubaflub: We really are going to need 6model for stage 1, and we can't do anything fun like that before stage 1
17:59 benabik joined #parrot
17:59 mls Is it ok to insert an extra SET_LINE_NUMBER in the "return SUB" case, so that ops profiling knows where subs begin?
18:00 mls (I mean return SUB in imcc.l)
18:00 bubaflub whiteknight: yeah, i'm looking forward to that.  hopefully cardinal will take off again.
18:00 whiteknight mls: please do. So long as it doesn't break any important tests
18:00 whiteknight where "important tests" are those tests that wouldn't have broken anyway
18:00 whiteknight bubaflub: yeah, 6model should do a hell of a lot to help several languages
18:00 benabik o/
18:03 whiteknight I would love to see cardinal get moving again
18:05 dalek rakudo/nom: 6c9731a | moritz++ | src/Perl6/Grammar.pm:
18:05 dalek rakudo/nom: die on use of match variable in declaration
18:05 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6c9731a031
18:08 bubaflub whiteknight: a lot of Rubinius is written in Ruby and I wonder how much we can steal / borrow from them
18:08 whiteknight bubaflub: if we go the bootstrapped way, probably a lot
18:09 whiteknight write stage 0 in NQP/PCT or whatever, write stage 1 mostly in Ruby
18:09 bubaflub whiteknight: yeah, that's what i'm hoping for.  i know we'll have to put in a lot of work for the Object system and probably clean up some of the syntax stuff, but i'm thinking we won't have to focus on the stdlib stuff
18:09 whiteknight the boundaries where Rubinius had to resort to lower-level code probably forms the same boundary that we have to use
18:10 Tene We can't really borrow anything from rubinius until we have a good object model.
18:10 whiteknight right, that part is first
18:10 sorear what's Rubinius again?
18:10 whiteknight and 6model plays a big part of that
18:11 tadzik sorear: Ruby implementation in Ruy
18:11 Tene We can't even really steal from any of the test suites, as they all use test libraries that rely on monkey patching and other advanced object model features.
18:11 tadzik iirc
18:11 Tene they're not really intended for implementations in-progress.
18:12 whiteknight Tene: did you have anything written for moving Cardinal to 6model? I know you were talking about it, I don't know if you have any fragments around half written
18:12 Tene but, that's approximately the plan I intended.
18:12 benabik Tene: The conversation started with "after 6model we could…"  Also, somewhat sadly, monkey patching is a somewhat basic part of ruby
18:12 Tene whiteknight: yeah, I have some notes in ink in a notebook, and an implementation that fails pretty badly.
18:12 tadzik sorear: oh pardon, the vm itself seems to be in cpp
18:12 whiteknight Tene: more than the rest of us have. you have a scanner?
18:13 Tene The implementation is doing it wrong, and isn't properly factored.
18:13 Tene whiteknight: I'm not on-call anymore, and I actually got some sleep last night, so I might be able to work on Parrot stuff tonight.
18:13 Tene Hopefully.
18:14 Tene First is documenting HLL interop stuff for dukeleto.
18:14 whiteknight Tene: I don't want to put you straight to work if you need some time to relax!
18:14 Tene After that, I'd love to talk about cardinal 6model stuff
18:14 whiteknight awesome, definitely something good to talk about
18:18 dukeleto Tene++ # docs please
18:19 whiteknight dukeleto: you're insatiable!
18:20 Tene whiteknight: would you really prefer that he said "You know what?  I think we have enough Parrot.  Let's stop here."
18:20 whiteknight I didn't say it was a bad thing!
18:20 Tene :)
18:26 dukeleto whiteknight: indeed, i am
18:36 jsut_ joined #parrot
19:15 cotto_work whiteknight: didn't you have a 64-bit windows install somewhere?
19:18 whiteknight cotto_work: I have one here at work. I haven't had a chance to play with the branch
19:25 alvis joined #parrot
19:31 dukeleto i have been coding in perl for 15 years and I still can't ever remember if it is ~= or =~
19:32 arnsholt Which is why Perl 6 replaces it with ~~ =)
19:34 tadzik Perl 6 has ~= too :P
19:35 TimToady and =~ :)
19:35 benabik Blarg?
19:35 NotFound Spain is different. Amazon has opened here, and is cheaper to buy in amazon.uk than locally.
19:35 tadzik oh noes, what have we done!
19:36 cotto_work and ==
19:36 NotFound Do you have =: ?
19:36 TimToady not yet
19:36 NotFound Good, winxed innovates something.
19:37 benabik What's =:?
19:37 NotFound assign
19:38 NotFound (in pir terms)
19:38 benabik What's that make = ?  (Pardon if I'm being slow today.)
19:38 NotFound set
19:38 benabik set is register, assign is VTABLE?
19:39 cotto_work there's not a very good distinction between those concepts in pir
19:41 whiteknight I didn't know Winxed had =:
19:41 NotFound And its related vtables are aslo funny.
19:43 NotFound whiteknight: ack for it the winxed sources and you'll see is heavily used.
19:44 NotFound Is what allows to write fast closures.
19:45 whiteknight what do you mean?
19:46 NotFound It avoids the need to constantly load and store var lexicals.
19:46 NotFound (if you are careful)
19:48 cotto_work most interesting
19:48 benabik Grab the PMC and assign to it, rather than store a new value?
19:48 NotFound benabik: yeah
19:52 NotFound More funny: a[1] = 1; => $P1[1] = 1 ----- a[1] =: 1; => $P2 = $P1[1] \n assign $P2, 1
19:57 whiteknight ...is that actual code, or did you break your keyboard?
19:57 cotto_work +1
19:57 cotto_work I especially like the ----- operator
19:57 whiteknight really, really decrement
19:58 whiteknight also "subtract like crazy"
19:58 cotto_work quit being so negative ;]
20:00 NotFound --++
20:07 dalek rakudo/nom: 16bb82c | jnthn++ | src/core/Enum.pm:
20:07 dalek rakudo/nom: key and value should declare themselves as public attributes in Enum.
20:07 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/16bb82c184
20:09 whiteknight NotFound: so how do you use that for cheap closures?
20:09 NotFound More cleanly:
20:09 NotFound winxed:
20:09 NotFound a[1] = 1;
20:09 NotFound gives
20:09 NotFound $P1[1] = 1
20:09 NotFound winxed:
20:09 NotFound a[1] =: 1;
20:09 NotFound gives
20:09 NotFound $P2 = $P1[1]
20:09 NotFound assign $P2, 1
20:11 whiteknight ...and somewhere in there is an optimization for closures?
20:12 NotFound whiteknight: no, this is just a more funny usage without lexicals involved.
20:12 whiteknight okay
20:12 benabik whiteknight: Instead of having to many find_lex, store_lex he uses one find_lex and many assign.
20:13 NotFound For closures, using =: avoids the need to use the 'volatile' modifier and don't generate store_lex
20:13 NotFound You just need to avoid mixing inadvertently = and =: with the same var.
20:13 whiteknight there's a volatile modifier?
20:13 whiteknight are we talking about the same language?
20:14 NotFound whiteknight: yes, plobsing asked for it short after introducing lexicals.
20:14 whiteknight Are there examples about what it does?
20:14 NotFound I've not used it yet.
20:15 whiteknight I want to add ops for typed attribute access soon. I was going to prototype them last night but ran out of time
20:15 whiteknight I might add them tonight
20:16 NotFound It just generates a find_lex in all lexicals reads, instead of using just one at the start of the sub.
20:16 whiteknight oh, okay
20:18 NotFound That makes closures more complicated in winxed than in javascript, but allows shorter and faster generated code.
20:32 ttbot joined #parrot
20:43 contingencyplan joined #parrot
21:43 Coke any reason for =: over := for winxed binding?
21:47 dalek Heuristic branch merge: pushed 85 commits to rakudo/optimizer by jnthn
22:00 dalek Heuristic branch merge: pushed 40 commits to rakudo/nom-buf by tadzik
22:00 dalek rakudo/nom: ed8f4ad | tadzik++ | t/spectest.data:
22:00 dalek rakudo/nom: Run S06-advanced_subroutine_features/callsame.t
22:00 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ed8f4ad131
22:41 dalek rakudo/nom: 867ebd6 | jnthn++ | src/core/control.pm:
22:41 dalek rakudo/nom: Add a cheating &warn to avoid meta-errors.
22:41 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/867ebd6473
22:41 dalek rakudo/nom: 12b860d | jnthn++ | src/core/Temporal.pm:
22:41 dalek rakudo/nom: A doubled 'or' caught by work in the optimizer branch.
22:41 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/12b860df69
22:44 Coke joined #parrot
22:45 dalek rakudo/optimizer: ed8f4ad | tadzik++ | t/spectest.data:
22:45 dalek rakudo/optimizer: Run S06-advanced_subroutine_features/callsame.t
22:45 dalek rakudo/optimizer: review: https://github.com/rakudo/rakudo/commit/ed8f4ad131
22:45 dalek rakudo/optimizer: 1088154 | jnthn++ | src/Perl6/Optimizer.pm:
22:45 dalek rakudo/optimizer: Stub in some check-time failure/warning infrastructure. First cut of detecting undeclared routines; fails on building CORE.setting.
22:45 NotFound Coke: = is used for the more javascriptish meaning, I thought that using := for assign will be confusing. Using a symbol with less history forces users and readers to think about it.
22:45 dalek rakudo/optimizer: review: https://github.com/rakudo/rakudo/commit/108815461c
22:45 dalek rakudo/optimizer: e98e205 | jnthn++ | src/Perl6/Optimizer.pm:
22:45 dalek rakudo/optimizer: Don't complain about calling routines that we get passed as &foo style parameters.
22:45 dalek rakudo/optimizer: review: https://github.com/rakudo/rakudo/commit/e98e205735
22:45 dalek rakudo/optimizer: 867ebd6 | jnthn++ | src/core/control.pm:
22:45 dalek rakudo/optimizer: Add a cheating &warn to avoid meta-errors.
22:45 dalek rakudo/optimizer: review: https://github.com/rakudo/rakudo/commit/867ebd6473
22:45 dalek rakudo/optimizer: 12b860d | jnthn++ | src/core/Temporal.pm:
22:45 dalek rakudo/optimizer: A doubled 'or' caught by work in the optimizer branch.
22:45 dalek rakudo/optimizer: review: https://github.com/rakudo/rakudo/commit/12b860df69
22:45 dalek rakudo/optimizer: 6bb0357 | jnthn++ | / (3 files):
22:45 dalek rakudo/optimizer: Merge branch 'nom' into optimizer
22:45 dalek rakudo/optimizer: review: https://github.com/rakudo/rakudo/commit/6bb0357811
22:53 whiteknight joined #parrot
22:54 dalek parrot/whiteknight/native_attrs: 1d08697 | Whiteknight++ | / (5 files):
22:54 dalek parrot/whiteknight/native_attrs: Add in getattribute and setattribute variants for native types. Right now they do the same boxing/unboxing, but in the future they might be replaced with real native attribute access
22:54 dalek parrot/whiteknight/native_attrs: review: https://github.com/parrot/parrot/commit/1d08697da2
22:56 dalek Rosella/harness_test: 4daeac7 | Whiteknight++ | / (6 files):
22:56 dalek Rosella/harness_test: Implement several tests. Only the FileResult tests are failing now, and I'm setting up a mock infrastructure for them
22:56 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/4daeac77f6
22:56 dalek Rosella/harness_test: 353cd78 | Whiteknight++ | / (2 files):
22:56 dalek Rosella/harness_test: Tests for FileResult
22:56 dalek Rosella/harness_test: review: https://github.com/Whiteknig​ht/Rosella/commit/353cd78cad
22:56 dalek Rosella: 03dbe25 | Whiteknight++ | s (2 files):
22:56 dalek Rosella: Fix setup.winxed to exit(1) when there is an error. +some comments and small cleanups in Harness
22:56 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/03dbe2527b
22:56 dalek Rosella: c45a149 | Whiteknight++ | / (2 files):
22:56 dalek Rosella: +tests for the new Harness accessor methods
22:56 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/c45a149912
22:56 dalek Rosella: ec3b15d | Whiteknight++ | / (2 files):
22:56 dalek Rosella: Stub tests for TestExecutor
22:56 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/ec3b15d569
22:56 dalek Rosella: e2a8dc5 | Whiteknight++ | / (8 files):
22:56 dalek Rosella: Several Harness test fixes
22:56 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/e2a8dc5309
22:56 dalek Rosella: 09c731e | Whiteknight++ | src/harness/TapParser.winxed:
22:56 dalek Rosella: Change TapParser to be more intelligent about whitespace, which fixes error reporting.
22:56 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/09c731e672
22:56 dalek Rosella: 68c099b | Whiteknight++ | / (3 files):
22:56 dalek Rosella: Trim out newlines from the tap stream. Fix parsing of TODO to be more robust.
22:57 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/68c099baa6
22:57 dalek Rosella: ca77256 | Whiteknight++ | t/tap_harness/Harness/TestRun/Factory.t:
22:57 dalek Rosella: +tests for TestRun.Factory
22:57 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/ca7725653b
22:57 dalek Rosella: 4daeac7 | Whiteknight++ | / (6 files):
22:57 dalek Rosella: Implement several tests. Only the FileResult tests are failing now, and I'm setting up a mock infrastructure for them
22:57 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/4daeac77f6
22:57 dalek Rosella: 353cd78 | Whiteknight++ | / (2 files):
22:57 dalek Rosella: Tests for FileResult
22:57 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/353cd78cad
22:57 dalek Rosella: 16f8ad3 | Whiteknight++ | / (31 files):
22:57 dalek Rosella: Merge harness_test branch. Fix conflicts
22:57 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/16f8ad33b3
22:57 dalek Rosella: e036fbf | Whiteknight++ | s (3 files):
22:57 dalek Rosella: Add 'winxed_debug' arg to Distutils wrapper to build winxed files with --debug
22:57 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/e036fbf0b7
22:57 dalek Rosella: 03e6688 | Whiteknight++ | src/harness/ (2 files):
22:57 dalek Rosella: Fix up TestFile.PIR
22:57 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/03e668874b
23:08 dalek rakudo/optimizer: bfcf87a | jnthn++ | src/Perl6/Compiler.nqp:
23:08 dalek rakudo/optimizer: While in the optimizer branch, turn optimizer on by default, for easier testing.
23:08 dalek rakudo/optimizer: review: https://github.com/rakudo/rakudo/commit/bfcf87a507
23:08 dalek rakudo/optimizer: 4cba402 | jnthn++ | src/Perl6/Optimizer.pm:
23:08 dalek rakudo/optimizer: Avoid some inlines when we might screw up the workings of control exception handlers (like with given/when). This fixes the spectest regressions that we got from the immediate block inlining optimization. (We now fail some tests, but it's because the tests have undeclared routines, which the optimizer complains about at CHECK time; these tests need fixing.)
23:08 dalek rakudo/optimizer: review: https://github.com/rakudo/rakudo/commit/4cba402000
23:10 plobsing joined #parrot
23:14 whiteknight msg cotto can you please take a look at the whiteknight/native_attrs branch? We talked about it at #ps a few weeks ago
23:14 aloha OK. I'll deliver the message.
23:15 cotto_work whiteknight: I'm here
23:16 whiteknight okay. It's a simple patch.
23:17 cotto_work no it's not.  It nearly killed firefox.
23:17 cotto_work ;)
23:18 cotto_work hard to criticize the implementation
23:19 cotto_work I'll give it a better look and remind myself of its motivation this evening.
23:19 whiteknight the motivation is that we can start making HLLs, like winxed, pretend to have native-typed attrs
23:19 whiteknight until we get 6model and make it a reality
23:25 cotto_work whiteknight: looks good to me.  Throw it at NotFound and see what happens. ;)
23:25 whiteknight also, we're merging kill_threads soon, unless somebody yells "stop" real soon
23:25 whiteknight NotFound: ping
23:28 * sorear wonders how the 6model merge will work
23:28 whiteknight sorear: you want to get involved? We need more hands
23:30 whiteknight you can help shape it
23:30 NotFound whiteknight: pong
23:30 whiteknight NotFound: I just pushed a whiteknight/native_attrs branch. Very straightforward. Adds ops to getattribute/setattribute native types. How quickly could Winxed add support?
23:30 whiteknight I could put together a patch if you think it would be easy
23:31 whiteknight or, you can do it if you want it done the right way :)
23:31 NotFound Too late for me now, I'll take a look tomorrow
23:32 sorear whiteknight: sadly I'm quite booked for now
23:33 whiteknight NotFound: okay, no rush. Let me know if you like these ops. I can merge them soon if they are good
23:35 whiteknight ANYBODY HAVE ANY RESERVATIONS ABOUT A kill_threads MERGE? It's happening in a few minutes if nobody says anything
23:36 cotto_work whiteknight: http://www.youtube.com/watch?v=u6ALySsPXt0
23:37 whiteknight bleh, I just got a failure in t/pmc/timer.t after the update from master
23:41 plobsing whiteknight: yes, I have a reservation for kill_threads. I beleive it was for now o'clock under the name "killy mckillerson".
23:41 whiteknight :)
23:43 cotto_work You don't want to get that guy's name wrong.
23:44 whiteknight Timers!!!
23:44 * whiteknight hates
23:45 whiteknight if hate for software was a muscle, mine would be ripped
23:46 plobsing if software didn't suck, we'd be out of a hobby.
23:46 cotto_work whiteknight: is it failing reliably?
23:46 whiteknight yeah
23:46 cotto_work Timers, I am disappoint.
23:50 cotto_work whiteknight: passes for me
23:51 whiteknight yeah, I updated the branch from master and the failure appeared
23:56 dalek parrot/whiteknight/kill_threads: 405ca66 | NotFound++ | ext/winxed/driver.pir:
23:56 dalek parrot/whiteknight/kill_threads: update winxed snapshot to 0de3efd454
23:56 dalek parrot/whiteknight/kill_threads: - option -X
23:56 dalek parrot/whiteknight/kill_threads: review: https://github.com/parrot/parrot/commit/405ca66d6d
23:56 dalek parrot/whiteknight/kill_threads: 0223380 | Whiteknight++ | / (73 files):
23:56 dalek parrot/whiteknight/kill_threads: merge master into kill_threads. fix conflicts
23:56 dalek parrot/whiteknight/kill_threads: review: https://github.com/parrot/parrot/commit/0223380f2c
23:56 dalek parrot/whiteknight/kill_threads: 4f52f73 | Whiteknight++ | src/interp/inter_cb.c:
23:56 dalek parrot/whiteknight/kill_threads: Merge branch 'whiteknight/kill_threads' of github.com:parrot/parrot into whiteknight/kill_threads
23:56 dalek parrot/whiteknight/kill_threads: review: https://github.com/parrot/parrot/commit/4f52f73231
23:57 whiteknight cotto_work: pull that, and see if you get a failure
23:58 cotto_work doing so naough

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced