Camelia, the Perl 6 bug

IRC log for #parrot, 2012-06-11

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:15 kid51 whiteknight: No fight.  I just think getting other users' feedback would be good.
00:20 benabik On threads, I get a lot of failures in t/src/embed.t that it can't find _Parrot_new_string
00:22 kid51 On threads branch, darwin/ppc repeatedly hangs at test 37 of t/pmc/nci.t
00:27 kid51 In master, 'make test' continues to PASS on darwin/ppc.  So what's happening in threads is a regression.
00:37 benabik WTF?  Now I can't compile master.  It can't find the library for -licuuc.
00:37 benabik It was working yesterday...  any libicuuc.dylib is where it should be.
00:39 benabik But the linker line doesn't have the right path.  WTF Configure?
00:40 benabik moritz: ping
00:44 benabik Okay, I think I can fix it for me without breaking it for moritz
00:49 benabik And I'd love to fix that bignum thing while optimized
01:43 dalek parrot/whiteknight/io_cleanup1: 81003fd | Whiteknight++ | / (4 files):
01:43 dalek parrot/whiteknight/io_cleanup1: Fix the way we pretend to set buffers on StringHandle. A few other small fixes. t/pmc/stringhandle.t now passes.
01:43 dalek parrot/whiteknight/io_cleanup1: review: https://github.com/parrot/parrot/commit/81003fdf6b
01:43 dalek parrot/whiteknight/io_cleanup1: 3d5c787 | Whiteknight++ | src/io/buffer.c:
01:43 dalek parrot/whiteknight/io_cleanup1: Fix small issue reading multi-byte characters out of the buffer
01:43 dalek parrot/whiteknight/io_cleanup1: review: https://github.com/parrot/parrot/commit/3d5c787e60
01:43 dalek parrot/whiteknight/io_cleanup1: c71ec0c | Whiteknight++ | src/io/buffer.c:
01:43 dalek parrot/whiteknight/io_cleanup1: Start peppering a few assertions through the buffering code to start trying to track down a memory corruption error I'm seeing in rare circumstances. Untested (might not even compile in this state)
01:43 dalek parrot/whiteknight/io_cleanup1: review: https://github.com/parrot/parrot/commit/c71ec0c57c
01:46 kid51_at_dinner threads branch does not look good.
01:47 dalek parrot: e446d96 | benabik++ | MANIFEST.SKIP:
01:47 dalek parrot: Regenerate MANIFEST.SKIP
01:47 dalek parrot: review: https://github.com/parrot/parrot/commit/e446d96d00
01:47 dalek parrot: 399bb0e | benabik++ | config/auto/icu.pm:
01:47 kid51 On Darwin/PPC, t/pmc/nci.t, I had to comment out 8 tests between lines 1310 and 1985 to get that file to complete.
01:47 dalek parrot: Fix auto::icu for non-standard ICU paths
01:47 dalek parrot:
01:47 dalek parrot: While --ldflags was too much, --ldflags-libsonly isn't enough.  So
01:47 dalek parrot: also ask icu-config for --ldflags-searchpath
01:47 dalek parrot: review: https://github.com/parrot/parrot/commit/399bb0e499
01:48 kid51 benabik: Would that commit require modifying t/steps/auto/icu-01.t?
01:48 benabik kid51: No.  moritz already made it slightly more liberal.
01:48 kid51 thanks
01:49 benabik It checks for --ldflags.  Any suffix is ignored
01:52 ttbot Parrot 399bb0e4 i386-linux-thread-multi make error http://tt.taptinder.org/cmdinfo/89553
01:54 kid51 benabik:  that looks like a Makefile error
02:04 kid51 benabik:  Something is wrong with your commit to config/auto/icu.pm.
02:04 kid51 It's resulting in a value for 'icu_shared' in lib/Parrot/Config/Generated.pm that has a newline in it.
02:04 kid51 which is what's probably causing Makefile error
02:11 benabik *sigh*
02:12 kid51 The output you're capturing at line 282 has an embedded newline
02:12 kid51 icushared:  captured -L/usr/lib
02:12 kid51 -licui18n -licuuc -licudata
02:13 benabik Doesn't do that locally, but alright.
02:13 kid51 try: perl Configure.pl --verbose-step=auto::icu
02:14 benabik I checked the output of icu-config.  Must be something different between versions.
02:15 nopaste "kid51" at 70.85.31.226 pasted "perl Configure.pl --verbose-step=auto::icu" (106 lines) at http://nopaste.snit.ch/143734
02:16 benabik Working on a fix.
02:16 kid51 thanks:
02:19 kid51 msg whiteknight threads branch: had to comment out 8 tests in t/pmc/nci.t to get 'make test' to complete on Darwin/PPC
02:19 aloha OK. I'll deliver the message.
02:20 kid51 That's significant regression from master on same machine
02:22 dalek parrot: e69972a | benabik++ | config/auto/icu.pm:
02:22 dalek parrot: Fix auto::icu for newline from icu-config
02:22 dalek parrot: review: https://github.com/parrot/parrot/commit/e69972a105
02:23 kid51 makefile error fixed
02:23 benabik Woo.
02:23 benabik It should have worked, but wasn't 100% sure since I didn't see the error.
02:25 kid51 We'll have to see whether your fix works on Windows or not.
02:25 benabik u.u
02:26 kid51 Alright, that's enough for tonight
02:26 benabik I suppose icu-config might use \r\n in Windows.  Blah.
02:37 crab2313 joined #parrot
03:15 kurahaupo joined #parrot
03:50 dalek m1: 3bba565 | jimmy++ | src/gencode.c:
03:50 dalek m1: added more unuse_reg
03:50 dalek m1: review: https://github.com/parrot/m1/commit/3bba565282
05:05 dalek m1: 534079a | jimmy++ | src/gencode.c:
05:05 dalek m1: add more unuse_reg
05:05 dalek m1: review: https://github.com/parrot/m1/commit/534079a894
05:43 dalek m1: bda1b6a | jimmy++ | src/gencode.c:
05:43 dalek m1: some codes clean up, also reset the register allocator for each chunk
05:43 dalek m1: review: https://github.com/parrot/m1/commit/bda1b6a8c8
05:54 dalek m1: a08ef9d | jimmy++ | src/gencode.c:
05:54 dalek m1: registers of each type is 61
05:54 dalek m1: review: https://github.com/parrot/m1/commit/a08ef9d569
06:09 dalek mod_parrot/cleanup: 697da51 | (Bart Wiegmans)++ | p (6 files):
06:09 dalek mod_parrot/cleanup: Move provish to pudding
06:09 dalek mod_parrot/cleanup: review: https://github.com/bdw/mod​_parrot/commit/697da51343
06:09 dalek mod_parrot/cleanup: 926798d | (Bart Wiegmans)++ | pudding/ (2 files):
06:09 dalek mod_parrot/cleanup: This has become more of a messy than a cleanup branch. But 'pudding' actually works, which is nice.
06:09 dalek mod_parrot/cleanup: review: https://github.com/bdw/mod​_parrot/commit/926798d976
06:23 dalek m1: 1fad6ef | jimmy++ | src/gencode.c:
06:23 dalek m1: add more unuse_reg
06:23 dalek m1: review: https://github.com/parrot/m1/commit/1fad6efcc2
06:43 pipebrain joined #parrot
06:44 pipebrain hello
06:44 moritz hello pipebrain
06:45 pipebrain i made a simple language with just the if statement by following the tutorial
06:45 pipebrain executing files is fine
06:45 pipebrain but when i try a file that contains unicode i get this
06:45 pipebrain (wait I'll get the error)
06:46 pipebrain Invalid character in ASCII string
06:46 pipebrain the file is encoded utf8 without BOM
06:47 moritz I'm not familiar with the tutorial
06:47 pipebrain the tutorial is beside the point
06:47 moritz but somewhere you must have a main file, ie the entry point
06:47 sorear don't you need to explicitly specify utf8?
06:47 pipebrain it works in ascii
06:47 moritz and it contains a call to .command_line on the compiler object
06:47 pipebrain this works: if 1 {say "hello"; }
06:48 moritz there you can pass the option :encoding('utf8')
06:48 pipebrain ah
06:48 pipebrain ok
06:48 pipebrain command line argument?
06:48 moritz pipebrain:  https://github.com/rakudo/rak​udo/blob/nom/src/main.nqp#L35
06:48 moritz that's how rakudo does it
06:48 pipebrain ok
06:48 pipebrain thanks for your help
06:49 moritz wait to see if it actually works :-)
06:50 moritz if your entry point is written in PIR; it might need to be   'utf8' :named('encoding')
06:50 moritz I can never remember which syntax PIR uses for named arguments :/
06:51 pipebrain i am using the skeleton generated by mk_language_shell.pl
06:51 pipebrain i am gonna search the docs
06:51 pipebrain i will post when done
06:51 moritz ok
06:51 moritz I'll be afk for ~40 minutes
06:51 pipebrain thanks
06:51 pipebrain take your time
07:04 brrt joined #parrot
07:09 pipebrain nice
07:09 pipebrain my program works in utf8
07:11 pipebrain the only remaining issue is that parrot cannot handle utf8 files that start with BOM (\ufeff)
07:19 sorear not even if you strip the BOM after decoding?
07:22 pipebrain if you strip it it's fine
07:22 pipebrain it's just an inconvenience
07:23 pipebrain since a lot of editors (Windows based mostly) add the BOM
07:24 pipebrain i think parrot should consider it whitespace for parsing purposes
07:35 moritz you can write your own token ws { } that does that
07:48 contingencyplan joined #parrot
08:23 dalek m1: 2b9a939 | jimmy++ | src/gencode.c:
08:23 dalek m1: avoids magical number
08:23 dalek m1: review: https://github.com/parrot/m1/commit/2b9a939b6e
08:27 kjs joined #parrot
08:29 cosimo joined #parrot
08:44 dalek m1: e239f13 | kjs++ | src/gencode.c:
08:44 dalek m1: add a few clarifications and cleanups.
08:44 dalek m1: review: https://github.com/parrot/m1/commit/e239f13acf
08:55 dalek m1: 97da8da | kjs++ | src/gencode.c:
08:55 dalek m1: some comments and simplification.
08:55 dalek m1: review: https://github.com/parrot/m1/commit/97da8daad9
09:05 dalek m1: ea1faf8 | jimmy++ | src/gencode.c:
09:05 dalek m1: add more unuse_reg
09:05 dalek m1: review: https://github.com/parrot/m1/commit/ea1faf832a
09:05 dalek m1: 69d6757 | jimmy++ | src/gencode.c:
09:05 dalek m1: fixed conflcts
09:05 dalek m1: review: https://github.com/parrot/m1/commit/69d67578f8
09:22 dngor joined #parrot
09:55 kjs joined #parrot
10:16 mtk joined #parrot
10:51 aloha (parrot/parrot) Issues closed : 784 (m0 C interpreter cleanup (work in progress)) by gerdr : https://github.com/parrot/parrot/issues/784
10:53 JimmyZ joined #parrot
10:56 aloha (parrot/parrot) Issues closed : 784 (m0 C interpreter cleanup (work in progress)) by gerdr : https://github.com/parrot/parrot/issues/784
11:10 JimmyZ kjs: ping
11:10 kjs JimmyZ: poing
11:12 JimmyZ kjs: Do you how to do unuse_reg once a variable is unused?
11:12 * JimmyZ dinner&
11:12 kjs ?
11:12 kjs i don't get the question
11:20 brrt joined #parrot
11:30 kjs JimmyZ: not sure if this is what you mean, but a key to this is that m1_reg objects now have a is_symbol field. If a register is allocated to a symbol (see the code that checks for NO_REG_ALLOCATED_YET), then that flag is set. When calling unuse_reg() on such a reg, it won't do anything.
11:50 kid51 joined #parrot
11:53 not_gerd joined #parrot
11:53 not_gerd hello, #parrot
11:53 not_gerd anyone around with gcc 4.6 or newer?
11:56 not_gerd joined #parrot
12:11 kjs joined #parrot
12:13 kjs JimmyZ: ping
12:13 JimmyZ kjs: pong
12:13 kjs hi there
12:14 kjs i wasnt sure what you meant before you left for dinner
12:16 JimmyZ kjs: I want to reg.is_symbol = 0 when a variable is not used.
12:17 kjs JimmyZ: sorry I still dont' get it.. you want to ...?
12:17 JimmyZ hmm
12:18 kjs there's a verb missing which is essential :-)
12:18 JimmyZ kjs: you set reg.is_symbol when you  gen a var, right?
12:19 JimmyZ set reg.is_symbol = 1
12:19 kjs you ONLY set reg.is_symbol is IF the register is assigned to a symbol = a declared variable.
12:19 kjs so, whenever a register is retrieved through "use_reg" (a new reg is generated), then it MAY be assigned to a variable that is declared.
12:20 kjs in THOSE cases, you "freeze" that register; you set its is_symbol flag, so it won't be "unused" when unuse_reg() is called on it.
12:20 kjs when that flag is NOT set, then it's just a temporary register.
12:20 JimmyZ kjs: but when a declared variable  is not used any more, we should set reg.is_symbol = 0, so we can freeze that register
12:20 kjs you mean unfreeze?
12:21 JimmyZ kjs: yeah, unfreeze
12:21 kjs JimmyZ: You would be right, but in the current setup of this system, I don't think there is a way to find out, YET
12:21 kjs I know what you mean
12:21 kjs you want to know when a symbol is no longer used, in the rest of a function
12:21 kjs and then you can reuse that register
12:22 kjs it's not that advanced yet :-()
12:22 kjs :-)
12:22 kjs I'll think about how to do that, but it's not very high priority at the moment. This is already a great improvement over what it was before.
12:23 JimmyZ kjs: yeah
12:24 kjs The number of registers is not so much a problem; a bigger problem is that there are a lot of instructions generated that are redundant. I think quite some improvements can be made there, but I'm not sure yet how to do that.
12:25 JimmyZ kjs: I got a easy idea, but I'm not sure whether  it's right or not
12:25 kjs JimmyZ: it's not very likely that there will be 60 variables in a single function anyway.
12:26 kjs JimmyZ: perhaps you can propose it on the mailing list?
12:26 kjs get some input from a wider audience? Not everybody reads IRC logs.
12:27 kjs One way to do it is a register allocator algorithm. Linear Scan Register allocation is one of them. Parrot/PIRC (google that for the repo.) has that.
12:33 kjs JimmyZ: how comfortable are you with programming in M0?
12:33 kjs M0 assembler coding, that is.
12:33 JimmyZ kjs: not comfortable, but can give it a try
12:34 JimmyZ m0_assembler.pl?
12:34 kjs oh well it's easy learn, but also easy to make mistakews
12:34 kjs no, I mean using M0 as a programming langauge
12:34 kjs its instruction set.
12:34 kjs writing .m0 files :-)
12:35 kjs I'm thinking about figuring out how to implement a vtable (an array) that contains functions/methods
12:36 JimmyZ kjs: I think you're more comfortable than me now :P.
12:38 kjs there's also a refactoring job to do in gencode.c if you're looking for some hackign to do
12:38 kjs gencode_binary_{plus, minus, div, mult, mod, isgt, isge} share a lot of code. That needs to be refactored.
12:40 not_gerd bye, #parrot
12:40 not_gerd left #parrot
12:53 bluescreen joined #parrot
12:58 PacoAir joined #parrot
12:59 PacoAir joined #parrot
13:02 kjs JimmyZ: ping
13:03 JimmyZ kjs: pong
13:04 kjs JimmyZ: do you know how we can use a test harness or tool like Test::Harness or the "prove" commandline tool with M1's tests?
13:06 JimmyZ kjs: I don't know yet :(
13:06 kjs ok, no worries. I'll figure it out
13:09 Psyche^ joined #parrot
13:17 dalek m1: 53feeb9 | jimmy++ | / (2 files):
13:17 dalek m1: changed make test to use prove
13:17 dalek m1: review: https://github.com/parrot/m1/commit/53feeb9d20
13:17 JimmyZ kjs: ^^
13:17 kjs does that work?
13:18 JimmyZ kjs: test needs to be improved too. I did not do that.
13:18 kjs jimmy++ # it works
13:18 kjs JimmyZ: let me do that.
13:19 JimmyZ kjs: 'make test' or 'make test-v'
13:27 dalek m1: f93dc28 | kjs++ | t/ (4 files):
13:27 dalek m1: fix up some tests.
13:27 dalek m1: review: https://github.com/parrot/m1/commit/f93dc28e62
13:27 kjs JimmyZ: that's  beautiful. works now.
13:29 JimmyZ kjs++
13:29 kjs jimmy++ #making tests useful!
13:34 dalek m1: 46c19c6 | jimmy++ | run_m1.sh:
13:34 dalek m1: improve run_m1.sh
13:35 dalek m1: review: https://github.com/parrot/m1/commit/46c19c6fce
13:46 kjs JimmyZ: There's something wrong with the unuse_reg system. I think we call it a bit too often :-)
13:46 kjs or don't set the flags of is_symbol yet.
13:47 kjs I think I'll do the "freezing" in the registers data structure rather than the individual registers, that seems safer.
13:57 dalek m1: a8f73b2 | kjs++ | src/gencode.c:
13:57 dalek m1: add a freeze_reg funciton (unused still). Cleanup some code, disable reg. optimizer as it's not working yet.
13:57 dalek m1: review: https://github.com/parrot/m1/commit/a8f73b20f6
14:04 whiteknight joined #parrot
14:04 whiteknight good morning, #parrot
14:09 JimmyZ kjs: aye.
14:09 kjs JimmyZ: yeah, it's broken alright :-(
14:09 brrt morning, whiteknight
14:10 whiteknight msg kid51: Do you have a list of the nci.t tests you had to comment out on PPC? I suspect they might not be a problem. Rakudo doesn't use our NCI anyway, so problems there are not a severe and can be managed through fixes or deprecations
14:10 aloha OK. I'll deliver the message.
14:10 whiteknight hello brrt
14:15 jsut joined #parrot
14:15 brrt hows life
14:15 whiteknight brrt: blah.
14:16 brrt oh, why?
14:16 whiteknight brrt: Did a lot of yard work this weekend. When I was tired I went inside where my son was watching brightly colored childrens cartoons. So get angry, go back out side
14:16 dalek m1: fd6ef89 | jimmy++ | run_m1.sh:
14:16 dalek m1: improve run_m1.sh again
14:16 dalek m1: review: https://github.com/parrot/m1/commit/fd6ef89193
14:16 whiteknight now I'm sore from doing yard work
14:17 brrt :-) well, its better than sore from doing nothing
14:18 brrt i'm halfway through the goal of the 'cleanup' branch
14:20 brrt but that does mean the intermediate part is kind of.. troubling
14:25 dalek m1: 8d6d364 | kjs++ | / (2 files):
14:25 dalek m1: add test for pre/post inc. fix up reg. optimizer.
14:25 dalek m1: review: https://github.com/parrot/m1/commit/8d6d3644e7
14:26 whiteknight brrt: What do you mean?
14:28 brrt well, its in this intermediate-nothing-working-right-now state
14:28 brrt which i'd like to get out of ASAP
14:33 kjs joined #parrot
14:38 PacoAir joined #parrot
14:38 JimmyZ joined #parrot
14:38 dngor joined #parrot
14:38 cosimo joined #parrot
14:38 alvis joined #parrot
14:38 bacek_at_work joined #parrot
14:38 hernan joined #parrot
14:38 pjcj joined #parrot
14:38 autark joined #parrot
14:38 rblackwe joined #parrot
14:38 Util joined #parrot
14:41 Util joined #parrot
14:41 rblackwe joined #parrot
14:41 autark joined #parrot
14:41 pjcj joined #parrot
14:41 hernan joined #parrot
14:41 bacek_at_work joined #parrot
14:41 alvis joined #parrot
14:41 cosimo joined #parrot
14:41 dngor joined #parrot
14:41 JimmyZ joined #parrot
14:41 PacoAir joined #parrot
14:42 dalek m1: e20a972 | kjs++ | src/gencode. (2 files):
14:42 dalek m1: cleanup.
14:42 dalek m1: review: https://github.com/parrot/m1/commit/e20a972615
14:45 JimmyZ kjs++,  unuse_reg system is fixed!
14:45 kjs JimmyZ: it seems so. We still need to test this to death, because I don't trust it completely.
14:46 dalek m1: b3e6cd1 | jimmy++ | src/gencode.c:
14:46 dalek m1: add more unuse_reg
14:46 dalek m1: review: https://github.com/parrot/m1/commit/b3e6cd1c9c
14:46 kjs JimmyZ: not sure if you know this, but the general rule is, you can unuse() a reg if it's not pushed onto the regstack
14:47 kjs that is, if it's not passed onto the regstack for other functions to use.
14:47 kjs that's the most important thing to keep in mind.
14:47 JimmyZ kjs: yes, I know. and I know I can unuse a poped reg too
14:48 kjs JimmyZ: yes.
14:48 kjs unless it's pushed back :-)
14:48 JimmyZ aye
14:56 contingencyplan joined #parrot
15:07 dalek m1: d99c937 | jimmy++ | src/gencode.c:
15:07 dalek m1: small clean up
15:07 dalek m1: review: https://github.com/parrot/m1/commit/d99c937797
15:07 dmalcolm joined #parrot
15:16 lateau joined #parrot
15:17 dalek mod_parrot/cleanup: a4f66c2 | (Bart Wiegmans)++ | / (8 files):
15:17 dalek mod_parrot/cleanup: Updated Configure for the new setup
15:17 dalek mod_parrot/cleanup: review: https://github.com/bdw/mod​_parrot/commit/a4f66c27ff
15:51 dalek m1: b8f91af | jimmy++ | src/gencode.c:
15:51 dalek m1: add more unuse_reg
15:51 dalek m1: review: https://github.com/parrot/m1/commit/b8f91af95c
16:02 brrt left #parrot
16:03 dalek mod_parrot/cleanup: 5a97330 | (Bart Wiegmans)++ | / (2 files):
16:03 dalek mod_parrot/cleanup: Deleted uneccesary files
16:03 dalek mod_parrot/cleanup: review: https://github.com/bdw/mod​_parrot/commit/5a973304d5
16:03 dalek mod_parrot/cleanup: c2f9e52 | (Bart Wiegmans)++ | pudding/ (2 files):
16:03 dalek mod_parrot/cleanup: Add module loading to pudding
16:03 dalek mod_parrot/cleanup: review: https://github.com/bdw/mod​_parrot/commit/c2f9e52de0
16:25 jashwanth joined #parrot
16:28 jashwanth hello whiteknight
16:28 whiteknight hello jashwanth
16:28 jashwanth did you update PLA
16:31 whiteknight no, I have not
16:32 jashwanth actually even I did not no that night so could you tell me where I should do it
16:45 whiteknight src/pmc/nummatrix2d.pmc
16:45 whiteknight Add it somewhere inside the pmclass { } block
16:45 Coke what week are we in on GSOC?
16:45 whiteknight Coke: I have no idea. 4 or 5 maybe?
16:46 whiteknight jashwanth: In the future if you don't know where something goes, do experiments to find a good place
16:46 jashwanth ok
16:56 Coke whiteknight: I haven't seen any blogs. what's up?
16:56 Coke except brrt++
16:57 kjs good afternoon, #parrot
16:57 * moritz has a feeling that we'll have a lower success rate than usual
16:57 Coke I see initial posts from benabik  jharper1 Jashwanth and brrt , but followups only from brrt++
17:00 kjs Quick survey on preferred M1 features. Does anyone like M1 to have heredocs?
17:00 moritz -1
17:02 kjs well that's clear then :-)
17:02 kjs saves a lot of work.
17:07 Coke -1
17:10 whiteknight Coke: That's a good question. I've sort of forgotten about that
17:10 Coke msg cotto: how is benabik doing
17:10 aloha OK. I'll deliver the message.
17:11 Coke msg benabik: as your backup mentor, I'm not seeing any blog entries. what's up?
17:11 aloha OK. I'll deliver the message.
17:13 Coke msg dukeleto: I'm not seeing any blog entries from jashwanth. Everything ok with GSOC?
17:13 aloha OK. I'll deliver the message.
17:17 rurban Coke: my student 2 years ago did lots of blogging but bad coding. Got totally stuck in OO hell
17:19 Coke rurban: mmm, no blogging is not an an indicator of success. but this year, whiteknight said:
17:19 Coke "Weekly blog posts and weekly meetings with your mentor are REQUIRED.
17:19 Coke is very serious."
17:19 Coke If you miss these things repeatedly, you will FAIL THE PROJECT. This
17:19 Coke so, we have 3 failures already by that standard.
17:20 Coke my initil send there is misleading: "Blogging is not an indicator of success". ... but it is required.
17:20 whiteknight Coke: Yes, we've been giving some leeway because some students had prior committments (finals, etc) for the first few weeks
17:20 whiteknight I've just sent out an email reminder that even though we had been relaxed on it until now, blogging needs to pick up
17:22 jashwanth Coke:sorry for the mistake will write one right away
17:25 Coke jashwanth: I'm not your mentor, but thank you!
17:25 Coke jashwanth++
17:25 jashwanth :-)
17:36 dalek website: Jashwanth++ | The initial step
17:36 dalek website: http://www.parrot.org/content/initial-step
17:47 cotto moritz++ coke++
17:48 cotto benabik, ping
17:48 moritz what did I do?
17:48 cotto -1 to heredocs in m1
17:48 moritz oh :-)
17:48 moritz easily earned karma then :-)
17:53 kjs joined #parrot
17:55 kjs Another quick survey on M1: do you want to have native M1 support for writing PMCs, including inheritance and vtable method overriding? (not sure if possible, but think so)
17:59 moritz that depends on your goals for M1
17:59 moritz if you want it to be what stuff that we would currently use PIR or C for, then yes
17:59 kjs my goal for M1 is to become adopted as Mole :-)
18:03 cotto kjs, before answering that, try writing one and see where the holes would be.  PMCs as we now know them should eventually be implemented in either M1 or something above it.
18:11 kjs ok, I'd better bite the bullet and do this.
18:20 dalek m1: 8629aa2 | kjs++ | / (6 files):
18:20 dalek m1: Start working on PMCs.
18:20 dalek m1: review: https://github.com/parrot/m1/commit/8629aa2afb
18:25 dalek rakudo/module-trace: 1b00e64 | moritz++ | src/Perl6/ModuleLoader.pm:
18:25 dalek rakudo/module-trace: note end of module loading in debug mode
18:25 dalek rakudo/module-trace: review: https://github.com/rakudo/rakudo/commit/1b00e6425e
18:25 dalek rakudo/module-trace: 339771a | moritz++ | src/ (2 files):
18:25 dalek rakudo/module-trace: start to trace module loading in a dynamic variable
18:25 dalek rakudo/module-trace: review: https://github.com/rakudo/rakudo/commit/339771a1a8
18:50 dalek rakudo/module-trace: b5d0d37 | moritz++ | src/ (2 files):
18:50 dalek rakudo/module-trace: enable basic module traces in X::Comp
18:50 dalek rakudo/module-trace: review: https://github.com/rakudo/rakudo/commit/b5d0d37b75
18:59 dalek m1: 55e2921 | kjs++ | src/ (3 files):
18:59 dalek m1: more infrastructure for PMCs.
18:59 dalek m1: review: https://github.com/parrot/m1/commit/55e292146a
19:04 dalek m1: 8f09ce6 | kjs++ | / (4 files):
19:04 dalek m1: PMC defs parse. Add a test.
19:04 dalek m1: review: https://github.com/parrot/m1/commit/8f09ce651c
19:05 pmichaud anyone know the steps we would go through to test something on the gcc compile farm?  (more)
19:06 pmichaud larger question:  NQP's serialization doesn't work on bigendian boxes; we're trying to get jnthn++ access to a bigendian box to test code updates.  Any suggestions?
19:08 sorear do any of the parrot/rakudo core devs have a ppc mac lying around?
19:10 whiteknight pmichaud: dukeleto can probably set you up with it
19:13 dalek m1: a1e78f6 | kjs++ | / (2 files):
19:13 dalek m1: AST is generated correctly, apparently, as code is spit out for methods. Need to do a lot of mangling though. WIP!
19:13 dalek m1: review: https://github.com/parrot/m1/commit/a1e78f63d5
19:14 pmichaud whiteknight: yes, I just saw his message referencing where to sign up for the compile farm
19:14 pmichaud so I pointed jnthn++ at that.  If anyone has other suggestions, forward them :)
19:14 whiteknight I haven't used it myself, so dukeleto is probably the best source of information
19:17 NotFound I have uses it a few times, but don't know if dukeleto has stablished some way to automate tests.
19:18 pmichaud we're not needing automated testing at this time; we just need something to develop and test on
19:19 dalek nqp: b1226fb | jnthn++ | src/6model/serialization.c:
19:19 dalek nqp: Try to unbust things on big endian platforms. Untested, but should do the trick.
19:19 dalek nqp: review: https://github.com/perl6/nqp/commit/b1226fb1bf
19:19 NotFound Then the references provided for sign up and use are fine.
19:19 whiteknight the parrot packfile serialization/deserialization code contains some helper functions and macros for dealing with the differences between platforms. We can try to expose more of that through an API if necessary
19:20 whiteknight or, we can try to absorb some of the low-level bit-fiddly stuff from NQP into Parrot to deal with that too
19:21 whiteknight probably what we want in the long run, considering our serialization leaves much to be desired
19:30 not_gerd joined #parrot
19:30 not_gerd hello, #parrot
19:30 kid51 joined #parrot
19:30 whiteknight hello not_gerd
19:30 not_gerd anyone around with a gcc version >= 4.6
19:31 whiteknight alester usually plays with the bleeding edge
19:31 not_gerd I believe the warning produced by https://gist.github.com/2909666 on 4.5 is a bug
19:32 brrt joined #parrot
19:32 not_gerd a related issue got fixed for 4.6, so I'm hopeful that has gone away as well
19:33 kid51_ joined #parrot
19:33 whiteknight ok
19:34 not_gerd this issue generates noise when building https://github.com/gerdr/m0 with gcc
19:34 bluescreen joined #parrot
19:41 jashwanth joined #parrot
19:47 dalek rakudo/nom: 901f2a7 | jnthn++ | src/Perl6/Metamodel/MethodContainer.pm:
19:47 dalek rakudo/nom: Fix duplicate method detection; don't boolify. Fixes the crappy error reporting for duplicate regex definitions in a grammar.
19:47 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/901f2a7ee1
20:02 dalek rakudo/nom: b12854a | jnthn++ | src/Perl6/Actions.pm:
20:02 dalek rakudo/nom: Make .foo behave the same way as $_.foo does with regard to perl6ization of types.
20:02 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b12854a978
20:15 not_gerd left #parrot
20:20 alester joined #parrot
20:43 dalek mod_parrot/cleanup: b947d2d | (Bart Wiegmans)++ | / (17 files):
20:43 dalek mod_parrot/cleanup: Simplified the makefile and the structure
20:43 dalek mod_parrot/cleanup: review: https://github.com/bdw/mod​_parrot/commit/b947d2d7ba
20:43 dalek mod_parrot/cleanup: 91c18b7 | (Bart Wiegmans)++ | / (4 files):
20:43 dalek mod_parrot/cleanup: Add recursive makefile. Nearly done
20:43 dalek mod_parrot/cleanup: review: https://github.com/bdw/mod​_parrot/commit/91c18b7ec9
21:04 dalek rakudo/module-trace: f97e562 | moritz++ | src/Perl6/ModuleLoader.pm:
21:04 dalek rakudo/module-trace: restore GLOBAL if module loading failed
21:04 dalek rakudo/module-trace: review: https://github.com/rakudo/rakudo/commit/f97e562eee
21:04 dalek rakudo/module-trace: 3d9fa49 | moritz++ | src/Perl6/ (2 files):
21:04 dalek rakudo/module-trace: pass the line number to the module loader
21:04 dalek rakudo/module-trace: review: https://github.com/rakudo/rakudo/commit/3d9fa49592
21:14 alester joined #parrot
21:16 brrt left #parrot
21:20 dalek m1: 9a139ec | kjs++ | / (6 files):
21:20 dalek m1: various changes towards method calls. temporary hacks included.
21:20 dalek m1: review: https://github.com/parrot/m1/commit/9a139ec023
21:29 kurahaupo joined #parrot
21:44 dalek m1: 6399a7f | kjs++ | / (3 files):
21:44 dalek m1: more work towards PMC vtables.
21:44 dalek m1: review: https://github.com/parrot/m1/commit/6399a7fb27
21:47 dalek m1: 949a34d | kjs++ | src/ (2 files):
21:47 dalek m1: Store the registers usage table in the compiler object.
21:47 dalek m1: review: https://github.com/parrot/m1/commit/949a34d87b
21:57 dalek m1: d7b8411 | kjs++ | src/gencode.c:
21:57 dalek m1: add all methods to the const segment, so they can be copied from there in the vtable that is allocated in mem.
21:57 dalek m1: review: https://github.com/parrot/m1/commit/d7b8411780
22:16 kid51 joined #parrot
22:25 dalek m1: a4bf32a | kjs++ | / (3 files):
22:25 dalek m1: fix up chained assignments. a = b = c.
22:25 dalek m1: review: https://github.com/parrot/m1/commit/a4bf32a900
22:27 dalek m1: 362c1b3 | kjs++ | t/assignchain.m1:
22:27 dalek m1: fix up assign chain test.
22:27 dalek m1: review: https://github.com/parrot/m1/commit/362c1b39ca
22:30 whiteknight joined #parrot
22:34 dalek m1: 6fb9eb7 | kjs++ | src/gencode.c:
22:34 dalek m1: refactoring gencode for bitwise ops.
22:34 dalek m1: review: https://github.com/parrot/m1/commit/6fb9eb7eda
22:35 whiteknight good evening, #parrot
22:35 kjs hi whiteknight
22:35 whiteknight hello kjs
22:35 kjs how are things?
22:36 whiteknight things are good. Glad to be home from work
22:37 kjs :-)
22:42 dalek rakudo/module-trace: 22f512d | moritz++ | src/Perl6/ModuleLoader.pm:
22:42 dalek rakudo/module-trace: add line numbers where module loading occurs
22:42 dalek rakudo/module-trace: review: https://github.com/rakudo/rakudo/commit/22f512dc24
22:42 dalek rakudo/module-trace: 7812723 | moritz++ | src/ (3 files):
22:42 dalek rakudo/module-trace: p6ize module trace in World and pass it to the exception
22:42 dalek rakudo/module-trace:
22:42 dalek rakudo/module-trace: it almost works, but the hashes seem to get flattend out
22:42 dalek rakudo/module-trace: review: https://github.com/rakudo/rakudo/commit/7812723944
22:48 dalek parrot-linear-algebra: ec828bc | Whiteknight++ | pmcs/examples/nummatrix2d.md:
22:48 dalek parrot-linear-algebra: remove a directory that shouldn't exist on this branch
22:48 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/ec828bcebb
22:48 dalek parrot-linear-algebra: 46c9b37 | Whiteknight++ | s (5 files):
22:48 dalek parrot-linear-algebra: add get_pointer vtables to the matrix types, along with big warnings against using them.
22:48 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/46c9b37a84
22:49 whiteknight msg jashwanth I just added get_pointer vtables to all the matrix types in PLA. This should help you out.
22:49 aloha OK. I'll deliver the message.
22:51 dalek m1: 8cd613f | kjs++ | / (3 files):
22:51 dalek m1: the bitwise or operator was never added to lexer. Fix this. Remove whitespace from gencode. Add testfile for bitwise ops.
22:51 dalek m1: review: https://github.com/parrot/m1/commit/8cd613f05a
22:54 dalek rakudo/module-trace: d53986b | moritz++ | src/ (2 files):
22:54 dalek rakudo/module-trace: itemize hashes in p6ize_recursive; remove debugging output
22:54 dalek rakudo/module-trace:
22:54 dalek rakudo/module-trace: Now the output looks like I wanted it
22:54 dalek rakudo/module-trace: review: https://github.com/rakudo/rakudo/commit/d53986b173
23:03 dalek parrot/pmc_emitter_document: 95e6f6c | jkeenan++ | lib/Parrot/Pmc2c/Emitter.pm:
23:03 dalek parrot/pmc_emitter_document: Write preliminary documentation in POD format for find(), subst() and replace()..
23:03 dalek parrot/pmc_emitter_document: review: https://github.com/parrot/parrot/commit/95e6f6ca24
23:11 dalek m1: b0e3588 | kjs++ | / (2 files):
23:11 dalek m1: minor clean up in m1.y. add test file for math ops.
23:11 dalek m1: review: https://github.com/parrot/m1/commit/b0e358849e
23:22 dalek m1: d25b9c0 | kjs++ | docs/TODO.txt:
23:22 dalek m1: update status in TODO. not too bad.
23:22 dalek m1: review: https://github.com/parrot/m1/commit/d25b9c0013
23:24 travis-ci joined #parrot
23:24 travis-ci [travis-ci] parrot/parrot#419 (pmc_emitter_document - 95e6f6c : jkeenan): The build passed.
23:24 travis-ci [travis-ci] Change view : https://github.com/parrot/par​rot/compare/500836e...95e6f6c
23:24 travis-ci [travis-ci] Build details : http://travis-ci.org/parrot/parrot/builds/1596608
23:24 travis-ci left #parrot
23:29 jsut joined #parrot

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

Parrot | source cross referenced