Parrot 2.10.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Long live Git! github.com/parrot/parrot | git clone git://github.com/parrot/parrot.git | Please test rakudo with bleeding edge parrot!
Set by moderator on 2 December 2010.
00:00 Kristaba left
sorear dukeleto: I think you should be able to create multiple wholly independant interpreters 00:00
with possibly overlapping lifespans
dukeleto: the "parent pointer" thing is, iiuc, a legacy of ithreads. 00:01
00:18 bacek_mobile left 00:30 bluescreen joined 00:33 hercynium left 00:34 whiteknight joined 00:45 lucian left
whiteknight damnit 01:09
I HATE IMCC
if it were a person, I would beat it's damn intestines out of it's body with a stick
sorear what did ey do to you this time? 01:10
whiteknight it's throwing an unhandled exception, and it isn't going through the new API to do i 01:12
and it's calling exit() directly 01:14
01:14 kid51 joined
cotto_work hears the chainsaw of progress approaching from a distance 01:14
01:23 stilgar left
whiteknight the hack will be strong with this one 01:23
cotto_work: no, that's just me cursing like a pirate
cursing because I wish we were using PIRATE
if this new fix works, I am a freaking wizard 01:24
....I think it might have worked 01:34
dukeleto whiteknight: immc makes PL/Parrot segfault PostgreSQL, because it doesn't throw proper exceptions
whiteknight: when there is a syntax error
01:34 dmalcolm left
whiteknight dukeleto: I just fixed that 01:34
dukeleto whiteknight: oh really? 01:35
whiteknight: committed? I would really like to see that.
whiteknight I added two new functions, Parrot_api_try and Parrot_api_end_try that can be used to wrap things like IMCC that throw out-of-bound exceptions
I'm testing it locally right now
01:35 davidfetter left
whiteknight the new API is going to bring all sorts of wonderful gifts to people 01:35
01:38 fbrito left
dukeleto the links for PaFo on parrot.org/foundation/legal have been fixed 01:39
whiteknight: I can't wait to update PL/Parrot with the new api. It sounds much nicer
whiteknight dukeleto++
dukeleto how does one check that a packfile is valid? 01:44
whiteknight I have no idea. That's a plobsing question 01:45
01:48 silug joined
dalek rrot/leto/embed_grant: 9612866 | dukeleto++ | docs/embed.pod:
Add another function signature example

The 'constants' section was removed because the interpreter flags are now documented and there do not appear to be other constants to document.		 01:54
dukeleto aloha msg plobsing how does one check that a packfile is valid ?
aloha dukeleto: OK. I'll deliver the message.
02:06 Matt_ joined, Matt_ left 02:07 Matt-GCI joined
Matt-GCI Hello all 02:08
I just wanted to introduce myself, I'm a high school student working on a Parrot GCI (Google Code-In) task 02:09
02:13 jasonmay left 02:15 jasonmay joined
whiteknight hello Matt-GCI 02:15
Matt-GCI Hey
The task I'm working on is renaming functions in src/sub.c to meet naming standards 02:16
whiteknight nice 02:18
Matt-GCI sadly it's due tomorrow afternoon.. haha
dukeleto Matt-GCI: howdy!
dalek rrot/leto/embed_grant: 8d5dca1 | dukeleto++ | t/src/ (7 files):
Make t/src tests skip_all if src/parrot_config.o is not built
dukeleto Matt-GCI: if you are close but not done, we can give you an extension 02:19
Matt-GCI thank you for the offer, i'm going to do my best to do this timely however
dukeleto Matt-GCI: welcome to the parrot community! I am the person organizing Parrot's involvement with GCI.
whiteknight yes, extensions are easy
Matt-GCI dukeleto: thanks! i'm really excited to be working on this 02:20
most of my experience has been with Java, so this is new
dukeleto Matt-GCI: and we are excited to have you working on it! Learning new stuff is fun.
whiteknight Matt-GCI: what OS areyou on? 02:22
dalek rrot/tt532_headerizer_refactor: e76c132 | jkeenan++ | / (3 files):
Refactor two blocks of code out of function_components_from_declaration(). Place in Functions.pm. Test them in 01_functions.t.
Matt-GCI whiteknight: i'm running mac osx
whiteknight ok
does that have sed? 02:23
kid51 Yes. 02:24
Matt-GCI i assume there's a way to install it, since it's unix-based
02:24 bacek left
Matt-GCI I was originally going to use araxis find and replace to make the replacements but if you recommend sed then i'll use that instead 02:25
whiteknight sed -i 's/<oldname>/<newname>/' srcc/*.c
kid51 Matt-GCI: You should have 'sed' by default, probably in /usr/bin/sed
whiteknight it doesnt matter, any tool is better than doing it manually
Matt-GCI kid51: you're right, i just which'd it and it's there 02:26
whiteknight: amen to that
kid51 It's one of the oldest of the Unix tools ... ancestor of Perl's regexes
Matt-GCI when you guys edit the code, what IDE do you use? Eclipse?
kid51 vi
;-) 02:27
Matt-GCI lol, hardcore
kid51 Actually, I would like to know if any of our developers are using Padre.
It would be plausible, since Padre's creator drops in here from time to time. 02:28
whiteknight medit here 02:29
i would use padre if it had better C support
dukeleto Matt-GCI: i mostly use vim 02:32
Matt-GCI: with many macros and shortcuts
Matt-GCI and are most of you running linux? 02:34
rfw heh, everyone in gci sure loves parrot
dukeleto Matt-GCI: people use lots of different OS's, but many use linux 02:37
rfw i use lolwindows :(
dukeleto Matt-GCI: Parrot supports linux, *bsd, OS X, windows and solaris and we have developers that use all those
whiteknight: there is a todo'ed test in t/src/embed.t for the IMCC bug that you said you fixed 02:44
whiteknight: does it pass for you?
whiteknight dukeleto: no, my fix doesn't work yet and causes segfaults that I have to debug still
it's a pretty big change, and it's going to take a while to get it working correctly 02:46
dalek rrot/leto/embed_grant: 0694baf | dukeleto++ | t/src/embed.t:
[t] Add an embed test that creates multiple interpreters
whiteknight I'm going to have to solve this all tomorrow. For now, sleep 02:49
02:50 whiteknight left
dukeleto sorear: gist.github.com/727859 <-- this is what happens when you create a new interp without using the 1st as an argument 02:51
dalek rrot/leto/embed_grant: 41ddd30 | dukeleto++ | t/src/embed.t:
[t][TT#1880] Add embed test for using Parrot_new to create multiple interps without chaining off first interp
 02:55
rrot/tt532_headerizer_refactor: 5045bf7 | jkeenan++ | / (3 files):
Refactor one block of code out of function_components_from_declaration(). Place in Functions.pm. Test it in 01_functions.t.
 02:56
rrot/tt532_headerizer_refactor: b8cf4e4 | jkeenan++ | / (3 files):
Fix syntax error.
 02:59
sorear dukeleto: I said the requirement was stupid, not that it was misdocumented
dukeleto sorear: yes, i agree 03:02
sorear: do you think it is a small change, or does a lot of stuff need to change to fix that?
sorear dukeleto: I don't know, it's tangled up in the classic thread system 03:06
dukeleto sorear: ok, just wondering if you had tuits to spare 03:07
dalek TT #609 closed by dukeleto++: increased portability
TT #609: trac.parrot.org/parrot/ticket/609
TT #1881 created by dukeleto++: Parrot on Android
TT #1881: trac.parrot.org/parrot/ticket/1881
dukeleto finds a gold mine of details about parrot function signatures that he didn't know about
kid51 dukeleto: A question about GCI: Are student participants explicitly matched up with mentors as they are in GSOC? 03:09
rfw kid51: no, we're just assigned a mentor based on what task we're doing
kid51 Okay, I just want to make sure that each GCI participant who shows up here is matched up with someone from the project. 03:10
dukeleto kid51: tasks have mentors associated with them, then students pick a task 03:13
kid51 We are now under 600 new/open Trac tickets: tinyurl.com/2a9xwyt 03:20
sorear \\o/
kid51 The most recent ticket closed was closed via GCI submission!
sorear Nice. Who? 03:21
kid51 fbrito++
kid51 suspects dalek is once again failing to track Trac ticket closings 03:22
trac.parrot.org/parrot/ticket/1876#comment:6
dalek TT #1876 closed by jkeenan++: Should 'make clean' remove blib/ folder? 03:23
TT #1876: trac.parrot.org/parrot/ticket/1876
kid51 Aha! So dalek is reporting closings; good.
Hope to see everyone at 2300 UTC Sunday for Parrot Developer Summit 03:24
Matt-GCI rfw: you're gci too?
rfw Matt-GCI: yeah
you probably won't like me very much once you realize who i am though 03:25
sorear fbrito, rfw, Matt-GCI, ?
rfw :)
Matt-GCI lol rfw why is that 03:26
rfw you'll see
Matt-GCI high up in rankings?
dalek rrot: e0a2fb8 | dukeleto++ | README:
Fix some of the README git instructions
rfw i think high is an understatement
Matt-GCI daniel...?
rfw hell no
lol
kid51 dukeleto: Any progress on RTEMS? 03:27
Matt-GCI fernando?
rfw Matt-GCI: might want to refresh that list
no, fernando's fbrito
he's not here though
he was ._.
Matt-GCI refreshed, and tony young is at top
rfw yeah
Matt-GCI hey
rfw hi 03:28
i take it you're matt rajca
dukeleto Rubinius just got a bytecode verifier: github.com/evanphx/rubinius/commit...ae74eb980a
Matt-GCI why would i not like you? im not gunning for first lol
dukeleto kid51: no
rfw oh i guess that's fine
Matt-GCI and hell no, im not even high up
rfw i just want top 10
sorear how many folks are doing gci?
rfw ah
matt habel?
Matt-GCI ~235 right now
(students doing gci)
and idt im in the rankings since i havent completed tasks yet
im doing this for fun 03:29
rfw oh
i want to go to google ._.
they have slides!
Matt-GCI plus im a senior, im busy as
rfw dem slides!
my school is over for now 03:30
Matt-GCI wtf, why?
rfw so i have a ton of free time
www.hoax-slayer.com/google-office-p...aphs.shtml <-- have you never seen this before?
dukeleto kid51: i've been concentrating on other things, but I would like to concentrate on RTEMS again soon
rfw oh but that's in zurich
Matt-GCI why dont u have school? 03:31
rfw i'm not american
Matt-GCI not fair
rfw :)
Matt-GCI where do you live? 03:32
rfw new zealand
Matt-GCI lucky 03:34
rfw not really
place is boring as hell
Matt-GCI american college process sucks
it's extremely competitive
rfw yes but most american colleges don't
Matt-GCI true
my top choice is carnegie mellon 03:35
rfw ah
Matt-GCI falls under the not-suck category lol
sorear is in the US, 20, and doesn't understand the college (uni0 process at all 03:36
Matt-GCI lol sorear are u in college? 03:37
so rfw how'd you gain the experience that's gotten you to top rank? 03:38
rfw i'm just as stumped as you are
Matt-GCI as in what programming experience do you have 03:39
rfw oh
i can write a bunch of things
dalek TT #1844 closed by dukeleto++: Add 'git' instructions to README
TT #1844: trac.parrot.org/parrot/ticket/1844
rfw python, C, C++, javascript, that parrot-nqp language and other stuff
Matt-GCI cool 03:40
what's your current gci task?
sorear Matt-GCI: community college, yes
rfw none yet
still huntin
Matt-GCI do projects add more tasks as they're completed? 03:41
rfw yeah
Matt-GCI cool
if it's any consolation ill root for you in the rankings lmao 03:42
rfw lol
Matt-GCI what program have u used so far to do coding edits? 03:43
rfw um 03:44
sometimes vim
sometimes notepad++
depends where i'm doing it i guess
Matt-GCI hmm
**waits while emacs compiles...
rfw enjoy your lisp interpreter
Matt-GCI im doing the cocoa gui version 03:45
rfw enjoy your lisp interpreter with cocoa* 03:46
Matt-GCI lol
something against lisp?
rfw (no (not (really)))
Matt-GCI if it has a nice gui and gets the job done, works for me lol 03:47
dukeleto so who needs more GCI tasks? 03:48
Coke belatedly answer "plain vi with no macros".
<redacted>
rfw dukeleto: what are you adding :D 03:51
dalek tracwiki: v73 | jkeenan++ | ParrotQuotes 03:55
tracwiki: trac.parrot.org/parrot/wiki/ParrotQ...ction=diff
dukeleto rfw: i have some tasks created, but i can't seem to approve any tasks now. Melange is borked. 03:56
rfw :(
has melange ever been not borked 03:57
03:57 jsut_ joined
dukeleto rfw: no 03:57
kid51 I recommend everyone read whiteknight's paper on Parrot Embedding API prior to Sunday's Summit meeting: tinyurl.com/2arp469
dalek rrot/leto/embed_grant: b6adfd6 | dukeleto++ | docs/embed.pod:
Note that spaces are not allowed in function signatures
 03:58
rrot/leto/embed_grant: 5edba45 | dukeleto++ | docs/embed.pod:
Improve example and reorder some POD
sorear Coke: nvi? 03:59
04:01 jsut left
dalek rrot/leto/embed_grant: 1fafcb6 | dukeleto++ | t/src/embed.t:
[t] Add an embed test for returning numerics
 04:09
dukeleto cotto_work: ping 04:11
cotto_work: also, why are you still at work? 04:12
04:25 kid51 left
Matt-GCI so even though the task says to rename functions in sub.c, i still have to rename the references to those functions in other files, right? 04:48
im gonna take that back because i just realized how stupid that sounded... it's been a long night 04:50
04:51 preflex left
dukeleto Matt-GCI: yes, you have to rename all occurences of those functions 04:54
Matt-GCI: "make test" would have answered that for you :) 04:55
Matt-GCI as i hit the last lines of configure i realized how much coffee had failed me lol
04:55 preflex joined
Matt-GCI brain function = low 04:55
dukeleto Matt-GCI: sometimes asking a question makes you realize the problem. Don't be afraid to ask us questions :)
Matt-GCI ok thanks :)
make passed w/o errors 05:17
time for make test :)
"All tests successful, 15 tests and 587 subtests skipped." 05:21
woohoo
05:36 theory left
Matt-GCI anyone still here? i have a question 05:47
05:48 preflex left, Matt-GCI left 05:51 preflex joined 06:13 preflex left 06:15 preflex joined
dukeleto Anybody have any book recommendations that I should read? 06:16
dukeleto is trying to use up an amazon gift card
06:40 wagle left 06:44 silug left 06:46 wagle joined
cotto dukeleto, pong (and I'm not) 06:59
I suppose cotto_work could pong you, but he wouldn't be happy about it. 07:01
dalek TT #1874 closed by dukeleto++: ports/ directory is listed in .gitignore, can't commit files to it 07:10
TT #1874: trac.parrot.org/parrot/ticket/1874
rrot: d2c3de0 | dukeleto++ | docs/project/git_workflow.pod:
[t] Add some info about git add --force
 07:11
dukeleto cotto: i was wondering if you could login to the gci site and see if you have to do something to a task i assigned to you before i can approve it 07:12
cotto: i created a task and assigned to you, but i can't approve it
cotto: i am thinking maybe you have to do something first
aloha msg moritz what happen to github.com/moritz/Date ? I mentioned it at trac.parrot.org/parrot/ticket/1650 but now it is gone 07:13
aloha dukeleto: OK. I'll deliver the message.
cotto dukeleto, doing so now 07:14
dukeleto, which task? 07:15
07:17 bacek joined
cotto PL/Perl6? 07:17
I don't have an approve button for that either
dukeleto cotto: darn. Melange is borked. 07:18
i can create tasks, but can't approve them so that students can take them
cotto srsly?
melange--
and melange-- again for being designed by people too smart to know that they're not very good at UIs. 07:19
dukeleto cotto: sigh. 07:26
cotto: I have been dealing with melange for 3 years now. It gets *slowly* better.
cotto dukeleto, is the ui that was previously used to approve projects now gone? 07:28
dukeleto cotto: there were 2 radio boxes that used to be on tickets and now they are gone 07:30
cotto: i poked some google people about it 07:31
cotto dukeleto++ 07:32
don't let me not buy you a beer next time I see you
dukeleto cotto: i will surely not let you don't do that 07:36
cotto: i did some TT wrangling
cotto: we are under 600 nowadays. Let's try to keep it that way
cotto: i feel like we need to just reject a lot of old tickets that are too vague, no longer apply, or both 07:37
cotto dukeleto, +1 07:40
dalek rrot: 72908f9 | doughera++ | / (5 files):
Make config/gen/platform/generic/sysmem.c more generic.

Include the two most likely methods (sysconf and sysctl) in the generic file, based on various #ifdef probes. This way, new systems are more likely to work without first needing their own sysmem.c file.
If neither sysconf() nor sysctl() seems to work, assume a default of 512 MB. One could certainly argue for putting an error message here instead.
Signed-off-by: Jonathan "Duke" Leto <jonathan@leto.net>		 07:41
cotto After last night, my minimum usefulness threshold for keeping tickets around raised quite a bit.
dukeleto cotto: yeah. I am sick of passing over old tickets that never change. 07:43
cotto one down 07:45
dukeleto cotto: i have found a middle ground with signing off on stuff. If someone gives you a patch generated with format-patch, it is just one command to sign off on a patch
git am --signoff foo.patch
cotto that's sensible 07:47
dukeleto cotto: and i will write a tool for making a patch from a branch. it is easy to automate
cotto we don't need diffs that originate with committers to be signed off
thanks
dukeleto cotto: but i have so many parrot projects going at once, i need to re-prioritize what is most important
cotto I guess that's a good problem to have. 07:50
dukeleto cotto: i guess.
cotto Maybe you can find some more community management volunteers and delegate some work to them. 07:51
dukeleto cotto: i agree.
cotto: what are your priorities? 07:54
07:54 fperrad joined
dukeleto cotto: hacking on a new MOP is getting pushed to the top of my list 07:54
cotto dukeleto, you're working on that too? 07:55
great!
Is it the same as what jnthn is doing? 07:56
dalek TT #1843 closed by dukeleto++: NetBSD fails compiling sysmem.c 07:58
TT #1843: trac.parrot.org/parrot/ticket/1843
TT #1026 closed by cotto++: Can't use } in TGE transform definitions.
TT #1026: trac.parrot.org/parrot/ticket/1026
dukeleto cotto: we are basically going to use what jnthn has a baseline, and see what parrot will need to tweak. I don't think it will be much 07:59
cotto dukeleto, is t/pmc/mappedbytearray.t not running any tests for you? 08:00
dukeleto cotto: jnthn said he has a 'nom' branch that adds MOPish things to nqp-rx
cotto I have difficult expressing how great of a name that is. 08:01
dukeleto cotto: it runs 4 tests on my machine
cotto y
(QED)
dukeleto cotto: we need to unblock Ruby and Javascript on Parrot development. That is my goal with our new MOP 08:04
cotto dukeleto, I fully agree. We're not much of a VM for dynamic languages if two of the big ones can't work with our object model. 08:05
different though they are
dukeleto, I committed a couple local changes and checked out a sha1 from before my changes took effect. How do I get back those commits? 08:12
nm
checkout master did it
dalek rrot: 6acd7b2 | cotto++ | src/pmc/class.pmc:
[pmc] remove unneeded use of PARROT_IN_OBJECTS_C
 08:15
rrot: 6702284 | cotto++ | / (2 files):
[oo] remove some unneeded #defines
sorear how important is it that parrot languages use the parrot Object and Class PMCs?
cotto sorear, why do you ask? 08:17
are you thinking that it might be a better idea for differently-objectified languages to roll their own? 08:18
sorear that's part of it
also I'm thinking in terms of porting niecza to parrot, which already has some rather peculiar ideas of how a class should behave 08:19
dukeleto sorear: please send an email to parrot-dev about what you would need from a meta object protocol for niecza 08:23
dukeleto takes it easy
sorear takes it easy?
also I'm pretty sure I'll need to try before I know 08:24
it's not something that a MOP can provide
a MOP lets you manipulate classes. I need to create classes with different behavior
niecza's classes implement a very odd form of vtable overrides 08:25
they keep methods in an ordered dictionary
dalek TT #892 closed by cotto++: PARROT_IN_OBJECTS_C used only in one file 08:30
TT #892: trac.parrot.org/parrot/ticket/892
TT #550 closed by cotto++: t/pmc/codestring.t failing 08:46
TT #550: trac.parrot.org/parrot/ticket/550
moritz dukeleto: re Date, it's now in core 08:53
09:06 contingencyplan left
dalek TT #1472 closed by cotto++: Namespace.get_pointer_keyed is TOTAL KLUDGE 09:19
TT #1472: trac.parrot.org/parrot/ticket/1472
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1495) fulltest) at 6702284 - Ubuntu 10.10 i386 (g++-4.5) 09:24
09:24 rfw left 09:39 preflex left 09:42 preflex joined
bacek aloha 09:56
barely here
When is PDS?
aloha, pds?
aloha bacek: Dunno.
bacek stupid girl...
aloha, developer summit? 09:57
aloha bacek: developer summit is www.doodle.com/q4nwdt9i5k929n2s
sorear bacek: sunday iirc 09:58
ah
bacek indeed. Looks like Monday morning from my side :) 09:59
aloha, time
clock?
aloha, clock?
aloha bacek: bacek: LAX: Sat, 01:59 PST / CHI: Sat, 03:59 CST / NYC: Sat, 04:59 EST / UTC: Sat, 09:59 UTC / LON: Sat, 09:59 GMT / BER: Sat, 10:59 CET / TOK: Sat, 18:59 JST / SYD: Sat, 20:59 EST
bacek afk # back to kids 10:00
10:38 lucian joined 11:12 plobsing left, plobsing joined 11:22 bacek left 12:40 cognominal left 12:44 cognominal joined 13:18 kid51 joined 13:27 preflex left 13:31 fbrito joined, preflex joined 13:32 fbrito1 joined 13:33 fbrito left 13:34 fbrito1 is now known as fbrito 13:42 fbrito left, Psyche^ joined, Patterner left, Psyche^ is now known as Patterner 14:22 whiteknight joined
whiteknight good morning, #parrot 14:25
14:28 Kristaba joined 14:29 bluescreen left 14:39 bluescreen joined 14:44 cognominal left 14:46 preflex left, cognominal joined 14:50 preflex joined 14:53 spinclad left
kid51 whiteknight good morning 15:03
whiteknight: I like the Embed API doc, though much of it was about issues I don't understand well. 15:10
When I went to re-test in the embed_api2 branch last night, I saw that there had been no further commits, so I didn't re-test.
Is there anything we can do today to prepare for tomorrow's Summit. 15:11
?
whiteknight kid51: I'm working on it right now. I have one nagging error that I am still trying to debug/fix 15:15
I was hoping to have it resolved by last night, but no such luck
kid51 k
15:29 smash joined
smash hello everyone 15:29
15:31 Coke left, Coke joined 15:37 Coke left
kid51 howdy smash 15:38
15:43 Coke joined 16:03 mikehh_ joined 16:07 mikehh left 16:11 GodFather joined 16:18 GodFather left
dalek rrot/tt532_headerizer_refactor: 8ffe773 | jkeenan++ | / (3 files):
[codingstd] Fix trailing whitespace and POD formatting errors.
 16:33
dukeleto 'ello 16:54
17:08 Matt-GCI joined
Matt-GCI g'morning everyone 17:09
whiteknight and dukeleto: i finished my GCI task, i just had a few last questions if that's ok 17:10
dukeleto Matt-GCI: wazzup? 17:11
Matt-GCI so i changed all the functions in sub.c to meet the standards, and it compiled perfectly!
dukeleto Matt-GCI: awesome! 17:12
Matt-GCI but i wanted to make sure i didn't change any private functions
dukeleto I would appreciate comments on this new RFC ticket: trac.parrot.org/parrot/ticket/1882
Matt-GCI: where is your diff? did you make a github pull request?
Matt-GCI i havent made the github pull req yet because i have a question about that too
i have my own fork of parrot at github.com/schallert/parrot that i'm working off of 17:13
but i accidentally made my changes in the master branch of my fork, is that ok?
17:14 Coke left
Matt-GCI (i haven't pushed back to my git fork yet, ill do that now so u can see) 17:15
dukeleto Matt-GCI: that is fine 17:17
Matt-GCI ok i just pushed
dukeleto Matt-GCI: make a pull request 17:18
Matt-GCI ok, i need to change one last little thing then i will
17:20 Coke joined 17:21 bluescreen left
dalek TT #1882 created by dukeleto++: gethostbyname failure doesn't throw an exception 17:23
TT #1882: trac.parrot.org/parrot/ticket/1882
Matt-GCI yey! "All tests successful," 17:24
dukeleto: I just made the pull request 17:28
dalek rrot/tt532_headerizer_refactor: 3fbc7cb | jkeenan++ | / (3 files):
Refactor some code into replace_pod_item(). Place that subroutine in Functions.pm and test it.
dukeleto melange still seems to be borked. sigh. 17:35
Matt-GCI ugh
do my changes look correct?
dukeleto Matt-GCI: looking 17:37
Matt-GCI: seems ok, i am running "make test" now 17:39
kid51 Matt-GCI: Was this work associated with a particular Trac ticket? (or just a GCI task?) 17:41
Matt-GCI a gci task that accomplished part of a larger ticket 17:42
to update all functions to standards in general
kid51 Do you recall which Trac ticket?
dukeleto kid51: it is a gci task, not a TT, i don't think
kid51 dukeleto: That's okay, I'm just wondering where those standards are documented, so I know what I'm "proofreading" against. 17:43
Matt-GCI kid51: heres the gci task: www.google-melange.com/gci/task/sho...9035234084 17:44
dukeleto kid51: whiteknight would know
17:45 zarchne joined
kid51 Matt-GCI: got it 17:46
dukeleto Matt-GCI: i am getting a PASS for "make test". Nice work!
Matt-GCI dukeleto: thank you! 17:47
dukeleto Matt-GCI: you need to do the "submit work" action for me to approve it
kid51 Matt-GCI Can you fix space alignment at line 108 src/ops/core_ops.c ?
Matt-GCI sure 17:48
dukeleto no 17:49
Matt-GCI ok, not doing
dukeleto Matt-GCI: i already merged your pull request
dalek rrot: 6233b6e | (Matt Schallert)++ | / (12 files):
Changed functions in sub.c to meet standards

Signed-off-by: Jonathan "Duke" Leto <jonathan@leto.net>
Matt-GCI what url do you want me to submit in melange?
rrot: 6aa336c | (Matt Schallert)++ | / (8 files):
Finished sub.c function renaming

Signed-off-by: Jonathan "Duke" Leto <jonathan@leto.net>
dukeleto Matt-GCI: doesn't matter, it is done already. Do they require a URL? 17:50
Matt-GCI: just put the one for the pull request, if they require something
kid51: we might need some deprecation wrappers for some of those functions, I am not sure 17:51
kid51: again, whiteknight is in charge of that
kid51 dukeleto: okay. I'll fix that space alignment problem in master.
Matt-GCI ok dukeleto i submitted through melange 17:52
cotto Matt-GCI, did you try to build Rakudo and some other HLLs?
dukeleto kid51: thanks, I appreciate it
cotto: i don't think that was in the task, although it would be nice
Matt-GCI: i marked your task as done, you can claim another task now. Good luck! 17:53
Matt-GCI thank you!
kid51 Hmm, running 'make bootstrap-ops', I'm getting a failure: 18:08
./parrot -o runtime/parrot/library/Regex.pbc ext/nqp-rx/src/stage0/Regex-s0.pir
error:imcc:No such file or directory
in file 'ext/nqp-rx/src/stage0/Regex-s0.pir' line 28
make: *** [runtime/parrot/library/Regex.pbc] Error 1
dukeleto kid51: make realclean 18:10
kid51 'make' succeeded 18:11
make realclean && perl Configure.pl && make bootstrap-ops failed with same error 18:13
dukeleto kid51: something is failing before that error 18:16
kid51: and why are you running bootstrap-ops by itself? perhaps that was broken already
kid51 I was exploring that column alignment problem cited above. 18:17
cotto Rakudo builds with Matt-GCI's changes. 18:18
dukeleto kid51: are you talking about whitespace?
cotto: sweet.
kid51 Yes.
I have to go now, but I opened a ticket (1883)
cotto bootstrap-ops worked for me just now 18:23
kid51, you might need make before make bootstrap-ops 18:26
istr the dependencies being lta
dukeleto perhaps melange unborked itself. I figured out how to approve tasks.
cotto: yeah, seems like it
dalek TT #1883 created by jkeenan++: 'make bootstrap-ops' target broken 18:27
TT #1883: trac.parrot.org/parrot/ticket/1883
cotto goes shopping
dukeleto just made this task for the Code-In students that are complaining that stuff is too easy: socghop.appspot.com/gci/task/show/g...9141853072 18:29
18:31 cognominal left 18:34 cognominal joined 18:36 Matt-GCI left 18:50 contingencyplan joined 19:14 s1n joined
moritz pointed to git-bisect as a comment 19:19
19:30 diakopter joined
diakopter hi; in a parrot-nqp backtrace, 'called from Sub xxxx pc nnnnn' what does nnnnn represent? 19:30
whiteknight that's a very good question 19:32
moritz example? 19:33
jnthn offset into the bytecode segment, iirc.
moritz ouch.
jnthn It's practically useless.
Unless you're debugging bytecode generation.
Or can get a disassembled output of the bytecode. 19:34
diakopter a HLL::Compiler subclass... that has many many by-name overloaded Subs
(so the Sub name isn't helpful)
sorear didn't pbc_disassemble work on parrot-nqp output at one point
diakopter all I have is a bunch of *.pir :) /me looks at jnthn 19:35
jnthn diakopter: wtf, it didn't give a PIR line number too? 19:36
diakopter current instr.: 'parrot;DNST2CSharpCompiler;cs_for' pc 143163 (gen_grammar.pir:8394)
called from Sub 'parrot;DNST2CSharpCompiler;cs_for' pc 140785 (gen_grammar.pir:7800)
so there are two bytecode offsets
jnthn gen_grammar.pir:8394 <-- this number is a PIR line number.
er but 19:37
diakopter right, but it doesn't appear to have anything to do with those cs_for subs
jnthn ...the file name and the...right...wtf. *sigh*
Can steal the Rakudo backtrace printer, or a simplified version, I guess. 19:38
diakopter really the thing that annotates 'parrot;DNST2CSharpCompiler;cs_for' might be brillianter about annotating the signature too 19:39
jnthn It uses the HLL line/file debug segment which is usually a bunch more accurate.
sorear so imcc debug info is still bogus? 19:40
I thought that was fixed
diakopter what's that
sorear debug info generated by imcc?
diakopter fixed as of when? this is parrot 2.9.1 19:41
er
yes
jnthn I'm sure it used to be more accurate at some point too. :| 19:43
diakopter sorear: but I don't have much of an idea where in the compilation chain imcc fits... so I don't know whether your question applies
19:44 Matt-GCI joined
diakopter I suspect it's the nqp compiler that's inserting annotations 19:44
oh, hm 19:48
no, that can't be it (was imagining the line number was correct but filename wrong)
19:49 rfw joined
diakopter unless the .pir are actually concat in memory somewhere... 19:49
19:56 whiteknight left
lucian diakopter: line numbers are just totally bonkers 19:57
diakopter ok
19:57 kid51 left
diakopter I can accept that 19:57
:) 19:58
(I don't need them to be fixed, long term)
lucian diakopter: it may improve with time, there's an alternative to imcc (pirc)
20:02 smash left
diakopter lucian: what is pirc 20:06
lucian diakopter: another pir compiler
20:08 bluescreen joined
Tene been in progress or stalled for years. 20:49
20:55 zby_home joined 20:57 Matt-GCI left 21:01 bacek joined 21:11 bacek left 21:16 gbacon joined
Kristaba Maybe someone can help me about arrays in NQP? 21:25
I would have a NQP variable which PMC type is "array" (return true to VTABLE_does([...], "array"); )
So, if I do something like "my $var := (2.0, 1.0);" PMC type of var will be "array" or not?
cotto That pmc "does" array. "array" a role. 21:26
it'll probably be a ResizablePMCArray at the PIR level.
Kristaba Ok, thank you cotto :) 21:27
cotto np. Are you just curious or is there a practical application?
lucian cotto: so does that work mixin-style?
Kristaba cotto: I work on a Parrot Linear Algebra task for the GCI 21:28
cotto I think you could get it to work like Ruby's mixins, but at the PMC level a role is just an arbitrary string that has meaning only by convention. 21:29
That's a lie. 21:30
The thing returned by VTABLE_does is an arbitrary string that has meaning by convention. Roles are separate thing.
lucian cotto: so more like python protocols 21:32
good
21:37 mikehh_ is now known as mikehh
cotto Confusingly, there's also a "does X" pmclass modifier that's supposed to have something to do with roles. I don 21:37
mikehh opbots, names
cotto 't think anything uses it, so it's probably extravagantly broken. 21:38
The "provides X" modifier is the one that adds "X" to what's returned by VTABLE_does. 21:39
mikehh cotto: lost track of things - when is the Dev meet, (here or #ps)
cotto It's kind of a miss.
mess
It's in 25h 40m 21:40
20m
in #ps, though I don't recall if that was publicized properly 21:41
mikehh well should be around then :-}
cotto nm. whiteknight mentioned that it's in #ps in the message he sent to parrot-dev 21:42
mikehh cotto: 'k, seem to have lost, or not getting messages from parrot-dev 21:43
cotto www.mail-archive.com/parrot-dev@lis...05021.html 21:44
mikehh cotto: think I saw that one (by looking at the archives) - do we have an agenda? 21:46
lucian cotto: i see. sad that it's a mess 21:50
cotto mikehh, kid51 posted something. whiteknight responded within the last few days with a summary of his api work. 21:51
lucian, yeah but this is exactly the kind of thing Lorito will help with. There, we'll only have one way to interact with the object model rather than C and PIR. 21:52
cotto goes afk
lucian cotto: i don't think that's the only issue. the object model itself needs to be carefully designed to be able to efficiently compose all the behaviours HLL object systems need 21:53
like single vs multiple dispatch
22:18 preflex left 22:19 bacek joined 22:20 preflex joined 22:30 preflex left 22:32 preflex joined 22:34 fperrad left 22:43 zby_home left
sorear diakopter: NQP[String] -(NQP::Grammar and NQP::Actions)-> PAST[PCT::Node] -(PAST::Compiler)-> POST[PCT::Node] -(POST::Compiler)-> PIR[String] -(IMCC)-> PBC(struct Packfile_Segment) 22:45
diakopter ok. nqpclr isn't using that last step (to pbc), at least to disk.
Kristaba Someone know when Whiteknight come back? 22:46
sorear Kristaba: how long have you been waiting? 22:49
diakopter: imcc is the only way to load PIR that currently works
(except pirate, but that doesn't count since it's dependant on imcc) 22:50
diakopter I just meant (to pbc), at least to disk
22:54 rfw left
Kristaba sorear: I'm waiting from few time, but I'd like to know when I'll probably see him ;) 22:59
sorear "morning" or something like it 23:03
23:04 rfw joined
Kristaba sorear: Yes, I guess, but as I live in France, my "morning" isn't necessary his "morning" (if he live in US, Australia, etc...) 23:08
sorear right. 23:09
also we have a couple people who do most of their IRCing after work 23:10
so really it's better just to wait 24 hours
Kristaba Yes, you're right 23:12
23:20 rfw left 23:31 rfw joined 23:40 rfw left 23:44 rfw joined 23:57 theory 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!