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 28 November 2010.
fbrito (or a task to create more unit tests) 00:02
00:10 fbrito1 joined 00:14 fbrito left
whiteknight bluescreen: pong 00:14
bluescreen hey... good evening
I'm setting up a Gist to track embed_api's TODO list 00:15
so far this is what I've got gist.github.com/719422 00:16
we would need some cleanup phase 00:17
00:19 theory joined
whiteknight bluescreen++ 00:20
dukeleto: you can comment on that pull request too
bluescreen i had to do a nasty debug print since the test case rely on that
whiteknight urg
bluescreen "*** Parrot VM: slow core ***", hate doing that 00:21
so at some point we need to sit down and define what kind of debug information we would put in the api
if what I'm testing is OK, we will have only one failing test case 00:22
the profiling one
BTW IMCC_warnings ommits warnings when you add the "-w" switch and prints them we you remove that switch 00:23
whiteknight wtf? 00:24
stupid IMCC
bluescreen yeah...
try this... create some junky pir and run ./parrot -w /tmp/junk.pir
and without -w
whiteknight ok 00:25
bluescreen or read compilers/imcc/debug.c: IMCC_warnings
i was like.... WTF++
whiteknight yeah
bluescreen I've talked to plobsing about it, he told me to leave it like that 00:26
can you take a look at the t/profiling/profiling.t test case ? 00:28
whiteknight i might be able to
bluescreen I've no clue what's going on there
thats the only test case failing at this point
i think the problem is with ProfTest.pbc 00:29
either is not able to read it or its compiled for another version of libparrot.so
whiteknight ok
src/exceptions.c:die_from_exception() is printing the error and backtrace info out to stderr 00:30
we need to take out that behavior and move it into the API
bluescreen yeah 00:31
let me write that down into the TODO
also imcc/main.c has a nice fprintf(STDERR,...) in it 00:32
whiteknight the embedding application should be in charge of where/how to display error information, not libparrot
bluescreen new funct? 00:33
I don't think we would make it for 2.11 00:34
00:39 kid51 joined
whiteknight stupid IMCC 00:44
bluescreen PARROT_PROFILING_NAME=/tmp/test.pprof ./parrot --hash-seed=1234 --runcore profiling /tmp/test.pir
that is what is causing the last test case fail
--hash-seed=1234 is making parrot.exe blow
whiteknight that's in the profiling test? 00:45
bluescreen yes 00:46
PIRProfile.nqp:84
I've removed the hash-seed option and it passes 7 out of 13 test cases 00:47
whiteknight damnit. That must be the only test that covers the --hash-seed option 00:52
bluescreen for some reason is failing in void * const ptr = malloc(size); // where size is 4096 00:54
whiteknight ... 00:55
bluescreen parrot: malloc.c:3097: sYSMALLOc: Assertion 00:56
whiteknight sounds like a mem corruption thing\\ 00:57
bluescreen do you know how to use valgrind? 01:00
whiteknight I've used it a few times. Probably need a refresher 01:01
bluescreen haven't used in my life... 01:02
whiteknight ah, so you can't teach it to me :) 01:09
bluescreen nop.. sorry just run it... I feel like reading the matrix :) 01:10
fbrito1 i used valgrind once (cachegrind for benchmarking) 01:11
with c++ 01:12
all I can remember is that there is a tool called "kcachegrind" that generates beautiful graphs of called functions 01:13
and that cachegrind warns you if you forget to free something 01:14
whiteknight yeah, that's what I've done with it. Generated a profile and used kcachegrind to view it 01:15
I'm trying to figure out why all the test snippets are trying to create "Interger" objects 01:19
bluescreen i think i go it 01:20
main.c: initargs->hash_seed = strtoul(arg, NULL, 16);
while api.h: Parrot_Int hash_seed 01:21
whiteknight so hash_seed should be an unsigned long instead of a Parrot_Int? 01:22
bluescreen in interpreter.h is an UINTVAL
whiteknight testing now
bluescreen atoi? 01:24
whiteknight: it think changing hash_seed in parrot/api.h will do the trick 01:32
valgrind helped in some way
whiteknight I changed it to unsigned long, not fixed 01:38
bluescreen UINTVAL is already long i guess 01:40
ok... time to go. I'll leave that commit to you my friend :) 01:42
whiteknight ok 01:44
01:50 shockwave joined
shockwave ping @whiteknight 01:51
whiteknight pong
how are you shockwave? I wasn't expecting you till next week
Coke I am bemused that the cross platform "cmake" is failing to build on my os x box. 01:52
shockwave Yep. I was just dropping by casually.
whiteknight shockwave: oh, okay.
shockwave I was just curious if the API stuff was going smoothly?
whiteknight the embedding API work is going very well and is almost 100%. We're still failing some tests with the profiler, but that's a small issue
shockwave awesome 01:53
So you feel that by next Monday it should be in a usable state?
whiteknight easily
shockwave That's great! 01:54
01:56 particle left
whiteknight yeah, I'm excited about it 01:58
cotto: ping 01:59
shockwave Cool. I'll drop by next Sunday. Maybe we can talk about the change that were made a bit, and then on monday I can start the embedding process.
whiteknight that will be awesome 02:00
Sunday evening there is a big meeting of Parrot devs
so I will definitely be online for that 02:01
and before, most likely
shockwave About what time?
6ish?
whiteknight I think 8PM EST is the meeting 02:02
that won't affect our meeting
shockwave Cool. Do you mind if I log on a little before that. If you're online, maybe we can talk then? 02:03
Or afterwards, if you preffer.
I mean, I'll probably eaves drop on the Parrot meeting. I was refer to the API talk. 02:04
Ok. g2g. See you, guys. 02:06
02:06 shockwave left 02:18 bluescreen left, jsut_ joined 02:23 jsut left
sorear What was the valgrind option to display executed instruction counts again? 02:40
02:40 bluescreen joined
bluescreen whiteknight: I just push a commit to fix the profiling test... all test pass now 02:41
dukeleto sorear: --dump-instr=yes ? 02:42
02:44 bluescreen left
dalek rrot/tt532_headerizer_refactor: c998b1e | jkeenan++ | / (5 files):
To demonstrate all execution paths, encapsulate some code within

Parrot::Headerizer::Functions and add tests for them.		 02:45
whiteknight bluescreen++ I'll check it out tomorrow 02:48
bed time now
sorear dukeleto: thanks 02:50
02:53 whiteknight left
dalek rrot/tt532_headerizer_refactor: 2aa5bc9 | jkeenan++ | / (2 files):
Add tests for Parrot::Headerizer::Functions::shim_test().
 03:04
03:07 bacek left
dalek rrot/leto/embed_grant: 2c941c4 | dukeleto++ | docs/embed.pod:
Add a 'nothing in, nothing out' type signature example
 03:19
cotto ~~ 03:30
dukeleto cotto: wazzup
cotto leftovers
<3 leftovers 03:31
cotto wonders what whiteknight was pinging him about
I guess we'll never know. 03:32
03:32 kid51 left 03:39 jasonmay left, jasonmay joined 04:26 davidfetter joined
dukeleto nope. 05:17
05:19 rurban_ joined 05:23 rurban left, rurban_ is now known as rurban 06:15 silug left 06:17 fbrito1 left
cotto msg whiteknight pong 06:31
aloha OK. I'll deliver the message.
06:34 fbrito joined 07:11 jsut joined 07:15 jsut_ left 07:49 fbrito left 07:54 davidfetter left
dalek rrot: 51cfb38 | jimmy++ | config/gen/makefiles/root.in:
apply patch from TT #1865, it allows more windows environments with git to work correctly, ronaldws++
 07:55
rrot: 6138a06 | jimmy++ | / (229 files):
Merge branch 'master' of github.com:parrot/parrot
07:58 lucian joined 08:02 jsut_ joined 08:07 jsut left 08:11 fperrad joined
dalek TT #1865 closed by jimmy++: Allow more windows environments with git to work correctly with build ... 08:11
TT #1865: trac.parrot.org/parrot/ticket/1865
moritz oh hai 08:20
aloha: msg bacek that rakudo on parrot RELEASE_2_10_1-525-g8aa40c9 just had a clean spectest run - will retest to verify, but looks good so far. Thanks! 08:21
aloha moritz: OK. I'll deliver the message.
08:37 theory left 08:42 bacek joined
dukeleto moritz: good to hear 08:44
08:44 mj41 left, mj41_ joined, mj41_ is now known as mj41
dukeleto mj41: howdy 08:45
mj41 dukeleto: winter mood 08:46
dukeleto mj41: i know it well.
mj41: how hard would it be to get TapTinder working with git? 08:47
mj41: people miss the taptinder bot
mj41 dukeleto: I know. I have busy months in the work. But I'm back probably. 08:48
dukeleto mj41: if you can describe what needs to be done, others can help you
mj41: wait! github has svn access! 08:49
mj41: we just need to point taptinder at github svn mirror of parrot
mj41 dukeleto: I'm familiar with Git a bit now. So next step is to change DB schema and get repo-to-db script working for Git. 08:50
dukeleto mj41: it is a lot easy and a lot less work to point ttbot at the Github svn URL 08:51
mj41 dukeleto: Ok. Will look at it too as fast temporary solution. 08:52
dukeleto mj41: svn checkout svn.github.com/parrot/parrot.git
GeJ dukeleto: I just read that like a weird DC-Marvel crossover. 08:56
dukeleto GeJ: i know it looks crazy, but it actually works 08:59
bacek seen moritz 09:29
aloha moritz was last seen in #parrot 1 hours 7 mins ago saying "msg bacek that rakudo on parrot RELEASE_2_10_1-525-g8aa40c9 just had a clean spectest run - will retest to verify, but looks good so far. Thanks!".
bacek msg moritz, I suspect that I slowed down parrot at least by 10%. Can you compare performance? 09:30
aloha OK. I'll deliver the message.
dalek TT #1871 closed by moritz++: Rakudo randomly segfaults (probably GC related) 09:32
TT #1871: trac.parrot.org/parrot/ticket/1871
bacek aloha, (200-124)/200*100 09:45
aloha bacek: 38
bacek moritz, nm, looks like it's still about 35% faster on building core.pir in rakudo. 09:46
moritz \\o/
bacek moritz, testing on different box is highly welcome :) 09:54
mikehh bacek: what are you testing on? 10:04
bacek mikehh, linux/ia32/2GB
mikehh bacek: 'k, I'm on that at the moment - Ubuntu 10.10 i386, but will switch to amd64 later 10:08
bacek mikehh, ok. I tested HEAD vs b23259670 (just before merge)
10:09 cotto left 10:12 lucian_ joined
lucian_ i've been reading trac.parrot.org/parrot/wiki/HLLInteroperability. it doesn't look up to date 10:21
fperrad msg bacek your fix in default.pmc saves lua. but there are still many segfault with the full test suite (./setup.pir test). 10:27
aloha OK. I'll deliver the message.
bacek fperrad, sigh...
fperrad, can you add "+1" to trac.parrot.org/parrot/ticket/1875? I'm tired of "constant" pmcs. 10:30
Tene lucian_: No, it's not up to date. 10:35
lucian_ Tene: good, because it has some scary assumptions 10:36
Tene lucian_: as far as I'm aware, since I stopped pushing HLL stuff about a year ago, I haven't seen anyone else do anything at all on it. 10:39
And, even then, I felt like I was the only person who cared about it.
It would be great to have someone new pushing HLL interop. :)
lucian_ Tene: I'm still trying to figure out what's missing from calling ruby AND common lisp stuff from, say, python
so far as I can tell, the basic PMCs assume you're using a mostly imperative language with single-dispatch OO bolted on 10:40
Tene lucian_: That's worked to at least some degree about three times in the past.
lucian_ if you deviate from that, you add to much for the PMCs to be useful for interop
Tene: oh, and mutable data structures]
if you try to implement haskell or clojure on parrot, using return values from them would be very, very hard 10:41
Tene: this isn't fully clear to me, but so far I think in the first place, there should be both mutable and immutable versions for most (all?) data types 10:42
that way, HLLs can more directly use PMCs
Tene lucian_: Sure. 10:43
lucian_ so a Python tuple would be an ImmutableBlahArray
it could even specialise at creation, since it's immutable 10:44
and afaik, there's no linked list implementations. that also seems like a basic thing HLLs might want
Tene Sure.
The specific choice of PMCs to ship with Parrot isn't actually relevant to HLL interop, fwiw. 10:45
lucian_ Tene: how so? wouldn't thinner wrappers over PMCs help interop?
Tene: in any case, HLL interop is a really hard problem 10:48
moritz a common basic MOP would help
lucian_ Tene: it sort of, almost works on CLR and JVM because everyone works around the Java/C# types
moritz: but only to a point. what about languages where there are no objects? 10:49
Tene lucian_: It may be useful to interop, but you'll never have a strict subset of all languages fully implemented in core parrot.
moritz lucian_: all languages have objects, in some sense 10:50
Tene You're always going to end up having to play by the rules of the language you want to interoperate with.
Languages are going to want to implement their own data structures outsied outside of Parrot core.
etc.
lucian_ Tene: yeah, i know. but the common cases could be much easier
Tene It will *help* an dmake some tasks *easier*.
It won't actually be a solution to any of the hll interop problems, though. 10:51
As long as you're aware of that. :)
lucian_ moritz: it depends on the MOP
Tene: i know it's not a solution, but it seems like a prerequisite to me
moritz: can you design one that allows someone to call methods on a python object from common lisp + clos? and do so without losing too much efficiency in either language? 10:53
Tene lucian_: The issue is, you'll *need* to deal appropriately with cases that fall outside the boundaries anyway, so that's the more important problem, by far. If you happen to also have a nice useful set of core PMCs, that's a bonus. 10:55
moritz lucian_: you can define a base meta class that contains stubs (or maybe default implementations) for the most common operations (add a method, add an attribute, call a method), and hav all languages inherit from that
lucian_ Tene: yes, of course, you need to deal with the myriad of other cases. but i don't think the problem is tractable in the first palce without the "bonus" 10:56
moritz: so the equivalent for parrot's Integer for objects? 10:57
moritz lucian_: I don't understand that question
lucian_ moritz: i'm concerned about the extra indirection, which you'll pay for even when not using code in another HLL
Tene lucian_: There are no nontrivial cases, working with real languages, where only core PMCs will be involved, I assert.
(sorry, intermittently afk
moritz lucian_: well, all languages can chose to work with parrot's integer registers, or chose not to 10:58
lucian_ moritz: if they choose not to, other HLLs can't use their ints
but Integer is extra indirection and there's a cost to pay for it 10:59
moritz lucian_: unless they implement a common low-level API for numbers
lucian_ and my point was that the cost is payed in HLL code in that single language too, not just when calling into other HLLs
moritz well, if a language doesn't use native ints, it's already paying a price
Tene lucian_: realistically, this isn't a problem you're going to solve for users. When you're a user working with two languages, it's up to you to choose how you want to handle it. You can either pay the complexity cost when using foreign objects, or you can convert them into more-comfortable types, but neither of those are problems that have a general solution. 11:00
lucian_ moritz: exactly. but if ints are immutable in the language, you can use unboxed ints much of the time
Tene You are not going to convince me than an M to M type mapping registry, for example, is a sane solution.
lucian_ Tene: no, that's silly
Tene: (m*m mapping) 11:01
Tene lucian_: in previous hll interop discussions, it was suggested more than once. :) 11:02
lucian_ Tene: oh. but it obviously doesn't scale 11:03
Tene lucian_: It was never discussed for *long*, no
lucian_ Tene: i see
Tene lucian_: Mostly, as far as I've been able to tell, Parrot interfering with inter-language objects is not a desirable behavior, and would cause more confusion than it would solve. 11:05
If you're working in python and you run ruby code, you're getting back a ruby object with ruby's methods, etc. If you want to convert it into a similar python object, that's something you do explicitly.
lucian_ Tene: yes. you don't need to convince me of that "Explicit is better than implicit." - The Zen of Python 11:06
Tene There's also the issue of centralization. Making everything go through Parrot is not a reasonable option, afaict. 11:07
If a language wants to offer facilities for managing things like this, it can choose to ship a library that does so. 11:08
import extraperlsupport
or whatever
lucian_ hmm
but then you end up with crappy default interop and M*M work needed for nice support
Tene I'm sorry, but I disagree. 11:09
You end up with *sane* default interop.
If you can think of examples to the contrary, I'd like to hear them.
To be clear, I am *not* convinced that my conclusions here are obviously right. I just haven't been able to find anything better yet, and nobody has really contradicted me. 11:10
lucian_ Tene: I can't think of a better way to deal with it, at least right now
Tene lucian_: I recommend that you work up some specific concrete examples.
lucian_ Tene: that would be helpful 11:11
Tene "I want to use Perl's DBI from Python. What would that look like?"
lucian_ Tene: and I need to have it work from as different languages as possible
Tene "I run some ruby code from python that returns a number; how do I deal with that number object I get back?"
lucian_ Tene: yes, the examples are good. but they should be more general 11:12
Tene "I'm working in a functional language without OO support, but I'm using a library that returns objects I need to call methods on; how do I do that?"
lucian_ perl, ruby and python are quite similar
Tene lucian_: concrete and general examples are both important.
lucian_ Tene: true
Tene Trust me, there are things you can get wrong even between python and perl, etc.
lucian_ Tene: but yeah, by providing sane defaults, you can layer an awesome solution for solving the M*M niceness issue on top 11:13
Tene: yeah, i know. but they're still much more similar between themselves than to, say, OCaml
Tene lucian_: I know. I'd love to see what you come up with. 11:14
Parrot HLL interop is very desirable to me, and I've spent a lot of time thinking abotu and working on it, and I'd love to see it succeed.
lucian_ Tene: don't get your hopes up, i'm probably the wrong person to do this
Tene lucian_: So am I, apparently. :) 11:15
lucian_ Tene: let's hope parrot likes double negatives, then :)
Tene lucian_: In the past, I've never seen any proposal involving converting things between langauges, or anything other than just passing foreign objects straight through, that didn't cause more problems than it solved. 11:16
lucian_ Tene: right
Tene So, I'm not sure what your "providing sane defaults" means here, then. 11:17
I expect we're speaking too abstractly. :)
lucian_ Tene: probably
Tene I'm trying not to get too involved, as I hope to be able to go to bed any minute now. 11:18
lucian_ Tene: ok. sleep tight
Tene I'm up waiting for a third party to deal with a failure for work. ><
'night 11:19
lucian_: If you do come up with something, please notify me here on IRC by name. 11:20
lucian_ Tene: will do
11:59 lucian left
lucian_ i find it weird that documentation everywhere refers to pynie as a serious/usable language :) 12:01
12:02 cotto joined 12:50 cotto left 12:54 cotto joined 12:58 lucian_ left 13:19 rurban_ joined 13:22 rurban left, rurban_ is now known as rurban 13:24 dngor left 13:28 dngor joined 13:42 Patterner left 13:46 whiteknight joined 13:49 lucian joined 13:51 fperrad left 13:52 Psyche^ joined, Psyche^ is now known as Patterner
whiteknight good morning, #parrot 13:54
13:58 fperrad joined
whiteknight msg cotto ping 13:58
aloha OK. I'll deliver the message.
whiteknight doubts the efficacy of using aloha to send ping/pong messages
lucian whiteknight: aloha the bot or aloha the protocol? 14:09
14:19 nwellnhof joined
whiteknight lucian: both are equally inferior 14:20
14:33 bluescreen joined
lucian Tene: still awake? gist.github.com/720017 14:34
14:41 bluescreen left, particle joined 14:49 particle left 14:53 Andy left 14:54 particle joined 15:18 particle left
dalek nxed: r693 | NotFound++ | trunk/winxedst1.winxed:
improve handling of typecast predefs in stage 1
 15:25
15:29 particle joined
dalek tracwiki: v9 | whiteknight++ | StringsTasklist 15:30
tracwiki: trac.parrot.org/parrot/wiki/Strings...ction=diff
nxed: r694 | NotFound++ | trunk/winxedst1.winxed:
rename checkResult methods in stage 1 to checkresult for consistency
 15:41
whiteknight msg dukeleto: There are several suggested tasks which I think can be approved/opened so the students can get their hands on them. Our list of open, unclaimed tasks is dwindling. I'm going to try to add at least 5-10 new tasks today (I've already added 2) 15:50
aloha OK. I'll deliver the message.
Coke I had a patch that could use some eyes slash be applied. 15:51
whiteknight Coke: sure. ticket?
Coke trac.parrot.org/parrot/ticket/1716 15:52
aloha, msg moritz ad a patch that could use some eyes slash be applied.
10:51 <@whiteknight> Coke: sure. ticket?
aloha Coke: OK. I'll deliver the message.
Coke whoops.
whiteknight :) 15:53
Coke aloha, msg moritz trac.parrot.org/parrot/ticket/1716 was yours, does the patch work for you?
aloha Coke: OK. I'll deliver the message.
moritz uhm, how do I actually download that patch? 15:56
Coke you want the raw patch? trac.parrot.org/parrot/raw-attachme...-dir.patch
(bottom of the trac patch screen says "... original format")
moritz ah, thanks 15:57
Coke it is hidden. 15:58
how goes, moritz ?
... hey, is your name pronouned "eats" or "its" ? 15:59
16:00 theory joined
whiteknight that's a very good question 16:00
moritz "its"
dukeleto whiteknight: ok, sounds good. Will take a look in a few minutes.
moritz stress is on the o
Coke poor o. 16:01
whiteknight dukeleto: some of these students are a lot more qualified as coders than I originally thought. I'm comfortable putting together some tasks that require a little bit more know-how than I have been doing
dukeleto puts on his lOllerskates
whiteknight: yes, i agree
moritz Coke: the patch seems to work fine
whiteknight In fact, I think I can trick these students into implementing most of the IOTaskList on trac wiki :)
dukeleto whiteknight: works for me 16:03
16:04 fbrito joined 16:09 particle left
Coke moritz: just need someone to test it on windows, I think. 16:09
moritz indeed
Coke if taptinder comes back soon, can just apply it and hope. ;) 16:10
dukeleto I talked to mj41 last night and told him how to use SVN access on github, so hopefully it comes back soon. 16:11
whiteknight: i just approved a bunch of tasks
fbrito: new tasks! Take a look.
fbrito Yeah! I am going to take a look for sure. 16:12
atrodo wishes he was back in school so he could participate 16:13
dalek nxed: r695 | NotFound++ | trunk/winxedst1.winxed:
add a few internal checks
 16:14
whiteknight atrodo: have children. Force them to become programmers. Code vicariously through them
fbrito hahahahah!
moritz s/Force/motivate/ 16:15
atrodo whiteknight> Already working on it. However, that has a 15+ year lead-time
whiteknight s/Force/motivate with a stick/
atrodo: yeah, lazy kids don't grow very fast 16:16
atrodo seriously.
16:17 particle joined
dalek nxed: r696 | NotFound++ | trunk/pir/winxed_ (2 files):
update installable files
 16:19
tracwiki: v17 | whiteknight++ | IOTasklist
tracwiki: trac.parrot.org/parrot/wiki/IOTaskl...ction=diff
Coke I have a 13 year old who could be a better dev than I am, but he has other ideas. 16:21
fbrito Ok, I requested this task: www.google-melange.com/gci/task/sho...9035423390
But now I have to lunch :S. Be right back 16:22
whiteknight fbrito: that should actually be a pretty easy one. Let me know if you need any help. 16:28
nwellnhof whiteknight: Thanks for pointing me to the IOTasklist document. I didn't know about it.
But I had a look at the IO code and came to similar conclusions. 16:29
whiteknight nwellnhof: Ha! It's actually on the front-page of trac, but that page has gotten so cluttered that nobody knows about things there
Infinoid and myself had worked on it a few months ago, but he left and I moved to other things
dukeleto whiteknight: where did Infinoid go? 16:30
nwellnhof I once had a look at most of the items there, but somehow I must have missed the IO task list.
whiteknight He got busy with other things and hasn't been working much on Parrot
he's still around. Was maintaining dalek for a while
that IO system really needs a complete ground-up redesign
because it's not really sane how sometimes PMC methods call the API routines, and sometimes the API routines call PMC methods 16:31
sometimes both for the same API calls
this situation improves when the open/close opcodes disappear 16:33
16:33 dmalcolm joined
whiteknight (have they been removed yet?) 16:33
dukeleto whiteknight: i think they have been deprecated, but not removed, but don't trust my memory 16:36
whiteknight: are there any easy deprecation items that we can turn into tasks?
whiteknight blah. So we can't just do that now then
NotFound They are in dynops 16:37
whiteknight Okay, I'll look at that. That may make a good task too, to update them to suck less
dukeleto whiteknight: doing a deprecation is a good learning experience. You have to delete code, regen the manifests, update a wiki page and add something to NEWS
whiteknight dukeleto: yes, it is good
let me take a gander
dukeleto whiteknight: it gets people seeing the whole process, instead of "i wrote a line of code, i am done" 16:38
NotFound dukeleto: update *two* wiki pages.
dukeleto NotFound: you are indeed correct
NotFound Not correct, but true ;) 16:39
whiteknight: you want to delete things? Take a look at TT #1682 16:42
whiteknight :) 16:43
16:43 hudnix left
Infinoid Oh hey, I remember the IO stuff 16:44
Specifically, making sockets not suck 16:45
whiteknight Infinoid: you had a rough implementation of a new PipeHandle type at one point. Would you by chance have that still around anywhere? 16:46
I would *love* to complete that particular part of the refactor
I think I had a copy at one time, but my laptop got wiped 16:47
dukeleto Infinoid: i have been working on making sockets suck less by actually testing them 16:49
Infinoid: i have a branch for IPv6 sockets, and I am adding many tests and finding bugs
whiteknight dukeleto: there was a time when sockets were much, much worse 16:50
hell, only a few years ago there were no sockets
nwellnhof One of the problems with the IO code is that there's a lot of functionality in the FileHandle PMC that should be moved to the Handle base class.
whiteknight nwellnhof: exactly! I'm creating several GCI tasks so students can work on that
nwellnhof Especially buffering and handling of encodings. 16:51
Infinoid dukeleto++
whiteknight YES
Infinoid whiteknight: I can look. I don't remember it being a particularly huge patch
whiteknight Infinoid: me either, but I would hate to have to reproduce any amount of good code
Infinoid If I don't have it, I can remember how it was set up. One pipe container object (factory?), one pipe endpoint object
dukeleto Coke: how do we get different versions of various parrot.org pages rendered in different languages? We have the translated text, I am talking about how to actually tell Drupal about it 16:52
whiteknight right
a Pipe (2-way) and a PipeHandle (1-way) are what I remember
Infinoid If I do have it, it's more than a year out of date with respect to parrot at this point
whiteknight the IO system hasn't changed at all in that time, really
which is the big shame of it
dukeleto Infinoid: it could still be useful, even if it doesn't apply
whiteknight plus, we have to wait an entire deprecation cycle to move the pipe-like logic out of FileHandle 16:53
so we have time to make anything work
Infinoid As it happens, I might have a few cycles to spend on this anyway :) 16:54
dukeleto ++Infinoid
whiteknight: i just took a look at the Parrot+RTEMS logo. Looks a lot better than the nothing we have now.
whiteknight dukeleto: isn't it? I was very very impressed by that
Infinoid++ Indeed! Having you back hacking would be a big boost, if you have the tuits for it 16:55
Infinoid I don't have the development tree any more, but it looks like I had checked it into the io_rewiring branch, way back when 16:56
PipeHandle and Pipe were r39265 and r39266
dukeleto likes the #perl6-ism of a prefix increment to encourage doing stuff, and a postfix increment for actually doing stuff 16:57
whiteknight ah, I may be able to dig those out then.
NotFound ++stuff
whiteknight likes the parrot-ism of giving out lots of karma
Infinoid: got them! 16:58
Infinoid Oh, cool 16:59
whiteknight cotto++ on the trac plugin to link svn revs to github diffs!
Infinoid dalek++ for generating irc logs so I could search for it 17:05
17:11 PacoLinux joined
cotto_work ~~ 17:19
whiteknight: reping
whiteknight cotto_work: whadup? 17:20
cotto_work you pinged me earlier
whiteknight cotto_work: because aloha said you pinged me earlier than that 17:21
cotto_work whiteknight: irclog.perlgeek.de/parrot/2010-11-29#i_3039366 17:24
whiteknight ha
maybe we have nothing to talk about, just playing ping-tag
nwellnhof Question about the IO cleanup: If we move functionality from FileHandle to Handle, the code in Handle will need to call back into the subclasses. What's the best way to do that? 17:25
NotFound Call back? Why? 17:27
nwellnhof For example if we have a single buffering implementation in the Handle PMC, it would implement the read method and then only call the subclasses to fill the buffer. 17:28
NotFound I'm not sure I'd like that way. 17:31
whiteknight we should move the buffer into Handle
and all types get a buffer 17:32
child types call the instance in the parent type
fbrito I am having some problems in understanding my task (rename global_setup functions). Which functions (beside Parrot_set_config_hash_internal) should I rename? 17:33
17:35 hercynium_ joined
dukeleto fbrito: what is the link for your task? 17:35
fbrito www.google-melange.com/gci/task/sho...9035423390 17:36
There are some functions auto-generated somewhere else. Should I also rename that? (ex: Parrot_initialize_core_pmcs)
dukeleto fbrito: config/gen/core_pmcs.pm is involved 17:40
fbrito: you should ask whiteknight for clarification. I am not sure.
fbrito: if you form your pull request as a few different commits, we can choose to use some commits and not others 17:41
fbrito: that way, if one thing you do is wrong, we can still merge your branch
fbrito: that is why it is good to make many small-ish commits instead a few huge commits
fbrito dukeleto: ah, ok, I will do that. thank you :D 17:42
nwellnhof whiteknight: yes, child types call the buffered read code in the parent type at some point. but that code has to call back into the child types if it has to refill the buffer. 17:45
whiteknight fbrito: let me look 17:53
fbrito github.com/fernandobrito/parrot/commits/master
whiteknight Parrot_set_config_hash_internal, Parrot_set_config_hash_interpreter, init_world_once, init_world, parrot_global_setup_2 17:55
All of them need to be renamed so they start with "Parrot_gbl_"
so Parrot_gbl_global_setup_2 17:56
or, even Parrot_gbl_setup_2
Parrot_gbl_set_config_hash_interpreter, etc
cotto_work Why not "Parrot_global_"... 17:57
dukeleto cotto_work: i agree.
I don't like 'gbl' as a shortening of 'global'. Non-english speakers will be confused. gbl sounds like our turkey subsystem or something...
17:58 hercynium__ joined
dukeleto Parrot_global_setup_2, Parrot_global_init_world, etc... 17:58
NotFound WTF is 'global'?
dukeleto NotFound: stuff that changes global state 17:59
NotFound I see that functions as stuff only for internal usage. Be global or not is secondary.
whiteknight cotto_work: the standard we've been going by has been a 3 or 4 letter abbreviation 18:00
dukeleto NotFound: nope. Parrot_set_config_hash is needed by embedded projects
18:00 hercynium__ left
whiteknight there may be a better abbreviation than gbl, but once they are named by a standard it becomes trivial to do a mass find+replace later 18:00
cotto_work whiteknight: sure, but that doesn't mean we need to keep doing that if it makes for ugly function names.
fbrito gbl, glb, glbl or global?
dukeleto votes a +1 on Parrot_global_* 18:01
whiteknight cotto_work: gs for "global setup"?
Parrot_gs_?
cotto_work gl�bal
dukeleto That seems non-intuitive.
whiteknight I don't like Parrot_global because it breaks the standard. They will be the only ones not using a short acronym
dukeleto whiteknight: we have a standard?
whiteknight: where is it written down? Or is it just convention? 18:02
NotFound init_world_once and the like should never be used by any one other than parrot internals.
Using the same prefix as public api functions is confusing.
whiteknight dukeleto: the original ticket is here: trac.parrot.org/parrot/ticket/443
18:02 hercynium__ joined
whiteknight There are more docs for it. I will need to find them 18:02
NotFound: the difference is the use of the PARROT_EXPORT 18:03
NotFound: but those two functions *are* public from the point of view of the global_setup.c file
that is, they are visible outside it, so need to be prefixed accordingly
18:03 hercynium_ left
whiteknight if they were static, that would be different 18:03
cotto_work whiteknight: there were some other gci tasks that resulted in some function names that were similarly ugly, iirc. 18:04
whiteknight fbrito: do Parrot_gbl for now, so we don't make you wait. We'll figure out what we want and update it after you are done
cotto_work whiteknight: wfm
dukeleto fbrito: what whiteknight said
whiteknight cotto_work: I'm sure. Again, we need to figure out what we want to change them to
NotFound whiteknight: PARROT_EXPORT is a mess, we need to diferentiate functions accesible from inside the library and from the outside world.
fbrito ok, thank you :D
whiteknight and once we have like functions with a unique prefix, it's an easy replace job 18:05
cotto_work there's no need to hold up a gci project
18:05 hercynium_ joined
dukeleto cotto_work++ # jfdi 18:05
whiteknight NotFound: that's a problem with PARROT_EXPORT, not a problem with the functions it decorates
cotto_work: I do agree with you though, some of them have become a bit ugly. We should review them for aesthetics eventually
NotFound I fail to understand that logic. 18:06
nwellnhof I think it's basically a problem with headerizer. There should be a way to put headers of extern functions in different header files for internal and external use.
whiteknight NotFound: the function prefixes are used to show the subsystem where a function lives, not whether that function is exported 18:07
PARROT_EXPORT and PARROT_API (in the embed_api branch) show whether a function is exported by the .so
dukeleto whiteknight: i like that 18:08
NotFound I can live with that, but I think that initializing the basics of the interpreter is not the same league as changing its state.
dukeleto NotFound: that also sounds about right. What about "Parrot_init_*" functions ? 18:09
18:09 hercynium__ left
NotFound Maybe kho, Keep Hands Out ;) 18:10
dukeleto NotFound: that is just asking for trouble...
NotFound Parrot_aft_ ? 18:11
Parrot_init_ will be fine for me. 18:12
Coke dukeleto: I have no idea how to do anything complicated with drupal. 18:16
dukeleto Coke: who does? Who in the parrot community knows the most about how parrot.org works? 18:17
18:18 hercynium_ left 18:19 hercynium_ joined
whiteknight Parrot_init_* works fine for me 18:19
Coke dukeleto: do you want to know about drupal, or do you want to know about parrot.org? 18:20
whiteknight we might want to move some files around though. src/global_setup.c could probably be moved into src/interp/
src/interp/inter_create.c is very similar
(and inter_create.c should be renamed. That name has always bugged me. Should be interp_create.c, or even create.c, since we're in the interp folder) 18:21
dukeleto Coke: i want to know how parrot.org uses Drupal and how we add internationalization to parrot.org pages via Drupal 18:24
cotto_work whiteknight: +1 to both Parrot_init_* and create.c (just don't spell it "creat")
Coke dukeleto: particle is probably an excellent person to talk to about parrot.org in general, as he's on the board & has had interaction with OSU as much as anyone else.
dukeleto: SFAIK, www.parrot.org is just a braindead drupal instance.
dukeleto Coke: sadly, I have a lot more interaction with OSUOSL than particle these days
18:24 hercynium_ is now known as [hercynium]
Coke SFAIK, parrot community has no drupal experts. 18:24
dukeleto which is a sad state of affairs. 18:25
Coke: do you know which version of drupal we are running?
Coke dukeleto: I presume that www.parrot.org could tell you. It's whatever OSUOSL gave us, SFAIK.
whiteknight cotto_work: don't worry. I wouldn't ever name anything "creat"
NotFound 'krt' 18:26
whiteknight NotFound: :)
NotFound And put it in src/sms/
NotFound thinks about writing a PASM variant that looks like 8080 assembler. 18:27
whiteknight That would be fun
Notfound: you know what I would really like to do someday? I would like to write an LALR parser generator in Winxed 18:28
we could get some pretty fast parsers from it, and we would be able to build new compilers without using NQP
I know some people, like the python people, have an aversion to things that are perlish
Coke dukeleto: there's a "drupal-5.9" directory on the VM, so that's my guess. 18:29
NotFound whiteknight: I'm also thinking about that, but I don't know enough about PCT grammars yet.
Coke dukeleto: ls /var/www/www.parrot.org 18:30
NotFound whiteknight: or do you mean to write them on another way?
Coke on parrotvm.osuosl.org
dukeleto: I'm not sure anyone was particularly happy with drupal. It was just what we got after we switched from perl.org to parrot.org 18:31
18:32 davidfetter joined
cotto_work My dayjob is hacking on Drupal. 18:34
whiteknight NotFound: I'm thinking about a new grammar engine, not PCT
dukeleto Coke: i am not sure i have ssh access to that machine
whiteknight we could use the same algorithm as YACC
dukeleto Coke: do i have to give my pubkey to somebody to get access to that machine? 18:35
Coke: do you know about i8n for drupal?
cotto_work: ^^^
Coke: mis-tab, mea culpa
NotFound whiteknight: it may be a nice gsoc.
cotto_work dukeleto: not atm though I could read up on it tonight. 18:36
dukeleto cotto_work: that would rock. I also asked OSUOSL about it. 18:37
whiteknight NotFound: yes, it would make a very nice GSoC for a student with some compiler design experience
Coke dukeleto: pubkey is for the ftp server. parrotvm is actual logins with your own account. (and if you want to setup a pubkey, go for it.)
here's a list of homedirs on that box: 18:38
dukeleto Coke: "actual logins with your own account" meaning what?
Coke actown allison coke cotto jeff jhorwitz particle rspier russell_h
dukeleto: as opposed to the ftp server where everyone logs in as "parrot"
NotFound thinking about posting a job offer asking for at least 5 years of winxed experience.
Coke you don't seem to have an account. You should get one. 18:39
dukeleto Coke: and how do I go about that? ping OSUOSL ?
Coke NotFound: I have nearly 2 decades worth of experience in the winxed family of languages, does that count?
dukeleto: yup.
then you'll have access to the trac, drupal instances, doc site... 18:40
NotFound Coke: only if you apply for a unpaid job. 18:42
Coke I am already not making enough money, so nevermind. ;)
+1 on reverting the change Lithos noted. 18:43
NotFound +1 18:44
cotto_work regexen gone wild
dalek tracwiki: v23 | whiteknight++ | BigProjectIdeas 18:45
tracwiki: +LALR parser generator
tracwiki: trac.parrot.org/parrot/wiki/BigProj...ction=diff
whiteknight GSOC next year should have some awesome projects 18:46
they should do one every season: a winter of code, autumn of code, spring of code
NotFound An april fool's of code will be nice, too. 18:48
whiteknight NotFound: we only offer one project: Solve the halting problem using only Parrot 18:49
Notfound: but we introduce a GC bug so the answer is always "yes"
dukeleto just laughed pretty hard 18:51
does Parrot have a security list? 18:52
for people to send 0day exploits to so we can fix them? 18:53
cotto_work I remember bringing up the idea.
dukeleto cotto_work: let's make one. Nowishly.
cotto_work I don't have that kind of power.
NotFound I'd like it better if the answer were '42'
cotto_work but while you're talking to the right people, could you also set up a parrot-architecture list? 18:54
davidfetter mmm...sploitz
Coke parrot-architecture? 18:56
is there some class of email discussion for which parrot-dev is not sufficient?
moritz parrot-dev is fairly low traffic these days, why do we need another list?
dukeleto cotto_work: how is that different than parrot-dev ?
dukeleto agrees with moritz
moritz he, three people, one thought
fbrito why there is a function called parrot_global_setup_2 but no parrto_global_setup_1 or?
dukeleto fbrito: a dingo ate your function 18:57
whiteknight fbrito: global_setup_1 is generated, I think
cotto_work dukeleto: specific to the architecture team
fbrito ah, ok. and should I rename it to Parrot_gbl_setup_2? or Parrot_gbl_global_setup_2? 18:58
whiteknight fbrito: you are the developer! You can pick
either works for me
18:59 PerlJam left, PerlPilot joined
fbrito ok then. i will pick Parrot_gbl_global_setup_2_Fernando_Brito_rocks() 18:59
whiteknight :)
tadzik ...who introduced the idea of April Fool's Code?
cotto_work tadzik: the original incarnation of Parrot was a joke combination of Perl and Python
tadzik yeah, I know that ;) 19:00
maybe it should have a function run_perl_or_python()
cotto_work Sure. All we need is valid perl and python compilers 19:01
dukeleto tadzik: april fools stuff is very old. There are RFC's that are april fool's jokes as well 19:02
fbrito but if I am renaming parrot_global_setup_2 I should do the same with setup_1, but I can't find it :P
whiteknight apparently the python people are hung up on the object metamodel problems
the ruby people might be too
dukeleto tadzik: people have been writing april fools code since at least the 70's and probably the 60's
cotto_work whiteknight: what do you mean?
dukeleto fbrito: do you know how to use 'ack'
fbrito: if not, you should
fbrito: betterthangrep.com
whiteknight cotto_work: From the people I have talked to, problems with the object system are a big impediment to getting pynie to work, and getting more python devs working on it 19:03
dukeleto fbrito: it will raise you to the next level :)
whiteknight: i plan on helping chromatic with our metaobject protocol
whiteknight I know that Cardinal went dormant in part because of frustrations with the object system
dukeleto: me too. It's very important
19:04 PerlJam joined 19:07 PerlJam left, PerlJam joined 19:09 PerlPilot left
fbrito ok. as soon as I finish running tests, I will rename the last function, run the tests again and make the pull request, so you can check what I have done 19:17
whiteknight fbrito++
19:17 PerlJam left, PerlJam joined
whiteknight fbrito: You're being very helpful and productive! Thanks for all the work 19:17
cotto_work whiteknight: are you referring to the Parrot versions of those languages?
fbrito the first time I renamed a function I forgot to make clean, so when I ran the tests some of them failed
whiteknight cotto_work: yes 19:18
Pynie and Cardinal
cotto_work I'm glad chromatic will be working on a new one then. 19:19
whiteknight cotto_work: I'm not sure how else to say it. deficiencies in the Parrot object model are dissuading people from working on Pynie and Cardinal 19:27
oh... I was looking at the wrong part of the channel
hah, it looked like you were asking me what I was talking about again 19:28
disregard
dukeleto just created a parrot-security mailing list for the report of Parrot security issues (OSUOSL++) 19:29
dalek tracwiki: v70 | cotto++ | ParrotQuotes 19:34
tracwiki: trac.parrot.org/parrot/wiki/ParrotQ...ction=diff
fbrito when I do a pull request with lot of commits, can you choose which ones do you accept?
a lot*
dukeleto fbrito: yes. We can either cherry-pick the commits we want, or we can merge your branch, and then revert individual commits 19:36
fbrito ok. I will try to revert myself the commit that I don't want to be pulled 19:38
whiteknight dukeleto: do we have an ETA on the merge of the ipv6 branch? 19:44
I would really like to see that one merged by 2.11
fbrito Ok. Just did the pull request 19:45
whiteknight awesome. Let me go take a look at it 19:46
dukeleto whiteknight: when it actually works :)
whiteknight: i have a TPF grant to work on improving the embed/extend stuff, so ipv6 will take a back seat to that, but it is the next in line in terms of priority 19:47
whiteknight: i will hack on that when i am against a wall on embed stuff
whiteknight: some stuff needs to be re-org'ed on the ipv6 branch. Like Addrinfo PMC needs to become SockaddrInfo, or not exist at all. 19:48
whiteknight oaky
I won't hassle you then, I'm just trying to take stock of things that I think are awesome
dukeleto whiteknight: it is a thin wrapper for keeping track of a pointer, and creating a whole new PMC to do that seems bad 19:49
whiteknight: if you can take a quick look at tt1798_ipv6 and tell me how to reorganize the ipv6-related PMC's to something more sane, that would be Super.
whiteknight I may just do that 19:50
dalek rrot/global_setup_func_rename: 6735bb1 | (Fernando Brito)++ | / (4 files):
renaming Parrot_set_config_hash_internal to Parrot_gbl_set_config_hash_internal
 19:53
rrot/global_setup_func_rename: 2f7ae4e | (Fernando Brito)++ | README (2 files):
adding .markdown extension to README for GitHub
rrot/global_setup_func_rename: 4e7ba26 | (Fernando Brito)++ | / (4 files):
renaming Parrot_initialize_core_pmcs to Parrot_gbl_initialize_core_pmcs
rrot/global_setup_func_rename: ed542fb | (Fernando Brito)++ | / (3 files):
renaming Parrot_register_core_pmcs to Parrot_gbl_register_core_pmcs
rrot/global_setup_func_rename: 6c6bfd6 | (Fernando Brito)++ | src/global_setup.c:
renaming parrot_global_setup_2 to Parrot_gbl_setup_2
rrot/global_setup_func_rename: ad66f87 | (Fernando Brito)++ | src/ (3 files):
renaming init_world_once to Parrot_gbl_init_world_once and parrot_set_config_hash_interpreter to Parrot_gbl_set_config_hash_interpreter
rrot/global_setup_func_rename: 92cf618 | (Fernando Brito)++ | include/parrot/global_setup.h:
forgot to rename a init_world_once function
whiteknight fbrito: Pulled to a branch. I'll test/evaluate it and then merge it to master and close your ticket 19:54
give me a few minutes
thanks
dukeleto fbrito++ 19:57
karma fbrito
aloha fbrito has karma of 2.
dukeleto fbrito: you need to add yourself to CREDITS
fbrito: then aloha will know that (Fernando Brito)++ and fbrito++ are the same person, and will lump the karma together
whiteknight fbrito: yes, add yourself to CREDITS 19:58
dukeleto fbrito: send a pull request with a patch that adds yourself. You will see the format from the other entries
whiteknight and don't rename it so it looks better on github!
20:00 preflex left
fbrito :o 20:02
20:03 Andy joined 20:05 preflex joined 20:07 rurban left
fbrito whiteknight: I was thinking about getting another Parrot task on GCI. do you consider it a good idea or should I open space for new contributors? 20:07
dukeleto fbrito: take as many as you think you can do. We will create more tasks. 20:08
whiteknight fbrito: yes, take as many as you want. we have MANY tasks
there are many that haven't been made public yet, and many more that are in my head and I haven't created yet 20:09
fbrito nice :D 20:10
whiteknight fbrito: testing now. Cross your fingers! 20:13
fbrito has all fingers crossed
whiteknight dukeleto: that README task from vmax looks much better. Do you want to pull it or should I? 20:14
dukeleto whiteknight: go for it 20:15
whiteknight: and if you want to create a new task for converting README to POD or markdown, i will approve it 20:16
Coke -1 on POD conversion. 20:17
fbrito take a look on that: github.com/fernandobrito/parrot/tree/wip#readme 20:18
whiteknight fbrito: Complete. Good work
fbrito only by adding .markdown extension to the README file you already get this 20:19
dalek rrot: 6735bb1 | (Fernando Brito)++ | / (4 files):
renaming Parrot_set_config_hash_internal to Parrot_gbl_set_config_hash_internal
rrot: 2f7ae4e | (Fernando Brito)++ | README (2 files):
adding .markdown extension to README for GitHub
rrot: 4e7ba26 | (Fernando Brito)++ | / (4 files):
renaming Parrot_initialize_core_pmcs to Parrot_gbl_initialize_core_pmcs
rrot: ed542fb | (Fernando Brito)++ | / (3 files):
renaming Parrot_register_core_pmcs to Parrot_gbl_register_core_pmcs
rrot: 6c6bfd6 | (Fernando Brito)++ | src/global_setup.c:
renaming parrot_global_setup_2 to Parrot_gbl_setup_2
rrot: ad66f87 | (Fernando Brito)++ | src/ (3 files):
renaming init_world_once to Parrot_gbl_init_world_once and parrot_set_config_hash_interpreter to Parrot_gbl_set_config_hash_interpreter
rrot: 92cf618 | (Fernando Brito)++ | include/parrot/global_setup.h:
forgot to rename a init_world_once function
rrot/new_readme: 9a0f1b5 | vmax++ | README:
Added "how to get parrot from github" section
 20:22
rrot/new_readme: 4f9d434 | vmax++ | README:
Added some information, formatted
rrot/new_readme: fccfb06 | vmax++ | README:
review: github.com/parrot/parrot/commit/fccfb06493
rrot/new_readme: 1cfc9b9 | vmax++ | README:
review: github.com/parrot/parrot/commit/1cfc9b9f74
Coke who broke the build? 20:24
Rant: github emails not showing us who committed something, just who wrote it. 20:25
whoof. neither does github?
or does fernandobrito have commit privs to master?
dukeleto Coke: whiteknight just merged fbrito's pull request 20:26
whiteknight shit, did something break? I tested it here before I pushed 20:27
Coke: fernandobrito wrote it and committed it. I merely pushed those commits into the repo 20:28
anybody has commit privs, it's push privs to github that are controlled
moritz it would be nice if such merges had a Signed-Off-By line
whiteknight I'm not sure how to do that, otherwise I would be 20:29
fbrito hm
moritz if you cherry-pick, the syntax is just
git cherry-pick -s $sha1
I don't know if pull requests support that too
Coke whiteknight: he said in the pull request "don't take this commit".
looks like 2f7ae4ef4c96e51b3b39579731af34668727f20c
fbrito yes, its the README commit 20:30
"Please do not accept the commit that changes de README file extension. It make it look pretty nice on GitHub but $ perl Configure.pl will complain about not founding the README file."
Coke sorry, "broke" is too strong a word. 20:32
whoops! no, it is broke, my build script is stupid.
dukeleto There is no need to change the README extension, methinks. 20:33
whiteknight I thought I added another commit to undo that README change 20:35
I might not have pushed that one
okay, I found it 20:36
pushed
dalek rrot: 414892d | Whiteknight++ | README (2 files):
undo README rename
whiteknight my apologies
fbrito my apologies too
20:37 lucian_ joined
dukeleto fbrito: learning is fun :) 20:39
dalek rrot: 3e1dedc | Whiteknight++ | README:
merge in the README improvements with new git information from GCI student vmax++
20:42 lucian left
whiteknight I have decided that I absolutely love GCI 20:42
fbrito me too :D 20:45
beside the problems that I had doing a task that involved kernel compiling with xen virtualization modules and so on :P 20:46
whiteknight I've never used Xen. I've heard it isn't very user-friendly 20:47
I like VirtualBox a lot personally
fbrito yes, me too. but it was a task for ganeti (virtualization clusters management) web-interface 20:50
sorear whiteknight: I'm suprised at how well it's going
fbrito a really easy task that just required some jQuery knowledge, but setting the environment is like 10 times harder than the task it self 20:51
sorear whiteknight: I had no idea people like fbrito existed. I thought that all members of the age bracket were shortsighted and immensely irritating 20:54
whiteknight sorear: and they all listen to stupid music! 20:55
:)
fbrito whaaat? don't call mozart stupid, please
Coke thought sorear was in the same age bracket! 20:56
sorear Coke: not anymore 21:01
I was 19 when I joined here
fbrito should I also have a single commit for each different function that I rename? 21:04
(on this new renaming task that I am working on) 21:06
Coke thinks it's overkill but isn't applying your patches 21:15
21:21 Andy left 21:22 lucian_ left 21:28 whiteknight left 21:41 zostay joined, TypeNameHere_____ left 21:43 bluescreen joined
Infinoid Hmm. Do I still have a commit bit? 21:48
21:50 nwellnhof left
dalek rrot: f6b063c | infinoid++ | README:
[README] Minor nit: the Debian/Ubuntu package is called git-core, not core.
 21:52
Coke Infinoid: no, we setup an elaborate hoax to make it look like it, though! 21:53
Infinoid Coke: As long as that elaborate hoax includes the change magically appearing in other people's repos when they update, I'm okay with that. :) 21:54
22:01 PerlJam left, PerlJam joined
fbrito hmm... I probably renamed something wrong on my function renaming task -.- getting 22:08
ar: blib/lib/libparrot.a: Malformed archive
cotto_work fbrito: try make clean 22:11
That might result from an overeager sed command 22:12
fbrito I have also tried that. Now I have to do a binary search on my commits to find which one broke the build :D
Coke git makes that easy, at least. 22:16
dalek tracwiki: v6 | plobsing++ | PlobsingTaskList 22:18
tracwiki: trac.parrot.org/parrot/wiki/Plobsin...ction=diff
Infinoid (git-bisect)++
fbrito interesting. it seems that make clean doesn't delete blib/lib/libparrot.a
is that the expected behavior? should I mark that as a bug/try to make a patch? 22:22
Coke that seems like something that clean should get (as opposed to realclean0 22:25
dalek rrot/generational_gc: e128286 | bacek++ | src/gc/gc_ms2.c:
Increase gc_threshold to speed-up testing.
 22:30
Coke so, why bother fighting with the gc_threshold parameter. wouldn't a flat "use no more than this amount of memory" suffice? 22:31
where "fighting" means "that looks hard and confusing." 22:32
cotto_work fbrito: looks like make clean isn't as clean as one would expect. Can you file a ticket?
bacek_at_work Coke, branch is old and doesn't have all new features of master 22:33
fbrito cotto_work: ok, sure
Coke bacek_at_work: I meant in general, not in your real world case. ;)
bacek_at_work Coke, master uses fraction of available memory already. 22:34
Coke bacek_at_work: in your case, perhaps. 22:35
so I guess it's a different use case then. hokay.
fbrito hope to see whiteknight here soon. I have completed another GCI task :)
dalek TT #1876 created by fbrito++: Should 'make clean' remove blib/ folder? 22:51
TT #1876: trac.parrot.org/parrot/ticket/1876
cotto_work fbrito: do you feel like fixing that bug?
fbrito oh, sorry for taking so long 23:01
was on telephone
I am going to have dinner and then when I come back I will request a pull with my attempt to fix it :D
brb
cotto_work fbrito++
fbrito: I'll take a look at the pull request when I get home if nobody beats me to it 23:03
23:08 preflex left 23:10 preflex joined 23:17 fperrad left 23:21 whiteknight joined 23:25 [hercynium] left
whiteknight good evening, #parrot 23:40
GeJ Bonjour whiteknight. 23:45
whiteknight hello GeJ, how are you today? 23:46
you haven't fallen off the bottom of the planet yet?
dalek rrot/hll_func_rename: 056265d | (Fernando Brito)++ | CREDITS:
Adding myself to CREDITS file
 23:48
rrot/hll_func_rename: 5e02476 | (Fernando Brito)++ | / (3 files):
Renaming Parrot_init_HLL() to Parrot_hll_init_HLL()
rrot/hll_func_rename: 82337d1 | (Fernando Brito)++ | / (6 files):
Renaming Parrot_register_HLL() to Parrot_hll_register_HLL()
rrot/hll_func_rename: cff8e25 | (Fernando Brito)++ | / (5 files):
Renaming Parrot_register_HLL_type() to Parrot_hll_register_HLL_type()
 23:49
rrot/hll_func_rename: c412704 | (Fernando Brito)++ | / (3 files):
Renaming Parrot_get_HLL_type() to Parrot_hll_get_HLL_type()
rrot/hll_func_rename: e515afa | (Fernando Brito)++ | / (26 files):
Renaming Parrot_get_ctx_HLL_type() to Parrot_hll_get_ctx_HLL_type()
rrot/hll_func_rename: 8e80030 | (Fernando Brito)++ | / (6 files):
Renaming Parrot_get_ctx_HLL_namespace() to Parrot_hll_get_ctx_HLL_namespace()
rrot/hll_func_rename: 35090a7 | (Fernando Brito)++ | / (6 files):
Renaming Parrot_get_HLL_namespace() to Parrot_hll_get_HLL_namespace()
rrot/hll_func_rename: f7134de | (Fernando Brito)++ | / (4 files):
Renaming Parrot_regenerate_HLL_namespaces() to Parrot_hll_regenerate_HLL_namespaces()
rrot/hll_func_rename: f89a69a | (Fernando Brito)++ | / (5 files):
Renaming Parrot_get_HLL_name() to Parrot_hll_get_HLL_name()
rrot/hll_func_rename: 99bcc83 | (Fernando Brito)++ | / (4 files):
Renaming Parrot_get_HLL_id() to Parrot_hll_get_HLL_id()
fbrito hope that I don't break the build again, ahhaha
whiteknight Once I pull the code, I'm responsible. I hope *I* don't break it again :) 23:50
dukeleto fbrito: nice atomic commits!
fbrito :D 23:52
whiteknight commit early, commit often
fbrito i love how lots of opensource projects has integration between repository, IRC bot, ticket system, build bots, etc :) 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!