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
Set by moderator on 18 November 2010.
00:03 Themeruta joined 00:04 NotFound left
dalek rrot/luben/gc_threshold_adjust: 410aadd | luben++ | / (4 files):
use --gc-threshold commandline parameter to specify percentage of total
 00:18
00:31 davidfetter joined
dalek rrot/luben/gc_threshold_adjust: a226bdc | luben++ | include/parrot/gc_api.h:
adjust default gc_threshold to 5% of total memory
 00:51
00:53 lucian left 01:06 davidfetter left 01:25 bluescreen joined 01:30 bluescreen left 01:34 bluescreen joined 01:37 bluescreen left
cotto ~~ 01:38
01:40 bluescreen joined 01:49 bluescreen left 01:57 Coke left 02:03 Coke joined 02:06 bluescreen joined 02:17 TypeNameHere_ joined 02:21 whiteknight joined
dalek rrot: ba88b03 | Whiteknight++ | NEWS:
change the NEWS entry. Too many files are having their functions updated to list individually
 02:45
rrot: 139944d | Whiteknight++ | / (28 files):
Merge branch 'exit_func_rename'
rrot: 3097140 | Whiteknight++ | docs/project/ (2 files):
Merge branch 'master' of github.com:parrot/parrot
02:45 TypeNameHere_ left
whiteknight dukeleto: ping 02:56
msg dukeleto: I don't know if you saw the submitted Parrot+RTEMS logo. I was hoping you would get to look at it before I approved it, but I made the student wait for a long time and I approved the task tonight. I have all the files here in a .zip. I'll send it to you for perusal later 03:04
aloha OK. I'll deliver the message.
03:06 bluescreen left
whiteknight msg cotto you have a few GCI tasks you're mentoring up for review. Let me know if you're too busy and would like somebody else to help dealing with them all. 03:10
aloha OK. I'll deliver the message.
03:13 mikehh left 03:15 hercynium joined 03:21 whiteknight left
cotto msg whiteknight I'll get to them later this evening. 03:41
aloha OK. I'll deliver the message.
03:47 plobsing_ left 03:55 plobsing joined 04:00 plobsing left 04:03 mikehh joined 04:37 hercynium left 04:42 TypeNameHere_ joined
cotto aloha, clock? 05:12
aloha cotto: cotto: LAX: Thu, 21:12 PST / CHI: Thu, 23:12 CST / NYC: Fri, 00:12 EST / UTC: Fri, 05:12 UTC / LON: Fri, 05:12 GMT / BER: Fri, 06:12 CET / TOK: Fri, 14:12 JST / SYD: Fri, 16:12 EST
dalek rrot: 153eb84 | mikehh++ | src/exit.c:
fix item detail to pass c_function_docs.t
 05:18
05:19 rurban_ joined 05:21 rurban left, rurban_ is now known as rurban
cotto bacek_at_work, ping 05:36
05:39 plobsing joined
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1343) fulltest) at 153eb84 - Ubuntu 10.10 i386 (g++-4.5) 05:53
05:59 mikehh left 06:01 TypeNameHere_ left 06:09 mikehh joined
cotto dukeleto, ping 06:27
dukeleto, unping (found the info where you posted it) 06:28
06:49 theory left 07:49 bacek_mobile joined
bacek_mobile Cotto, pong 07:49
cotto bacek_mobile, the packfile PMCs need to be updated to use plobsing's dynop_mapping changes before any internals can start to use them. What's the right approach? 07:52
bacek_mobile I think of mirroring new bytecode related structs as pmc 07:53
cotto that's what the opmap_aware_pmcs branch tries to do 07:54
You mentioned that the branch needs to take a different approach, didn't you? 07:55
bacek_mobile Nope. On branch you are trying to do it differently. Im talking about replicating as is. 08:01
cotto ok 08:04
dalek rrot/generational_gc: 2f55862 | cotto++ | src/gc/api.c:
[gc] clean up and update api POD docs
 08:06
rrot/generational_gc: a8a8341 | cotto++ | src/gc/gc_ms2.c:
try to improve readability in generational gc notes
08:08 cosimo_ joined, cosimo left
cotto bacek_mobile, please check that I didn't change the meaning of anything important. 08:11
08:14 rfw joined
rfw hey, i'm a gci student and i'm not bugging anyone to approve a task for me; i'm just here to ask a few questions about parrot and winxed 08:15
sorear Go ahead
rfw what's the equivalent of INIT in nqp in winxed?
i'm looking at www.google-melange.com/gci/task/sho...9033192711 this task atm and just getting to know the rails before i pick it up 08:16
sorear I don't know very much about winxed
whiteknight will
rfw ah, i chose it because it wasn't perl
Themeruta What do you mean by 'INIT'? 08:17
08:17 Themeruta is now known as NotFound
rfw well, i'm not sure where to load bytecode in winxed 08:17
bacek_mobile Cotto, looks good. You can cherrypick first commit into master right now. 08:18
rfw so i get syntax errors because my types are undefined
at least, i think that's why
sorear INIT blocks are how you include code to run in a file
rfw yeah
sorear especially. a module file
rfw i'm just looking how to include src/nqp/pla.prc into a winxed file
+at
NotFound You can add a function with the init modifier.
sorear they're somewhat tied up with the Rakudo build mangler
rfw NotFound: ah, so that would be [init]? 08:19
NotFound rfw: Better [init,load] to cover all usages.
bacek_mobile Cotto, diff contains french quotes. Did you use them or it's github? 08:20
rfw ah
08:20 fperrad joined
rfw hm no dice 08:20
cotto bacek_mobile, I did. They were to make it easier to distinguish from ->
rfw NumMatrix2D still seems to be undefined
i think i'm missing something 08:21
NotFound For load_bytecode you also have the $load directive,
bacek_mobile Cotto, bad idea to use them for POD markup :)
rfw NotFound: i'm sorry about knowing absolutely nothing about this all (and the winxed TFM is non-existent), butwhere does that go? 08:22
cotto bacek_mobile, are they only valid for perl 6 pod?
bacek_mobile Cotto, i think so
cotto I don't particularly care what's used as long as it's valid.
NotFound rfw: my fault, I'm a lazy documentation writer.
rfw hehe, fair enough
bacek_mobile Just try perldoc on file 08:23
NotFound $load should be at main lavel, out of any function, class or namespace.
rfw so i just throw $load on a line by itself? 08:24
cotto bacek_mobile, you're right. Those quotes confuse perldoc.
NotFound Mmmm... you are trying to use PLA, isn'it? That's a different problem, it has a dynamic PMC lib.
rfw ah
i'm completely lost with this all
forgive me
NotFound You need: using extern 'linalg_group';
rfw hm, NumMatrix2D still seems undefined 08:25
nopaste "NotFound" at 192.168.1.3 pasted "A winxed PLA test I wrote some weeks ago" (47 lines) at nopaste.snit.ch/26190 08:26
cotto night. I'll cherry-pick the commit into master tomorrow.
rfw NotFound: does it have to be var mat or can i have NumMatrix2D mat? 08:27
NotFound rfw: don't worry if it emits a warning about not locating a class, this is a compile time warning, it may work at runtime anyway.
rfw: winxed data types are: int, float, string and var. There are no static types for objects, all are var. 08:28
08:28 bacek_mobile left
rfw oh 08:28
ah that helped _a lot_.
NotFound This mimic pir types. 08:29
sorear oh, so var = Pxxx?
ah
opbots names
rfw thanks NotFound
NotFound You can think about winxed as a higher level pir.
rfw i don't know much at all about pir
NotFound If you know some winxed, use winxed -c, and you already know pir ;) 08:30
rfw hehe
now all i need is my moinmoin task approved and i'll be able to get on this right away
NotFound I must go, feel free to msg or email me any doubt about winxed. 08:31
rfw sure
also: hi Kovensky
08:41 wesjdj joined
rfw oh, another question 08:42
VTABLE void set_number_keyed(PMC * key, FLOATVAL value)
what does PMC * correspond to in winxed?
08:43 lucian joined
rfw or on whatever higher level parrot language there is 08:43
wesjdj Hi. I'm a student working on Google Code-in and I need edit priviledges on the website to see the page source code 08:44
Is there someone that can give me edit permissions?
My username is wesjdj there
08:51 contingencyplan left
moritz Coke: ping (see above) 08:52
wesjdj: I've found that the wiki uses HTML as source format, so you can probably just use the HTML source that "View -> Source" gives you 08:53
wesjdj moritz: thanks 08:54
rfw oh man i've already done a lot of that task which i haven't even accepted yet 09:05
i hope nobody's going to be interested in that one 09:09
09:14 rfw left 09:37 lucian left 09:44 jsut_ joined 09:49 jsut left 10:01 kurahaupo joined 10:07 kurahaupo_mobi joined, kurahaupo_mobi left, kurahaupo_mobi joined 10:12 kurahaupo left 12:10 whiteknight joined 13:02 darbelo joined 13:15 bluescreen joined 13:19 rurban_ joined 13:21 rurban left, rurban_ is now known as rurban
bluescreen good morning 13:23
13:25 Patterner left, Psyche^ joined, Psyche^ is now known as Patterner
darbelo o/ 13:44
14:01 rurban left 14:40 rurban joined 15:04 jan left
whiteknight hello 15:05
15:11 masak joined
masak oh hai 15:11
dalek rrot/util_func_rename: dcb1b5b | (Matt Rajca)++ | / (16 files):
Rename public functions in src/utils.c
 15:16
rrot/util_func_rename: 88a1a0c | Whiteknight++ | /:
Merge branch 'master' of github.com:parrot/parrot
rrot: 88a1a0c | Whiteknight++ | /:
Merge branch 'master' of github.com:parrot/parrot
 15:23
15:29 kid51 joined 15:41 jsut_ left 15:42 jsut joined
dalek rrot/embed_api: 6f8c9ec | Whiteknight++ | src/global_setup.c:
this commit got put on the wrong branch
 15:45
rrot/embed_api: e9934d7 | (Mariano Wahlmann)++ | src/embed/ (2 files):
Made it compile with gcc
rrot/embed_api: abab373 | (Mariano Wahlmann)++ | src/exit.c:
Fix segfaults when calling Parrot_exit twice
rrot/embed_api: 81e2599 | (Mariano Wahlmann)++ | src/e (2 files):
Fixed return value from api calls when executing parrot exit op
rrot/embed_api: e796f63 | (Mariano Wahlmann)++ | / (3 files):
Changed api_jmp_buf to *api_jmp_buf so we can verify its valid or not, programs using Parrot_exit directly without the API ( i.e pbc_merge ) will have a segfault if api_jmp_buf don't have a valid jmp_buf
whiteknight bluescreen+
bluescreen++ even
bluescreen lol
I'm not really happy about ++status before longjmp, but i think its a tradeof 15:47
whiteknight bluescreen: can you add yourself to CREDITS?
bluescreen longjmp won't return 0 to setjmp it will force it to 1
sure why not
whiteknight I want to make sure you get attribution for the work you are doing 15:48
bluescreen how was turkey btw?
still plenty of it?
whiteknight turkey was good. Yesterday was very crazy because we have so many people to visit. We didnt keep any left-overs 15:49
bluescreen oh... no left overs :(
whiteknight it looks like pbc_merge and pbc_disassemble are still having build errors 15:53
bluescreen longopt?
I was able to get pbc_disassemble working
whats not working is ops2c
tough it runs, it doesn't pickup command line args 15:54
whiteknight damnit 15:55
bluescreen and also I think parrot debugger needs to be re-work to use the api
whiteknight we should be able to #include "parrot/longopt.h", and link it with src/longopt${O)}
bluescreen which might bring some changes to it
whiteknight yes, I haven't touched the debugger yet
bluescreen i did changes to root.in
to make those compile
whiteknight did you reconfigure? 15:57
bluescreen perl Configure.pl, yes many times
maybe i forgot to upload something
whiteknight I don't think I have reconfigured in a while 15:58
they build on your machine?
bluescreen yes
i tried pbc_dissasemble and worked for me
whiteknight ok 15:59
I'm rebuilding now
bluescreen ./ops2c --dynamic src/dynoplibs/obscure.ops --quiet
asdas:
This is ops2c, part of the Parrot VM's build infrastructure.
normal options:
-c --core generate the C code for core ops (must be run from within Parrot's build directory)
-d --dynamic <file.ops> generate the C code for the dynamic ops in a single .ops file
-q --quiet don't report any non-error messages
-h --help print this usage information
-n --no-lines do not print #line directives in generated C code (line numbers are not currently supported)
debugging options:
-g --debug perform all processing but do not write to any files
sorry... about the flooding
you see its printing the help again
whiteknight blah 16:00
bluescreen its really hard to follow ops2c is bootstrapped x 2
nqp -> pir -> pbc -> c
whiteknight I don't see longopt$(0) being linked into pbc_dissassemble in the makefile. Are you sure you committed that? 16:01
bluescreen mmmm
is not there 16:03
and git diff doesn't show it
16:04 mikehh left
bluescreen stupid me.. forgot to push it 16:04
whiteknight :)
bluescreen the hardest part will be wrap everything up... write tests, docs, clean up 16:05
whiteknight hey, if you send in a signed CLA we can vote to make you a committer on parrot/parrot
bluescreen CLA?
whiteknight yes. I added a GCI task for a student to write up the documentation
Contributor License Agreement. Basically, it's just an agreement that you follow the license, etc 16:06
bluescreen lol... didn't know there was one
whiteknight www.parrot.org/files/parrot_cla.pdf
bluescreen but in short, I'm organs donor, marrow donor and code donor 16:07
whiteknight what the CLA says is that you own everything you contribute, and the Parrot Foundation can use your work so long as we are good with it 16:08
bluescreen great, I'm going to sing that
whiteknight :)
if you send in a signed CLA, I'll nominate you to be a committer 16:09
bluescreen thanks
brb 16:13
whiteknight okay, I just checked and the ops2c programs are receiving the commandline args we are passing them 16:34
so now it's a question to find out what that program doesn't like about them
moritz what language is ops2c written in? 16:35
compile rakudo programs don't receive @*ARGS, programs run from source do 16:36
I don't know if NQP-rx programs suffer from the same problem
16:40 jan joined
whiteknight NQP 16:42
they don't receive @*ARGS? That's extremely weird 16:43
they should still be available in the interpreter
bluescreen: I just fixed ops2c 16:50
the test suite runs. But many tests fail 16:56
lots of coredumps
16:59 masak left
whiteknight blah. EXCEPT_exit is causing most of these problems, I think 17:03
17:04 mikehh joined
dalek rrot/tt532_headerizer_refactor: c9bf5a0 | jkeenan++ | / (5 files):
Place methods in more logical order within module. api_first_then_alpha() was

subroutine. Add tests.		 17:06
wesjdj moritz: When I'm done with my GCI project, how should I get it to you?
moritz wesjdj: is there no submission process? if not, mailto:parrot-dev@lists.parrot.org 17:08
17:08 kid51 left
dukeleto hola 17:10
whiteknight hello dukeleto
dukeleto whiteknight: how goes it? 17:11
whiteknight: wow, i see a lot of gci-related activity on github 17:16
whiteknight: that is good to see
17:22 M_o_C joined
bluescreen whitekinght it is not that bad 17:29
is the string's api
dukeleto just found the Addrinfo PMC and is horrified 17:31
cotto dukeleto, it's used t hold the gc root set. 17:32
17:34 contingencyplan joined
dukeleto cotto: not from what I see. Are you thinking Addrregistry? 17:35
cotto: maybe you are right 17:36
cotto: but it is used in sockets. I can't quite tell yet
dalek rrot: 7ad73bd | dukeleto++ | / (2 files):
[t] Add a basic test for the Addrinfo PMC
 17:39
cotto dukeleto, yes 17:45
my mistake
dukeleto also, cloning an Addrinfo coredumps Parrot. 17:49
cotto where's addrinfo? I remember the name but I don't see it in src/pmc. 17:52
wesjdj moritz: One Question. How would you translate Version Control? 17:53
17:53 contingencyplan left
cotto dukeleto, the only place I can find the string "Addrinfo" is in those tests you just added. Am I missing something or is this a phantom PMC? 17:55
moritz wesjdj: yes, "Versionskontrolle" :-)
wesjdj moritz: Thought so, thanks. 17:56
17:56 contingencyplan joined 18:02 GodFather joined 18:04 kid51 joined
NotFound Where is that Addrinfo PMC? 18:07
I can't see it in master
$ prove -v t/pmc/addrinfo.t 18:08
18:08 GodFather left
NotFound t/pmc/addrinfo.t .. 18:08
1..1
Class '[ 'Addrinfo' ]' not found
18:08 GodFather joined 18:10 GodFather left
wesjdj moritz: Sorry for the questions, I just have one last one.. How would you translate "To check out a local clone"? 18:20
mikehh NotFound: same here 18:27
and I was trying to test a patch to TT #1872, so I was con-fuze-ed 18:29
cotto NotFound, it looks like it's in the 1798_ipv6 branch 18:34
dukeleto, found it 18:35
18:43 PacoLinux left
moritz wesjdj: "eine lokale Kopie herunterladen" or so 18:44
18:45 PacoLinux joined
wesjdj moritz: OK, thanks! In that case I'm ready to hand in my project. 18:45
moritz wesjdj++ 18:47
wesjdj moritz: There is an official way to finish the project, I'll do it that way. 18:49
dukeleto sees his error
I didn't realized that Addrinfo was added in my branch
wesjdj: are you a GCI student? Welcome! 18:50
wesjdj dukeleto: Yes I am! Thanks!
dalek rrot: 61d023e | dukeleto++ | / (2 files):
Revert "[t] Add a basic test for the Addrinfo PMC"

This reverts commit 7ad73bd58ca2a092c67061891b8c4e7f91c1f9bc.
This commit should have gone into the tt1798_ipv6 branch. I did not releaze that the Addrinfo PMC was added in that branch.		 18:52
moritz dukeleto: regarding the translated parrot.org pages... is there any plan for how to publish them?
dukeleto wesjdj: it is really awesome to see very smart young students getting involved in Parrot and open source in general
moritz dukeleto: just add a -<lang> to the pages? or will there be a different domain?
dukeleto moritz: i think drupal supported pages in many languages 18:53
moritz: i am not sure about the exact details, but I've heard that Drupal will detect the language from the browser and can present a page in that language, if it exists
wesjdj dukeleto: Yea GCI has been fun so far!
dukeleto moritz: we probably want a manual way for people to choose which language as well. It hasn't been ironed out yet. 18:54
wesjdj: that is great to hear!
wesjdj: how did you find out about Parrot? Did you just look through all the tasks? Or did you hear another way?
wesjdj My Dad introduced me to GCI, then I looked through the tasks and saw the translation and then decided to have a look! 18:55
dalek rrot/tt1798_ipv6: ccffd27 | dukeleto++ | / (2 files):
[t] Add a basic test for the Addrinfo PMC
rrot/tt1798_ipv6: 42f16ac | dukeleto++ | t/pmc/addrinfo.t:
[t] Add a test for cloning an Addrinfo, which currently coredumps
 18:56
wesjdj dukeleto: And whilst you translate you kinda learn a lot about the whole project :)
dukeleto wesjdj: that is the point :) 18:58
wesjdj dukeleto: Yep! 18:59
dukeleto wesjdj: you are working on the german translation now? 19:00
wesjdj dukeleto: I just finished about 20 minutes ago
19:02 tcurtis joined
wesjdj dukeleto: I'm actually just looksing at some other tasks. There are a couple of Parrot tasks I'm quite interested in. 19:02
looking*
dalek rrot/tt1798_ipv6: 4107908 | dukeleto++ | src/pmc/addrinfo.pmc:
Fix coredump when cloning an Addrinfo PMC that hasn't had pointer set
 19:03
dukeleto wesjdj: i have a bunch more, waiting :)
wesjdj dukeleto: Sounds good! Any more translation? I'm also interested in the 5 minute movie task..
moritz wesjdj: I'm currently preparing a p5 task for a Perl 6 compiler. After that I can look into another translation task 19:04
wesjdj: any preferences which page you'd like to translate this time? 19:05
wesjdj moritz: parrot.org/dev looks good!
dukeleto: Looking forward to it! 19:06
moritz wesjdj: the submission URL is a 404 19:10
wesjdj oh. Let me check what's wrong. 19:11
dukeleto cotto: our Socket PMC doesn't allow asking for IPv6. The address family is hardcoded and is not an argument to the 'sockaddr' method 19:12
cotto: do i need a dep cycle to add an argument to that method?
whiteknight: ^^^
bluescreen msg whiteknight I think failing test are because of the output format of the messages when "die"
aloha OK. I'll deliver the message.
wesjdj moritz: Seems to be working for me. I'll check it's not local. 19:13
moritz: Otherwise I'll submit the HTML
moritz wesjdj: is pjcj.sytes.net/de/download-de/ the correct URL? 19:14
says "The requested URL /de/download-de.html/ was not found on this server."
wesjdj moritz: Yes, that's what I have
OK, I'll submit the HTML
19:16 M_o_C left
whiteknight bluescreen: no, several tests dump core or segfault 19:19
bluescreen whiteknight: I've just fixed that 19:20
the thing was something like: char *msg = "Hello\\n"; free(msg) 19:21
we were freeing a non-allocated string
dukeleto wesjdj: i believe many translation tasks have not yet been input to Melange. We have a few more german translation tasks, definitely.
wesjdj dukeleto: OK :) 19:22
moritz I just put up another translation task - needs approval
whiteknight bluescreen: ah, that was a temporary debugging thing 19:23
thanks for fixing it
dalek rrot/tt532_headerizer_refactor: c232b9f | jkeenan++ | / (4 files):
Add tests for Parrot::Headerizer::Object methods.
 19:24
wesjdj moritz: OK, found the problem. 19:26
moritz: Resubmitted. 19:28
moritz wesjdj: a few small nits, s/Verpackungen/Packete/ for example, but otherwise it looks quite good 19:32
cotto dukeleto, no deprecation is required for adding something. They're only needed if you'll break existing code that depends on something we support.
wesjdj moritz: Do you need me to change it? 19:33
moritz wesjdj: no, I can fix it myself before publishing, just wanted to give you some feedback
wesjdj OK, thanks!
19:33 theory joined
NotFound cotto: if you add an argument, it must be optional, else surely breaks something. 19:33
moritz wesjdj: thank you
wesjdj moritz: You're welcome! :=) 19:34
cotto misread that as "adding a method". NotFound is correct.
darbelo It could be made a multi-method 19:36
NotFound Yeah
TIMTO.. you know ;)
cotto We have the technology.
darbelo And the technology to build the technology. If it ever comes to that. 19:37
NotFound We are the world...
dukeleto cotto, NotFound: yes, I will add an optional argument to a function 19:42
which will hopefully not break anything
19:42 kurahaupo joined
whiteknight bluescreen: I'm thinking we need to add a flag to Parrot_api_get_last_error to determine whether we have an error message 19:43
bluescreen you're think in the die case?
thinking
NotFound dukeleto: I don't think we should support doing method calls with wrong number of arguments on purpose. So if it breaks something, too bad for who is doing it.
whiteknight bluescreen: there are several exit conditions 19:45
bluescreen yes!
dukeleto NotFound: i think you misunderstand
whiteknight "exit 0" is a normal exit. "exit 1" usually indicates a problem, and unhandled exception is the worst
bluescreen: and then there are rare, fatal problems like memory corruption and stuff
dukeleto NotFound: currently, when you make a socket in Parrot, it defaults to IPv4. I think that is sane. But there needs to be a way to create an IPv6 socket 19:46
NotFound whiteknight: Did you see that guy? : <rfw> i'm looking at www.google-melange.com/gci/task/sho...9033192711 this task atm and just getting to know the rails before i pick it up
whiteknight so a non-zero exit may be an "exit 1" or an unhandled exception
NotFound: no, I didn' see him
dukeleto NotFound: so i will add "family" as another, optional, argument to the function
NotFound whiteknight: about 12 hours ago
whiteknight oh, I definitely didn't see them
19:47 kurahaupo_mobi left
bluescreen and the flag will indicate "normal exit" vs. doomed? 19:48
darbelo whiteknight: I'm not sure if it's what you mean, but calling exit() from inside libparrot is a bad idea.
whiteknight darbelo: no, the exit_i opcode
NotFound whiteknight: He's learning winxed for the task. I haven't thinked about code-in as a winxed promotion tool.
whiteknight which should exit the interpreter, but should NOT exit the embedding rogram
NotFound: it is! It can promote anything you want
NotFound: If you have any ideas of things a student can do for Winxed, let me know and I will add a task 19:49
darbelo whiteknight: Excellent. Please disregard my comment.
whiteknight documentation, examples, etc
NotFound Even if you don't know yet that you want :D
dukeleto: fine for me. We can later think if we should make that argument mandatory. 19:52
dukeleto NotFound: yes, there are other optional flags that are IPv6-specific that we may want to support in the future 19:57
19:57 patrickas joined
dukeleto NotFound: so we will need to think about that 19:57
What kind of exception should be thrown on "gethostbyname" errors ? 19:58
Currently, something is printed to STDERR and that is it. LTA.
patrickas hello, I was trying to build latest parrot on Windows7 and got an error from config\\gen\\platform\\win32\\begin.c ... I had to remove the check "if WINVER < Windows2000" in order for it to build 19:59
19:59 silug joined
dukeleto patrickas: can you make a Trac ticket on trac.parrot.org and send an email to parrot-dev with a link to the ticket? 20:00
patrickas: that will get more eyes on the bug
patrickas ok ... will do
whiteknight bluescreen: I think we're getting close
dukeleto patrickas: thanks!
NotFound That may be the same problem I have with XP home
whiteknight very few failures left on my machine
dukeleto patrickas: try to include as much info in the ticket as you can, i.e. OS version, compiler version, etc...
bluescreen whiteknight: really... it was awesomed that it finished compilation
dukeleto Can I make a vtable method a multimethod? Or do I just make normal C optional arguments? 20:01
bluescreen whiteknight: what will be the procedure to merge that back in the trunk ( I know I'm jumping ahead )... will this be subject of a code review from an architect or something... because things really changed in embed_api branch 20:02
NotFound dukeleto: vtable method?
dukeleto NotFound: yes, something with METHOD ... 20:06
NotFound: in C
NotFound dukeleto: don't call it vtable, is confusing. 20:07
patrickas dukeleto: I created an account on trac.parrot.org ... when I login I get redirect to the main page without loggin in (I still have a button to login and still have no provilege to create tickets) ... I am not sure if I am doing something stupid or the server is !
NotFound dukeleto: add :optional, as in PIR.
20:07 kurahaupo left
dukeleto NotFound: i am incorrect. It is not a vtable method 20:07
NotFound dukeleto: is not C, is PMC preprocessing language.
dukeleto NotFound: just realized that. It is just a method. 20:08
NotFound: I like to think of it as "C with sugar", but yes.
NotFound It has no official name, so name as you like ;)
dukeleto patrickas: Trac will redirect you to the main page, does it say you are logged in at the top right?
patrickas no .. and I still have the links for [login] and [register] 20:09
dalek rrot: a34696c | mikehh++ | / (3 files):
add patch (modified) from bacek++ to fix TT #1872
dukeleto patrickas: and what happens if you reload? Still the same? 20:10
NotFound Cookies blocked?
patrickas dukeleto: yes same I reloaded and tried a differnt browser 20:11
just tried using ingognito mode
seems to have worked ... there must be something messed up with my cookies! 20:12
dalek rrot/tt532_headerizer_refactor: 237a741 | jkeenan++ | config/gen/makefiles/root.in:
Add 'headerizer_tests' target.
20:13 kid51 left
cotto What'd be a good name for the policy that concerns what goes into core? "Inclusion"? 20:14
NotFound Coredumping? ;)
whiteknight bluescreen: when the time comes we 1) bring the branch up to date with master 2) Send out messages to all the developers for review. 3) incorporate feedback, 4) figure out what the deprecation issues and work out a merge timeline
bluescreen gr8 20:15
whiteknight bluescreen: I think a Parrot_api_get_exit_status function will be needed with two flags: The first is a boolean to say whether we have an unhandled exception or error message, and the second will be the exit code 20:16
And if we have an error message, we call Parrot_api_get_last_error to retrieve it
or, we merge the two functions together and only have Parrot_api_get_result that has three returns: exit type, exit code, and error message 20:17
20:17 TypeNameHere_ joined
bluescreen to me the simpler the better... I'm thinking in people using the API 20:18
20:18 gbacon joined 20:26 ZeroForce joined
patrickas dukeleto: Wrote the bug report and clicked [preview] and got the error that I am not logged in ... and now I am unable to log in even in incognito mode ... so I'll just send the report itself to parrot-dev, apologies for failing to do it the right way! 20:27
dukeleto Is Parrot_io_sockaddr_in a public facing function? I think not, but I want verification. 20:29
whiteknight, NotFound : ^^^
patrickas: no worries. Sounds like you have browser issues. Are you using NoScript or something?
20:29 patrickas left 20:30 patrickas joined
dukeleto patrickas: sending to parrot-dev is fine 20:30
patrickas dukeleto: I am not using no script .... I have a really slow and often flaky internet connection but I don't usually have such problems, (usually I just get random timeouts and I just need to refresh) 20:31
dukeleto: Thanks for your help anyways. 20:32
wesjdj moritz: When you publish a new translation task, please tell me :)
NotFound dukeleto: I don't think so, is a platform interface internal detail.
moritz wesjdj: I've published it, but it needs aproval from dukeleto
wesjdj Oh ok.
20:33 patrickas left
dukeleto NotFound: thanks! 20:35
dukeleto goes to approve the gci task
wesjdj: just approved the task 20:36
moritz: ^^
wesjdj dukeleto: OK I'll have a look. 20:38
Looks like I have to register as a student before I can take it... better get that done now-- 20:40
dalek rrot/tt1798_ipv6: 2f63e9f | dukeleto++ | / (3 files):
Allow creation of IPv6 sockaddr structs, whilst defaulting to IPv4
 20:41
tadzik is ResizableIntegerArray filled with zeroes by default?
dukeleto tadzik: i don't think so 20:45
tadzik: but you should check :)
tadzik hmm
if the function name is void set_integer_keyed_int(INTVAL key, INTVAL value), what is the opcode name? 20:46
in PASM 20:47
moritz you're writing PASM? not PIR?
tadzik bf is written in PASM
at least the basic part
moritz in pir you'd say $P0[1] = 2
tadzik hello world prints "Ðíôô÷ World!", looks like a notzeroed memory to me 20:48
cotto if anyone has ideas about what should and shouldn't go into core, msg me. 20:49
NotFound tadzik: quick trick: write the op you need in pir and 'parrot -t1' it. 20:50
dukeleto NotFound: that is a good trick!
wesjdj moritz: I requested the new task :)
tadzik > echo '$P0[1] = 2' | parrot -t1 - 20:52
error:imcc:syntax error, unexpected PREG ('$P0')
pardon my paste. Am I doing it right?
NotFound tadzik: pir needs that the code is placed inside a sub 20:53
moritz as I said I only know PIR :(
whiteknight dukeleto: to answer your question, I don't know if that function is public or not 20:54
tadzik Yay, I fixed brainfuck
whiteknight tadzik++ 20:55
NotFound Now you should fuck brainfix
tadzik or not
oh, it's the optimizing compiler that's broken 20:56
alright, I fixed the optimizing one 21:03
yay, my first Parrot related commit
dalek rrot: 9a7cfab | bacek++ | include/parrot/pointer_array.h:
Properly calculate CELL_PER_CHUNK.
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1362) fulltest) at a34696c - Ubuntu 10.10 amd64 (g++-4.5) 21:07
dukeleto 'make headerizer' has a lot of warnings. Are others seeing this? 21:15
21:17 bacek joined
mikehh bacek: did I mess that one up 21:19
bacek mikehh, hio. Mess what?
21:19 rurban_ joined
mikehh bacek: ref your last commit 21:20
bacek mikehh, ah. No. It was actually me :)
In original patch.
moritz, ping.
mikehh bacek: well it seemed to test ok 21:21
moritz bacek: pong
bacek mikehh, it should be ok now. I basically restored bit of functionality from master before gc_ms2_faster_list
moritz, can you retest rakudo on bleeding edge parrot? 21:22
21:22 rurban left, rurban_ is now known as rurban
moritz bacek: is RELEASE_2_10_1-509-g61d023e new enough? 21:22
bacek moritz, nope.
moritz bacek: then I'll rerun spectests 21:23
bacek moritz, it will be really helpful 21:24
21:25 rfw joined
bacek moritz, how you generate "RELEASE_2_10_1-509-g61d023e"? 21:25
rfw hi, i'm here about www.google-melange.com/gci/task/sho...9033196801
i take it whiteknight is my mentor?
whiteknight let me look
yessir, that's me 21:26
rfw :D
whiteknight Hello rfw, how are you today?
rfw i'm doing fine and on a roll
whiteknight that's great to hear! I would love to see that reverse Laplace expansion code you have done already
rfw it says recursive :p
moritz bacek: git describe --tags
whiteknight ah, right 21:27
dalek rrot/tt1798_ipv6: a34a716 | dukeleto++ | t/pmc/socket_ipv6.t:
Specify the protocol family when created sockaddr's for the IPv6 tests
whiteknight I've never actually used PLA with Winxed, so I'm very interested to see the code
dalek rrot/tt1798_ipv6: e262295 | dukeleto++ | / (2 files):
Add more IPv6 Socket tests
bacek moritz, RELEASE_2_10_1-511-g9a7cfab is what I get on my box. Looks like 2 most valuable commits difference from your tag :) 21:28
rfw whiteknight: pastebin.com/nC7YCjzC
i haven't tested it too thoroughly, though
moritz bacek: segfaulted while trying to compile Test.pm 21:29
bacek moritz, sigh...
moritz, Linux/amd64 + optimize? 21:30
moritz bacek: correct
moritz -> sleep
running spectest now
(compilation succeeded on second attempt)
bacek moritz, good night
moritz will report results in a copule of hours
bacek moritz, thanks! 21:31
NotFound whiteknight: i can co-mentor if winxed help is needed.
rfw oh, ni NotFound
hi*
21:31 tcurtis left
rfw thanks for the help yesterday 21:31
whiteknight NotFound: thanks! 21:32
bacek afk # breakfast
NotFound rfw: it was this morning for me. Parrot world is big ;)
rfw hehe
whiteknight rfw: awesome!
That's a great example
rfw heh, too bad that task's gone already
i was wanting to do that but i had a moinmoin task tied up 21:33
whiteknight what task?
rfw the nummatrix2d user examples one
i've gone for the test complexmatrix2d one instead
close enough :D
whiteknight oh. Nobody else is doing it I don't think. You can take it again if you want it
yes, that's a good one too :)
rfw again? you have to make one
since there was only one
whiteknight: should i withdraw from the complexmatrix2d one? 21:35
whiteknight NO! We need tests too!!
rfw well
can i do it after?
i'm only allowed to take one task at a time :(
whiteknight Let me look. I don't think anybody else is doing it
rfw it's already closed, lol 21:36
er, accepted
rather
NotFound rfw: how do you learnt winxed? Looking at the examples? 21:37
rfw NotFound: yeah
well, and bugging^Wasking you
whiteknight That task still exists. Looks like it is claimed by somebody named Fernando
rfw yeah
which is why i've opted for complexmatrix2d test cases
21:37 darbelo left
NotFound I think I can say that the objective of making it easy to learn is acomplished :) 21:38
whiteknight rfw: I just accepted it. Sorry for the delay
this google interface is not always easy to navigate
rfw NotFound: i wouldn't say easy :3
whiteknight: am i required to write the tests in nqp?
or is any language that runs on parrot fine 21:39
NotFound rfw: I know, I must write some doc and tutorials.
whiteknight rfw: That's a good question. I assumed NQP, but there is value in having tests in other languages too
rfw whiteknight: i don't know perl (i loathe perl)
but i could pick it up if you want 21:40
whiteknight Hah!
you would prefer Winxed?
rfw yeah
21:40 bluescreen left
rfw i'd be even happier if there was a working implementation of python on parrot, lol 21:40
whiteknight Okay. Winxed is fine. I'll have to modify the harness eventually to be able to run winxed tests as part of the harness
rfw ah cool!
so, 15 tasks? 21:41
whiteknight there is a python implementation, but I don't know if it's currently working
rfw er
tests*
yeah it looked iffy
whiteknight yes. if you take a look at the current tests you can see some TODO notes where we need more coverage
rfw parrot-linear-algebra / t / methods / complexmatrix2d / conjugate.t ?
er
not the conjugate.t part
i don't see any TODO notes 21:42
whiteknight hmmm... I thought there were more
there are two in gemm.t
rfw ah 21:43
i see them
whiteknight You can decide what to test, just try not to test things that are already tested.
rfw sure
Kovensky <@whiteknight> this google interface is not always easy to navigate <-- melange is a buggy mess :(
whiteknight Most tests just cover basic behavior: What are the results when all the inputs are perfect?
rfw hi Kovensky
Kovensky hi rofflwaffls
whiteknight Kovensky: Thanks for saying that!
rfw caws Kovensky
Kovensky why did you drop your full nick for a three letter one anyway
rfw i dunno
it seemed more pretentious 21:44
Receiving objects: 7% (183/2373), 604.00 KiB | 11 KiB/s
wtf
Kovensky whiteknight: I had to use it for gsoc early this year... definitely not nice indeed :E
21:44 tcurtis joined, tcurtis left
Kovensky rfw: :3 21:44
whiteknight Kovensky: what was your GSoC project? Who did you work with?
Kovensky github.com/Kovensky/x264-audio 21:45
it's kinda unfinished tho :/
rfw Kovensky: does it encode aac
21:45 bacek left
Kovensky yes, using qtaac or libfaac through libavcodec's wrapper 21:45
you can use libavcodec's aac encoder too if you want, but don't complain about the result :> 21:46
rfw my parrot working environment is a lolmes
s
i'm doing it in a virtualbox shared folder with a terminal tunneled from my vm :<
Kovensky why so complicated
rfw because >windows 21:47
Kovensky what about cygwin
rfw cygwin is really lol 21:48
i tried it with pla/blas/lapack before
it didn't work very well at all :(
Contextual $*FileSystem not found 21:50
wat mean D:
oh
hurr i can't read
whiteknight I've never tried to install BLAS on windows or cygwin 21:51
rfw whiteknight: This is resolved by using the special forked version of Kakapo, as discussed in the README <-- i don't see this
whiteknight you don't? damnit
rfw i just cloned gitorious.org/kakapo
and built it
whiteknight github.com/Whiteknight/kakapo.
rfw and 21:52
ah*
thanks
whiteknight yeah, the master one is all b0rked
I made a series of ugly but functional fixes to my fork
rfw Receiving objects: 8% (202/2386), 868.00 KiB | 235 KiB/s
YES NO MORE 11KB/S
Kovensky: is rakudo star really that bad 21:53
Kovensky I haven't used rakudo yet
rfw i heard it's really _really_ slow and a huge memory hog
whiteknight: it runs :) 21:54
i might as well write nqp, since the rest is in nqp
Kovensky: you win; i'm writing perl :(
dukeleto waves
dukeleto then becomes a particle
Kovensky rfw: :3 21:55
rfw :<
Kovensky which perl
rfw nqp
well
i'm not writing perl
!!
whiteknight: is method test_METHOD_gemm_aABbC() { missing an assert_equal at the end 21:56
NotFound Not quite.
whiteknight nqp is perl6, not perl5, if that makes any difference
rfw i don't know either so it's all good 21:57
whiteknight rfw: yes! it is missing an assert
damnit
rfw lol
NotFound Maybe someone should write nqp5
rfw this doesn't count as 1/15 does it :(
whiteknight it can. Since technically it's adding a test I didn't have before
rfw lol
whiteknight though the other 14 better be really good 21:58
rfw looks for more missing asserts
parrot-nqp: symbol lookup error: ./dynext/linalg_group.so: undefined symbol: cblas_zgemm
is this broken? 21:59
or is this my cblas's fault
hm i'll just not touch gemm 22:00
22:04 ZeroForce left 22:05 ZeroForce joined
rfw whiteknight: question: how do i add a new test to a case? 22:07
also fancyvalue doesn't work with complex 22:14
not ok 1 - test_fill_with_resizing_complex 22:19
test failed
dalek nxed: r690 | NotFound++ | trunk/winxed_installed.winxed:
add options -I and -L to installable compiler
 22:26
nxed: r691 | NotFound++ | trunk/pir/winxed_ (2 files):
update installable files
nxed: r692 | NotFound++ | trunk/examples/xlibtest.winxed:
rename main function to 'main' in xlibtest example to avoid the weakness of the
22:29 bacek joined
rfw whiteknight: complexmatrix2d.conjugate is the _conjugate_ of all the numbers in the matrix, not the hermitian transpose of the matrix? 22:29
$m = 1-1i 2-2i 3-3i 4-4i 22:30
$n = 1-1i 2-2.81617e-322i 3-3.60739e-313i 4-0i
might want to review that :)
er, i put m and n the wrong way around 22:31
dukeleto cotto: i enjoyed your blog post about dynop_mapping. I understand how it effects PL/Parrot now
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1363) fulltest) at 9a7cfab - Ubuntu 10.10 amd64 (g++-4.5 with --optimize) 22:34
dukeleto whiteknight: your links to github are broken on whiteknight.github.com/2010/11/11/p..._imcc.html
whiteknight: they refer to the fork you deleted
22:34 fperrad left
dukeleto aloha: msg whiteknight I enjoyed your blog post about the trials of a product manager. Good stuff! 22:37
aloha dukeleto: OK. I'll deliver the message.
dukeleto Does POD that is in src/*.c files actually go anywhere? Or is it just for people reading the C source? 22:48
I don't see them on docs.parrot.org/parrot/latest/html/ 22:49
sorear I think make codetest verifies that it exists 22:54
rfw whiteknight: are you there yet? 23:06
i have a truckload of questions :(
plobsing dukeleto: it doesn't go anywhere. it is nifty to run perldoc on sourcefiles to get the gist of what's going on. 23:08
rfw how do i define a float in nqp? 23:12
tadzik Well, you can just bind the value to a variable, or do you need strictly a float? (It's Num btw) 23:13
rfw ah 23:14
okay
dalek rrot/leto/embed_grant: d2c740c | dukeleto++ | docs/embed.pod:
[docs] Add some basic docs about type signatures and improve description of some datatypes
 23:16
rrot/leto/embed_grant: 9771db2 | dukeleto++ | docs/embed.pod:
Add a more user-friendly description of Parrot type signatures
mikehh rakudo (82f3de0) - build on parrot (511-g9a7cfab) - make test PASS, make spectest_smolder[(#1364) roast (52c1e70)] PASS - Ubuntu 10.10 amd64 (g++-4.5 with --optimize)
t/spec/S02-magicals/pid.rakudo - TODO passed: 1
27,041 ok, 0 failed, 609 todo, 1,854 skipped and 1 unexpectedly succeeded
23:21 gbacon left 23:23 rfw left 23:25 rfw joined
cotto dukeleto, great! I'm glad it was helpful. 23:27
consider me encouraged 23:28
23:32 mikehh left
whiteknight dukeleto: thanks for the feedback! Sucks about the broken links. I will try to update them 23:33
rfw: sorry, dinner 23:35
rfw whiteknight: ah okay 23:38
(sorry, was at lunch)
23:38 ZeroForce left
rfw whiteknight: i think non-square conjugate is broken 23:38
or just really inaccurate 23:39
1-1i 2-2.81617e-322i 3-2.80872e-237i 4-4.10654e-289i
whiteknight rfw: hence the need for more tests!
rfw hehe
well
currently i'm just porting the tests in testlib 23:40
to use complex numbers
i can't seem to get gemm to work
whiteknight conjugate should just be the straight conjugate of all elements in the matrix. Not the hermitian transpose
rfw ah
so it is broken
whiteknight I haven't run the tests in a while
rfw parrot-nqp: symbol lookup error: ./dynext/linalg_group.so: undefined symbol: cblas_zgemm
can you offer any help with that?
whiteknight weird. What blas library are you linking with? 23:41
rfw not atlas
hold on'
atlas wouldn't compile
whiteknight what OS are you on?
rfw archlinux
whiteknight I've never been able to compile atlas myself either
hmmm...
rfw www.archlinux.org/packages/extra/i686/blas/ <-- whatever this is 23:42
23:42 jan left
whiteknight www.archlinux.org/packages/extra/i686/blas/ 23:42
gotcha
rfw yeah i linked against that
[root@myhost parrot-linear-algebra]# ldd dynext/linalg_group.so
linux-gate.so.1 => (0xb7879000)
libblas.so.3 => /usr/lib/libblas.so.3 (0xb7806000)
hm whiteknight can i just give you my gemm tests and you try them out? 23:44
whiteknight yes
rfw what do you mean by # Test that we can call GEMM with alpha of various types
23:44 jan joined
rfw just try GEMM with a complex alpha and a float alpha? 23:44
whiteknight rfw: pass in various PMC types for alpha. Numeric types (Int, Num and Complex) should work. Weird types probably should throw an exception 23:45
rfw ah
so
i should also throw in a string :)
whiteknight so there are 4 tests right there: three with numeric types that should work, and another with a non-numeric type that should fail
a string may auto-convert to an integer
rfw so i should make one that doesn't like
"this is absolutely totally not an integer" 23:46
whiteknight so that will be an interesting test, but won't necessarily fail
yes
rfw how do i define an int? (sorry for all the stupid questions, i'm so new at this :(
23:47 mikehh joined
whiteknight in NQP or Winxed? 23:47
rfw nqp
whiteknight my $int := pir::new__PS('Integer'); 23:48
rfw i'm doing this all in nqp since all the other tests are nqp, lol
whiteknight that's good. Learn something new
put "perl" on your resume
make lots of $$
rfw yes, perl is full of funny symbols like $$
how do i set the value of it 23:49
can i just do $int := 1 and it'll be an int?
whiteknight no, all numeric literals in NQP are Floats
it's a very very small language
rfw ah
so how do i create an integer with a value of, say, 2? 23:50
whiteknight pir::set__PI($int, 5), I think
tests with Integers are always hard
rfw The opcode 'set_p_i_i' (set<3>) was not found. Check the type and number of the arguments
whiteknight at least in NQP
jnthn my $int := pir::box__Pi(5);
whiteknight jnthn++ 23:51
he's the grandmaster at this kind of stuff
rfw ah
thanks
also i can't help but feel i'm writing pascal, with all the :=
jnthn rfw: NQP is a subset of Perl 6. := is the binding operator. Perl 6 has = for assignment. However, NQP doesn't have assignment. 23:52
rfw ah
jnthn (Deliberate design decision.)
rfw this is the first time i've ever touched perl
jnthn NQP is generally good for the stuff it was designed for, but probably a little restrictive for general purpose programming. 23:53
e.g. I love working with it for access to the grammar engine and writing compiler and meta-model bits.
But wouldn't want to write e.g. a web app in it. :)
rfw so rakudo is intended to be the more general purpose version of nqp? 23:54
jnthn Rakudo aims to be a full Perl 6 implementation.
NQP is aimed at being a small, bootstrappable language ideal for compiler implementation.
It is more restricted than Perl 6
rfw ah 23:55
jnthn But can make various extra optimizations as a result and generate simpler code.
rfw and at the lowest level, you have pir?
jnthn These days, some of NQP is also cross-compilable to the .Net CLR.
Yes, in fact NQP gets compiled into PIR. 23:56
rfw parrot .net? awesome
jnthn Well, not Parrot on .Net
rfw well
NQP compiling into ILASM
at least
jnthn More that writing compilers using NQP, PAST and so forth means that you'll be able to get a VM-portable compiler.
With a modest investment. 23:57
rfw awesome
jnthn (e.g. it won't "just run" but you'll be much of the way)
rfw whiteknight: pastebin.com/xfLJgNdd can you take a look at that
jnthn Somebody is working on a JVM port too :)
mberends++
rfw whoops, how did i type 2+01 instead of 2+0i 23:58
jnthn No, you just imagined it. ;)
rfw i see what you did there
whiteknight rfw: nice. I'm building now, I'll run these in a minute 23:59
rfw whiteknight: oh no, don't run them yet
i'm just asking if they look fine
whiteknight oh, okay! that's even less work for me :)
rfw also, no fancyvalue with imaginary numbers?
i get funny results when i do that

Powered by Cro and the Raku® Programming Language.

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

Thank you!