www.parrot.org/ | Parrot 1.4.0 "Mundo Cani" Released!
Set by moderator on 21 July 2009.
00:01 Whiteknight joined
dalek TT #108 closed by jkeenan++: remove Data::Replace? 00:02
rrot: r40206 | jkeenan++ | trunk (3 files):
Remove Data::Replace per trac.parrot.org/parrot/ticket/108.
rrot: r40207 | jkeenan++ | trunk/config/gen/makefiles/root.in:
Remove 'make' target fulltest_all. Cf.: trac.parrot.org/parrot/ticket/862.
 00:06
00:06 Austin joined
Austin Hello, parrot.
ping pmichaud 00:07
purl I can't find pmichaud in the DNS.
Austin Thanks, purl.
purl no worries Austin
00:08 hercynium joined
Austin purl message pmichaud Did your commit #39895 (trac.parrot.org/parrot/changeset/39895) -- a fix for tt#803 (bad register generation for attribute lookups of register vars) make it into trunk? 00:09
purl Message for pmichaud stored.
Austin Thanks purl
dalek TT #862 closed by jkeenan++: remove fulltest_all make target
pmichaud Austin: I'm pretty sure it did, as I don't recall committing it to a branch 00:10
00:10 Austin joined
pmichaud 39895 shows that it was changed in trunk 00:10
Austin Did I just miss pmichaud?
Yes.
Hello, pmichaud.
Hmm. 00:11
cotto nice 00:12
Austin Hello, cotto. 00:13
kid51 Coke: no make testb failures for me at r40286 Linux i386. 00:15
cotto Coke, same here
Coke alrightee. must be related to removing jsr. 00:17
thanks.
dalek kudo: 78644a3 | pmichaud++ | (2 files):
Add Junction.Str, update Makefile.
 00:23
00:25 patspam joined
treed Is there a way to tell, from a method in NQP, what method is the next highest? (IOW, how it got there?) 00:26
I'm seeing an odd issue where according to --target=parse, a rule is only being used once. But with debug output from the actions.pm, I'm seeing that it's getting called twice. 00:27
Whiteknight hello Austin 00:28
hello treed
pmichaud treed: that often means there are multiple {*}'s in the rule 00:29
it also may depend on what is happening inside the action method
treed Getting called twice with different data.
Hi Whiteknight. 00:30
GeJ Good morning porters
pmichaud what about --target=past ?
treed: anyway, you're asking for a backtrace, and no, we don't quite have it yet.
treed Nothing indicating why it would run lhs again.
Damn.
Processing a as an lhs. 00:31
Got a as an lhs.
Processing puts as an lhs.
That's the debug output.
"puts" is never given to assignment as an lhs
GeJ cotto: FYI about the intermittent failure I told you yesterday on my amd64 box (FreeBSD), I wasn't able to reproduce it on an i386 setup.
treed And assignment is the only rule that uses lhs.
pmichaud this is one reason why I'm hoping to get backtraces from non-exceptions.
Whiteknight good morning GeJ
treed Can I get a backtrace if I throw an exception?
pmichaud yes, but currently it requires some extra code to display it
treed Ah.
pmichaud I'm hoping to have that extra code built-in to PCT soon.
I've added it to my task list. 00:32
GeJ heya Whiteknight. How's life treating you?
Whiteknight GeJ: absolutely fantastic 00:33
GeJ glad to hear that. 00:35
Whiteknight how bout you? Things going well? 00:37
00:37 wayland76 joined 00:38 Austin joined
wayland76 I have a question. I downloaded the parrot 1.4 release, and noted that there's no "ports" directory 00:39
Is that because it's a release, or has something happened to "ports"?
dalek rrot: r40208 | whiteknight++ | trunk (5 files):
[gc_options_remove] apply patch from kid51++ for the gc_options_remove branch. Required some tweaking of file names to match things that have changed in the GC in recent months. For TT #490
 00:42
GeJ Whiteknight: I've known better. $job is currently about porting parts of a C# app to Perl. The only comforting part is to have the confirmation that barely-competent-monkey'ing is not a Java exclusivity.
Whiteknight oh, that sounds fun. Most of what I do at work is C#.
00:42 bacek joined
GeJ wayland76: ports still seems to be in svn. I'd say that the directory is removed from the release tarball. better ask cotto or a previous release manager for a definitive answer. 00:43
wayland76 Hmm. Well, it makes the releases useless to me. 00:44
Ok, I'll continue investigating. Thanks :)
Whiteknight I think the releases get packaged for mac ports separately and released for that platform separately by a packager 00:45
Infinoid: ping
dalek rrot: r40209 | whiteknight++ | branches/gc_options_remove:
Removing branch created by kid51++ and committed to trunk in r40208
TT #490 closed by whiteknight++: Remove support for non-default config options for Configure.pl's --gc
wayland76 Well, ok, but I'm using Fedora
and the most recent Fedora packages are 0.9 00:46
Whiteknight oh, that's what I meant. It's packaged for different systems independently of the release
not just mac
I thought there was a more recent Fedora package, I wonder who has been responsible for making packages on that system
wayland76 Ok. I'm hoping to get it to the point where a release can be automatically turned into a package. 00:47
Infinoid Whiteknight: hi!
wayland76 There have been more recent RPMs built, but the Parrot site doesn't link to them
Whiteknight Infinoid: I want to get io_cleanups locked down and committed ASAP, it's stagnating 00:48
maybe we need to create a diff and take a look at what all has really changed 00:49
dalek rrot: r40210 | whiteknight++ | trunk/DEPRECATED.pod:
[DEPRECATED] remove the GC-related stuff, since that was just removed from trunk
Infinoid Whiteknight: You're right. Unfortunately I'm pretty buried in work this week, how's the weekend look for you? 00:51
Whiteknight On saturday I'm in NYC for a wikiconvention, but should have some sporadic hacking time then. Plus, free all day Sunday 00:52
cotto wayland76, make release VERSION=x.y.z is the target that creates the tarball. I know it removes DEVELOPING, so it probably also removes ports. 00:53
wayland76 cotto: I see the line in the makefile that removes DEVELOPING, and there's no mention of ports 00:55
config/gen/makefiles/root.in
purl config/gen/makefiles/root.in is, like, what generates parrot/Makefile
cotto It looks like ports isn't in MANIFEST 00:56
pmichaud japhb: ping 00:57
wayland76 Right. So does that mean it's just an oversight?
darbelo wayland76: It might not be. I think the tarball is targeted at users, not packagers. 01:00
dalek kudo: 21066f1 | pmichaud++ | src/setting/Any- (3 files):
Fix boxing of strings in .flip, .join, .lc, .uc, and others.
wayland76 Ah, they're in MANIFEST.SKIP 01:01
darbelo: Well, as a user, I like to get my tarball and go "rpmbuild -ba <tarball>" and have it build automatically 01:02
So as a user, I'm finding the releases completely useless :( 01:03
dalek rrot: r40211 | whiteknight++ | branches/bsr_jsr_ret:
[bsr_jsr_ret] creating a branch to remove these three opcodes
pmichaud afk (returning to OSCON venue) 01:05
bacek Whiteknight: don't forget to update vim syntax file after removing opcodes :)
cotto wayland76, file a tt. I'm not familiar with what's in ports, but it sounds like its exclusion could use some discussion.
That and the emacs file should probably be generated. 01:06
That might make a good newbie task.
darbelo wayland76: In the mean time, you culd check out the release as tagged in svn and generate your rmps from there. 01:07
wayland76 cotto: Ok, will do 01:08
darbelo: Good idea; I'll try that
jdv79 where are the pcc rework plans? 01:10
01:11 theory joined
dalek rrot: r40212 | whiteknight++ | branches/bsr_jsr_ret (3 files):
[bsr_jsr_ret] remove jsr opcode
 01:11
rrot: r40213 | whiteknight++ | branches/bsr_jsr_ret (3 files):
[bsr_jsr_ret] remove more references to jsr
 01:14
Whiteknight jdv79: inside Allison's head, I think 01:15
dalek TT #873 created by wayland++: Blanket ports exclusion needs discussion
rrot: r40214 | whiteknight++ | branches/bsr_jsr_ret/src/ops (2 files):
[bsr_jsr_ret] on my signal, unleash hell: Remove the bsr and ret opcodes
 01:18
cotto waits for Whiteknight's signal
The release is out. Let's blow stuff up! 01:19
mikehh Whiteknight: just checking out your wiki/blog, haven't checked it for a few days - really like your Scheduler object concept
cotto well, more stuff. ;)
kid51 Hmm, all my 1.5 milestone tickets are closed. 01:20
mikehh cotto: I thought the release went particularly well
cotto Thanks. I liked it when other release managers said what they were going to do, so I tried to do a lot of that.
kid51 thought the release went particularly well because he slept right through it ;-)
Whiteknight this bsr/ret stuff is a lot more prevalent then I thought it was 01:21
and removing them from PIR code isn't exactly a straightforward transformation
cotto Yeah. I love waking up to a release too.
afk
wayland76 kid51: Have you any further thoughts on trac.parrot.org/parrot/ticket/712 ? :) 01:22
mikehh transformations are *never* as straightforward as you think at first
01:23 jimmy joined
jimmy hello 01:23
Function Renaming , # cx: concurrency scheduler. is it wrong?
I think it is cs 01:24
ps: from trac.parrot.org/parrot/wiki/APIFuncRenaming
japhb pmichaud: pong 01:25
kid51 wayland76: I was hoping that that would get cleared up as part of the larger work on install issues that pmichaud and particle were discussing at YAPC.
I know that in the following week's discussion on #parrotsketch, there was heavy discussion between pmichaud and allison about those issues. 01:26
And I was hoping for some clear resolution.
But I don't have clarity yet.
wayland76 ok
kid51 I believe allison got *some* issues resolved, but I don't know how much agreement they had on the issues pmichaud raised in his post on list. 01:27
mikehh they are both at OSCON are they not?
kid51 yes
though allison may be there in her capacity as an O'Reilly person
pmichaud is there as speaker/attendee 01:28
mikehh yeah - but they might just get some face-to-face discussion
wayland76 Any idea roughly when the parrotsketch discussion was? 01:29
ie. which month, and whether it was towards the start or the end? 01:30
kid51 let's see ... probably June 30
week after YAPC
wayland76 looking...
kid51 Patrick's July 8 post to parrot-dev is also important 01:31
Whiteknight purl msg jimmy the concurrency scheduler is "cx", not "cs".
purl Message for jimmy stored.
kid51 wayland76: Try this: irclog.perlgeek.de/parrotsketch/2009-06-30 01:33
... starting at about 18:45
wayland76 I was up to 18:39 when you posted this :). Thanks 01:35
kid51 wayland76: Added 'component: install' to your ticket. pmichaud will pick it up when he searches
This is a problem we face which is probably endemic to distributed open-source projects ... 01:36
viz., that when the core project leaders don't yet have a consensus about how to proceed on some issue, ... 01:37
... it's difficult to break down that issue into steps that can be tackled by other developers
wayland76 ah, thanks :) 01:38
kid51 I agree with you that I hope allison and pmichaud can meet up at OSCON. allison's injury prevented her from coming to YAPC and that was a big loss in what was otherwise a very productive (for parrot) experience 01:40
01:40 dukeleto joined
kid51 afk 01:41
wayland76 I still haven't gotten a handle on what they don't agree on. Will keep reading tickets... 01:42
01:50 darbelo left
Whiteknight goodnight #parrot 01:52
dalek rrot: r40215 | whiteknight++ | branches/bsr_jsr_ret (2 files):
[bsr_jsr_ret] remove bsr/ret from two PCT-related library files. Build still doesn't complete but gets further. Need to test these things to make sure I haven't ruined anything
 01:53
pmichaud (install) I think allison and I are in basic agreement on the direction parrot will take 01:58
it's summarized in my Jul 8 post to parrot-dev
message Whiteknight note that it's also possible to remove bsr/ret by using local_branch and local_return, instead of creating separate Parrot subs for it 02:00
purl Message for whiteknight stored.
kid51 What's the command in gdb to run a .pir file to get a backtrace? 02:07
pmichaud run x.pir 02:09
02:09 mokurai joined
kid51 (gdb) run t/compilers/imcc/syn/macro_25.pir 02:09
Starting program: t/compilers/imcc/syn/macro_25.pir
No executable file specified.
Use the "file" or "exec-file" command.
Why doesn't that work?
pmichaud how did you invoke gdb? 02:10
it should be "gdb parrot"
normally I do: 02:11
$ gdb ./parrot
(gdb) run foo.pir
kid51 thx. updating rt.perl.org/rt3/Ticket/Display.html?id=60926 02:13
02:19 tetragon joined
japhb pmichaud: are you still looking for me? 02:20
pmichaud japhb: just wanted to give you a "heads up" to my response on parrot-dev :-) 02:28
japhb ah. I'm just finishingmy response to you. :-) 02:29
pmichaud okay!
we can haz profitable dizcussion?!
is that allowed on parrot-dev?!?
:-)
japhb heh 02:30
02:35 theory joined 02:42 janus joined
wayland76 kid51: What needs to happen to trac.parrot.org/parrot/ticket/788 ? Do we just need to update the PDD? And if so, am I right in guessing that it's PDD30 that needs updating? 02:44
pmichaud: Just out of interest, what's your objection to multiple make files? Is it that it makes it harder to search? Or that there is duplicated information? Or something else? 02:46
kid51 wayland76: AFAICT, if PDD30 were updated to reflect allison's concerns in her 2nd paragraph, the ticket would be closable 02:48
kid51 must sleep
purl $kid51->sleep(8 * 3600);
japhb pmichaud: After many interruptions, reply sent. :-) 02:49
pmichaud japhb: it's likely to be very difficult to optimize the for loop the way you've done it 02:53
(in Perl 6, at least)
not impossible, but not something we're likely to be able to do by 2.0 02:55
japhb pmichaud: Oh sure. My point was that I got 24x out of the work I did on that loop. Even getting only a few of those optimizations should show significant improvement, more than just a few percent. I'd be very surprised if PCT optimizations could not get 2x to 4x.
pmichaud sure, but you're also measuring far more than the cost of a sub
because in Rakudo that sub involves a *huge* amount of binding overhead due to the PCC issues
japhb Even if someone managed to just get 2x (averaged across the board) by the time 2.0 came out, I'd be happy -- because not only is the improvement in and of itself valuable, but we'd then have the structure in place to do more significant work during 2010. 02:56
pmichaud okay. But I don't think avoiding the sub call is going to be likely for Rakudo. 02:57
It's very likely for other languages.
NQP will be doing it.
NQP will probably be doing it before 1.5.
japhb pmichaud: Don't focus too much on just my single point about scope == sub. The other optimization types are at least as important in my mind.
Awesome!
purl hmmm... awesome is a window manager or at awesome.naquadah.org or awesome!
pmichaud But it requires more information than PCT has available at the moment 02:58
and I'm not sure it can be generically done for dynamic languages
we'll have to see
japhb fair enough. I still think you'd be surprised about the value of constant propagation, avoiding double loads, etc. 03:00
The PIR produced by the Perl version of that loop was eye-opening to me. 03:01
pmichaud rakudo is very non-representative there, though.
because of all of the dynamic things it has to be able to support. 03:02
yes, I do plan to move constants out of loops also soon -- likely before 1.6. It got a DEPRECATION notice.
anyway, to your overall point -- yes, I totally agree we should look at optimizations. 03:07
but what I've discovered in working with dynamic languages and a dynamic environment like Parrot is that the "obvious" places for optimizations actually turn out to not be optimizations, or require a heck of a lot of high-level language knowledge 03:08
I think I should shut up now. :-) 03:09
wayland76 ...or maybe *before* you proved that you were the ideal person to do it :) 03:19
bacek pmichaud: (in continue of optimisations) How hard to add self-hosted bootstrap to PCT?
japhb pmichaud: sorry, called away by $family
03:19 tetragon joined
bacek e.g. writing any optimisations in PIR will be painful. 03:20
So, instead of doing it, implement "optimisations steps" for PCT using NQP.
treed actually prefers to write in PIR vs. NQP. 03:21
pmichaud Oh, I expect PCT to be rewritten in NQP at some point.
Possibly even by 2.0.
NQP is going to get a *ton* of changes over the next few months.
wayland76 Only one ton? :) 03:22
pmichaud Only one.
bacek Yeah. And I've got few kilograms for this ton :)
japhb pmichaud: How much value and effort would be involved in having HLLs provide PCT with more info about what optimizations are allowed? A friend of mine posited last week being able to tell a dynamic language compiler that you explicitly *weren't* going to be dynamic in particular ways in a particular lexical/dynamic scope pair, so optimizations that used to be off the table could then be done.
03:22 tetragon joined
pmichaud japhb: that's the direction I think we'll ultimately be heading, yes. 03:23
For example, I expect PAST::Val to have flags on it that indicate when it's safe to re-use a register and avoid extra find_lex calls. 03:24
(and blocks can say "all of my variables are bind-safe, so fetch once and re-use the register afterwards") 03:25
japhb pmichaud: This was brought on by the blog post by (I believe) a Ruby-on-Java implementor who posted an extended explanation of why the maximum performance of a Ruby implementation was inversely proportional to the compliance of said implementation to the "spec" of standard Ruby. To which he said, "Well, why couldn't I put a pragma in my code to tell the compiler I didn't *need* strict adherence to the spec, that in fact I was willing to g
ive up guarantees A, B, C, and D, and to give me back the performance lost when providing those guarantees instead?"
pmichaud: Ah, good.
pmichaud anyway, isntead of looking at optimizing what Rakudo produces, I think a simpler stage would be to optimize what NQP produces :-) 03:27
and yes, NQP will have sub-avoidance
japhb good.
03:28 tetragon_ joined
pmichaud NQP has a lot more options for optimization than Rakudo does 03:29
otoh, Rakudo is the place where optimizations (for other dynamic languages) will matter
japhb I'm not sure I understand, unless you meant s/Rakudo/PCT/g 03:30
pmichaud Rakudo is at the other extreme
NQP is highly optimizable, because it's the "least dynamic"
bacek Convert PAST to SSA form can give more optimisation opportunities.
pmichaud it also has the most constraints
japhb Nod, but what did you mean by "otoh, Rakudo is the place where optimizations (for other dynamic languages) will matter"? 03:31
pmichaud Rakudo has the least optimization opportunties and the fewest constraints
bacek "Constant folding", CSE, Function inlining, etc.
pmichaud therefore, figuring out optmizations that can work for Rakudo will be useful to lots of languages 03:32
japhb gotcha
pmichaud basically, if we have a spectrum
NQP ..... other dynamic languages ..... Rakudo
03:33 tetragon joined
pmichaud then it's easiest to come up with potential optimizations for NQP, some of which will help some dynamic languages and many of which won't help Rakudo 03:33
if we can come up with optimizations that work for Rakudo, then they should be applicable to many dynamic languages, but they're harder to come up with
japhb yup, I understand now. 03:34
03:40 tetragon_ joined
wayland76 There's someone called "jerry" who I'd like to comment on ticket 92 ( trac.parrot.org/parrot/ticket/92 ). Does anyone know how I get in contact with him? 04:42
cotto wayland76, you mean particle aka jerry gay? 04:43
wayland76 Probably :) 04:45
particle: Are you able to comment on trac.parrot.org/parrot/ticket/92
Coke is the one who posted his name into ticket 92, so he's the most likely person to be able to comment sensibly on this 04:46
05:02 particle1 joined
wayland76 particle1: Are you able to comment on trac.parrot.org/parrot/ticket/92 ? 05:03
05:15 TiMBuS joined 05:39 tomyan joined 05:47 theory joined 05:50 szabgab joined 06:03 chromatic joined 06:16 mokurai joined
dalek rdinal: 4e749b6 | (Ted Reed)++ | (2 files):
Implement Array.collect!; test included.
 06:36
rdinal: 04e7817 | (Ted Reed)++ | (2 files):
Implemented Array.delete_at and Array.delete_if. Tests included
06:52 flh joined 07:09 s1n joined
wayland76 do we have a messaging bot? 07:13
TiMBuS wayland76, yes, purl can leave messages for people 07:14
moritz wayland76: aye, purl msg $nick $msg
iirc 07:15
wayland76 purl: msg particle1 Are you able to comment on trac.parrot.org/parrot/ticket/92 ?
purl Message for particle1 stored.
wayland76 Thanks :)
dalek kudo: 7b81c05 | moritz++ | src/setting/Num.pm:
the :Trig tag is gone from spec, remove it.
 07:18
07:19 iblechbot joined
mikehh codetest FAIL, All Others PASS (pre/post config, smolder, fulltest) at r40215 - Ubuntu 9.04 amd64 07:20
wayland76 Who do I ask whether I can have more powers in Trac? 07:25
nopaste "mikehh" at 86.149.232.155 pasted "codetest failures at r40215" (45 lines) at nopaste.snit.ch/17356
wayland76 (I just want to be able to change some of the ticket metadata)
szabgab I need some help here , I am finally trying to integrate the PBC based code highlighting into Padre. we have it for perl 6 already but I have not tried it with anything else 07:32
can you recommend me which project is the easiest to compile ?
is it possible to compile them with a parrot that was built by rakodo ? 07:33
moritz last time I looked, checking out cardinal (ruby) into parrot/languages/cardinal and then 'perl Configure.pl && make' went smooth 07:36
don't know if that matches your idea of "easy"
szabgab yeah that sounds good 07:37
nopaste "mikehh" at 86.149.232.155 pasted "first PATCH for codetest failures in r40125" (13 lines) at nopaste.snit.ch/17357 07:38
mikehh podchecker passes for the patch 07:40
szabgab moritz, that worked, great, thanks 07:43
07:44 s1n joined
szabgab is there an accepeted env variable such as CARDINAL_DIR that could/should point to the directory of cardinal? 07:44
cotto mikehh, thank 07:55
thanks, even
dalek rrot: r40216 | cotto++ | trunk/config/auto/gc.pm:
[codingstd] fix some broken POD, patch courtesy of mikehh++
 07:57
mikehh the second failure I think is the test data rather that the c 08:02
checking now 08:03
yup - passes test - pasting now 08:08
nopaste "mikehh" at 86.149.232.155 pasted "second PATCH for codetest failures in r40125" (12 lines) at nopaste.snit.ch/17358 08:10
mikehh that's interesting - if you add a c prog you need both the c "pod" and need to add data to the test 08:12
cotto mikehh, thanks again 08:28
dalek rrot: r40217 | cotto++ | trunk/t/codingstd/c_function_docs.t:
[codingstd] add pbc_info to c_function_docs.t, patch courtesy of mikehh++
 08:30
08:32 dukeleto joined
mj41 Hi. Sorry, TapTinder and ttbot are gone until Friday. Server Room Air Conditioning failed. :-( 08:37
cotto ouch 08:44
08:53 mj41 joined 09:11 wayland76 joined 09:21 Patterner joined 09:27 donaldh joined 09:39 MoC joined 09:45 darkwolf joined 09:46 rob joined, mikehh joined
darkwolf hi 09:47
mikehh hi 09:48
rob hi 09:49
purl niihau, rob.
darkwolf I have problem compiling last parrot from svn rep 09:50
mikehh what platfornm - what dis it complain about 09:51
s/dis/did/
darkwolf it ends with ./md4.c:61:75: error: invalid suffix "PARROT_DYNEXT_EXPORT" on integer constant
make[1]: *** [md4.o] Error 1
amd64
09:54 bacek_ joined
bacek_ o hai 09:54
mikehh did you make clean/realclean
09:54 mokurai left
darkwolf i just made rm -rf 09:55
it's lame but works ;)
so it's in clean new snap 09:57
mikehh I just built on r40217 - Ubuntu 9.04 amd64 09:59
darkwolf mine is 40185 10:00
ok
bacek_ darkwolf: can you nopaste md4.c?
darkwolf I'll check it again
bacek_ nopaste?
purl nopaste is at nopaste.snit.ch/ (ask TonyC for new channels) or poundperl.pastebin.com/ or paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at www.extpaste.com/ or paste.scsys.co.uk (for #catalyst, #dbix-class, #moose and others) or gist.github.com/ or paste or gtfo
bacek_ purl: nopaste?
purl hmmm... nopaste is at nopaste.snit.ch/ (ask TonyC for new channels) or poundperl.pastebin.com/ or paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at www.extpaste.com/ or paste.scsys.co.uk (for #catalyst, #dbix-class, #moose and others) or gist.github.com/ or paste or gtfo
darkwolf ok 10:01
bacek_ Bah!
darkwolf just a moment
bacek_ nopaste.snit.ch/
mikehh I had no problem with r40185 - built ok - All tests PASSed inc fulltest 10:08
darkwolf the file md4.c has wrong encoding so gedit can't open it 10:14
i have to use nano less or something like that
bacek_ darkwolf: use cat. 10:15
mikehh All tests PASS (pre/post config, smolder, fulltest) at r40217 - Ubuntu 9.04 amd64 10:16
darkwolf one stupid question 10:17
how to get all the file from cat and copy it
mikehh cat filename > out.file 10:18
darkwolf it's the same 10:19
some of the characters r not valid
bacek_ darkwolf: hmm... md4.c shouldn't have any "weird" characters. It's plain ASCII... 10:21
darkwolf i know
it should be c code
mikehh I think that that file is corrupt 10:24
darkwolf yes 10:25
is md4.c generated? 10:27
bacek_ yes. 10:28
double generated actually
first is src/dynpmc/md4.pmc 10:29
second is md4.c by itself
darkwolf: do you have openssl devel libs installed?
darkwolf i have openssl 0.9.8k-2 10:31
is that dev version
mikehh md4.c is generated from md4.pmc in the build process 10:32
you alsao need libssl.dev 10:34
s/alsao/also/
darkwolf my computer is acting weird
sorry bodering u 10:35
now it's fine 10:36
mikehh not a problem - always try to help if I can
10:54 mj41 joined 11:01 mj41 joined 11:04 masak joined 11:06 mj41 joined 11:22 donaldh joined 11:28 clinton joined 12:07 whiteknight joined 12:08 payload joined
whiteknight good morning #parrot 12:10
pmichaud good morning $whiteknight
whiteknight thanks about the local_branch note, I've never used that opcode before
pmichaud PGE uses it heavily
whiteknight okay, I'll look into that then. Thanks! 12:11
pmichaud essentially you just need a resizable integer array that it can use to store program counters in 12:12
whiteknight after I get rid of jsr/bsr/ret, I'm ripping out the last vestiges of stacks in Parrot 12:15
moritz MANIFEST? 12:22
purl MANIFEST is, like, used for a ton of things, yes
whiteknight it will be happy days
purl, no MANIFEST is a major hassle that's designed to break Configure.pl 12:26
purl OK, whiteknight.
12:37 payload joined
Coke (mac ports) that's me. 12:49
Happy to take an updated ports file that works.
japhb: Feel free, btw, to try to optimize partcl. =-) 12:51
ports are done post release, not pre-release, so including the ports files in the release is problematic. 12:55
12:57 iblechbot joined 12:58 whiteknight joined 13:05 whiteknight joined 13:10 mikehh joined 13:22 patspam joined
wayland76 Who do I ask if I want more Trac permissions (so I can edit ticket metadata) 13:24
Coke what's your trac id? 13:25
wayland76 It appears to be "wayland" 13:28
Coke I just gave you TICKET_MODIFY 13:29
wayland76 Thanks! :) 13:31
Coke See? I share my toys. =-) 13:35
NotFound I want TICKET_MAGICALLY_SOLVE
wayland76 NotFound: If that works, you could rent it to people for $10,000 a day :) 13:37
dalek kudo: 5901f2a | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 416 files, 11343 passing, 651 failing

   S32-num/exp.rakudo aborted 40 test(s)
   S32-trig/trig.rakudo aborted 611 test(s)
whiteknight The "Magically Solve" button resolves the ticket as "WONTFIX" and adds a note to the documentation that the problem is actually a "feature" 13:38
NotFound That's darka magic. 13:40
13:40 tomyan joined
Coke wishes he could ack over a 'network place' on windows. 13:46
NotFound buliding parrot 1.0
Coke: Why tou can't? 13:47
s/tou/you
Coke you get "no such file or directory" 13:48
network place != mapped share drive.
NotFound Ah, sorry.
13:49 darkwolf left
Coke (thankfully, these particular ones are mountable as drives, so I'm not sol) 13:52
NotFound Example in TT #872 segfaults with parrot 1.0, adding comment to ticket. 13:53
14:09 whiteknight joined 14:18 mj41 joined
Coke hurls thedailywtf.com/Articles/Programmin...ation.aspx for folks looking to submit something in perl6. 14:29
whiteknight yeah, I read that 14:30
Coke rakudo: say 'hi' 14:34
polyglotbot OUTPUT[Parrot VM: Can't stat languages/perl6/perl6.pbc, code 2.ā¤main: Packfile loading failedā¤]
Coke nice. 14:35
flh Coke, I'm still looking at TT#840, and found a strange thing: the test passes if I run it from (i.e., exec it with perl -e "system 'parrot t/op/io.t';") 14:36
masak rakudo: sub russian-mul($a, $b, $sum? is copy) { $sum += $b if $a % 2; return $a == 1 ?? $sum !! russian-mul(int($a/2), 2*$b, $sum) }; say russian-mul(18, 23)
polyglotbot OUTPUT[Parrot VM: Can't stat languages/perl6/perl6.pbc, code 2.ā¤main: Packfile loading failedā¤]
masak Coke: there you go.
flh Coke, and exactly the same thing happens with the old (perl) version of the test: directly the generated .pir file does not give the expected output 14:37
masak rakudo: sub russian-mul($a, $b, $sum? is copy) { $sum += $b if $a % 2; return $a == 1 ?? $sum !! russian-mul(int($a/2), 2*$b, $sum) }; say russian-mul(18, 23)
polyglotbot OUTPUT[Parrot VM: Can't stat languages/perl6/perl6.pbc, code 2.ā¤main: Packfile loading failedā¤]
purl i already had it that way, masak.
masak something wrong with polyglotbot... 14:38
oh well, works in #perl6.
jonathan purl, rakudo: sub russian-mul($a, $b, $sum? 14:50
purl i guess sub russian-mul($a, $b, $sum is copy) { $sum += $b if $a % 2; return $a == 1 ?? $sum !! russian-mul(int($a/2), 2*$b, $sum) }; say russian-mul(18, 23)
14:57 mj41 joined 15:05 payload1 joined 15:07 HG` joined 15:14 payload joined 15:15 MoC joined 15:20 donaldh joined
dalek kudo: e506d36 | jnthn++ | perl6.pir:
Make sure we pass along the correct exit code if we have an exit exception.
 15:29
kudo: 23e8f09 | jnthn++ | perl6.pir:
Add a null check to avoid exceptions in the backtrace printing; need this for the release, but Parrot probably should not be putting them there in the first place, so investigate that too.
15:33 barney joined
japhb Coke: Oh sure, I'll just get right on that. ;-) 15:39
Actually, if I knew Tcl at all, I might be tempted ....
japhb loves me some optimizations 15:40
Coke japhb: knowledge of tcl not entirely required. =-) 15:41
there's certainly a lot of generated PIR to wade through.
japhb Where is Partcl's design / project pages? 15:42
s/is/are/ # mmmm, Engrish
Coke partcl?
purl i think partcl is tcl on parrot or code.google.com/p/partcl or completely dead since before parrot 1.0
japhb Nothing like speaking your native tongue as if you had just learned it ....
Coke purl, partcl =~ s/or completely dead since before parrot 1.0
purl Coke: sorry...
Coke purl, partcl =~ s/or completely dead since before parrot 1.0//
purl OK, Coke
japhb Nice! Had no idea purl understood that 15:43
Tcl still uses CVS? Seriously? Wow.
Coke it works.
japhb Must be influenced by those crazy OpenBSD guys ...
Coke and in some ways is better than svn.
(conflating tags and branches is not considered by all to be a win. =-) 15:44
japhb Sure, but "My kingdom for an atomic commit" and all that. :-)
Coke mmm. 15:45
NotFound Many people have problems with the C language because of the lack of native teachers. 15:52
dalek kudo: 254b3a5 | jnthn++ | src/pmc/perl6multisub.pmc:
When we have a named slurpy param, we don't need to do a bindability check to see if there's matching named params since it accepts 'em all anyway. This lets us cache many multi-method dispatches again now methods get their *%_, winning back some performance.
 15:57
16:07 iblechbot joined 16:08 theory joined 16:12 whiteknight joined 16:41 whiteknight_ joined 16:51 Psyche^ joined 17:05 MoC joined
Coke odd. partcl's t/cmd_array.t is failing sporadically. 17:19
17:21 mikehh joined
Coke ls 17:22
moritz no such file or directory 17:23
17:41 jan joined 17:53 chromatic joined 17:55 theory joined 18:01 theory joined
dalek TT #719 closed by coke++: out of memory running a tcl spec test 18:06
18:20 theory joined
Coke msg kid51 FYI, just because a ticket is a tcl blocker, doesn't mean it's an HLL_INTEROP issue. 18:21
purl Message for kid51 stored.
18:24 darbelo joined
dalek TT #773 closed by coke++: Can't override vtables of a PMC from PIR 18:26
18:34 sekimura joined 18:48 theory joined
Coke en.wikipedia.org/wiki/Parrot_Foundation - freaky. 19:02
mikehh I think some editin' due there 19:09
as Honourable Member of Board :-}
japhb Coke: I don't see the TT 719 code in the partcl repo ... can you tell me where you put it? (I ask because you said you had Perl 6 and Perl 5 versions as well, and I want to make sure I use the same code as you.) 19:10
Coke moment. 19:14
nopaste "coke" at 72.228.52.192 pasted "japhb" (15 lines) at nopaste.snit.ch/17365 19:15
Coke the tcl version is i the ticket. 19:19
19:21 donaldh joined, Andy joined 19:26 brbrooks joined
brbrooks !seen Whiteknight 19:27
seen Whiteknight
purl Whiteknight was last seen on #parrot 4 hours, 56 minutes and 54 seconds ago, saying: yeah, I read that
japhb Coke: Ah, nice, using the subset of Perls 5 and 6 that is identical. :-) 19:36
dalek kudo: bd1e7cb | moritz++ | docs/announce/2009-07:
[release] feature list for the announcement text
 19:40
Coke japhb: just lazy 19:50
for added laziness, convert the say to a print. 19:51
dalek kudo: e898146 | moritz++ | docs/announce/2009-07:
[release] typos spotted by masak++ and japhb++
 19:53
19:54 Eevee joined
NotFound FYI rakudo build an run successfully with parrot 1.4 built with c++ 20:09
moritz great! 20:10
Coke tries a spec test run for tcl against an /optimized/ parrot 1.4.0, thanks to whoever reminded me of that. 20:11
dalek kudo: fd6cb4a | jnthn++ | src/classes/Signature.pir:
When .perl'ing a Signature, put the : there for named parameters.
 20:19
kudo: 4ee9623 | jnthn++ | src/p (3 files):
First cut implementation of .^roles for introspecting what roles a class does, plus some other corrections to various bits of role introspection.
kudo: d8505f8 | jnthn++ | :
Merge branch 'master' of git@github.com:rakudo/rakudo
20:25 flh joined
dalek kudo: 03aa56f | jnthn++ | (3 files):
Add an Attribute class that will be the descriptor for an attribute. For now it only has a .name, will add more later. Also implement first cut of .^attributes.
 20:38
20:46 theory joined
mikehh latest rakudo fails to build on parrot r40217 - perl6.ops multiple errors 20:52
builds on parrot r40185 20:53
Tene Yeah, a lot of people are seeing that, it seems. I heard something about it going to be fixed before the rakudo release, I think.
mikehh well r40185 was the parrot release 20:54
NotFound Why? Isn't 1.4 the targeted parrot?
Tene In that case, what i heard was probably "rakudo won't work against the latest Parrot until after the rakudo release" 20:55
mikehh Yeah - but I keep testing with the latest parrot - which passes all parrot tests
moritz what Tene++ said
mikehh ok - I'll wait until after the release before I complain again :-} 20:57
dalek kudo: 94b741a | jnthn++ | src/ (2 files):
Give Attribute the type and build attributes, and set them up in .^attributes.
 20:58
20:59 bacek_ joined 21:06 payload joined
dalek kudo: 8c79a13 | jnthn++ | src/ (3 files):
Implement accessor attribute for Attribute descriptor class.
 21:23
kudo: fb0601a | jnthn++ | docs/ChangeLog:
Extra changelog entry plus tweaks.
 21:29
rtcl: r536 | coke++ | wiki/SpecTestStatus.wiki:
Mark some more skips from our latest run that we haven't figured out yet.
 21:39
Coke partcl has regressed quite a bit. frustrating. 21:41
though running it optimized shaved a bunch of time. rerunning with the new skips in place... 21:42
dalek kudo: 5530469 | moritz++ | docs/announce/2009-07:
improved introspection warrants an entry the annonce text
 21:48
kudo: 6062092 | moritz++ | docs/announce/2009-07:
[docs] improved wording by jnthn++
 21:54
22:09 Whiteknight joined
jonathan OK, getting an early night...hopefully some sleep will shake off whatever I've got... 22:09
night all
22:10 dduncan joined
jonathan erm, that was for #perl6... 22:12
jonathan clearly needs his early night :-/
22:13 dduncan joined 22:15 Limbic_Region joined 22:30 theory joined 22:38 kid51 joined
dalek rtcl: r537 | coke++ | wiki/SpecTestStatus.wiki:
Record more unknown failures. LOTS of regressions here. :(
 22:38
22:38 rg joined 23:00 tomyan joined 23:10 Whiteknight joined
dalek rrot: r40218 | bacek++ | branches/orderedhash_revamp/src/pmc/orderedhash.pmc:
Initial draft how OrderedHash implementation can look like.
 23:14
pmichaud Good afternoon, #parrot 23:18
dalek rrot: r40219 | bacek++ | branches/orderedhash_revamp/src/pmc/orderedhash.pmc:
[pmc] Totally broke OrderedHash.

  - Add stubs for new methods.
  - Build is broken.
  - But 2 first OrderedHash tests are passed!
bacek_ good morning pmichaud
23:20 donaldh joined
bacek_ wish to have Tuple PMC... 23:21
and pony. 23:22
anyway, $dayjob time
dalek rrot: r40220 | whiteknight++ | branches/bsr_jsr_ret (2 files):
[bsr_jsr_ret] undo my previous commit on suggestion from pmichaud++. Will resolve these issues a different way
 23:29
rtcl: r538 | coke++ | wiki/SpecTestStatus.wiki:
... Everytime I run the spec test, more files are failing.
 23:30
Coke this is damn depressing. 23:34
every time I rerun the spec test suite, more files are reporting as failing. :|
Whiteknight so stop running it 23:40
Coke ^_O 23:41
At this point, I just want to get a steady state. :| 23:42
dalek rrot: r40221 | jkeenan++ | trunk/docs/pdds/pdd30_install.pod:
Applying patch submitted by wayland++ in trac.parrot.org/parrot/ticket/788 (plus a few other small POD-formatting improvements).
 23:43
23:45 wayland76 joined
dalek rrot: r40222 | petdance++ | trunk/compilers/imcc/optimizer.c:
removing unexecutable code. Thanks, Coverity!
 23:47
rrot: r40223 | whiteknight++ | branches/bsr_jsr_ret (4 files):
[bsr_jsr_ret] replace bsr/ret with local_branch/local_return on suggestion from pmichaud++. Parrot builds, but fails plenty of tests
 23:51
kid51 Whiteknight: About that 'wikiconvention in NYC' you mentioned yesterday -- this was hit #1 when I googled: www.actorsandcrew.com/press/archives/4044 23:53
Was that what you were referring to?
Whiteknight no, I don't think so 23:54
dalek rrot: r40224 | whiteknight++ | branches/bsr_jsr_ret (4 files):
[bsr_jsr_ret] convert a few more files
 23:59

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!