Devel: 0.6.1 | parrotcode.org/ | 696 new/open tix
Set by moderator on 12 May 2008.
pmichaud I may be able to merge tonight. If not tonight, then tomorrow morning, certainly. 00:01
chromatic Looks like r27450 killed Pheme. 00:02
Sadly, it was a branch merge from pgeupdates.
pmichaud ohhhhhhh!
I bet I know what it is.
pmichaud checks.
yes.
PGE doesn't like having rules names 'list'
because Match objects have a 'list' method on them already 00:03
chromatic I can fix that.
pmichaud same for 'hash', 'item', and 'chars'.
It'll be fixed when either (1) I switch PGE to use MMD or (2) I redesign PGE with LTM
perhaps I'll put some code into PGE to warn about methods that collide with the existing ones. 00:07
that would save a lot of grief. 00:08
or I could just try switching to MMD.
dalek r27492 | chromatic++ | trunk: 00:09
: [Pheme] Changed 'list' rule to 'pheme_list' to avoid conflicts with PGE's Match
: object methods.
diff: www.parrotvm.org/svn/parrot/revision?rev=27492
chromatic MMD sounds good, but I'm of the inclination to wait just over a week for that.
00:09 nopaste joined 00:10 TonyC joined
pmichaud (wait a week) okay. 00:21
00:24 particl1 joined 00:27 particl1 left, wknight8112 joined 00:46 AndyA joined 00:48 wknight8112 left 00:55 kid51 joined 00:57 wknight8113 joined 01:19 AndyA joined 01:48 rdice_ joined 01:49 AndyA joined
wknight8111 what was the number on that ticket with configure problems on ubuntu? 01:50
kid51 rt.perl.org/rt3/Ticket/Display.html?id=50068 01:51
wknight8111 thanks 01:52
i just installed ubuntu, and i'm having that same exact problem
kid51 It seems the ticket has been resolved twice already! 01:54
So you may need to post to re-open it.
wknight8111 no, it's actually a different issue. I seem to remember a different ticket 01:55
configure is failing with "Compilation failed with 'cc'" 01:56
pmichaud fwiw, I'm on kubuntu and it seems to work for me
wknight8111 i literally installed ubuntu less then 3 hours ago, so i'm still learning my way around. I probably need a setting or something 01:57
pmichaud what error are you getting?
oh, I see it there. 01:58
you may need to install an extra package.
kid51 This one? rt.perl.org/rt3/Ticket/Display.html?id=49023
pmichaud have you installed the 'build-essentials' package? 01:59
01:59 particle joined, particle left
kid51 wknight8111: FWIW, I build on an older Ubuntu every day, with just 'perl Configure.pl' 01:59
wknight8111 probably not. I'll head out to cpan now and pick up all the stuff I need
pmichaud 'build-essentials' is a ubuntu package
kid51 Can you paste failure?
pmichaud sudo apt-get install build-essentials 02:00
02:00 particle joined
pmichaud (or whatever package manager you prefer using. I don't know if ubuntu uses adept, but that can be used also) 02:00
kid51 Hmm, I probably have that, but probably added it so long ago I've forgotten that I have it.
pmichaud I have a setup script that I use nowadays
so, after installing kubuntu, I just run my script and it grabs any packages I need :-) 02:01
wknight8111 i'm not organized to have a script, i am enjoying the learning experiene 02:02
chromatic aptitude 02:03
purl hmmm... aptitude is www.cme.com/market/cfot/quiz/ or members.aol.com/jocrf19/ or members.aol.com/daveander/
pmichaud I really like kubuntu -- it's my preferred distro at the moment.
(and I'd use ubuntu if I were a gnome person.)
wknight8111 it's easy to find KDE apps, they all start with L 02:09
K
kid51 kubuntu? 02:10
purl i think kubuntu is ubuntu + kde (instead of gnome)
kid51 Thanks, purl
kid51 is busy with revamping one of his CPAN modules. 02:11
dalek Jim Keenan | YAPC::NA::2008: 02:18
link: www.perlfoundation.org/parrot/index...pc_na_2008
wknight8111 apt-get can't find package "build-essentials" 02:35
pmichaud sorry, no 's'
build-essential
purl rumour has it build-essential is the psuedo package on debian iirc
wknight8111 i'll try that
ok, that solved my problem perfectly. pmichaud++ 02:40
i love having a console that expands to full-screen. that feature alone was worth the switch 02:42
kid51 must sleep 02:51
purl $kid51->sleep(8 * 3600);
dalek r27493 | pmichaud++ | pctcoerce: 03:09
: [pct]:
: * Optimize register handling for while, until, repeat_while, repeat_until.
diff: www.parrotvm.org/svn/parrot/revision?rev=27493
03:25 grim_fandango joined 03:33 Zaba_ joined
Eevee would it be reasonable to have a haskell compiler for parrot? I'm talking to a haskell fan and he worries about performance, static language on a dynamic vm, etc 03:47
03:50 Andy joined
chromatic Sure, why not? 04:20
dalek will@coleda.com | YAPC::NA::2008: 04:26
link: www.perlfoundation.org/parrot/index...pc_na_2008
04:31 Psyche^ joined
Eevee chromatic: he said, re how ghc works, "but in a nutshell it consists of rewriting code as it executes, jumping to and fro, without a traditional call stack / the sheer number of continuations, for example, would kill the stacks of most VMs" 04:31
chromatic Most VMs aren't register based. 04:37
On the other hand, I do want to optimize our continuation scheme (no pun intended). 04:38
Eevee groan.
japhb chromatic: Would his comment remain true if you simply s/stacks/GC/
er s/remain/become for Parrot/
Eevee I know it's register-based but I don't know haskell or parrot well enough to know if that's enough to mitigate the concern
japhb ghc is deeply magical. That doesn't mean we can't make something that is reasonably source-compatible. 04:39
Just because you can't beat Machine Forth on a P51, doesn't mean it's not fun to make Forth for Parrot. Parrot has more library support, for one thing. ;-) 04:41
chromatic This all depends on the GC scheme for Continuations. 04:42
You can probably make a very valid assumption that most continuations have either zero or one references.
Coke (the script for manifest should work if you svn add first!)
chromatic Thus if you have a reference barrier for continuations, you can flip one bit and when you destroy continuations, you can immediately add them to the free list if that bit is off. 04:43
japhb nodnod
chromatic ... and because we try to keep libparrot's internals opaque to the outside, we mandate that all access to continuations from the external interface goes through the barrier.
A similar scheme (there's that word again) might even work for continuations. 04:44
I mean PMCs.
... especially if we need a write barrier for compacting/copying collection anyway. 04:45
Or whatever kind of barrier it is.
japhb a big brick wall 04:48
chromatic Some sort of mediated access anyway. 04:49
Eevee a big brick wall with a brick missing
japhb Bob Geldof? 04:50
05:13 Debolaz joined 05:25 Debolaz_ joined
Tene I'm having a lot of trouble getting the cardinal grammar to parse indexed variables. 05:58
I guess part of this is what LTM is supposed to help with... 05:59
By reordering rules in an alternation I can get it to parse incorrectly or fail. 06:00
I'm having trouble figuring out why it's failing. 06:01
Maybe I need to add some lookaheads at the end of some of these rules to get them to fail at the right time instead of matching inappropriately and causing other failures later. 06:03
That sounds like a decent theory for now. 06:07
Now sleep.
purl i heard now sleep was impossible!
06:15 mire joined
dalek r27494 | pmichaud++ | pctcoerce: 06:20
: [pct]:
: * Register coercion gets smarter (and a bit more complex).
diff: www.parrotvm.org/svn/parrot/revision?rev=27494
r27495 | chromatic++ | trunk: 06:24
: [GC] Reversed conditional from check added in r27484 to perform a DOD run only
: *if* any buffer allocations have occurred since the last DOD run, not *unless*.
: This gives another 10% performance improvement to the Rakudo-building
: benchmark.
diff: www.parrotvm.org/svn/parrot/revision?rev=27495
japhb looks into the face of IMCC and feels the insanity reaching into his addled brain 06:27
dalek r27496 | pmichaud++ | pctcoerce:
: [pct]:
: * Immediate blocks have to be invoked with 'newclosure' in order
: for lexicals to work.
diff: www.parrotvm.org/svn/parrot/revision?rev=27496
06:31 uniejo joined
dalek r27497 | pmichaud++ | pctcoerce: 06:58
: [pct]:
: * In general, let the caller handle coercions instead of the callee.
diff: www.parrotvm.org/svn/parrot/revision?rev=27497
r27498 | fperrad++ | trunk: 07:15
: [Lua]
: - remove some get_class
diff: www.parrotvm.org/svn/parrot/revision?rev=27498
r27499 | fperrad++ | trunk: 07:16
: [Lua]
: - bytecode translation : small fix
diff: www.parrotvm.org/svn/parrot/revision?rev=27499
r27500 | pmichaud++ | pctcoerce: 07:17
: [pct]:
: * Slightly improve code generation for single-ended if/unless.
diff: www.parrotvm.org/svn/parrot/revision?rev=27500
japhb wheee, r27500 07:20
07:21 peepsalot joined
Patterner only 972500 left. 07:27
08:39 masak joined
dalek r27501 | allison++ | trunk: 09:51
: [pdd25] Deprecate 'internal_exception'.
diff: www.parrotvm.org/svn/parrot/revision?rev=27501
09:55 rdice joined 09:56 Zaba joined 10:03 ruoso joined 11:02 tetragon joined 11:16 DietCoke joined 12:21 wknight8111 joined 13:18 Zaba joined 13:27 rdice joined 13:35 wknight8111 joined
pmichaud ...something broke APL. 13:36
DietCoke check.s
pmichaud I'm guessing it was r27450 (the pge updates) 13:42
yes. 13:43
okay, well that narrows it down a lot.
aha
we have a 'list' rule. That would do it.
dalek r27502 | pmichaud++ | trunk: 13:46
: [APL]:
: * Fix rule error introduced by PGE updates in r27450.
diff: www.parrotvm.org/svn/parrot/revision?rev=27502
DietCoke doesn't get around to checking fast enough. 13:52
pmichaud++
jonathan hi all 13:55
pmichaud hello, jonathan
jonathan is arrived at his new appartment and about set up again
pmichaud yay!
DietCoke checks the '24' style countdown clock - you have 32 minutes left to commit a patch! hurry!
moritz hi jonathan ;) 13:56
pmichaud I don't think I'll ever get those 'beep beeps' out of my head from that series
13:57 cognominal joined 13:58 cognominal joined 13:59 Andy joined
jonathan pmichaud: How does it work for you if I have Rakudo day on Friday this week? 13:59
pmichaud should be just fine
anything you want me to work on between now and then?
jonathan OK, I'll go for that.
return ;-)
pmichaud heh
okay. Since I'm in the guts of pct anyway I might be able to do that one.
jonathan I can't think of anything right off.
Well, I'd love to have return, but I guess there's some infrastructure work needed. 14:00
pmichaud a little, yes, but I don't know that it's too much
jonathan OK
Will it let us do other control exceptions too?
pmichaud if I can do it easily, yes.
jonathan Are you thinking something general, or something specific to return?
OK.
pmichaud I always aim for general, unless general is going to take a long time :-) 14:01
work code trumps design
*working
DietCoke Note that there are now control style exceptions (at least in the exceptions branch.)
pmichaud mmmmm 14:02
I'll look at those.
just defining control constants, or actually doing stuff with them?
DietCoke at the moment, the former. If we need to hang more formality off them, we can do that.
but for now, using those allows us some base HLL interop. 14:03
pmichaud I'm only really interested in control constants, so that works fine for me
DietCoke return, break, continue, error, and ok are the 5 tcl states, and I think we naively have all five right now. 14:04
pmichaud might need a 'redo' for Perl 14:05
DietCoke I think ok is just "fall off the end", so that can probably go. and "error" is ... well, that's exceptional. I don't think we need a placeholder to say exception.
(redo) sure.
(wonder if you could decompose something with a redo into something with just breaks and continues.) 14:06
nopaste "pmichaud" at 76.183.97.54 pasted "pct register coercion code gen improvement" (13 lines) at nopaste.snit.ch/12963 14:12
jonathan we can haz compilez the LOLCODE fasta! 14:13
14:14 cjfields joined
moritz that's what? 10% gain? 14:14
not bad
pmichaud also we reduce the number of unneeded object creations 14:16
DietCoke yah, anything that doesn't have to be GC'd is a big win at this point. 14:17
I suppose I'm going to have to get tcl on the pct bus at some point.
nopaste "pmichaud" at 76.183.97.54 pasted "pct code generation improvement (calls to 'new')" (26 lines) at nopaste.snit.ch/12964 14:20
DietCoke wonders when we're getting the panie, penie, and pinie languages. 14:21
pmichaud I call "pwnie". All your languages are belong to me. 14:22
DietCoke I suspect prnie will be more fun.
(282-272.9)/47 14:24
purl 0.193617021276596
DietCoke Geoffrey Broadwell? 14:35
infinoid?
purl infinoid is Mark Glines <mailto:mark@glines.org>
Andy perlbuzz.com/2008/05/parrot-bug-day...urday.html 14:38
shorten Andy's url is at xrl.us/bkiyx
DietCoke Andy: ah, thank you.
particle andy++ 14:45
Infinoid DietCoke: hmm? 14:51
14:54 wknight8111 joined 14:56 IllvilJa joined, NotFound joined
NotFound Hello. 14:56
DietCoke Infinoid: you're not GB. =-) 14:57
NotFound?
NotFound: hello.
NotFound: are you Geoffrey?
Infinoid DietCoke: you're looking for japhb
purl, japhb? 14:58
purl japhb is probably not japhy.
Infinoid no, japhb is Geoffrey Broadwell
purl okay, Infinoid.
NotFound DietCoke: no, I'm NotFound 14:59
15:00 sjansen joined
NotFound is Julián Albo 15:00
aacute
15:02 donaldh joined
DietCoke Infinoid: danke. 15:03
15:16 donaldh joined 15:23 Theory joined 15:32 ambs joined 15:45 paco joined 16:06 Theory joined
japhb DietCoke: You are looking for me? 16:06
particle japhb: two things, likely, that coke wanted. 16:12
japhb particle: listening ....
particle 1) to make sure the bot had your name and nick and email 16:13
2) to discuss the patch you sent to the list
japhb particle: is it standard to seed that info into purl? 16:14
particle yes.
particle?
purl mailto:jerry.gay@gmail.com
particle particle is also jerry gay
purl okay, particle.
japhb no, japhb is Geoffrey Broadwell, mailto:geoff@broadwell.org 16:15
purl okay, japhb.
japhb As for the patch, do you know if there was anything other than what he sent in the ticket response? 16:16
ambs purl, ambs? 16:17
purl you are Alberto Simoes, one of the organizers of YAPC::EU::2005 and available at mailto:ambs@cpan.org
ambs purl++
Infinoid the reverse mapping is also important
Mark Glines?
purl somebody said Mark Glines was Infinoid
ambs Alberto Simoes? 16:18
purl Alberto Simoes is ambs
purl OK, ambs.
japhb no, Geoffrey Broadwell is japhb , mailto:geoff@broadwell.org
purl okay, japhb.
japhb japhb?
purl you are, like, Geoffrey Broadwell, mailto:geoff@broadwell.org
japhb Geoffrey Broadwell? 16:19
purl i heard Geoffrey Broadwell was japhb , mailto:geoff@broadwell.org
ambs me?
purl you are Alberto Simoes, one of the organizers of YAPC::EU::2005 and available at mailto:ambs@cpan.org
japhb me?
purl you are Geoffrey Broadwell, mailto:geoff@broadwell.org
ambs lol
NotFound you?
purl you is ;)
ambs us?
purl it has been said that us is humped, on debt. I'm speaking specifically of the social security retirement system.
ambs we?
purl assimilates ambs into the borg
pmichaud pmichaud?
purl i guess pmichaud is www.pmichaud.com/ or "Patrick R. Michaud" <mailto:pmichaud@pobox.com>
Eevee me?
purl i heard Eevee was kidding ^^;
Eevee :(
moritz me?
purl rumour has it moritz is creating a form in Customers/signup with formbuilder and formbuilder::source::file ; now this always tries to render a template called customers/signup but i called your template customers/signup.tt2 ; how can i set a different template for your form? i tried $c->form->template({ template => '/customers/signup.tt2' }) but that doesnt work
ambs Eevee++ 16:20
Tene tene?
Eevee that is terrible where did you pick that up purl
moritz wtf did perl store for me?
s/perl/purl/
Eevee moritz: wow
yours is the best!
moritz never did anything like that
16:20 allison joined
moritz purl is telling lies, damned lies! 16:21
Infinoid and statistics
purl hmmm... statistics is 71.5% of statistics are made up or Staff, it's come to my attention that 40% of sick days are being taken on Mondays and Fridays. This must stop immediately. or "Trust no statistic which you did not falsify." or www.statsoftinc.com/textbook/stathome.html or downlode.org/Miscellany/PoundPerl/Stats/
Tene purl: no, moritz is telling lies, damned lies!
purl okay, Tene.
moritz lol
Eevee I like to think raising children is like having a purl of my very own
NotFound purl?
purl NotFound?
Eevee (obviously I do not have children) 16:22
Infinoid purl needs a "can't hear you, listening to headphones" mode
purl Infinoid: huh?
Tene purl: tene? 16:23
purl tene: bugger all, i dunno
Eevee "the moon is a rock in space!" "no, the moon is made of green cheese." "okay, eevee."
japhb needs breakfast ... anyone know if DietCoke reads scrollback? 16:26
particle japhb: no, i don't know if coke had anything more specific
ambs purl, Tene is Stephen Weeks
purl OK, ambs.
japhb particle: thanks; I can deal with his ticket reply after I have some nutrients in my blood. :-) 16:27
Tene go to work
donaldh is still blushing after purl got the lube and followed him to bed last night. 16:29
particle :)
16:32 davidfetter joined
ambs purl, DietCoke ? 16:51
purl i don't know, ambs
ambs purl, DietCoke is Coke under diet.
purl OK, ambs.
16:55 paco joined
dalek afbach@gmail.com | YAPC::NA::2008: 17:01
link: www.perlfoundation.org/parrot/index...pc_na_2008
particle paco: you've run some builds recently on funky platforms. think you can throw together a patch for PLATFORMS that we can put into next week's release? 17:02
paco ok, particle, just now im Installing tur64/alpha :) 17:03
dalek afbach@gmail.com | YAPC::NA::2008: 17:04
link: www.perlfoundation.org/parrot/index...pc_na_2008
particle (paco's computer museum)++
japhb Proving once again that old Internet writings live forever, I got someone asking me to write more OpenGL in Perl 5 tutorials. I told him I would write more OpenGL in Perl tutorials someday, they'd just be for Perl 6. :-) 17:15
DietCoke: new dump_pbc.pl patch posted to address your nits 17:17
NotFound I suspect that what many people ready wants is a step by step tutorial on how to write his complete application.
17:18 mj41 joined
NotFound s/ready/really 17:18
japhb NotFound: in this particular case, he was asking about a set of tutorials that I was writing for Perl.com that I did not get a chance to finish -- I stopped writing before teaching texturing, and that's what he wanted to know how to do. (Because that's one of the places where C and Perl are going to be very different.) 17:19
I miss writing those, but unfortunately my OCD interfered heavily with my hourly rate (since it was fixed payment per article). 17:20
NotFound japhb: I was blaming the world in general, not a particular case ;) 17:21
17:24 mj41 joined
NotFound nopaste? 17:28
purl nopaste is at nopaste.snit.ch/ (ask TonyC for new channels) or rafb.net/paste or paste.husk.org/ or poundperl.pastebin.com/ or paste.scsys.co.uk/ or don't bother me while I'm eating
DietCoke japhb: I do read scrollback, but you can't rely on it. =-) 17:29
japhb DietCoke: well OK then. :-)
DietCoke always best to write things down on list or in rt. 17:30
nopaste "NotFound" at 213.96.228.50 pasted "Is this a good test for removeparent?" (33 lines) at nopaste.snit.ch/12967 17:38
particle NotFound: are you trying to add a test to the parrot test suite? 17:39
NotFound particle: I'm trying to implement removeparent 17:40
particle ok. so it sounds like this test is destined to make its way into the repo
then i suggest you use Test/More.pir
NotFound Actually pass all test, including this :) 17:41
particle it's a reasonable test, however i'd write the whole test file in pir 17:43
NotFound Some example I can borrow from? 17:44
particle t/oo/isa.t 17:45
DietCoke japhb: added yet another nit!
a simple happy path test is all I'm looking for.
Sorry I didn't think of it sooner. :| 17:46
nopaste "NotFound" at 213.96.228.50 pasted "removeparent patch" (91 lines) at nopaste.snit.ch/12968
DietCoke I blame particle.
NotFound Someone want to give it a try? 17:47
DietCoke NotFound: feel free to remove the ticket # if the patch resolves the ticket.
(in the docs in the code)
NotFound DietCoke: yeah, and rewrite the _item
particle NotFound: patch looks sensible. please add to the documentation what happens if the parent isn't found, or an invalid parent is passed, etc
NotFound The "This currently does nothing" part ;) 17:48
particle: yes, but I'm only testing for anow.
particle eg. can PMCNULL be a valid parent pmc?
NotFound particle: good question, but I don't know the answer. 17:49
particle NotFound: i mean the documentation for the remove_parent VTABLE
DietCoke NotFound: you loop over the parents 2x. the first time you're seeing if it's a parent that is removable. the second time is to remove it?
NotFound DietCoke: yes, i copied the first part from add_parent, will simplify it later. 17:51
DietCoke NotFound: don't return after a real_exception.
(avoid unreachable code exceptions)
s/exceptions/compiler warnings/
(will simplify later) ok. 17:52
dalek r27503 | pmichaud++ | pctcoerce:
: [pct]:
: * Improve register handling for PAST::Var nodes, 'if'
: * Update comments and documentation
diff: www.parrotvm.org/svn/parrot/revision?rev=27503
NotFound DietCoke: I suspected this will be the correct way, but seen other code that adds the return.
particle no return. it's bad form
NotFound: you should test the diamond pattern in your tests 17:53
A
B C
D
also a Y pattern
A B 17:54
C
D
DietCoke what about the X-wing!
particle that's episode IV 17:55
NotFound Well, but don't blame me, I copy that part from add_parent. 17:56
pmichaud blames chromatic.
NotFound Put the blame on chromatic, boy!
davidfetter o/` blame canada o/` 17:59
NotFound add_parent does not check explicitly for pmc null, but "STRING * const parent_name = VTABLE_get_string(interp, parent);" will fail in that case, right?
particle yes, should do 18:01
dalek r27504 | tene++ | trunk: 18:14
: Start to allow method calls.
diff: www.parrotvm.org/svn/parrot/revision?rev=27504
r27505 | pmichaud++ | trunk: 18:16
: [pct, nqp]:
: * Merge pctcoerce branch into trunk.
: * All tests pass (parrot, nqp, perl6, pynie, punie, APL, lolcode)
diff: www.parrotvm.org/svn/parrot/revision?rev=27505
18:17 askie joined 18:20 ambs joined
DietCoke pmichaud++ 18:20
dalek r27506 | tene++ | trunk:
: [cardinal]
: * Start of a class hierarchy
: * Completely stolen from Rakudo
diff: www.parrotvm.org/svn/parrot/revision?rev=27506
r27507 | tene++ | trunk:
: [cardinal]
: * do blocks start working now
: * add the 'downto' method for Integers
diff: www.parrotvm.org/svn/parrot/revision?rev=27507
r27508 | tene++ | trunk:
: [cardinal]
: * ARGS support
: * AST generation for indexed variables works, but parsing doesn't
diff: www.parrotvm.org/svn/parrot/revision?rev=27508
ambs cadinal? 18:21
cardinal?
purl cardinal is mail.freesoftware.fsf.org/pipermail...dinal-dev/ or the Ruby-on-Parrot project. or xrl.us/uyz3
Tene ambs: ruby compiler
ambs purl, thanks
purl de rien ambs
ambs Tene, thanks ;)
long time I do not see commits on cardinal, so, forgot it :)
Tene Also I usually forget to actually tag my commit messages. 18:22
ambs :D
that might explain it
18:25 iblechbot joined
DietCoke pmichaud: more work planned on pctcoerce, or is that going away now? 18:34
pmichaud pctcoerce going away shortly 18:37
dalek r27509 | pmichaud++ | pctcoerce: 18:41
: Removing obsolete pctcoerce branch.
diff: www.parrotvm.org/svn/parrot/revision?rev=27509
18:49 cjfields joined
moritz perlpunks.de/paste/show/482b34a0.3c12.374 fix two syntax errors in languages/PIR/src/pir.pg 18:51
weird, the language-smoke shows lua all in read, but when I cd languages/lua && make && make test it's (nearly) all fine 18:59
same for tcl 19:01
19:05 Zaba joined
DietCoke some tests depends on their running dir. I thought tcl was free of that. 19:06
moritz I suspect that my smoke was somehow borked
will try again later 19:07
19:09 Theory joined
ambs DietCoke, domm just helped in your task of voting on TPF grants 19:13
moritz patch for languages/PIR/ sent 19:17
19:30 apple-gunkies joined 19:35 Ivatar joined
nopaste "NotFound" at 213.96.228.50 pasted "Better test for removeparent" (123 lines) at nopaste.snit.ch/12969 19:38
ambs DietCoke, ok :) 19:39
DietCoke Being the new guy is liberating!
DietCoke gets an email that says "but I can't talk about it right now". AIGH TELL ME NOW NOW NOW! 19:40
NotFound Some comment?
19:40 allison joined
ambs .za? 19:40
purl hmmm... .za is south africa
particle NotFound: it's a good start. i'd add a lot more tests, but it's definitely a good start 19:42
good enough to commit
NotFound Thanks, I will do some more testing and prepare a patch. 19:43
DietCoke particle: can you test out japhb's tool on windows on a simple .pbc ? 19:44
phone in 0:15 19:45
particle DietCoke: i'll put it on my list, try to do today 19:46
DietCoke danke. anyone else with windows, feel free to jump in. =-)
NotFound Jump through the window? 19:49
ambs better not
DietCoke self-defenestration. 19:50
19:52 desertmax joined
DietCoke Andy, you've spoiled me. You're going to have to write ack-diff. 19:52
particle ooh, right, i'll use my windows xp vm. that'll speed things up 19:55
DietCoke wonders why there isn't a linux distribution named "Defenestrate"; Someone go get on that right now; I want my name in the credits. 19:56
Eevee defenestrate is the best word in the english language 19:59
Tene sed -ir s/fedora/defenestrate/ig /
20:03 japhb joined 20:04 purl joined
cotto_work It is said that some Catholics ascribed the survival of those defenestrated at Prague Castle in 1618 to divine intervention, while it is said that some Protestants claimed that it was due to their landing in a large pile of manure. 20:04
20:08 desertmax_ joined
NotFound Patch sended to #50646 20:13
Oooopppssss.... forget to remove the return. 20:14
Fixed. 20:18
cotto_work: those people are always protesting. 20:21
Infinoid bovine intervention 20:22
20:26 Zaba_ joined 20:32 donaldh joined
particle help! 20:40
there's a patch, [perl #54148]
from japhb
i'm trying to apply it to a windows xp virtual machine
but it's not unified diff format, so svn doesn't understand it
and i don't have 'patch'
could somebody convert the patch to unified diff format for me? 20:41
ah, forget it, it's small enough that i can probably apply manually
Tene Do you know the args to patch to convert?
particle no idea. can patch covert? 20:42
i figured somebody could apply the patch, then do svn diff
Tene Oh, you wanted patch to *apply*, not to convert so that svn could apply
Sure.
Link?
particle rt.perl.org/rt3/Ticket/Display.html?id=54148
parrot-dump-pbc-2008-05-14.patch 20:43
NotFound particle: in the meantime, you can take a look at mine ;)
Latest cardinal patches fail several t/distro tests here. 20:45
Tene Ack. That's unfortunate.
NotFound Trailing spaces and metadata.
Tene I'll fix them.
castro.iodynamics.com/~tene/parrot-...icle.patch 20:46
shorten Tene's url is at xrl.us/bkjhb
particle tene, good, but it's missing tools/util/dump_pbc.pl 20:47
maybe need to svn add? 20:48
Tene Oh. Sure.
updated
particle tene++ # works great! 20:50
now, can somebody write me a webapp that does that conversion for me? :) 20:51
Tene particle: how serious is that request? How often do you need that?
particle well, it's a common windows problem
however, i can find unix utilities compiled for windows, so i *can* install patch 20:52
dalek r27510 | tene++ | trunk: 20:58
: [cardinal]
: * Metadata and whitespace fixes.
: NotFound++
diff: www.parrotvm.org/svn/parrot/revision?rev=27510
particle frack. i need a c compiler 20:59
there, that's better.
20:59 peeps[work] joined
NotFound Put a TODO ticket to write a diff utility in pir 20:59
diff/patch 21:00
particle some time ago i did write a grammar for patch....
i wonder if it's in examples/pge/grammars
cotto_work still there
particle it's likely out of date, anyway, and either works with universal diff or the regular, but not both
yay! it's a start, anyway :) 21:01
patch on parrot would be a great tribute to larry
21:01 peepsalot joined
NotFound Tene: fine now. 21:15
japhb particle: What were you trying to do the patch with, if not 'patch'? (FWIW, the patch was generated with `svk diff`) 21:33
particle japhb: tortoisesvn, a windows svn client, has a utility to apply universal diff formatted patch files 21:34
japhb huh. Too bad they don't handle svk diffs properly ... which are not-quite-unified-diff format 21:35
particle is there a way to have svk diff provide universal diff format? i run into this problem with chromatic's patches frequently
japhb ponders
particle yeah, without "Index: ..." my svn client gets confused 21:36
japhb Anyone know where clkao hangs out?
particle try freenode's #svk
though he may be busy with yapc::asia, iunno the schedule
japhb lemme check docs first .... 21:37
Hmmm, the only thing that looks relevant is setting SVKDIFF in the environment to change to an external diff tool, and then setting that to diff -Naur or something 21:39
japhb trundles over to freenode
Eevee I think that's the intended way to "fix" it yeah
particle yay, drew carey will be in town friday 21:41
japhb Sorry, no responses on #SVK yet. I'll leave the session open for a few hours; if I get anything, I'll pass it on. 21:46
21:47 rdice joined
particle japhb++ 21:48
japhb Oh, did anyone answer the question I saw a couple days ago asking if git-svn can set SVN properties? I've been looking at trying it out for Parrot work, as it seems from the tutorials I could find that git has a better model for non-committers that contribute regularly than SVK does ....
particle only uses svn 21:49
Infinoid I was looking for a way earlier today, and couldn't find it. it has a propget but no propset
japhb went ClearCase -> SourceSafe -> CVS -> SVN -> SVK, and has stayed on that last one for a couple years now.
Infinoid: grr 21:50
21:50 slightlyoff joined, slightlyoff left
particle pvcs -> dimensions -> svn for me, with conversions from clearcase, vss, cvs, rcs along the way 21:50
Infinoid has been talking to therapists about cvs
donaldh japhb: do you miss ClearCase?
particle VOBs-- 21:51
japhb Oh yeah, I guess brief times with Perforce and (forgetting the name now ... RCCS?)
donaldh MVCS++ 21:52
s/MVCS/MVFS/
japhb donaldh: NOT A BIT. I was doing checkouts of multi-GB code bases from Germany to US before the company was Internet connected -- over a shared 64K Frame Relay line.
ClearCase ... not so internationally efficient. 21:53
donaldh Ah, but if you have a big fat intranet then ClearCase Multisite is fine.
SCCS -> RCS -> CVS -> ClearCase for the last 10+ years. 21:54
particle one of the first things i wrote in perl was a two-way synchronizer between a usa pvcs server with three international mirrors
purl be coke =~ ClearCase 21:55
purl particle: sorry...
particle purl, be coke
purl particle: sorry...
particle huh.
21:55 allison joined
japhb donaldh: I stopped using ClearCase about 10 years ago, and it probably was an old version even then. 21:57
donaldh :D
There's a lot to dislike about ClearCase, but I do like the multiversion filesystem. Versioned source files at the FS level seems so elegant. 21:58
japhb would not be surprised if in fact any efficiency increases in ClearCase actually came at the request of SAP, the company I was working for back then ....
Few things are quite as persuasive as Hasso Plattner having a hissy fit. 21:59
donaldh Yes, I believe a lot of Multisite bugs have been fixed as demanded by Cisco, who currently support me. Biggest multisite in the world I believe. 22:00
ClearCase build avoidance saved us a lot of cycles in the old days when CPUs were slow and disk was expensive. I suspect it's quite outmoded now. The distributed version control systems are winning. 22:02
22:17 mire joined 22:19 Limbic_Region joined
japhb particle: so did dump_pbc.pl work for you on Windows? As I recall, that was one of DietCoke's requests before committing. 22:24
particle japhb: i just finished building and testing parrot on my severely resource-constrained xp vm 22:25
japhb ah
particle there are test failures, but i suspect they're related to an old perl module, and not your code
i'll run dump_pbc.pl now
C:\\usr\\local\\parrot\\trunk>perl tools\\util\\dump_pbc.pl x.pbc 22:27
List form of pipe open not implemented at tools\\util\\dump_pbc.pl line 55.
japhb Oh for heaven's sakes
particle This is perl, v5.8.8 built for MSWin32-x86-multi-thread
japhb This is perl, v5.8.8 built for i486-linux-gnu-thread-multi 22:28
Limbic_Region can test on Win32/MinGW and Win32/Cygwin if someone tells me what needs to be ran
japhb Limbic_Region: the command line that particle just posted, with a newly-built parrot
particle ...and japhb's patch applied
that part's important :) 22:29
japhb particle: uh yeah, right.
Limbic_Region needs to know where said patch is
particle castro.iodynamics.com/~tene/parrot-...icle.patch
shorten particle's url is at xrl.us/bkjhb
japhb Patch first, then make ... because the makefiles/root.in is one of the patched files.
Limbic_Region too late
will have to re-make
particle reconfigure, actually
japhb Limbic_Region: no prob ... all it does is enable an extra compile
(It doesn't change libparrot, just enables compile/link of a new binary, 'disassemble', which is required by dump_pbc.pl) 22:31
particle well, disassemble.exe works as i expect it 22:32
japhb Anyway, I'm rather boggled by "List form of pipe open" working on one 5.8.8 and not another. I can understand if pipe open didn't work *at all* on Windows, but this is silly.
Limbic_Region has 5.10 btw
particle pipe open doesn't work on windows
it's noted in the docs
japhb particle: ah, so the error message just sucks.
or rather, it's misleading 22:33
particle yep
japhb crap.
Limbic_Region oh, should I bother?
japhb So what's the replacement? Backticks?
particle temp file?
purl well, temp file is good in some cases... in ytti's regulars.cgi ei use temp file because i need to know if user has oper status or not and from what ip he is coming
japhb Limbic_Region: I'd like to know if it has been made to work in Windows Perl 5.10, actually 22:34
particle uses IPC::Run in his production code 22:35
japhb particle: really didn't want to face that hobgoblin, but if it doesn't work in 5.10, and backticks are also broken, then that will be my fallback.
particle but we don't require IPC::Run for parrot
japhb particle: never used that one. Is it core?
particle backticks should work
no, it's not core :(
japhb particle: ah.
particle really wish it were
Limbic_Region japhb - I am fairly confident without testing it that pipe open did not change in 5.10
22:35 Zaba joined
Limbic_Region has read perldelta very carefully 22:36
particle Limbic_Region: agreed
japhb is generally willing to use portability libs, but only if they are core -- because otherwise, you're just begging the question when it comes to the populous at large
Limbic_Region well, there was a time when anything that was required to build parrot had to be p5 core
but that restriction has been lifted for quite some time 22:37
particle somewhat because perl 5 core has grown quite a bit in seven years
Limbic_Region the idea being that p5 wasn't intended to be part of the public release of parrot
japhb Limbic_Region: OK then. I've got some $paid_work to do right now, but if I get a break I will try to regen the patch using backticks instead. SIGH.
particle File::Temp could work, too 22:38
Limbic_Region k - I am rarely on the channel these days (Win32/Cygwin and Win32/MinGW) but there are others
japhb Apparently particle is one ... :-)
particle msvc++
purl msvc++ is developer.apple.com/qa/qtw/qtw94.html or lists.w3.org/Archives/Public/www-li...02916.html or lists.w3.org/Archives/Public/www-li...02913.html
japhb Limbic_Region: is there a channel you *are* regularly on?
Limbic_Region well, he is Win32/MSVC
but yeah
particle i have cygwin and mingw, but hate them :)
Limbic_Region japhb - normally I am on #perl and #perl6 on freenode and #parrot on irc.perl.org but lately life has taken a priority 22:39
japhb ah
Limbic_Region actually, that seems like a good project 22:40
japhb Hmmm, anyone mind if I tell purl about people's platforms?
The two of you and me, at least.
particle that's what bots are for
Tene japhb: go ahead
Limbic_Region I don't and since the channel is logged now - anyone who mentions it in the channel has already had it recorded
particle but i also have many other platforms in virtual machines 22:41
Tene japhb: perhaps also a list of people for 'windows' and 'macos'
japhb purl, MSVC is also used by particle
purl OK, japhb.
Tene purl: windows?
purl i think windows is a window switcher or total shite or fuck you bill gates you think you won but you only won for an eternity in damnation - get your fucking charity off it's anthropocentric arse and make it tackle issues that really matter for the long term good of the biosphere or an catapostrophe. or warezed or there's no hate like geek hate or if you're using windows it sucks to be you
Tene Or maybe not
Eevee hahaha
japhb purl, WIN32/MSVC is also particle
purl OK, japhb.
Tene I forget that purl isn't for this channel exclusively
particle purl, you're too full of hate. forget windows
purl particle: excuse me?
particle purl, forget windows
purl particle: I forgot windows
japhb purl, Cygwin is also Limbic_Region
purl okay, japhb.
japhb purl, MinGW is also Limbic_Region 22:42
purl okay, japhb.
Eevee I assume linux is not exactly uncommon amongst this channel's population
japhb purl, WIN32/Cygwin is also Limbic_Region
purl okay, japhb.
japhb purl, WIN32/MinGW is also Limbic_Region
purl OK, japhb.
japhb purl, WIN32 is also Limbic_Region and particle
purl okay, japhb.
japhb purl, Windows is also Limbic_Region and particle 22:43
purl OK, japhb.
Limbic_Region scratches his head trying to figure out what perl Configure.PL would look if/when parrot builds without perl 5
japhb purl, Linux/i386 is also japhb
purl OK, japhb.
japhb purl, Linux is also japhb
purl okay, japhb.
japhb purl, Debian/i386 is also japhb
purl OK, japhb.
japhb purl, Debian is also japhb 22:44
purl okay, japhb.
japhb OK, that should cover a few.
Cygwin?
purl i heard Cygwin was pretends to be Unix-like on Windows but doesn't know where to put its ass; try to use unlink() and rejoice or a le cul entre deux chaises or Limbic_Region
japhb well, better than nothing
Infinoid tries to use unlink() and rejoices
japhb LOL
Who runs Mac OS X?
Or something more obscure than the big 3? 22:45
Infinoid kid51 and tetragon use OSX
Limbic_Region japhb - have your read Platforms?
japhb purl, OS X is also kid51 and tetragon
purl okay, japhb.
Infinoid OS X?
purl somebody said OS X was not Unix. Unix doesn't crash. or hangs after toddler induced suspend resume cycles or kid51 and tetragon
Infinoid awesome. 22:46
Eevee it may be more useful (and less cluttersome) to just list this on a wiki
japhb Limbic_Region: yep, it just doesn't give any information about who in the channel could help with testing a patch for portability to some particular OS
Limbic_Region japhb - I was more interested in your comment about the big 3 22:47
japhb Eevee: well, I probably should have /msg'ed it, but I was thinking and not looking at the screen. Didn't realize how much a mess I had made
Limbic_Region other than zaurus
most of those platforms aren't very obscure
and since one of parrot's original goals was to build anywhere perl 5 does - we really need to get work on some other systems 22:48
I don't see any Amiga OS, VAX, etc
Eevee well I just mean a table or list would make it easier to pick out than "OS X is <long rant from #dbix-class> or <people who are useful to #parrot>"
Tene purl: parrot platforms?
purl tene: i haven't a clue
Tene There's an option.
Limbic_Region heck, I don't even see some major *nix like HPUX or AIX
japhb Limbic_Region: I didn't mean "obscure" in a pejorative sense. The emphasis was intended to be "who can we recruit to represent these other platforms, since we have just covered the biggest 3"?
particle we haven't made a push for platform porters yet 22:49
as we've been concentrating on getting core subsystems implemented
NotFound paco is working on aix.
particle when we hit parrot 0.50, we'll definitely be looking to expand beyond the big three 22:50
Limbic_Region well, I would hope that we expand prior to some of the upcoming milestones (or immediately after)
things like concurrency
native threads, etc
particle concurrency will merge to trunk just after next week's release
japhb purl, parrot platforms is Win32/MSVC: particle; Win32/Cygwin: Limbic_Region; Win32/MinGW: Limbic_Region; Mac OS X: kid51 and tetragon; Debian Linux/i386: japhb; AIX: paco 22:51
purl OK, japhb.
japhb How's that?
Limbic_Region would rather have that than all the OS specific ones 22:52
since I am definately NOT cygwin
I just happened to build parrot there regularly
hey particle - was it you that I pointed out that make spectest in rakudo on Win32/MinGW doesn't produce good summary results? 22:54
no, it was Jonathan
japhb OK, I just /msg'ed purl to forget everything I said before, except 'parrot platforms' 22:55
Limbic_Region purl Limbic_Region
purl well, Limbic_Region is coming in from behind a corporate proxy
Limbic_Region purl forget Limbic_Region
purl Limbic_Region: I forgot limbic_region
japhb debian? 22:56
purl, debian?
purl wish i knew, japhb
japhb Yes, you do purl. :-) 22:57
Tene Ack. I accidentally co
japhb ?
Tene Ack. I accidentally committed dump_pbc.pl.
japhb Just that one file?
or the whole patch?
purl i think the whole patch is really just a hack for use until the subreq'd action can be given it's own context object
Tene Along with my svn metadata and trailing whitespace fix.
japhb purl, forget the whole patch
purl japhb: I forgot whole patch
Tene I was converting a diff for particle and forgot that I was using svn. 22:58
japhb It's not appearing when I 'svk pull', so you may be in the clear
Tene, what's your OS?
Tene japhb: linux 22:59
japhb ah
NotFound $ find . -name dump_pbc.pl
./tools/util/dump_pbc.pl
japhb hmmm 23:00
23:00 cjfields joined
japhb shurgs 23:00
NotFound Tene: blame me for pressing you ;) 23:01
Limbic_Region rakudo is passing 1/3 of spectest on Win32/Cygwin 23:02
japhb still wondering how 'svk pull' is not getting it, if Tene committed using ordinary SVN
Tene japhb: dunno, but I just pulled it down with git-svn 23:03
japhb frowns at SVK with arms akimbo 23:04
dalek didn't say anything, either 23:05
Tene japhb: it wasn't just now, it was my last commit
r27510
japhb Freakish. I've got 27510, but SVK is swearing up and down that it still has the original diff against trunk. 23:07
WTF
gah. 23:08
japhb will fight this fight later
DietCoke svk is awesome when it doesn't fail in some way that I can't understand. 23:32
23:41 tetragon joined, cjfields left 23:55 Zaba joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!