Camelia, the Perl 6 bug

IRC log for #parrot, 2008-06-16

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 Zaba joined #parrot
00:04 Zaba_ joined #parrot
00:10 AndyA joined #parrot
00:18 dalek r28400 | tene++ | trunk:
00:18 dalek : Add a smalltalk grammar under the name 'chitchat'.
00:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28400
00:18 Zaba joined #parrot
00:22 dalek r28401 | pmichaud++ | trunk:
00:22 dalek : [rakudo]:
00:22 dalek : * Move postfix:<i> from builtins/op.pir to classes/Complex.pir .
00:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28401
00:23 particle[ventus] joined #parrot
00:28 Zaba_ joined #parrot
00:36 Whiteknight joined #parrot
00:45 dalek r28402 | Whiteknight++ | gsoc_pdd09:
00:45 dalek : [gsoc_pdd09] update to trunk r28401
00:45 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28402
00:57 Whiteknight joined #parrot
01:08 dalek r28403 | pmichaud++ | trunk:
01:08 dalek : [rakudo]:
01:08 dalek : * Fix prefix:<^>(Any) to return an exclusive Range instead of an inclusive one.
01:08 dalek : * Refactor prefix:<^>(Any) into src/classes/Range.pir .
01:08 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28403
01:13 TiMBuS joined #parrot
01:26 bacek__ joined #parrot
01:28 Whiteknight i never know, and i rarely ask
01:35 Zaba joined #parrot
01:57 bacek__ summon pmichaud :)
01:58 bacek__ pmichaud, http://nopaste.snit.ch/13235 this is changes to S29-list/minmax.t that require your opinion...
02:18 pmichaud first hunk (parenthesizing the @array.min : { ... } )   I agree with.
02:20 pmichaud second hunk  (changing $^a to 0+$^a)   shouldn't be needed.   <=> already imposes numeric context.
02:21 pmichaud karma pmichaud
02:21 purl pmichaud has karma of 1402
02:21 DietCoke karma coke
02:21 purl coke has karma of 1921
02:21 chromatic joined #parrot
02:21 pmichaud karma jonathan
02:21 purl jonathan has karma of 650
02:21 DietCoke karma allison
02:21 purl allison has karma of 743
02:21 pmichaud karma particle
02:21 purl particle has karma of 1352
02:21 DietCoke cheater.
02:23 pmichaud oh, I see, the 0+ are arguments to abs.  Still, shouldn't be needed.
02:23 bacek__ rakudo: min({ abs $^a <=> abs $^b }, (-2..2));
02:24 bacek__ pmichaud: 0+ is for abs()
02:24 bacek__ pmichaud: not for <=>...
02:25 pmichaud but abs already treats its argument as being in numeric context.
02:25 bacek__ hmm.. looks like it works now
02:26 pmichaud pugs: say ((0,1,2) x (0,1,2)).perl
02:26 pmichaud pugs: say ((0,1,2) X (0,1,2)).perl
02:26 pmichaud wrong chan
02:27 Whiteknight anybody here good with IMCC?
02:28 Whiteknight with "good" == "understands it at all"?
02:28 pmichaud maybe ask the imcc question?
02:29 Whiteknight well, i'm tyring to implement the .macro_local statement, and I'm trying to find the best place to do it
02:30 pmichaud I don't know that one, sorry.
02:30 Whiteknight it's okay
02:30 Whiteknight I'll monkey my way through it, and we can fix my mistakes later :)
02:36 dalek r28404 | pmichaud++ | trunk:
02:36 dalek : [rakudo]:
02:36 dalek : * More prefix:<^> refactoring.
02:36 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28404
02:40 dalek r28405 | chromatic++ | trunk:
02:40 dalek : [tools] Revised CONST_STRING handling so that occurrences inside
02:40 dalek : partially-generated files don't accidentally overlap hand-written occurrences.
02:40 dalek : This has been a problem occasionally in PMCs.
02:40 dalek : A better solution is to stop generating so much redundant code inside PMCs, but
02:40 dalek : this is an easier short-term fix until the long-term fix is available.
02:40 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28405
02:40 dalek r28406 | pmichaud++ | trunk:
02:40 dalek : [rakudo]:
02:41 dalek : * Fix abs() for non-numeric types.
02:41 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28406
02:42 Whiteknight rakudo progress has been inspiring
02:42 Whiteknight it's amazing to see things moving so quickly
02:42 bacek__ pmichaud: excellent. We can put S29-list/minmax.t in spectest_regression.data now :)
02:46 chromatic joined #parrot
02:57 diakopter someone looking for chromatic
03:02 dalek r28407 | duff++ | trunk:
03:02 dalek : [rakudo]:
03:02 dalek : * Add run builtin (Zev++ RT50558)
03:02 dalek : * patch courtesy Zev Benjamin <zev-perl6  at strangersgate.com>
03:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28407
03:03 dalek r28408 | pmichaud++ | trunk:
03:03 dalek : [rakudo]: Fix abs(Int) to return Int.
03:03 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28408
03:10 chromatic joined #parrot
03:20 dalek r28409 | pmichaud++ | trunk:
03:20 dalek : [rakudo]:
03:20 dalek : * Add S29-list/minmax.t to spectest_regression.data (bacek++)
03:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28409
03:28 Robrt joined #parrot
03:28 Robrt Sorry, I chose the very wrong moment to move RT to a different box.
03:28 Robrt Everything good now?
03:45 Andy joined #parrot
03:46 pmichaud Robrt: I'm still getting "Bad Gateway"
03:47 pmichaud I'm getting "Bad Gateway" on the https: interface.
03:47 pmichaud The http: interface appears to be working.
03:48 Robrt fixed.
03:49 pmichaud When I try to log into rt (via http:) and press "Log in with it now" I get back a "Bitcard data has expired"
03:49 pmichaud also via https:
03:49 pmichaud maybe I need to restart my browser?
03:49 Robrt Maybe I should just move it back to the other machine :)
03:50 pmichaud restarting browser didn't help either.
03:50 Robrt I'm seeing that now too.
03:50 Robrt Crud.
03:50 Robrt Weird.
03:50 Robrt I see.
03:51 Robrt Fixed.
03:51 pmichaud okay, good.  thanks!
03:52 Robrt It actually feels faster to me now.
03:52 Robrt (For multiple reasons.)
03:52 Robrt Sorry for interrupting the bugathon
03:56 pmichaud Works great.  Faster is definitely an improvement (and worth the interruption).
03:57 pmichaud I agree, it seems faster to me also.
03:57 PerlJam woo, it does seem speedy
03:59 dalek r28410 | chromatic++ | trunk:
03:59 dalek : [lib] Added CONST_STRING_GEN #define to dynpmcs.  This fixes a problem
03:59 dalek : introduced in r28405.
03:59 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28410
04:00 davidfetter joined #parrot
04:05 Robrt The slow speed was starting to get to me :)
04:06 Robrt It's still not fast enough, but it should be better
04:11 Tene ... oh, I probably should have kept the new language out of the repo until after the release.
04:11 Tene Yeah, that would have been a great idea.
04:11 PerlJam Tene: so ... what is chitchat?
04:11 Tene PerlJam: smalltalk
04:11 purl smalltalk is, like, very digestable or too One True Way (program text isn't a text file, wtf.)
04:14 Tene I guess I'll have to get it, and a small test suite, working well tonight, because tomorrow looks crazy.
04:19 coke joined #parrot
04:20 * coke ~~, having remembered another box he can ssh to and run irc on.
04:21 coke joined #parrot
04:21 coke bah. /quit ain't /win close.
04:30 coke ZOMG, am I an idiot.
04:31 coke I blame ... sunspots.
04:31 diakopter panel says...
04:31 coke hurm.
04:32 coke ok. low blood sugar. it's probably, and you have no way of checking!
04:32 PerlJam coke: perhaps there was too much interference from all of the other computers/brains in the room?
04:32 coke "probable"
04:32 coke especially pmichaud's
04:32 PerlJam well, next time don't sit so close to him!
04:32 diakopter pmichaud's brain needs a Faraday cage.
04:32 coke PerlJam: Basic diagnostics. I have some stuff I changed to try to add the new features. need to run an svn diff to find the one change that was not quite right. :|
04:33 coke I swear I have to teach this sort of thing at work. Very embarassing to then forget it myself.
04:33 PerlJam coke: it happens to all of us at some point.
04:34 coke ... This seems to have been a very long day of that happening. ^_^
04:36 dalek r28411 | allison++ | pdd25cx:
04:36 dalek : [pdd25cx] Bringing the pdd25cx branch up-to-date with trunk r28410.
04:36 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28411
04:40 dalek r28412 | coke++ | trunk:
04:40 dalek : [tcl]
04:40 dalek : Update tcl to use the now-standard CONTROL_* exceptions. This replaces
04:40 dalek : the original scheme of using our own types and an extra slot on the
04:40 dalek : exception object. It should also make our control exceptions interoperable
04:40 dalek : with rakudo's.
04:41 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28412
04:45 coke pmichaud: hey, patrick, can you make sure smash has whatever he needs in terms of PAUSE permissions?
04:46 pmichaud sure.
04:46 pmichaud tomorrow okay?
04:46 coke jaja
04:46 coke danke.
04:46 pmichaud what's smash's PAUSE id?
04:56 masak joined #parrot
04:56 coke SMASH
05:00 davidfetter yo
05:04 pmichaud okay, I'll check into it tomorrow.
05:06 pmichaud and I'm off to my room to write a journal post and sleep
05:08 davidfetter g'night, pmichaud
05:09 dalek r28413 | coke++ | trunk:
05:09 dalek : [codingstd]
05:09 dalek : Pass "make codetest"
05:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28413
05:11 Robrt left #parrot
05:13 cotto_yapc joined #parrot
05:16 dalek r28414 | coke++ | trunk:
05:16 dalek : [chitchat]
05:16 dalek : As long as it's in the repo, generate the makefile during config.
05:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28414
05:26 Infinoid joined #parrot
05:31 dalek r28415 | coke++ | trunk:
05:31 dalek : Build targets that differ only by case are not boss.
05:31 dalek : mingw32-make considers them the same targets.
05:31 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28415
05:41 coke allison++
05:42 particle[ventus] ...so does nmake...
05:42 coke and you -use- tags. why did I have to fix that. =-)
05:44 Zaba joined #parrot
05:49 masak it's been bugging me for a while that `./perl6 -e 'say $_ for <a b c>` doesn't work
05:50 masak I submitted a bug report about it, http://rt.perl.org/rt3/Tic​ket/Display.html?id=55184
05:50 masak but this has so far been unheeded
05:51 masak would someone be so kind as to review this patch, which resolves the issue: http://paste2.org/p/39784
05:53 masak I'm clearly batting in the dark, and I realize that it is a band-aid upon the couple of layers of band-aid already deployed in that area...
05:53 masak but it solves the $_ parsing problem, and all tests still pass
05:55 uniejo joined #parrot
05:55 DietCoke joined #parrot
05:56 particle[ventus] masak: i'm too tired to apply now. glad it solves the problem, but __SCARY_VARIABLES are too much for me to review now
05:57 particle[ventus] i'll look in the morning, but pmichaud will likely have beat me to it
05:57 particle[ventus] &
05:57 masak particle[ventus]: np. night
06:00 masak btw, this solution breaks for all evaluations except the first one when perl6.pbc is run in REPL mode
06:00 masak no idea why.
06:17 iblechbot joined #parrot
06:21 masak leaving now, will backlog later
06:23 * coke wonders why we roll our own c-level options handling.
06:24 coke masak, you might want to log the "breaks for all" on the ticket itself.
06:29 Zaba_ joined #parrot
06:43 pmichaud the patch is taking the wrong approach to solving $_ (but so is the existing code in actions.pm)
06:43 pmichaud I'll work on fixing $_ and friends tomorrow.  I had been waiting to resolve 'return' first, but now that is done so I can think I can do $_, $!, and $/.
06:44 pmichaud afk, sleep.
06:44 pmichaud (new journal posts at use.perl, btw)
06:44 dalek r28416 | coke++ | trunk:
06:44 dalek : RT# 45267 - Fix bug in parrot's handling of default args for long options.
06:44 dalek : (like --trace)
06:44 dalek : This allows parrot -trace foo.pir to work the same as parrot -t foo.pir
06:44 dalek : All tests pass, including the un-TODO'd one.
06:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28416
06:51 Psyche^ joined #parrot
06:51 Zaba joined #parrot
07:55 allison joined #parrot
08:23 cosimo joined #parrot
08:42 dalek r28417 | tene++ | trunk:
08:42 dalek : [chitchat]
08:42 dalek : * Take the grammar from gnu smalltalk instead of ST-80
08:42 dalek : * Add some very basic AST generation
08:42 dalek : * Add a single builtin, 'Transcript', with a single method
08:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28417
08:44 moritz pmichaud: your rakudo blog post has no title
08:44 polyglotbot joined #parrot
08:44 Tene chitchat: Transcript show: 'zomg hi'
08:44 polyglotbot OUTPUT[zomg hi␤]
08:50 TiMBuS how do i force a string copy without doing anything to the string?
08:51 TiMBuS maybe an empty concat..
08:52 TiMBuS nope
08:53 Tene TiMBuS: what are you trying to do?
08:54 TiMBuS reverse a string =/ but it doesnt copy
08:54 TiMBuS so, i copy the source too
08:54 TiMBuS uh
08:54 TiMBuS i reverse the source*
09:00 TiMBuS am i supposed to use something other than 'clone' perhaps?
09:01 IllvilJa joined #parrot
09:25 dalek r28418 | tene++ | trunk:
09:25 dalek : Mention chitchat in NEWS, LANGUAGES_STATUS.pod, and 'make languages'
09:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28418
09:26 mj41 joined #parrot
09:53 bacek evening
09:58 bacek karma bacek
09:58 purl bacek has karma of 44
10:01 bacek karma moritz
10:01 purl moritz has karma of 49
10:02 TiMBuS karma fight!
10:02 purl fight! has karma of 1
10:02 TiMBuS .. lol, what
10:02 bacek yak...
10:15 * bacek wander if there is something similar to C++ constructors in protoobjects...
10:17 Whiteknight joined #parrot
11:15 Eevee joined #parrot
12:03 iblechbot joined #parrot
12:39 Whiteknight joined #parrot
12:40 donaldh joined #parrot
12:41 Zaba_ joined #parrot
12:50 tetragon joined #parrot
12:59 gryphon joined #parrot
12:59 dalek r28419 | fperrad++ | trunk:
12:59 dalek : [docs]
12:59 dalek : - add language ChitChat
12:59 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28419
13:08 allison joined #parrot
13:11 dalek r28420 | Whiteknight++ | gsoc_pdd09:
13:11 dalek : [gsoc_pdd09] update to trunk r28419
13:11 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28420
13:15 dalek r28421 | Whiteknight++ | trunk:
13:15 dalek : [docs/book] adding missing image to chapter 7 for rt#55374.
13:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28421
13:17 mj41 joined #parrot
13:22 kid51 joined #parrot
13:26 kid51 YAPC::NA::2008 begins
13:29 Coke dun-dun-DUN!
13:34 allison coke: do your kids watch hoops and yoyo?
13:39 dalek r28422 | coke++ | trunk:
13:39 dalek : Add tcl and alphabetize language entries for this release.
13:39 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28422
13:39 smash joined #parrot
13:39 smash hello everyone
13:39 Whiteknight hello
13:40 Whiteknight I just had a test failure with t/dynpmc/rational, ended with a segmentation fault
13:41 paco joined #parrot
13:41 Whiteknight and a lot of segfaults in t/stm/basic
13:41 smash which OS are you running ?
13:41 Whiteknight debian
13:42 * smash fires up a debian virtual machine
13:42 Whiteknight I'll do a realclean and an update before I file any tickets. I dont think I've realcleaned this morning
13:42 Coke allison: I have no idea what that is.
13:43 confound Coke: you are so uncool
13:43 smash Whiteknight: i'll to a 'make fulltest' in a sec
13:43 Whiteknight okay, awesome
13:44 Coke "hoops" to me means basketball.
13:46 IllvilJa joined #parrot
13:50 smash any ideias now to simply replace 'saveall' and 'restoreall' opcode
13:50 smash ?
13:51 smash Whiteknight: what version of debian you using ?
13:52 Whiteknight Ubuntu 8.04
13:52 purl Ubuntu 8.04 is probably out, go get it
13:52 Whiteknight thanks purl
13:53 smash what test is failling ?
13:54 Whiteknight the first test in t/dynpmc/rational, several tests in t/stm/*
13:54 Whiteknight I closed it after that and started a make realclean
13:56 smash All tests successful, for those here.. i'm also running ubuntu
13:56 smash Coke: tomorrow release is going to be 0.6.3 correct ?
13:57 pmichaud oh, I need to update NEWS
13:57 smash pmichaud: please do that :)
13:57 Whiteknight okay, it must be a problem on my end. I'm rebuilding now
13:57 smash Whiteknight: ok, try it again afer rebuilding then
13:58 smash All tests successful, also on debian testing
13:59 allison coke: google search for "hoops and yoyo marshmallow", dun-dun-dun-DUN
14:06 smash_ joined #parrot
14:06 Coke smash: yes. I don't think we're ready for 0.7.0
14:06 TiMBuS joined #parrot
14:07 smash_ damn storms :\
14:07 masak joined #parrot
14:07 Whiteknight who is the release manager for tomorrow?
14:07 masak <coke> masak, you might want to log the "breaks for all" on the ticket itself.
14:07 masak
14:07 masak Coke: what does that mean?
14:07 purl That boy needs therapy.
14:07 smash_ Coke: roger that
14:08 Coke allison:Ah. those characters look vaguely familiar.
14:08 toddr joined #parrot
14:10 smash joined #parrot
14:11 smash (got disconnected again.. damn storms)
14:11 Whiteknight smash, all tests pass for me now.
14:11 smash great
14:11 Whiteknight every now and then realclean is a necessity
14:11 * smash nods
14:11 Coke masak you said you had a patch. then you said the patch failed in some conditions. wanted to make sure the fact that the patch failed in those conditinos was documented inthe ticketing system.
14:12 smash i could use some windows testing feedback tho
14:12 Coke pmichaud: did you setup smash in pause?
14:12 Coke smash: I have some test faiulres, but a mostly good run in strawberry perl.
14:12 * Coke will svn up and try again.
14:13 smash Coke: still trying to get parrot to build on my windows :\
14:13 masak Coke: the bug should be platform-independent
14:13 smash Coke: some codingstd tests are failling..
14:14 masak the bug is here: http://rt.perl.org/rt3//Publ​ic/Bug/Display.html?id=55184
14:14 masak how do I log it as "breaks for all"?
14:16 smash pmichaud: any hints on replacing 'saveall' and 'restoreall' opcodes ?
14:16 Coke masak: that's good enough: sounded to me like you had a -patch- that broke. I misunderstood.
14:16 masak no, I have a patch that fixes
14:17 masak I'd like to have it reviewed and commented
14:17 masak it's here: http://paste2.org/p/39784
14:17 Coke aha. Sorry. =-)
14:17 masak np
14:22 dalek r28423 | smash++ | trunk:
14:22 dalek : [release]
14:22 dalek :  * PLATFORMS update
14:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28423
14:26 Coke smash; the build is killing my battery. will complete later when I have power.
14:26 moritz masak: I'll test it, just a sec
14:26 smash Coke: sure, thanks
14:26 Whiteknight Anybody have an opinion on #55538, whether ".macro_const" should work in *.pir files?
14:27 moritz masak: it seems to fix it, but not in the interactive shell
14:28 masak moritz: that's my conclusion as well
14:28 Coke it would have been nice if it did 3 days ago. but you can always put them in a .pasm file, include that from your .pir.
14:28 masak moritz: or rather, the first evaluation in the REPL shell always works
14:28 masak but after that, another block is involved somehow and messes up the counting
14:28 moritz masak: to increase your chances of getting the patch applied, please send it to rakudobug@perl.org
14:28 masak moritz: will do
14:28 moritz the REPL does some funny stuff with scoping
14:29 masak seemingly.
14:29 Coke it should work. it's not a high priority blocker.
14:29 Whiteknight Coke, 3 days ago would have been nice, I have to wait now till after the release. The patch to fix it was trivial though
14:30 donaldh Calling NCI experts...
14:31 donaldh I'm writing some pir to call sqlite3 but am having problems with out parameters
14:31 donaldh Does anyone know what calling convention is required for:
14:32 nopaste "donaldh" at 144.254.89.228 pasted "sqlite3_open" (4 lines) at http://nopaste.snit.ch/13285
14:33 donaldh I'm reading through pdd16 and nativecall.pl but void** out parameters _seem_ to be an oversight.
14:34 MagNET joined #parrot
14:35 uniejo joined #parrot
14:36 donaldh I think I need the reciprocal for "p" but "P" isn't that. "B" seems close but is actually string specific.
14:37 smash joined #parrot
14:37 allison Whiteknight: on RT #55538, it was spec'd but not implemented (note, pdd19 is still in draft), the patch looks reasonable
14:42 Whiteknight thanks allison, that's why I wanted to open it up for discussion and see if I should implement it or not. I'm inclined to say that we should just because it's an easy fix and worthwhile functionality
14:42 Whiteknight but we've gotten along just fine without it too
14:43 dalek r28424 | pmichaud++ | trunk:
14:43 dalek : [rakudo]:
14:43 dalek : * Add docs/spectest-progress.csv to track progress on passing spectests.
14:43 dalek :   (I'll keep this file up-to-date for the time being.)
14:43 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28424
14:45 moritz pmichaud: any chance to include the graph in your blog post? people love charts ;)
14:45 * Whiteknight definitely loves charts
14:46 moritz it's attached to this mail: http://www.nntp.perl.org/group/perl.​perl6.compiler/2008/06/msg1942.html
14:46 shorten moritz's url is at http://xrl.us/b7hzx
14:46 moritz sadly nntp.perl.org doesn't show attachments
14:49 pmichaud moritz: yes, I'm working on that.
14:49 pmichaud Wanted to get the data out first.
14:50 moritz pmichaud: and I'm working on modifying my script to work with the CSV data directly
14:50 pmichaud okay, great.
14:50 pmichaud Tim Bunce asked for the data in CSV format, and that seems like a reasonable happy medium for now.
14:52 particle joined #parrot
14:52 toddr joined #parrot
14:55 Whiteknight seen chromatic?
14:55 purl chromatic was last seen on #parrot 17 hours and 38 minutes ago, saying: Wow, if only something could notify me every hour!
14:56 diakopter and that was that
14:58 smash joined #parrot
14:58 pmichaud grr, slow network.
15:03 davidfetter joined #parrot
15:07 davidrw joined #parrot
15:10 davidrw left #parrot
15:11 moritz pmichaud: I now updated the script. Do you want to add it to the repo? if yes, I'll add copyright information etc.
15:12 moritz nopaste?
15:12 clunker3 http://pasta.test-smoke.org/ or http://paste.husk.org/ or http://nopaste.snit.ch:8001/ or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/
15:12 purl rumour has it nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://paste.husk.org/ or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or don't bother me while I'm eating or App::Nopaste or tools/dev/nopaste.pl
15:12 particle Whiteknight: i'm sitting next to chromatic, what do you need?
15:12 toddr joined #parrot
15:13 contingencyplan joined #parrot
15:14 nopaste "moritz" at 89.13.234.141 pasted "tool for plotting t/docs/spectest-progress.csv" (40 lines) at http://nopaste.snit.ch/13286
15:15 Coke I need you to punch him in the face.
15:16 Coke crud, he's not sitting there any more.
15:22 Whiteknight particle, I just sent him an email with questions about PMC_EXT
15:23 Whiteknight not a big deal, don't bother him if he's busy
15:24 moritz why is rakudo.org so paranoid? first you have to log in to leave a comment, then it's still held for moderator approval
15:25 toddr left #parrot
15:25 toddr joined #parrot
15:27 Coke mebbe laster was getting a lotta spam. he's usually responsive if you want to ask him.
15:27 Coke s/laster/lester/
15:31 Coke smash:
15:32 Coke Files=585, Tests=11304, 752 wallclock secs ( 4.42 usr +  1.34 sys =  5.77 CPU)
15:32 Coke Result: FAIL
15:32 Coke Failed 9/585 test programs. 11/11304 subtests failed.
15:32 Coke mingw32-make: *** [test] Error 255
15:32 smash :\
15:33 Coke most of this there are already tickets for.
15:33 Coke I suppose we should TODO all those in time for the release.
15:35 smash ok, still trying to get this to build on windows
15:35 jhorwitz joined #parrot
15:37 particle joined #parrot
15:38 dalek r28425 | Whiteknight++ | gsoc_pdd09:
15:38 dalek : [gsoc_pdd09] lots of small fixes
15:38 dalek : * Fixed macros and moved definitions outside functions
15:38 dalek : * turned simple marking functions into macros
15:38 dalek : * Fixed error in pobject_lives logic for GC_IT_MARK_ROOTS state
15:38 dalek : * Updated Gc_it_hdr to cache card and flag offsets, instead of raw index numbers
15:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28425
15:42 particle sorry, Whiteknight, got disconnected. have anything for chromatic?
15:42 Whiteknight nothing important, I sent him an email he can read at his leisure.
15:42 nopaste "donaldh" at 144.254.89.228 pasted "patch: adding void** OUT parameter support" (30 lines) at http://nopaste.snit.ch/13287
15:43 donaldh Can someone check this patch for sanity?
15:44 Whiteknight donaldh, for my limited knowledge of NCI, the patch looks decent
15:44 donaldh Whiteknight: it certainly solves my problem with sqlite3
15:44 gryphon joined #parrot
15:46 NotFound I'm thinking that will be good to rename include/parrot/debug.h to pdb.h to avoid confusions with compilers/imcc/debug.h and reduce dependencies. Opinions?
15:51 pmichaud smash: you should now have co-maint permissions on all Parrot modules in PAUSE.  Let me know if you run into any problems.
15:52 smash pmichaud: ok.. thank you
15:54 smash pmichaud: as i look to my "view permissions" page everything looks good
15:56 * japhb is back from family-visit trip, just beginning to get caught up ...
15:57 japhb msg tetragon I plan for shapes.pir to be way more than a "teapot conveyor", but that's all I had time to get ready for the current release.
15:57 purl Message for tetragon stored.
15:57 dalek r28426 | Whiteknight++ | gsoc_pdd09:
15:57 dalek : [gsoc_pdd09] more small fixes
15:57 dalek : * Kill unused redirect functions
15:57 dalek : * Fix some comments
15:57 dalek : * Fix gc_it_add_arena_to_free_list to perform correct behavior
15:57 dalek : * Stop relying on sizeof(char) == 1 for pointer arithmetic (even though it's usually safe)
15:57 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28426
15:57 japhb Tene: You were getting a segfault in shapes.pir?  Is it still there, and do you have time to locate it with me?
15:58 smash Coke: you're building with 'nmake' or 'mingw32-make' ?
15:58 japhb Coke: were you looking for me, or just needed my email address?  (And if the latter, for what?  I haven't checked my mail yet.)
15:58 Whiteknight japhb, shapes.pir is awesome!
15:58 japhb Whiteknight: thanks.  :-)
15:58 NotFound MMmmm.... but sizeof(char) == 1 by definition!
15:59 japhb Whiteknight: Like I msg'ed to tetragon, I'm hoping to make it cooler still.
15:59 Whiteknight NotFound, I don't think so. I've heard it has to be "at least" that size, but no requirement that it must always be 1
15:59 sscaffidi joined #parrot
15:59 donaldh japhb: I never got a patch for cygwin for you - lightning took out most of my electronics.
16:00 NotFound Whiteknight: sizeof return the size in byte, and the definition of byte in the C standard is the size of a char.
16:00 japhb donaldh: OUCH.  I hope you were insured and backed up!
16:00 donaldh both, thanks.
16:00 japhb (offsite)
16:00 japhb good!
16:00 NotFound What is not required is that a byte has just 8 bits.
16:01 Coke d/smash: mingw32-make
16:01 japhb donaldh: some of your problems may have been fixed by now because of other people's requests.  Care to try again?
16:01 Coke japhb, double checking our commit bit lists/emails. nbd.
16:01 japhb Coke: Ah, OK.
16:01 * Coke will try to remember to do that in private, sorry.
16:01 donaldh japhb: it looks like we need a 3rd platform type for windows. VC++ != MinGW != cygwin
16:01 Coke My platform ATM is strawberry perl.
16:02 Coke I also have vc++ here, but don't use it by default.
16:02 japhb donaldh: nodnod.
16:02 donaldh japhb: I svn uped about 1 hour ago and still got errors. Give me a minute.
16:02 japhb donaldh: thanks.
16:03 Whiteknight NotFound, maybe that's what I'm thinking of, the 8-bits requirement
16:03 smash Coke: ok, thks
16:04 nopaste "donaldh" at 144.254.89.228 pasted "japhb: opengl configure error" (5 lines) at http://nopaste.snit.ch/13289
16:05 japhb donaldh: Oh, that's an easy one.  (Or, well, it should be ...)
16:05 donaldh Yep.
16:06 japhb donaldh: Should just be a matter of adding the standard include/GL directory for your compiler to the list in config/gen/opengl.pm
16:06 japhb in "sub runstep"
16:06 donaldh Yup. I thought I'd hacked through that.
16:07 cjfields joined #parrot
16:07 donaldh Maybe if I set INCLUDE it works. I forgot.
16:08 japhb donaldh: Which of the 3 compiler environments are you running?
16:09 donaldh cygwin gcc
16:09 donaldh japhb: I need /usr/include/w32api to be searched for GL/*
16:09 japhb mrf.  I thought cygwin used the unix-like directories?
16:09 japhb donaldh: oh.  OK, that's a fair change.
16:09 japhb give me a minute to svn up
16:10 smash i don't believe it !!!!! i'm actually running a test on windows !! sweet
16:10 Coke smash++
16:10 donaldh japhb: Awww. That's nice. shapes.pir is running.
16:10 japhb schweet
16:11 NotFound Whiteknight: 6.5.3.4 The sizeof operator. .... 3 - When applied to an a operand that has type char, unsigned char or signed char (or a qualified version thereof) the result is 1.
16:11 cjfields_ joined #parrot
16:11 donaldh so, yes, the link libs issues are already resolved.
16:11 NotFound From the draft of 1998, but I think in previous versions was the same.
16:12 donaldh japhb: I get a core when I quit. I'll investigate later.
16:12 japhb donaldh: perfect.  OK.  Give me a couple minutes to make the patch (I'm multitasking), or you can if you like.  Just comment that directory glob as being for cygwin.
16:12 japhb donaldh: That's interesting ... I wonder if that's where Tene was coring ...
16:13 japhb afk for just a sec
16:14 Whiteknight NotFound, thanks.
16:15 dalek r28427 | particle++ | trunk:
16:15 dalek : [core] remove warning for potentially uninitialized variable
16:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28427
16:16 japhb donaldh: bak
16:19 donaldh japhb: in glutTimerFunc () from /usr/bin/glut32.dll
16:19 donaldh japhb: I'd need to build glut with debugging to find the cause
16:20 japhb donaldh: odd, that
16:20 donaldh howso
16:20 japhb ... since I don't use glutTimerFunc in shapes.pir (yet)
16:23 dalek r28428 | Whiteknight++ | gsoc_pdd09:
16:23 dalek : [gsoc_pdd09] liberally apply "const" to variables, fix 16-bit bitfields to unsigned short ints (for ease and compatibility)
16:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28428
16:24 nopaste "donaldh" at 144.254.89.228 pasted "japhb: stack trace" (51 lines) at http://nopaste.snit.ch/13290
16:24 japhb donaldh: brb
16:27 japhb Interesting stack trace
16:29 japhb .oO( libglutcb has the nice (and unplanned) side effect of letting me see which GLUT callback occurred in C stack traces ...)
16:32 japhb donaldh: I'm confused by that trace ... it almost looks as if GLUT on Win32 is internally setting timers in order to keep idle events working.
16:32 dalek r28429 | Whiteknight++ | gsoc_pdd09:
16:32 dalek : [gsoc_pdd09] redo gc_it_get_free_object to use current data structures. initialize parent_pool pointer of all objects in initialization routine instead.
16:32 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28429
16:33 japhb Hmmm, I wonder if that's true cross-platform, and I just didn't notice before.
16:33 japhb summon Tene
16:34 donaldh japhb: that could be.
16:35 Coke who wants to write something that will email the user when summon is used?
16:37 donaldh japhb: sorry I need to go.
16:37 japhb donaldh: fair enough.  At least we got things to *run* for you, even if it crashes at exit
16:37 donaldh I'll look into the why when I get a chance.
16:38 jonathan pmichaud: Looks like we get all of our talking done on the first day of YAPC::EU.
16:41 smash jonathan: and i'm the the first one
16:42 jonathan Yes, I noticed.
16:42 jonathan That track is Perl 6 and Parrot for much of the day.
16:43 slightlyoff joined #parrot
16:43 * smash nods.
16:43 Tene japhb: I'll be around shortly.
16:44 japhb Tene: thx
16:45 Tene Just waking up.  Still need to shower, scrollback, email.
16:46 japhb Tene: I've got about 3 days of scrollback and email to address, so I feel your pain.
16:52 Tene I've got 7 hours.
16:52 Tene Okay, starting X, will test again shortly.
16:59 Tene japhb: Yes, it still segfaults, and I have a few minutes.
17:00 japhb Tene: at exit, or startup?
17:00 Tene startup.  does not display any window.
17:00 Tene lemme nopaste
17:00 Tene purl: nopaste?
17:00 purl i guess nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://paste.husk.org/ or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or don't bother me while I'm eating or App::Nopaste or tools/dev/nopaste.pl
17:00 japhb k
17:01 nopaste "tene" at 67.182.199.217 pasted "shapes.pir" (5 lines) at http://nopaste.snit.ch/13291
17:02 dalek r28430 | Whiteknight++ | gsoc_pdd09:
17:02 dalek : [gsoc_pdd09] several fixes:
17:02 dalek : * Simplify "sweep" phase mechanics in the state machine
17:02 japhb WTF?  That doesn't look like a bug in my code.  That looks like a Parrot bytecode bug of some sort
17:02 dalek : * Fix private function prototypes to use "static"
17:02 dalek : * Redo gc_it_enqueue_next_root for more efficient and sane handling of non-aggregate root items
17:02 dalek : * Update card resetting function name and comments
17:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28430
17:02 japhb Tene: are you running something other than trunk?
17:03 Tene japhb: plain vanilla trunk, just did a fresh checkout.
17:03 Tene I'm on x86_64
17:03 Tene fedora
17:04 japhb OK, that's really freakish
17:04 japhb Let me svn up and rebuild
17:04 Tene Hm, triangle.pir does the same thing.
17:08 Infinoid hmm.  I'm getting "freeglut (examples/opengl/triangle.pir):  ERROR:  Internal error <Visual with necessary capabilities not found> in function fgOpenWindow"
17:08 * Infinoid wonders what he broke.
17:08 japhb Given the failure, I'm not surprised.  As far as I can tell, the OpenGL code hasn't even begun to run at that point, it's the bytecode unpacker failing.
17:08 japhb (That was to tene)
17:08 japhb Infinoid: that's ... odd.
17:09 japhb Infinoid: I could sortof imagine that happening in shapes.pir, but not triangle.pir.
17:09 japhb triangle.pir is about as dead simple as possible.
17:09 Infinoid oh, glxinfo reports errors too.  let me fix my system libraries first, then I'll be back :)
17:09 japhb (intentionally -- it's supposed to be the 00-sanity of the OpenGL code)
17:09 japhb Infinoid: gotcha.
17:09 allison joined #parrot
17:09 * Infinoid is going through spinning triangle withdrawl
17:09 japhb Infinoid: heh
17:11 * Tene does actual clean checkout, just to try
17:11 smash pois
17:11 smash ups, sorry
17:14 Tene Ah.
17:14 Tene japhb: in a new checkout, not just realcleaned, I get:
17:14 Tene error:imcc:No such file or directory in file 'examples/opengl/triangle.pir' line 26
17:14 Tene Which is opengl_defines.pasm
17:15 Tene In Makefile.PL, I get: Generating OpenGL bindings.....................​....................skipped.
17:16 japhb Tene: OK, now your error is making sense
17:16 japhb What did the autodetection say?
17:16 Tene japhb: how can I see that?
17:16 japhb rerun Configure.pl
17:17 Tene What line am I looking for there?
17:17 japhb You should see a line "Determining if your platform supports OpenGL...."
17:17 Tene "no"
17:18 japhb That would be the problem, right there.  :-)
17:18 japhb Have you rebuilt your box recently?
17:18 * Tene tries "freeglut-devel"
17:18 japhb Tene, oh yeah, that would be needed, yes
17:18 Tene Yeah, works now.
17:19 Tene Sorry to waste your time.
17:19 japhb Tene: no problem!
17:20 japhb OK, lessee ... Tene
17:20 japhb er
17:21 japhb Tene's problem fixed, Infinoid fixing his box, donaldh getting weird segfault on exit (ooh, I think I have a workaround in my head for that one), ... Is that everyone with OpenGL problems?
17:22 moritz that shape example runs fine for me ;)
17:22 japhb Wish I'd had time to finish it for this release, but I guess that will happen in the next release cycle
17:33 toddr joined #parrot
17:44 cognominal joined #parrot
17:44 NotFound languages/perl6/docs/spectest-progress.csv is failing codetest
17:56 cotto_yapc joined #parrot
17:59 toddr joined #parrot
18:02 dalek r28431 | Whiteknight++ | gsoc_pdd09:
18:02 dalek : [gsoc_pdd09] add a scan minimum, in case we enqueue a very shallow tree. Separate out code for selecting serial/parallel modes into it's own function.
18:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28431
18:08 toddr joined #parrot
18:08 Ivatar joined #parrot
18:08 barney joined #parrot
18:10 packy joined #parrot
18:16 cognominal joined #parrot
18:17 Zaba joined #parrot
18:32 toddr joined #parrot
18:39 toddr joined #parrot
18:53 dolmen joined #parrot
18:57 dalek r28432 | pmichaud++ | trunk:
18:57 dalek : [rakudo]:
18:57 dalek : * SVN property fix on spectest-progress.csv .
18:57 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28432
19:07 AndyA_ joined #parrot
19:10 deven joined #parrot
19:16 purl joined #parrot
19:24 cotto_yapc joined #parrot
19:25 dalek r28433 | pmichaud++ | trunk:
19:25 dalek : [rakudo]:
19:25 dalek : * Update Complex as subclass of Any, add .perl method (dolmen++)
19:25 dalek : * Patch RT#55872 courtesy Olivier Mengu� <olivier.mengue@gmail.com>
19:25 dalek :   (with minor changes by pmichaud)
19:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28433
19:31 dalek r28434 | Whiteknight++ | gsoc_pdd09:
19:31 dalek : [gsoc_pdd09] bunch of small changes, mostly comments
19:31 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28434
19:34 cotto_yapc joined #parrot
19:36 smash joined #parrot
19:40 coke_ joined #parrot
19:55 integral joined #parrot
20:00 tewk pmichaud: ping.
20:00 pmichaud pong.
20:01 tewk I've got two questions.
20:01 tewk I want to write a CPP grammar that is only parses a line at a time.
20:01 tewk Should I split the input file into lines and then feed them to PGE a line at a time?
20:02 pmichaud that would probably work, yes.
20:02 tewk Would the interactive features of HLL, help at all here?
20:02 pmichaud whether that's the best approach or not -- I don't know.  I'd have to play with it.
20:02 avar You can define TOP as <line>+ and then have rule line {} define the real grammar
20:03 tewk Its not interactive, but because of #include the input stream can dynamically change. :)
20:04 pmichaud I don't think interactive parts of HLL would be of much use, no.
20:06 tewk Could I write a C preprocessor in rakudo as it stands today?
20:06 pmichaud Depends on what you want from the C preprocessor.  Certainly if there's a feature missing we could see about quickly adding it.
20:06 PerlJam tewk: someone could, I don't know if *you* can  ;-)
20:11 dalek r28435 | smash++ | trunk:
20:11 dalek : [release]
20:11 dalek :  * skip a couple of tests that are failling because deprecated opcodes are being used
20:11 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28435
20:13 tewk I basically need to read files, parse lines using grammars, and be able to call pir functions.
20:14 sjansen joined #parrot
20:14 tewk I was about to write my custom CPP in PIR, but that just seemed laborious given rakudos progress.
20:16 PerlJam tewk: so ... are you not going to use PCT but rather rakudo itself?
20:18 dalek r28436 | pmichaud++ | trunk:
20:18 dalek : [rakudo]:
20:18 dalek : * Make rakudo's grammar slightly closer to STD.pm (expect_term).
20:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28436
20:20 Coke2ElectricBoogal ... damnit, I just spilled my salad all over the chair and floor of my room. :|
20:22 tewk CPP isn't really a language.  I'm not sure if PCT will be much help.
20:22 cjfields_ pmichaud: how far away are we from implementing methods in p6?
20:22 cjfields_ s/methods/rakudo methods/
20:23 AndyA joined #parrot
20:23 PerlJam I find that statement odd "cpp isn't really a language", but since you're the one doing the work, however you want to look at it  :)
20:23 PerlJam tewk: btw, have I told you congrats on the grant yet?  if not, tewk++
20:25 tewk Its a text macro language, kinda like m4 or tcl.
20:26 * Coke glares at tewk.
20:27 Coke "IS TOO!"
20:28 PerlJam tewk: sure, but it's still a language that requires parsing and some semantics associated with the syntax.
20:28 PerlJam tewk: I'm just wondering where the dividing line is between "a langauge" and "not really a language"
20:29 Eevee by this definition I propose php is not really a language
20:31 moritz that reminds me of that "dynamic languages strike back" talk, where the definition of "dynamic languages" was "by definition, languages like perl, python, ruby, javascript, ..."
20:32 Eevee haha
20:32 Eevee dynamic languages are languages that are like other dynamic languages
20:33 cotto_yapc joined #parrot
20:33 NotFound So the only way to define dynamic languages is by listing all?
20:47 moritz NotFound: no, but it's the funniest way ;)
20:48 moritz common characteristics are run-time typing, eval(), closures, garbage collecting etc.
20:50 Casan joined #parrot
20:51 jhorwitz joined #parrot
20:58 pmichaud cjfields_: (methods in p6)  -- I thought I was going to have it working this past weekend, but there are some issues with parrot's :load :init semantics.
20:58 pmichaud so I need some time to map it all out in my head.
20:58 pmichaud (or in front of a white board or something.)
21:04 bacek morning everyone.
21:05 bacek pmichaud, Is there anything like 'C++ constructors' for protoobjects?
21:05 bacek e.g. code executed on "new 'List'".
21:06 PerlJam pm: I'm looking at rt55184.   The patch provided in the ticket does seem to fix the problem stated, but should parrot perl6.pbc -e '{ $_ }' do the same thing or does the presence of $_ automatically parameterize the code block?
21:09 PerlJam never mind, found the answer in S04
21:10 cjfields_ pmichaud: ok.  I have a couple of p6-based implementations (including trans).  I'll submit to them to RT when I get them passing tests.
21:23 pmichaud pj:  I'm rejecting 55184.  I don't like the way rakudo is (over)using the __VARIABLE_FLAGS .
21:24 IllvilJa joined #parrot
21:24 pmichaud it reminds me of 'infix:,' -- it's a meme that's being cargo culted in too many places to patch (not fix) the original problems.
21:25 tewk pmichaud: Can I prematurely end a parse with pge returning the current match and character position.?
21:25 pmichaud tewk: for what purpose?
21:26 pmichaud (short answer 'no', but depending on what you want to do there's likely another approach)
21:26 pmichaud bacek:  (constructors)   that's method .new
21:26 pmichaud e.g.,   List.new(parameters)
21:32 dalek r28437 | coke++ | trunk:
21:32 dalek : [examples]
21:32 dalek : Convert one of the older PASM example files (with opcodes that were removed)
21:32 tewk pmichaud: When I hit a #include I need to change the source stream,  Can I do that in a parse action?
21:32 dalek : into a functional PIR version.
21:32 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28437
21:32 pmichaud you could potentially alter the current string being matched and then pick up from the same location.
21:33 pmichaud it probably can be done from a parse action also.
21:42 bacek_ joined #parrot
21:45 hachi_ joined #parrot
21:46 cotto_yapc joined #parrot
21:46 Zaba_ joined #parrot
21:46 bacek pmichaud, thanks a lot!
21:52 Coke smash: I'm resurrecting your dead tests.
21:54 pmichaud (topic transfered from #hackathon) -- one of the things I was thinking we might cover during the hackathon was "brainstorming ideas for HLL interop, library management".  It didn't happen -- we were all busy.  (more)
21:54 pmichaud should we try to schedule a time to discuss that -- perhaps a miniBOF of some sort?
21:54 sjansen joined #parrot
21:54 askie joined #parrot
21:55 bacek pmichaud, .new doesn't work... ".namespace['List'] .sub '.new' :method say 'Hello' .end   list = new 'List'.
21:56 bacek pmichaud, what I did wrong?
21:56 pmichaud bacek:  That's because one has to use the protoobject to do that.
21:56 pmichaud $P0 = get_hll_global 'List'
21:56 pmichaud list = $P0.'new'()
21:57 dalek r28438 | coke++ | trunk:
21:57 dalek : Convert another very old, broken PASM test into modern pir.
21:57 dalek : The last converted test should have used PMCs, not ints, fix that up. Add
21:57 bacek pmichaud, ok... It's just too many places with 'new List'.
21:57 dalek : some comments to the description to clear that up.
21:57 dalek : All tests in t/benchmark/benchmarks.t now pass without TODO.
21:57 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28438
21:57 pmichaud sometimes we want  'new List'.  Sometimes we'll want listproto.'new'()
21:57 pmichaud it depends on the operation.
21:57 pmichaud (and sometimes we'll want  'list'(...)  )  :-)
21:58 bacek pmichaud, I always want 'listproto.new()' for initialize lazy list :)
21:58 pmichaud well, 'list'() will create a lazy list also.
21:58 pmichaud Even new List will create a lazy list.
21:59 bacek pmichaud, indeed.
21:59 pmichaud all lists will be lazy.
21:59 Coke pmichaud: We can bof it, but I doubt more than the usual crowd wills how.
21:59 Coke (doesn't hurt to ask.)
21:59 pmichaud well, I'm more interested in getting "the usual crowd to show" than trying to advertise widely.
21:59 bacek But I need some internal initialization to avoid checks during execution
21:59 pmichaud so I'm not sure I want to call it a BOF.
21:59 pmichaud bacek:  avoid checks?
22:00 Coke pmichaud: belt and suspenders we can do both. put it on the bof schedule and send an email to the list.
22:00 bacek pmichaud, yes. E.g. 'do we have initialized list of generators'?
22:00 Coke (like allison just did)
22:00 pmichaud I'm wondering a good time.
22:00 Coke ... ok. now that THOSE shiny tests are fixed...
22:00 Coke pmichaud: well, I have bofs from 5:30 to 7:30 tonight. after that "OMG HUNGREE"
22:01 pmichaud bacek: I don't think we need to worry about 'initialized list'
22:01 pmichaud coke:  I have the same bofs.  I could ask at the parrot foundation bof.
22:01 bacek pmichaud, another stupid question: how I can 'decorate' RPA methods in List?
22:01 pmichaud ...decorate?
22:01 bacek shift_pmc from List can generate new value.
22:02 pmichaud bacek: I don't understand.
22:03 pmichaud (I understand that shift from List can generate a new value... I don't understand what your question is getting at.)
22:03 bacek pmichaud, List is (evaluated, generators) tuple.
22:03 pmichaud yes.
22:03 dalek r28439 | smash++ | trunk:
22:03 dalek : [release]
22:03 dalek :  * skip some tests failling on win32 for scheduled release
22:03 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28439
22:03 bacek pushing item in List can be in any of this RPA depending of current state.
22:04 pmichaud pushing an item in list simply adds it to the generators tuple.
22:04 bacek Hey. What about premature optimisation!? :)
22:04 pmichaud hmmm?
22:05 bacek if we pushing scalar and list completly evaluated we can push directly to evaluated RPA
22:05 smash Coke: skipped a couple of tests on win32, think you can svn up and mingw32-make test on your machine ?
22:05 pmichaud do we need to do that at the time of the push?  Why not do it when we need it?
22:05 pmichaud i.e., is there an advantage to doing that at the time of the push?
22:06 bacek less work on get_pmc_keyed.
22:06 pmichaud why?
22:06 pmichaud I mean, how is that less work?
22:06 kid51 joined #parrot
22:07 bacek omit dummy 'evaluation'
22:07 pmichaud I still don't understand.
22:09 Tene So, what do I need to know about trying to bring files into parrot that are GPL-licensed?
22:09 bacek Example: if list is ((1,2), (3..5)) pushing element 'a'  will produce ((1,2), (3..5, 'a')). Its fine
22:09 PerlJam bacek: isn't it ((1,2), (3..5), 'a')  ?
22:09 bacek But if list is ((1,2,3,4,5), ()) pushing element 'a' can produce ((1,2,3,4,5,'a'), ()) and list[5] can directly access new element
22:10 pmichaud perljam:  I think he's looking at it as a (evaluated, generators) tuple.
22:10 bacek pmichaud, yes. It is
22:10 PerlJam oh
22:10 pmichaud bacek:  yes, but you're incurring the cost at the time of the push instead of at the time of the []
22:10 dalek r28440 | jkeenan++ | trunk:
22:10 dalek : Applying patch submitted in
22:10 dalek : http://rt.perl.org/rt3/Tic​ket/Display.html?id=55566.  config/init/hints.pm
22:10 pmichaud overall it's the same cost.
22:10 dalek : will no longer die when failing to find a hints file for an unknown OS.  It
22:10 dalek : will assume that this is an OS Parrot has never been configured on before and
22:10 dalek : attempt to DTRT.  Nathan Gray++ for our first attempt on GNU Hurd.
22:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28440
22:10 pmichaud on the push, we have to check that generators is empty and that the item being pushed is scalar
22:11 pmichaud however, if we just put it on generators, then on the [] we incur the cost only if evaluated is empty
22:11 pmichaud sorry, only if evaluated doesn't have enough elements
22:11 pmichaud but either way we have to check that either generated or evaluated is empty at some point
22:11 pmichaud I'd prefer to do it at the point of the [], since that has to do it *anyway*
22:11 pmichaud there's also the very real possibility that we'll never use [] after a push
22:12 bacek pmichaud, I actually check '$!pos' which is 'pos-after-evaluated-end'
22:12 pmichaud which would mean that we will have incurred the overhead of checking the generated array and the pushed value type for no reason
22:13 bacek pmichaud, ok. got the point.
22:13 pmichaud so in this case, it's definitely a "premature optimization."  In fact, it might not even be an optimization.  :-)
22:14 bacek :)
22:14 pmichaud one thing to remember is that array indexing really ought to be far less common than iterating and shift
22:14 bacek Ok.
22:14 pmichaud and push and the like
22:14 pmichaud and even in the case of using [] to look at elements in a List, there are many cases where we'll stop before we ever reach the end.
22:15 bacek It's just too many points in code, where "dummy C style iterating over List" used.
22:15 pmichaud I'm getting rid of those where I can -- I don't know how they got in there.
22:15 pmichaud (I didn't write them.)
22:16 pmichaud I already eliminated them for grep, map, and the like.
22:16 bacek (about []) yes... But almost all of this access invoke 'elements List' before. So it require full evaluation
22:16 pmichaud we can actually compute the number of elements without having to fully evaluate everything.
22:16 bacek pmichaud, they are not in List itself.
22:16 pmichaud oh, you mean in Perl 6 code?
22:16 bacek pmichaud, yes.
22:17 pmichaud I can only point to the synopses that say that 'for' is the proper way to iterate over things.  :-)
22:17 pmichaud if someone wants to write improper code, I'm not necessarily going to optimize for it.  :-)
22:17 PerlJam loop works :)
22:17 pmichaud so does     for @list.kv -> $index, $value { ... }
22:18 pmichaud so there's not much point in even using @list[$index] any more, except on something that has already been generated.
22:18 cotto_yapc joined #parrot
22:18 bacek pmichaud, +1.
22:18 bacek pmichaud, ok. I have almost fully implemented lazy list.
22:18 bacek It requires a bit of polish.
22:18 pmichaud bacek:  okay, I'll take a look at it when it's ready.
22:18 pmichaud Right now I need to rip out the $_ and $! guts and get those working.
22:19 bacek pmichaud, probably tonight. My tonight :)
22:19 PerlJam any of you going to the TPF grant "bof"?  It supposedly starts in 10 min.
22:19 pmichaud but lazy lists is going to be an area where I'm going to be pretty picky that it be done "the right way"  (according to my definition of "right"  :-)
22:19 pmichaud I am.
22:19 TimToady bacek: you might look at what I ended up in pugs/src/perl6/LazyMap.pm for reference; I had to polish it quite a while already...
22:19 bacek Have to go.
22:19 bacek TimToady, thanks. I'll take a look.
22:19 dalek r28441 | jkeenan++ | trunk:
22:19 dalek : Reverting patch applied in r28390.  Todd Olson's bug report is still valid,
22:19 dalek : but first attempt at remedying it didn't completely cleanly.  So am backing it
22:20 dalek : out until after 20080617 release.
22:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28441
22:20 bacek afk # kids->school, self->$dayjob
22:21 kid51 PerlJam: particle will be attending that BOF.  Am currently applying patches, but hope to go out for dinner after that BOF winds up.
22:21 pmichaud immediately after that BOF is the Parrot Foundation BOF.
22:23 pmichaud and time for me to go.
22:23 pmichaud afk # BOF
22:23 dalek r28442 | jkeenan++ | handledarwin:
22:23 dalek : Creating handledarwin in https://svn.perl.org/parrot//branches
22:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28442
22:23 Whiteknight does parrot support something like tied variables? callbacks on variable accesses?
22:24 PerlJam for some reason "bof" always makes me think of Beaker from the Muppets and I want to say "bof bof bof"
22:24 PerlJam But that might be confused with something else :)
22:24 dalek r28443 | jkeenan++ | handledarwin-28441:
22:24 dalek : Tagging trunk at r28441 so that the handledarwin can later be synched to it.
22:24 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28443
22:25 cbrandtbuffalo joined #parrot
22:31 teknomunk joined #parrot
22:31 Infinoid japhb: ping
22:32 nopaste "Infinoid" at 96.238.213.50 pasted "Parrot breaks my opengl!" (123 lines) at http://nopaste.snit.ch/13294
22:32 japhb Infinoid: "Here I am!"
22:33 japhb (Doing my best WildFire impression)
22:33 cjfields joined #parrot
22:33 Infinoid japhb: this is just weird...
22:34 japhb WTF?!?
22:34 japhb OK, does the same truly evil behavior happen with triangle.pir?
22:34 Infinoid seems to, judging from what happened last time, but I'd have to reboot to be sure.
22:35 Infinoid I doubt it's parrots fault, but I am fresh out of Clue
22:35 japhb So a reboot fixes everything, no need to make other system changes?
22:35 Infinoid a reboot fixes glxinfo, that's all.  parrot still doesn't work, and some side effect of it takes everything else out
22:36 japhb OS/Platform?
22:36 Infinoid linux (gentoo)/x86-64
22:36 japhb murf.
22:36 Infinoid after work I'll try futzing around with the xscreensaver opengl hacks, to see whether anything else runs.
22:36 japhb It sounds like somehow DRI is getting hosed.
22:36 Infinoid yeah.  I'd never seen anything like it.
22:37 japhb I do note that that is an RC version of Mesa you're running ... Intel DRI bug in the beta code, perhaps?
22:37 japhb What about glxgears?
22:37 Infinoid possibly.  its actually a git build
22:37 Infinoid same version that made your header parser barf :)
22:38 * japhb chuckles
22:38 Infinoid gonna reboot after work and try glxgears, but first, gotta work
22:38 japhb nodnod
22:38 japhb let me know.
22:38 Infinoid thanks!
22:38 japhb np!
22:56 dalek r28444 | smash++ | trunk:
22:56 dalek : [release]
22:56 dalek :  * make t/codingstd/pdd_format.t happy
22:56 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28444
22:57 kid51 seen particle?
22:57 purl particle was last seen on #parrot 7 hours and 15 minutes ago, saying: sorry, Whiteknight, got disconnected. have anything for chromatic?
23:00 dalek r28445 | smash++ | trunk:
23:00 dalek : * fix white trailling spaces
23:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28445
23:01 japhb .oO( It's the maroon trailing spaces you really need to avoid, of course ... )
23:02 * kid51 at this point is seeing polka dots in his trailing spaces :)
23:02 tetragon joined #parrot
23:02 kid51 Hola tetragon!
23:12 tetragon Hi
23:12 purl salut, tetragon.
23:15 japhb r28390?
23:15 japhb bah
23:15 tetragon Haven't tried that revision yet
23:16 japhb Anyone know if there's a syntax that one of the bots will respond to that will convert a revision number to a diff link?  (A parallel syntax for RT ids would be nice too.)
23:16 japhb tetragon: I was just trying to talk to a bot.  :-)
23:17 Whiteknight i had a bot that did it for a while, but it's unreliable
23:17 tetragon japhb: And your teapot conveyor needs more teapots
23:18 japhb tetragon: The next planned things are a particle effect and a 3D Lissajous animation ...
23:19 diakopter japhb: well, there's http://irclog.perlgeek.de/parrot/today which has the diff links from Rxxx
23:20 tetragon japhb: Make a commit, then edit the URL that gets output here
23:21 japhb The particle effect will test PIR-generated texture buffer contents, and the Lissajous (and possibly particle effect) will test NCI calls with pointers into the middle of a buffer.  I'm trying to make sure my OpenGL wrappers allow PIR to do anything C can do with OpenGL.
23:22 japhb diakopter: So just randomly refer to a rev number, then look at the irclog?  Nice idea ... bound to confuse the heck out of other live IRC users.  Perhaps if I prefix with 'logbait:'
23:22 japhb tetragon: I ended up doing something similar with someone else's commit.  ;-)
23:25 * japhb wonders how close we could get to passing a Turing test if we merged all the smarts from all of the IRC bots in existence ...
23:25 Whiteknight what does make fulltest do that make test doesnt?
23:29 japhb Whiteknight: from a grep, it looks like that answer varies depending on whether you are referring to Parrot or Rakudo.
23:29 Whiteknight japhb, parrot
23:30 japhb Well, if the docs are to be believed, Parrot's 'fulltest' reruns the test under different runcores.
23:30 Whiteknight okay, that explains why it keeps looping
23:33 Maddingue joined #parrot
23:43 kid51 Whiteknight: Yes, that is what 'make fulltest' does.  So it's something we need to do before releases, but not necessarily otherwise.
23:43 bacek_ joined #parrot
23:44 Whiteknight kid51, I'm running it now, doing my part for the release
23:45 kid51 Whiteknight: Very good.
23:47 Whiteknight t/benchmark/benchmarks is killing me! this is taking forever!
23:47 Whiteknight I think it's the gc-debug core, so that's not good
23:53 tetragon kid51: Any idea when the -arch patch can make it in?
23:54 kid51 I seem to remember that that test had failures last month ... but I don't think it's an obstacle to release.
23:55 japhb Has a version number been decided for the release tomorrow?
23:56 japhb It looks like the NEWS updates are very incomplete, but unless there's earth-shattering stuff, I'm guessing 0.6.3 ...
23:56 japhb s/stuff/stuff not yet entered in NEWS/
23:57 magnachef__ joined #parrot

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced