Camelia, the Perl 6 bug

IRC log for #parrot, 2008-04-17

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:02 GeJ FYI: r27007, there was one test removed from the execution, but the plan wasn't changed accordingly.
00:02 nopaste "GeJ" at 202.22.229.65 pasted "fix test failure in t/perl/Parrot_Distribution.t" (14 lines) at http://nopaste.snit.ch/12713
00:03 wknight8111 OMG too many emails!
00:04 wknight8111 at least RT never sends me any viagra spam
00:07 Infinoid no, you can count on the rest of the universe for that
00:07 Infinoid its the constant internet background radiation
00:12 wknight8111 the redshift in my spam folder proves the internet is expanding
00:12 Infinoid hah
00:13 dalek r27011 | rafl++ | trunk:
00:13 dalek : Fix plan for t/perl/Parrot_Distribution.t. GeJ++
00:13 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27011
00:20 GeJ rafl: thanks
00:33 rafl sure thing
00:56 Limbic_Region joined #parrot
01:00 kid51 joined #parrot
01:08 dalek r27012 | infinoid++ | pdd13pbc:
01:08 dalek : Create branch for PDD13 work.
01:08 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27012
01:17 * Coke returns from a porch elimination project.
01:29 rafl joined #parrot
01:31 Infinoid porchicide!
01:31 Debolaz Whiteknight++ # For the parrot wikibook
01:35 wknight8111 yay! somebody saw it!
01:36 wknight8111 i have A LOT more to write though
01:37 Debolaz :)
02:01 guru left #parrot
02:06 dalek r27013 | chromatic++ | trunk:
02:06 dalek : [PMC] Converted AddrRegistry PMC to PDD 17 storage instead of PMC_struct_val.
02:06 dalek : There's one small hack in init() thanks to the complete lack of encapsulation
02:06 dalek : in src/hash.c, but fixing that is a bigger problem than I want to tackle at the
02:06 dalek : moment.
02:06 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27013
02:45 Theory_ joined #parrot
02:49 Theory_ joined #parrot
03:32 AndyA joined #parrot
04:11 liona29 joined #parrot
04:20 markus joined #parrot
04:21 Psyche^ joined #parrot
04:26 tetragon joined #parrot
05:26 tetragon joined #parrot
05:44 Debolaz Yay, first language made.
06:08 Tene Debolaz: what did you make?
06:11 Debolaz Tene: A very minor modification of the default one. :)
06:13 * particle hopes to make a turing-complete language in five minutes
06:13 particle ...lightning talk...
06:32 iblechbot joined #parrot
06:42 Tene particle: should be doable.  depends on how good you are at talking while you're typing, though. :)
06:45 particle indeed. vim macros will likely help
07:21 mj41 hi, t/stm/runtime.t hangs occasionaly (last observed in r27012 in second smoke run) - i386-linux-thread-multi
07:58 purl joined #parrot
08:09 ilbot2 joined #parrot
08:09 Topic for #parrotis now #parrot Parrot 0.6.1 "Bird of Paradise" Released | http://parrotcode.org
08:16 ruoso joined #parrot
09:06 nopaste joined #parrot
09:11 iblechbot joined #parrot
09:29 IllvilJa joined #parrot
10:18 markus joined #parrot
10:21 markus_ joined #parrot
10:27 wknight8111 joined #parrot
10:47 Senaka joined #parrot
11:02 markus joined #parrot
11:26 muixirt joined #parrot
11:26 muixirt2 joined #parrot
11:26 muixirt2 left #parrot
11:31 kid51 joined #parrot
11:31 Senaka left #parrot
12:14 ask_ joined #parrot
12:29 Infinoid g++ has some special handling of "const" which makes unreferenced const global variables disappear.  this causes linker issues later on...
12:31 apple-gunkies joined #parrot
12:34 bphillips joined #parrot
12:48 dalek r27014 | rafl++ | trunk:
12:48 dalek : [examples] make nci/ls.pir work again.
12:48 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27014
12:59 spinclad particle: lightning language: define SKI combinators and application.  :)
13:09 jenny20 joined #parrot
13:23 skids joined #parrot
13:24 gryphon joined #parrot
14:01 dalek r27015 | infinoid++ | trunk:
14:01 dalek : [core] g++ likes to remove unreferenced globals that were declared as "const".
14:01 dalek : This causes linker failures later, because other objects wanted to reference
14:01 dalek : that global.
14:01 dalek : g++ will still emit the global, if it sees an "extern" declaration too.  So, do that.
14:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27015
14:37 DarkWolf84 joined #parrot
14:37 jq- joined #parrot
14:39 rdice joined #parrot
14:40 dalek r27016 | fperrad++ | trunk:
14:40 dalek : [win32]
14:40 dalek : - update gcc version
14:40 dalek : - update Inno Setup version
14:40 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27016
14:42 dalek r27017 | fperrad++ | trunk:
14:42 dalek : [docs]
14:42 dalek : - remove language TAP
14:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27017
14:46 dalek r27018 | coke++ | trunk:
14:46 dalek : [codingstd]
14:46 dalek : Enable another perl coding standard: prohibit function prototypes.
14:46 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27018
14:47 Coke Hurm. I just tried to install Perl::Critic::Bangs with a local perl, and it failed.
15:15 rafl i wonder what to read if i'm interested in writing bindings for c libs
15:15 nopaste joined #parrot
15:17 Infinoid rafl: perhaps start with docs/pdds/draft/pdd16_native_call.pod
15:19 rafl thanks, Infinoid.
15:21 Infinoid http://rt.perl.org/rt3/Tic​ket/Display.html?id=52988 has a nice example of binding libGL
15:22 rafl Infinoid: uhm.. callbacks need to return void? :-/
15:31 Tene rafl: also there are existing SDL bindings in the tree that you can look at for an example.
15:34 Infinoid rafl: I don't actually know.  if that's the case, it does look like a major shortcoming.
15:35 Infinoid there are 7 billion people on this planet, surely one of them will want to use NCI to call qsort() at some point :)
15:36 rafl Tene: seen that. to me it locked like most if it could be autogenerated. thought there might be a new way to do bindings given that is has been quite a while since the sdl stuff was written.
15:36 rafl Infinoid: or any other non-trivial library..
15:37 Tene rafl: autogenerating nci bindings was one of the goals of languages/c99
15:37 peeps[work] joined #parrot
15:38 rafl i see
15:39 Tene Hasn't been done yet, though.  I think one of the gsoc proposals was about that?
15:39 * Tene looks.
15:40 pmichaud I know that someone has been looking into reviewing NCI handling in Parrot
15:40 Tene Yeah, looks like that was one of the applications accepted for sponsorship.
15:41 rafl cool
15:51 rafl how many gsoc slots did tpf get this year?
15:52 pmichaud I don't think it's been officially announced yet
15:54 rafl projects have been notified already. thought someone might know it.
15:54 cognominal strangely there is support for try but not for CATCH in rakudo
15:55 cognominal how  do write a test, so as to catch an error message?
15:55 dalek r27019 | infinoid++ | trunk:
15:55 dalek : [IMCC] Fix up to compile cleanly with g++.  This is part of the patch from
15:55 dalek : RT#52874; Senaka++.
15:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27019
15:55 cognominal pmichaud?
15:55 purl pmichaud is probably http://www.pmichaud.com/ or "Patrick R. Michaud" <mailto:pmichaud@pobox.com>
15:55 Coke rafl: SFAIK, it's still not final.
15:55 cognominal do you have an idea?
15:55 Coke it's merely nigh-official.
15:55 rafl ok
15:57 ambs joined #parrot
16:12 Infinoid anyone here have access to a tru64 box?
16:12 Infinoid would be nice to get #39753 retested.
16:15 Infinoid also, #31806 could use a review.
16:15 cognominal http://rt.perl.org/rt3/Tic​ket/Display.html?id=53016  # can someone review my patch that defined a hash() composer for rakudo
16:17 moritz where's that thing specced?
16:19 Infinoid S04
16:21 moritz my %h = hash( @a, %a, 5, 6);
16:21 moritz I still can't find that syntax
16:22 moritz oh, I'm too dump
16:22 moritz ignore more ;)
16:26 rafl cognominal: s/use v6\K-alpha//
16:27 moritz what's \K?
16:28 rafl a 5.10 regex feature. keeps everything left of it, lik Regex::Keep
16:29 rafl Regexp::Keep, sorry.
16:29 cognominal rafl: anyway this test file would be really useful if I could comment out the tests for errors
16:29 cognominal any idea
16:29 moritz ah
16:30 rafl cognominal: no, sorry.
16:30 cognominal I am working to support my %h = {  @a, %a, 5, 6 }  but that will be ugly
16:31 Infinoid t/rakudo/hash_composer......Statement not terminated properly at line 7, near "my @a = (1"
16:31 cognominal arf, I change the 3 to 1 and erase the comma :(
16:32 pmichaud { @a, %a, 5, 6 } isn't a hash.
16:32 cognominal I misread S04?
16:33 cognominal it should be a hash composer
16:33 pmichaud possibly.  curly braces make a hash only if the closure is empty, or a comma-separated list starting with a pair or a hash
16:33 pmichaud @a isn't a hash, so it's code.
16:33 cognominal ha ok, I patches the example from hash() witout thinking
16:34 cognominal that the very reason the patch will be ugly
16:34 pmichaud it's not so bad.  Whenever we get a block in term position, we just check it to see if it's empty or contains a single infix:<,>  node
16:35 pmichaud or is a pair or hash
16:35 pmichaud if it is, then we convert it to be a call to the hash composer function
16:35 pmichaud instead of a block.
16:35 cognominal anyway if hash() makes in, we already have something clean enough.
16:36 pmichaud I need to look at hash() just a bit more but my first impression is that it's okay
16:36 pmichaud is it using Parrot's 'Pair' class?  Did Rakudo decide to use Parrot Pairs or make its own?
16:36 jq joined #parrot
16:37 cognominal it uses classes/Pair.pir
16:37 cognominal I think
16:37 cognominal hi jq
16:38 jq hey.
16:38 pmichaud ugh, Perl6Pair is using Parrot's Pair.  Bad.
16:38 pmichaud I'll probably refactor that.
16:38 pmichaud unless Parrot's Pair has been improved since the last time I looked at it
16:39 cognominal hum I forgot a test for pairs
16:40 cognominal in the test.
16:42 cognominal anyway the root problem is classes  derived from pmc that have same name as the pmc.
16:43 pmichaud I agree, that's a problem.  I don't know when the Parrot gods will fix that.
16:43 cognominal sure, I am not a god here
16:43 pmichaud me neither, as far as that goes.
16:43 pmichaud afk, lunch
16:44 cognominal for rakudo this is may be the hight priority problem
16:44 cognominal highest...
16:44 purl highest is 5 dan
16:44 cognominal not in go, purl
16:44 purl cognominal: huh?
16:45 pmichaud we can work around the class naming conflicts for now
16:45 pmichaud but yes, it's a high-priority thing
16:45 pmichaud (really gone now)
16:46 sjansen joined #parrot
16:51 spinclad no, purl, highest is 9 dan
16:51 purl okay, spinclad.
16:58 * Coke
16:58 ambs who is playing with readline?
17:18 Coke i didn't do it.
17:22 ambs :)
17:22 ambs probably is particle?
17:36 markus joined #parrot
17:46 askie joined #parrot
17:46 dalek r27020 | ambs++ | trunk:
17:46 dalek : Fix #52818: readline detection failing.
17:46 dalek : Needs some testing.
17:46 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27020
17:49 ambs working under two different linux distros, and darwin with macports
17:56 Theory joined #parrot
17:58 ambs oops... I need to read Damian's book
17:58 dalek r27021 | ambs++ | trunk:
17:58 dalek : Fixed a Coding Standard (cuddled_else)
17:58 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27021
18:16 cognominal on unix if I do make perl6 in the main Makefile, it does not rebuild the perl6.pbc if actions.pm has changed.
18:17 cognominal can't see what is wrong on the makefile though
18:24 PerlJam cognominal: check the timestamps of the files.  Maybe your clock is wonky?
18:25 * Tene is reminded of "Recursive Make Considered Harmful"
18:25 Coke cognominal: the main makefile doesn't depend on anything in languages/perl6
18:26 PerlJam oh, I was thinking of languages/perl6/Makefile as the "main makfile"
18:28 PerlJam "make perl6" just builds the perl6 executable from perl6.pbc right?  (I don't have a parrot source tree handy)
18:30 Tene perl6$(EXE): perl6.pbc
18:30 Tene $(PBC_TO_EXE) perl6.pbc
18:30 Tene PerlJam: what happens if you run 'make perl6.pbc' ?
18:32 Coke tene: yes, see my previous rant in here about all the various makefiles we have.
18:47 Senaka joined #parrot
18:50 Senaka left #parrot
18:55 Coke pmichaud: my simple hack to get perl6 tests running is, of course, not so simple.
18:55 Coke I'll beat my head against it some this evening.
19:04 cognominal this is no big deal, make perl6 justs avoided me to cd back and forth between the main and the perl6 folders
19:04 cognominal s/avoided/would have avoided me/;
19:14 markus Hi. I'm watching development on parrot for some time now and I would like to learn more about parrot. So I thought about looking into #45015 (implementing Rational-PMC), at first. Is using gmplib an acceptable way for implementing this in your opinion?
19:15 ruoso joined #parrot
19:15 Infinoid probably.  We're already using gmp for bigint.pmc, so that may make a good example to steal from.
19:16 Coke joined #parrot
19:16 barney joined #parrot
19:18 markus Yes, I saw that it is used in bigint as well, so I thought it would be the best choice.
19:18 Infinoid yeah, there's a lot less to reinvent that way.
19:19 markus Okay, than I'll probably work on it and tell you, when I'm confident that I can finish it. :-)
19:21 mj41 joined #parrot
19:23 Coke markus++
19:27 Infinoid there's one thing I've always wondered about.  Suppose gmp can't be found.  (Or suppose someone devious like me has disabled it locally because g++ barfs on gmp.h.)  Does that mean we don't get bigints or rats or any of that?
19:28 avar If this becomes a problem gmp could be bundled with parrot
19:28 pmichaud Coke: (perl6 testing)  yeah, I figured it was a bit more complex.  That's always been the brick wall I've run into, which is why I think "start over from scratch" is the most promising option.
19:28 markus As far as I've understood bigint.pmc, you would get an exception when using bigint and gmp is not available
19:37 * ambs 's back
19:42 lichtkind joined #parrot
19:44 Donald joined #parrot
19:50 particle joined #parrot
19:56 dalek Michal Jurosz | RFP: Parrot Needs Better Smoke Reports:
19:56 dalek link: http://www.perlfoundation.org/parrot/index​.cgi?rfp_parrot_needs_better_smoke_reports
19:56 rotty joined #parrot
19:56 shorten dalek's url is at http://xrl.us/bfeof
19:57 Coke win 2
20:03 szbalint irssi++
20:06 Tene win 4
20:09 Infinoid /you wins an internet!
20:09 moritz that's my second internet today
20:09 Infinoid they're hot-swappable
20:10 cognominal pmichaud, do you have any hint about http://rt.perl.org/rt3/Tic​ket/Display.html?id=53018 ?
20:11 cognominal one "just" need to generate the right name space for grammar token.
20:11 cognominal My problem is to understant what "just" entails...
20:12 cognominal I mean the right name space for grammar rules
20:15 dalek will@coleda.com | RFP: Parrot Needs Better Smoke Reports:
20:15 dalek link: http://www.perlfoundation.org/parrot/index​.cgi?rfp_parrot_needs_better_smoke_reports
20:17 Infinoid cool.
20:18 Theory joined #parrot
20:25 gryphon joined #parrot
20:29 MaDWorm joined #parrot
20:41 Theory joined #parrot
20:44 diakopter I suppose dalek could provide the wiki diff link: http://www.perlfoundation.org/parrot/index​.cgi?action=revision_compare&amp;page_name​=rfp_parrot_needs_better_smoke_reports&amp​;mode=source&amp;Button=Compare+Revisions
20:44 moritz that's a bit too long IMHO
20:45 diakopter oh, also, it's wrong :) sorry
20:45 diakopter (Compare Revisions 31 and 31)
20:45 diakopter never mind
20:45 purl Okie dokie
20:48 particle page link plus xrl diff link might be nice
20:48 purl Sorry, I don't know link's email address.
20:48 particle page particle
20:48 Infinoid xrl both of 'em :)
20:55 jan joined #parrot
21:05 spinclad page particle: a folion
21:06 purl Sorry, I don't know particle:'s email address.
21:18 jonathan I may have some time, but likely not much connectivity, over the weekend...got a family wedding to pop over to the UK for. Anyone got anything they'd like me to hack on Rakudo wise?
21:19 Infinoid if you run out of Rakudo tasks (unlikely), there's always the packfile PMC stuff :)
21:19 Tene jonathan: use SDL;
21:21 jonathan Tene: What specifically do you need?
21:21 jonathan Infinoid: I thought I'd delegated that? ;-)
21:21 Tene jonathan: Parrot has SDL libraries, but I can only use them from PIR (or NQP through inline PIR).
21:21 jonathan I did read your mail, though. :-)
21:22 jonathan It looks like a sensible, step-by-step plan.
21:22 Infinoid great, thanks
21:22 jonathan Though test wise, sure you'll need some new ones, but also making sure you havne't got any new ones from the main test suite failing is a good thing too.
21:22 Infinoid I might not have as much time to work on it as I had hoped, this weekend.  At some point, I have to go sort out a friend's mail server.
21:23 ambs night, folks
21:23 jonathan Tene: OK, so you need a way to use PBC libraries from Perl 6.
21:24 Tene jonathan: Yes.
21:24 jonathan I'll have a think about it, don't know that it's especially easy, but I might be able to make something workable.
21:25 Tene I'm not sure if PAST supports that yet.
21:25 jonathan Infinoid: Well, any progress on it is very welcome. :-)
21:25 jonathan Not sure that PAST needs to specifically support it.
21:26 jonathan Are the SDL libraries in PIR OO or just a functional interface?
21:26 Infinoid by "sort out", I mean "replace"... they don't make rubber gloves thick enough for me to touch Microsoft Exchange
21:26 jonathan Ouch.
21:27 Infinoid anyway, at the bare minimum I'll get the .pmc class files in place, and hopefully start fleshing them out
21:27 jonathan OK, that's a good start.
21:27 Tene jonathan: check out examples/sdl/blue_rect.pl
21:29 jonathan Aha.
21:30 jonathan OK, no promises I'll succeed, but I'll have a go.
21:30 Tene Fantastic.
21:31 jonathan Other plans are improve performance of assignments in the absence of type constraints to check.
21:47 Theory_ joined #parrot
22:06 particle Infinoid: make sure you run make testr
22:19 wknight8111 joined #parrot
22:24 Infinoid particle: its in a dev branch, I intend to break *everything* :P
22:25 particle Infinoid: i mean, testr will convert everything to a pbc before it runs it
22:25 particle it's a great way to test the pbc code
22:25 Infinoid cool!
22:35 Theory joined #parrot
22:56 tetragon joined #parrot
22:56 tetragon_ joined #parrot
23:04 skids joined #parrot
23:46 sjansen joined #parrot
23:47 TonyC joined #parrot
23:48 Limbic_Region joined #parrot
23:55 rdice joined #parrot

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

Parrot | source cross referenced