Camelia, the Perl 6 bug

IRC log for #parrot, 2008-07-28

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 dalek r29809 | Whiteknight++ | gsoc_pdd09:
00:02 dalek : [gsoc_pdd09] update to trunk r29808
00:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29809
00:05 bacek joined #parrot
00:07 ewilhelm joined #parrot
00:11 dalek r29810 | coke++ | trunk:
00:11 dalek : [tcl] http://code.google.com/p/p​artcl/issues/detail?id=29  --  first pass at [lsort -command], with a test.
00:11 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29810
00:21 dalek r29811 | Whiteknight++ | gsoc_pdd09:
00:21 dalek : [gsoc_pdd09] few changes
00:21 dalek : * Better maintenance of pool->num_free_objects in my allocators/deallocators
00:21 dalek : * remove some unused cruft from smallobjects.c
00:21 dalek : * Maybe some other stuff I don't remember
00:21 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29811
00:46 dalek r29812 | coke++ | trunk:
00:46 dalek : [tcl] don't catch and throw our own error, let the 'raw' error percolate out.
00:46 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29812
00:46 dalek r29813 | coke++ | trunk:
00:46 dalek : [tcl] adhere to the coding standards so we don't break the build. Sorry about that.
00:46 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29813
00:51 Schwern joined #parrot
00:52 DietCoke eek, it's a Schwern.
00:52 dalek r29814 | Whiteknight++ | gsoc_pdd09:
00:52 dalek : [gsoc_pdd09] add in a skip factor, so GC doesn't run when there is probably nothing to collect.
00:52 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29814
00:55 Schwern Rar!
01:10 dalek r29815 | Whiteknight++ | gsoc_pdd09:
01:10 dalek : [gsoc_pdd09] remove a stupid comment that I added earlier (I figured out what was going on, finally) and fix the return value in Parrot_dod_trace_root
01:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29815
01:22 Andy joined #parrot
01:23 dalek r29816 | Whiteknight++ | gsoc_pdd09:
01:23 dalek : [gsoc_pdd09] double-check (maybe unnecessary) that I am clearing out the pobj flags on allocation.
01:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29816
01:28 kid51 joined #parrot
01:38 kid51 ping DietCoke
01:38 purl I can't find DietCoke in the DNS.
01:38 kid51 DietCoke ping
01:45 Whiteknight ping www.dietcoke.com
01:45 purl /sbin/ping returned an error.
01:45 Tene purl: coke?
01:45 purl i think coke is mailto:will@coleda.com or just a figurehead. or http://coke-floats.blogspot.com/ or DietCoke or a pest.
01:45 Tene ping coleda.com
01:45 purl /sbin/ping returned an error.
01:46 Tene ping pleasedieinafire.net
01:46 purl /sbin/ping returned an error.
01:46 kid51 ping thenceforward.net
01:46 purl /sbin/ping returned an error.
01:46 kid51 purl your ping is broken
01:46 purl ...but purl's ping is ipv6 compliant...
01:46 Tene ping www.ipv6.org
01:46 purl /sbin/ping returned an error.
01:46 kid51 That doesn't matter, purl.
01:46 purl kid51: what?
01:46 kid51 ping www.google.com
01:46 purl /sbin/ping returned an error.
01:46 kid51 Q.E.D.
01:46 Whiteknight ping irc.perl.org
01:46 purl /sbin/ping returned an error.
01:47 Whiteknight ping 127.0.0.1
01:47 purl /sbin/ping returned an error.
01:47 Whiteknight IS B0RKED
02:09 dalek r29817 | jkeenan++ | parallel:
02:09 dalek : [configure] Fine-tuning of tests for configuration steps.  Clarify POD.
02:09 dalek : Provide inline comments functioning as subheads where tests previously found
02:09 dalek : in multiple files have been moved into a single file.
02:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29817
02:14 TiMBuS joined #parrot
02:23 dalek r29818 | jkeenan++ | parallel:
02:23 dalek : [configure] Fine-tuning of tests for configuration steps.  Clarify POD.
02:23 dalek : Provide inline comments functioning as subheads where tests previously found
02:23 dalek : in multiple files have been moved into a single file.
02:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29818
02:31 cesar joined #parrot
02:34 DietCoke yes?
02:34 DietCoke (tene, Whiteknight, kid51) pong
02:35 Tene ENOCLUE
02:41 Infinoid kid51: 1 pings transmitted, 1 received, 0% packet loss, time 56 minutes
02:43 DietCoke you funny, doctor jones.
02:44 kid51 DietCoke:  If you still have that Win32 box, can you do a checkout of the 'parallel' branch and see how the t/steps/*.t tests compare with trunk?
02:45 kid51 Am close to merging that branch into trunk.
02:45 kid51 Re:  http://rt.perl.org/rt3/Tic​ket/Display.html?id=56716
02:47 * bacek wonders what 'regex' in STD.pm means...
02:48 DietCoke kid51: Sure; won't be able to get to it probably until after work tomorrow; you'll probably want to hit the list.
02:49 teknomunk__ joined #parrot
02:51 kid51 Thanks, and now ...
02:51 * kid51 must sleep
02:51 purl $kid51->sleep(8 * 3600);
02:53 Theory joined #parrot
02:58 Andy joined #parrot
03:01 dalek r29819 | coke++ | trunk:
03:01 dalek : [tcl] implement [lreverse], and TODO all the remaining failing spec tests from that file.
03:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29819
03:13 dalek r29820 | coke++ | trunk:
03:13 dalek : [tcl] Fix a bug in [lsort -integer] than modified the values we were sorting (converted them to their decimal integer form instead of leaving them unmodified)
03:13 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29820
03:41 dalek r29821 | coke++ | trunk:
03:41 dalek : [tcl] TODO more tests from the spec test cmdMZ.test
03:41 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29821
03:43 bacek joined #parrot
03:57 ilbot2 joined #parrot
03:57 Topic for #parrotis now #parrot Parrot 0.6.4 "St. Vincent Amazon" Released | http://parrotcode.org/ | 15 new, 648 open, 80 stalled tix | logged in http://irclog.perlgeek.de/parrot/today
04:06 Theory joined #parrot
04:52 verve joined #parrot
05:34 Psyche^ joined #parrot
05:35 Theory joined #parrot
05:35 cesar left #parrot
06:07 Ademan joined #parrot
06:21 cotto_home I love how wikipedia can take me from register allocators to sudoku in 3 short hours.
06:23 moritz cotto_home: have you seen that paper on link path lengths on the wikipedia?
06:24 moritz there's a strongly connected set of articles, and you get from one to all other with 4 clicks most of the time ;-)
06:25 cotto_home no, and hopefully I won't before I go to sleep &
06:26 moritz hehe ;-)
06:26 moritz good night ;-)
06:27 ewilhelm pmichaud, ping
06:30 uniejo joined #parrot
06:33 Theory joined #parrot
06:51 iblechbot joined #parrot
07:22 barney joined #parrot
07:44 Schwern joined #parrot
07:49 dalek r29823 | fperrad++ | trunk:
07:49 dalek : [Lua] OpenGL
07:49 dalek : - handles exception in GLUT callbacks
07:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29823
07:52 dalek r29824 | fperrad++ | trunk:
07:52 dalek : [Lua] OpenGL
07:52 dalek : - add gl.VertexPointer()
07:52 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29824
08:05 dalek r29825 | fperrad++ | libs4php:
08:05 dalek : all merged in trunk
08:05 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29825
08:14 donaldh joined #parrot
08:22 contingencyplan joined #parrot
08:24 Debolaz joined #parrot
08:28 omega joined #parrot
08:35 masak joined #parrot
09:09 Whiteknight joined #parrot
10:19 tuxdna joined #parrot
10:32 timbunce joined #parrot
10:33 timbunce Parrot is "Is register based, with continuations, closures, and JIT compiling" ... what else can I add (especially buzzwords)?
10:34 timbunce I can't seem to find any (relatively) recent overview of parrot.
10:37 moritz "optimized for dynamic languages"
10:49 barney timbunce: includes a parser generator, PGE
10:50 timbunce moritz: and barney: thanks. Anything else that would contrast it with JVM ?
11:06 timbunce Is it fair to say that Parrot already supports all the features being explored as subprojects of the DaVinciMachine JVM project: http://openjdk.java.net/pro​jects/mlvm/subprojects.html
11:10 bacek joined #parrot
11:12 moritz I think so
11:13 barney I hope so. US-Westcoast Parrot hackers could check out http://openjdk.java.net/project​s/mlvm/jvmlangsummit/index.html
11:14 masak I can't open a file for reading in rakudo. surely that used to work at some point? could someone point me to a working way to do it?
11:16 moritz masak: Auzon tried yesterday, and failed. I think we need a ticket (or better multiple ones)
11:16 masak I'll see what I can do.
11:16 moritz especially since >500 regex tests depend on it
11:17 masak ouch
11:17 moritz there's one plain text file with one test on each line, and it contains 530 tests, and rakudo can't read them :/
11:18 mj41 timbunce: http://perl6.cz/wiki/Perl_6_and_Parrot_links probably you can find something in Jonathan's talks
11:21 jonathan Open to read is meant to work - my $fh = open("filename", :r);
11:21 jonathan No doubt we need tests for this, though.
11:51 bacek joined #parrot
11:53 dalek r29826 | bernhard++ | trunk:
11:53 dalek : [Pipp] Two more tests for double quoted string.
11:53 dalek : Skip the test that gets trapped in an infinite loop.
11:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29826
11:56 ruoso joined #parrot
12:20 kj joined #parrot
12:21 iblechbot joined #parrot
13:40 paco joined #parrot
13:59 bacek joined #parrot
14:25 daxelrod joined #parrot
14:26 gryphon__ joined #parrot
14:26 dalek r29827 | coke++ | trunk:
14:26 dalek : [testing] Enable simple parallel make test for TH3.
14:26 dalek : Patch courtesy Eric Wilhelm (RT #57358)
14:26 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29827
14:28 Infinoid ewilhelm++
14:31 rdice joined #parrot
14:46 rdice_ joined #parrot
14:48 bacek joined #parrot
14:59 jrockway joined #parrot
15:19 Auzon masak: regarding reading a file, slurp works if you just want the whole file as one string.
15:21 masak Auzon: that's good news. thank you!
15:22 Auzon opening and using $fh.readline works except you can't detect EOF.
15:22 moritz uhm, my $x = slurp('ROADMAP') gives me a "Unable to open file"
15:22 moritz what am I doing wrong?
15:22 masak Auzon: no, I'm pretty sure open() does not work
15:24 Auzon I had both working yesterday :-/
15:24 masak rakudo is a fast-moving project :P
15:24 Auzon ../../parrot perl6.pbc -e'say slurp "README"' # works for me
15:25 jhorwitz joined #parrot
15:25 masak works here too
15:26 moritz it works here too - wtf?
15:27 Auzon so does: ../../parrot perl6.pbc -e'my $fh = open "README"; say $fh.readline'
15:27 Auzon readline($fh) does not work
15:27 masak Auzon: the former works here too
15:28 ambs joined #parrot
15:28 masak but only with both statements, not with just the first
15:28 Auzon So we're really close to having basic IO working.
15:28 * moritz wonders why we don't have a Str.split(Str) working
15:33 jonathan I'm kinda tied up today, but I am doing Rakudo hacking tomorrw; if there are I/O issues, please file tickets and I will look at them then.
15:34 moritz Auzon++ filed one
15:34 Auzon I tried to pick the most useful one, but it's probably not the easiest
15:41 ambs left #parrot
15:41 jonathan OK, great, will try and sort it out tomorrow.
15:44 moritz greak
15:44 moritz *
15:44 moritz *great
15:44 moritz hard to type, isn't it ... ;-)
15:45 jonathan eys.
15:48 dalek r29828 | bernhard++ | trunk:
15:48 dalek : [Pipp] Handle special case: dollar without variable identifier
15:48 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29828
15:52 donaldh Oh dear. svn latest doesn't build for me.
15:53 moritz which platform?
15:53 purl I'm running on OS/2 on an Atari, can you help?
15:53 donaldh Cygwin.
15:54 donaldh It's a Makefile issue. Just looking into it.
15:54 donaldh It must affect more people than me.
15:54 moritz oh, fails for me as well
15:54 moritz tools/build/ops2pm.pl: Could not mkdir /home/moritz/src/parrot/lib/Parrot/OpLib: File exists!
15:55 donaldh Not the problem I'm seeing. undefined reference to `_Parrot_set_config_hash_internal'
15:55 donaldh src/null_config.c has a dependency on libparrot but appears at the end of the link line.
15:57 moritz after a (second?) make realclean it works
15:57 donaldh weird.
15:57 donaldh not for me.
15:58 moritz ah, I was having a completely differnt error
15:58 donaldh but make realclean is breaking for me (near the end I think)
15:58 cotto_home ewilhelm++ #parallel testing
16:10 tuxdna left #parrot
16:38 particle joined #parrot
16:42 nopaste "donaldh" at 144.254.89.228 pasted "patch for config/gen/makefiles/root.in" (26 lines) at http://nopaste.snit.ch/13680
16:42 donaldh That seems to fix it for me.
16:46 Infinoid donaldh: does that fix your build, or the realclean too?  (kinda scary that realclean would depend on a properly built parrot.)
16:47 donaldh That fixes the build.
16:52 Infinoid ok, testing on linux
16:56 nopaste "donaldh" at 144.254.89.228 pasted "patch updated to fix realclean" (35 lines) at http://nopaste.snit.ch/13681
16:57 donaldh Infinoid: that ^ fixes realclean too.
16:57 donaldh missing $(EXE)
16:57 donaldh (for Cygwin, Win32)
17:01 Infinoid makes sense
17:01 donaldh Infinoid: realclean doesn't depend on a properly built parrot.
17:02 Infinoid great.  clean rules should be as failsafe as possible :)
17:03 donaldh do you want me to raise a ticket?
17:04 dalek r29829 | infinoid++ | trunk:
17:04 dalek : [config] Apply patch from donaldh++ to fix cygwin builds
17:04 dalek : * undefined reference to `_Parrot_set_config_hash_internal'
17:04 dalek : (src/null_config.c has a dependency on libparrot but appears at the end of the link line.)
17:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29829
17:04 donaldh afk
17:06 Infinoid donaldh: thanks, all applied
17:07 dalek r29830 | infinoid++ | trunk:
17:07 dalek : [config] Apply patch from donaldh++ to fix realclean on cygwin.
17:07 dalek : (was missing $(EXE) suffix)
17:07 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29830
17:20 Theory joined #parrot
17:33 cjfields joined #parrot
17:41 Limbic_Region joined #parrot
17:56 ewilhelm where is runstep() defined for Parrot::Configure::Step?
18:00 donaldh joined #parrot
18:00 iblechbot joined #parrot
18:03 ewilhelm particle, ping
18:19 ewilhelm ah, config/${stepname}.pm
18:32 Ron joined #parrot
18:32 Ron_ joined #parrot
19:05 ewilhelm left #parrot
19:29 cotto_work joined #parrot
19:58 Schwern joined #parrot
20:22 gmansi joined #parrot
20:24 AndyA joined #parrot
20:30 Auzon pge?
20:30 purl pge is the Parser Grammar Engine
20:31 Whiteknight joined #parrot
20:38 cotto_work particle, http://www.wikiupload.com/images/gay.php
20:38 cotto_work it could happen to you
20:39 particle and oldie, but a goodie :)
20:39 cotto_work it's nice to know that I haven't seen everything on the tubes yet
20:39 cotto_work btw, no meeting cancellation so far
20:39 cotto_work I think it might actually happen this time
20:40 particle :)
20:40 particle perhaps because i harassed hank at oscon in person
20:40 cotto_work Anandeep said he talked with you too
20:41 particle i talked to garrett for a while, too
20:50 dalek r29831 | Whiteknight++ | gsoc_pdd09:
20:50 dalek : [gsoc_pdd09] small whitespace fix and rearranging some stuff for debugging
20:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29831
20:53 tewk joined #parrot
20:56 dalek r29832 | Whiteknight++ | gsoc_pdd09:
20:56 dalek : [gsoc_pdd09] update to trunk r29830
20:56 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29832
21:02 cotto-work joined #parrot
21:02 teknomunk joined #parrot
21:18 rurban joined #parrot
21:19 rurban How can I check the definedness of a sub in PIR?
21:20 Infinoid if you can assign it to a pmc variable, you could use the defined op
21:21 Tene rurban: find_sub on the appropriate namespace pmc
21:21 Tene is one way
21:21 Tene or just find_name
21:21 donaldh_ joined #parrot
21:22 slightlyoff joined #parrot
21:23 timbunce joined #parrot
21:24 rurban Great! I think I solved now the _config problem with already linked conf_hash
21:35 donaldh Infinoid++ # for applying the root.in patches. Verified on Cygwin.
21:35 rurban :)
21:36 rurban donaldh: You mean my massive halfway 56554-make-install-lang.patch ?
21:37 moritz rurban: no, donaldh's own patch to fix the build, I think
21:37 rurban Ah, ok.
21:37 Infinoid rurban: http://nopaste.snit.ch/13681
21:37 donaldh rurban: no, moritz: yes. You confused me for a moment there.
21:38 rurban I see, this failed for me also all the time, thanks!
21:39 rurban And I wondered what the LINKFLAGS have to do after the ALL_PARROT_LIBS. After the libs they are not effective
21:39 rurban @rpath_blib@ $(ALL_PARROT_LIBS) $(LINKFLAGS)
21:40 donaldh I think that might be for platform specifics.
21:41 rurban I suggested once to do @rpath_blib@ $(LINKFLAGS) $(ALL_PARROT_LIBS). But maybe those flags are only for $(LINK_DYNAMIC)
21:41 donaldh but then again, maybe not.
21:44 donaldh Another annoying feature is that perl Configure.pl --debugging=1 doesn't work on Cygwin because the installed perl is configured for stripped binaries.
21:44 rurban Well, you have to use perl5.10.0d Configure.pl --debugging=1 then.
21:45 dalek r29833 | Whiteknight++ | gsoc_pdd09:
21:45 dalek : [gsoc_pdd09] A few fixes:
21:45 dalek : * Simplify my flag setting/retrieving code, which appears to have been in error
21:45 dalek : * Improve compliance (and my personal understanding) with the lazy flag
21:45 dalek : * Improve recognition of objects needing timely destruction
21:45 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29833
21:45 rurban But then you have to build your own perl's
21:45 rurban Shoudl I provide a set of cygwin packages for some perl's. I have about 20 or so
21:46 donaldh I will eventually get around to patching Configure.pl to prefer my command-line options over the builtin perl defaults.
21:46 rurban Configure.pl accepts no -D: or -A: as perl5 does.
21:46 rurban Shouldn't that be a TODO?
21:47 donaldh yes, probably.
21:47 purl Really? Probably? Are you Certain it's not certain? Are you sure it's unsure? I think you need to look harder.
21:47 rurban prul, We've looked hard enough
21:47 donaldh probably as in certainly, but with less commitment.
21:48 rurban -A:ccflags=-g
21:48 donaldh Ah, no, that's not the problem.
21:49 donaldh I need to remove -s from the link line.
21:49 rurban You have an older perl. I removed that with 5.10 I believe already.
21:49 bacek joined #parrot
21:50 rurban You can always change your Config_heavy.pl to your liking. esp the -s strip flag is annoying.
21:50 rurban That would be easiest
21:51 donaldh Agreed. I thought it would be nice if Configure.pl actually honoured the command-line options tho.
21:51 rurban I'm not in a mood for a massive TODO Configure flags ticket now. Maybe tommorrow.
21:52 donaldh I'll take a look at it tomorrow. It's too late to start tonight.
21:52 rurban I'll leave also
21:52 Infinoid less commitment = fewer commits... :)
21:53 Tene too many commits = you get committed
21:53 rurban no commit bit, no commitment at all :)
21:54 donaldh commitment != time
21:55 donaldh :D
21:57 Infinoid good point.  we should issue a generous helping of time along with every commit bit
21:57 cotto_work 28 hour days for everyone with a commit bit!
21:57 Infinoid sadly, the POSIX standard for mktime() doesn't do what it ought to do
22:04 cotto-work joined #parrot
22:24 Limbic_Region joined #parrot
22:24 Schwern joined #parrot
22:25 donaldh S11 (modules) suggests that an external language can be imported with 'use Something:from<perl5>'.
22:26 donaldh I'm not sure how much sense that would make if Something has been compiled to bytecode.
22:27 moritz non, because perl 5 can't really be compiled to byte code ATM
22:27 donaldh Is there anything that specifies 'this bytecode is language X'.
22:27 moritz s/non/none/
22:27 donaldh moritz: perl 5 is a bad example (from S11)
22:27 moritz donaldh: the compiler knows which language it has to except from the use statement
22:28 donaldh But the installed libraries might be in bytecode so there's no longer a language to expect?
22:29 moritz oh yes, the languages make very different assumptions about bytecode
22:29 moritz s/bytecode/types/
22:29 moritz I should really go to bed
22:29 donaldh :)
22:29 donaldh also. I'm thinking on empty.
22:30 moritz btw you should never assume that you actually have a compiled copy of a module. If you have one, you're allowed to use it
22:30 donaldh Is that the philosophy?
22:32 donaldh So I'd start by wanting to 'use ICAN:from<LOLCODE>' and Perl will search the LOLCODE searchpath for bytecode followed by source?
22:33 donaldh I may have missed this in the PDDs. Is it covered?
22:33 moritz I don't know if it's actually specced
22:34 moritz but it fits the usual philosphy
22:34 TiMBuS joined #parrot
22:34 moritz that you can edit a module and run a script that uses it without having to recompile the module separately
22:38 jonathan I think it'll end up being a kinda "compilation cache".
22:38 jonathan "Has this module changed since we last compiled it? No? OK, use the bytecode version."
22:39 donaldh So will Perl 6 know how to search for modules from external languages, or will it need to delegate to parrot for that?
22:39 moritz I think that's what python does
22:40 moritz donaldh: it makes sense to delegate that to the particular implementation
22:40 moritz donaldh: otherwise all HLLs need to know how to find modules for each other HLL
22:40 donaldh Indeed.
22:40 * moritz -> bed
22:41 * donaldh seems to be asking leading or obvious questions
22:43 donaldh goodnight.
22:48 kid51 joined #parrot
23:00 DietCoke joined #parrot
23:00 DietCoke Tene: how goes?
23:06 Tene >.>
23:54 dalek r29834 | jkeenan++ | parallel:
23:54 dalek : [configure] Fine-tuning of tests for configuration steps.  Clarify POD.
23:54 dalek : Provide inline comments functioning as subheads where tests previously found
23:54 dalek : in multiple files have been moved into a single file.
23:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29834
23:59 kid51 DietCoke:  ping

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

Parrot | source cross referenced