Camelia, the Perl 6 bug

IRC log for #parrot, 2009-08-24

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 Whiteknight joined #parrot
00:06 dalek lua: 6699f63 | fperrad++ | src/pmc/lua (4 files):
00:06 dalek lua: remove useless need_ext
00:06 dalek lua: review: http://github.com/fperrad/lua/commit/66​99f63600418300db7323171e006b95c82c78a5
00:12 Coke is pcc_rewiring dead (for the other pcc_ branch?)
00:14 theory joined #parrot
00:15 Coke incoming
00:15 purl incoming is https://pause.perl.org/incoming/
00:15 michel joined #parrot
00:16 dalek partcl: r606 | coke++ | trunk/src/macros.pir:
00:16 dalek partcl: update include to avoid full path, remove unused interpinfo constants
00:16 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=606
00:20 Coke ... dalek?
00:21 dalek partcl: r607 | coke++ | trunk/runtime/builtin/ (50 files):
00:21 dalek partcl: Eliminate .prof macro - it's already obsoleted by cotto++ and
00:21 dalek partcl: chromatic++'s work on the profiling runcore.
00:21 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=607
00:21 dalek partcl: r608 | coke++ | trunk/runtime/builtin/ (3 files):
00:21 dalek partcl: Cleanup PIR
00:21 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=608
00:21 dalek partcl: r609 | coke++ | trunk/src/class/tclfloat.pir:
00:21 dalek partcl: Convert part of TclFloat into PIR.
00:21 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=609
00:21 dalek partcl: r610 | coke++ | trunk/ (2 files):
00:21 dalek partcl: Throw tcl-specific error on mod-by-zero.
00:21 dalek partcl: This avoids a segfault when running "expr 0%0" interactively.
00:21 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=610
00:21 dalek partcl: r611 | coke++ | trunk/runtime/ (5 files):
00:21 dalek partcl: Avoid root_new, which seems to take about 4 times as long as new
00:21 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=611
00:21 dalek partcl: r612 | coke++ | trunk/runtime/tcllib.pir:
00:21 dalek partcl: don't create explicit iterators
00:21 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=612
00:21 dalek partcl: r613 | coke++ | trunk/ (34 files):
00:22 dalek partcl: convert root_new back to new
00:22 purl I don't know how to convert root_new back to new.
00:22 dalek partcl: Initial run with profiler showed root_new op as taking about 4x as long as new.
00:22 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=613
00:22 dalek partcl: r614 | coke++ | trunk/src/macros.pir:
00:22 dalek partcl: Add .list() macro to initialize an empty list.
00:22 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=614
00:22 dalek partcl: r615 | coke++ | trunk/runtime/builtin/lrange.pir:
00:22 dalek partcl: cleanup PIR
00:22 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=615
00:22 dalek partcl: r616 | coke++ | trunk/runtime/builtin/ (50 files):
00:22 dalek partcl: add .argc() macro
00:22 dalek partcl: Be more consistent about parameter naming, update to use .argc() most places
00:22 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=616
00:22 dalek partcl: r617 | coke++ | trunk/runtime/builtin/l (2 files):
00:22 dalek partcl: cleanup PIR
00:22 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=617
00:24 kid51 joined #parrot
00:28 mikehh cOKE: TIED UP WITH SOMETHING ELSE AT THE MOMENT - WILL GIVE THAT A TEST IN A COUPLE OF HOURS
00:29 mikehh dammit
00:29 mikehh Caps Lock inverted
00:30 mikehh Coke: will test partcl in a couple of hours
01:33 mikehh All tests PASS (pre/post-config, smolder, nqp_test, fulltest) at r40744 - Ubuntu 9.04 amd64 (g++)
01:44 mikehh rakudo (c4c67da) builds on parrot r40744 - make test / make spectest (up to 28054) PASS - Ubuntu 9.04 amd64 (g++)
01:47 mikehh partcl r617 builds on parrot r40744 - make test same 6 tests FAIL but all subtests PASS - spectest is failing with - wrong # args: should be "lsearch ?options? list pattern"
01:47 dalek parrot: r40745 | allison++ | branches/pcc_arg_unify/src/call/pcc.c:
01:47 dalek parrot: [pcc] Build string signatures for calls initiated from ops. (Next step is
01:47 dalek parrot: to stop storing the constant flag arrays in the call signature object.)
01:47 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40745/
01:48 mikehh ok - that's me for a while - need sleep :-}
01:51 dalek parrot: r40746 | allison++ | branches/pcc_arg_unify/src/pmc/multisub.pmc:
01:51 dalek parrot: [pcc] Change MultiSubs to pull arguments from the current call signature
01:51 dalek parrot: object instead of from registers.
01:51 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40746/
01:52 kid51 joined #parrot
01:53 * kid51 builds 5.10.1
01:54 Coke (holy crap, someone is running the partcl spectest)
01:54 Coke mikehh++
01:55 * allison does the dance of joy, last two commits make pcc branch pass an additional 2500 core tests
01:57 Andy joined #parrot
01:58 dalek parrot: r40747 | allison++ | branches/pcc_arg_unify/src/multidispatch.c:
01:58 dalek parrot: [pcc] Better error checking for null signature string in call signature
01:58 dalek parrot: object.
01:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40747/
01:58 Coke allison: is pcc_rewiring killable?
01:59 allison Coke: yes
01:59 allison Coke: I kept it around for a while in case some changes didn't apply cleanly in the new branch, but we're safely past that point now
02:01 Coke want me to kill it?
02:05 dalek partcl: r618 | coke++ | trunk/runtime/builtin/lsearch.pir:
02:05 dalek partcl: fix [lsearch] so we check argc /after/ args processing. should fix spectest.
02:05 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=618
02:05 Coke mikehh++
02:06 Tene Coke is The Deleter.
02:06 Coke rm -rf ~tene
02:06 Tene removing deprecations, deleting old branches...
02:07 Coke deleting things is easier than creating things.
02:07 * Tene is reminded of PDS 08
02:07 Coke Coke didn't have anything to do with that. =-)
02:08 Tene There was at least one conversation when someone suggested assigning removing deprecations to you, iirc
02:09 Coke that would explain why I suddenly had tickets assigned to me.
02:09 Coke hey, new memory panics running 'make spectest'. (*&#>.
02:10 * Tene AFK flying.
02:13 * Coke grumbles at all the new memory panics.
02:13 * Coke blames whiteknight.
02:13 Coke you know, not all the commits I just made. =-)
02:19 allison Coke: sure (deleting pcc_rewiring)
02:23 Coke done
02:25 dalek parrot: r40748 | coke++ | branches/pcc_rewiring:
02:25 dalek parrot: obsolete; see pcc_args_unify
02:25 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40748/
02:27 allison Coke: you rock!
02:27 purl Dis is the drum
02:27 treed Tene: Is Cardinal's quote_expression.pir ganked from somewhere in particular?
02:27 treed (And is it edited?)
02:32 cotto joined #parrot
02:41 beta joined #parrot
02:42 janus joined #parrot
02:43 kay joined #parrot
02:45 tyler joined #parrot
02:59 kay left #parrot
03:00 Coke did recent commits (last week or so) change memory usage dramatically?
03:04 Infinoid Austin mentioned something about that this morning
03:05 Austin See my recent email to the list. I'm seeing consume-everything-and-die behavior in my compiler.
03:06 dalek partcl: r619 | coke++ | wiki/SpecTestStatus.wiki:
03:06 dalek partcl: With recent updates to parrot/partcl, we have an additional 9 tests failing due
03:06 dalek partcl: to memory panics. :(
03:06 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=619
03:12 Coke git-svn question: how can I say "show me what revision of the remote repository we're updated to locally?"
03:19 Coke (git svn info README works, I suppose)
03:34 Andy joined #parrot
03:41 cotto seen darbelo
03:41 purl darbelo was last seen on #parrot 1 days, 8 hours, 32 minutes and 49 seconds ago, saying: mikehh: tools/dev/gen_makefile.pl IIRC  [Aug 22 19:02:15 2009]
04:06 Infinoid Coke: I normally look at the URL at the bottom of "git show git-svn"
04:07 Infinoid hmm, I seem to have forgotten to go to bed.  night
04:15 dukeleto joined #parrot
04:19 cotto msg darbelo I'll be glad to continue working with you on the decnum dynpmcs if you're interested.  If you have some time later this week, I'd like to do a realtime code review.
04:19 purl Message for darbelo stored.
04:22 japhb Can dalek track gitorious?  And if so, how do we set that up?
04:23 kyle_l5l joined #parrot
04:24 cotto japhb, bug Infinoid (or get the source from github and write a plugin yourself)
04:25 japhb cotto, OK, thx
04:25 japhb Infinoid, still around?
04:28 dukeleto 'ello
04:29 TiMBuS joined #parrot
04:35 dalek tracwiki: v7 | japhb++ | ModuleEcosystem
04:35 dalek tracwiki: Add link to Gitorious project page at top
04:35 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Mod​uleEcosystem?version=7&action=diff
04:48 beta cotto: I don't think I am going to pursue parrot on our robot at this time. I will be back eventually though
04:49 beta Scala seems to be fitting my needs fine atm, and my wrapper works flawlessly since they all run on the JVM
04:56 cotto beta, I find that completely understandable.
04:57 cotto I do wish you luck with your robotic endeavors and will be glad when you decide that Parrot's a good fit for your project.
05:05 s1n what are the proper arguments to find_lex? i keep getting opcode find_lex_s_sc not found
05:06 Andy joined #parrot
05:10 Austin s1n: $P1 = find_lex '$foo'
05:12 cotto s1n, iirc only PMCs can be lexicals, not strings, ints or nums.
05:12 Austin Here's an example: http://nopaste.snit.ch/17654
05:12 s1n Austin: yes, i just found that myself :)
05:12 s1n thanks to the both of you guys
05:15 cotto np
05:17 chromatic joined #parrot
05:44 kyle_l5l where does the code for src/pmc/object.c's Parrot_Object_freeze() come from?  object.pmc never mentions a freeze().
05:48 uniejo joined #parrot
05:50 cotto kyle_l5l, Object is one of a couple special PMCs where VTABLE functions that are processed by different code from normal PMCs.
05:51 cotto The Perl code is in lib/Parrot/Pmc2c/Object.pm .
05:52 kyle_l5l cotto, ok, thanks
05:53 cotto an Object is an instantiation of a PIR-level Class, so that code allows VTABLE functions to dispatch to PIR.
06:00 theory joined #parrot
06:07 beta cotto: thanks
06:07 NotFound Where is Close respository?
06:08 NotFound purl: close?
06:08 purl it has been said that close is close file (or pipe or socket) handle or bad on sockets, use shutdown() instead or horseshoes or grenades or nuclear weapons or http://code.google.com/p/close/
06:28 mj41 joined #parrot
06:30 japhb http://use.perl.org/~geoffrey/journal/39516
06:31 allison Bryan O'Sullivan:  "git is popular because it makes everything so hard that you feel smart if you can do the simplest of things with it."
06:31 japhb allison, :-)
06:31 dukeleto so I notice that parrot has sinh but not asinh. is there any interest in adding inverse hyperbolic functions?
06:32 allison dukeleto: most of those obscure math ops are being moved out to dynop libraries
06:33 allison dukeleto: so, not as a core op, but maybe as a dynop
06:34 dukeleto allison: i agree that it should be a dynop
06:36 dukeleto allison: should I ask the list about it?
06:37 allison dukeleto: sure, see if anyone has a use for it
06:38 dukeleto allison: Ok. they are defined in <math.h> in C, so they are seemingly common
06:40 allison common for C and common for Parrot aren't the same
06:41 allison and, dynops are more equivalent to C (you have to load the lib)
06:41 allison core ops are like making it a primary keyword of the language
06:43 dukeleto allison: looks like Rakudo has an "asinh" the spec test suite
06:43 allison Rakudo can load a dynoplib
06:44 allison but yes, sounds like it would be a useful addition
06:44 dukeleto allison: I agree :)
06:44 moritz these days it can even load more than one ;-)
06:45 * allison sleeps
06:46 moritz (which was not possible with parrot-1.4)
06:47 szabgab joined #parrot
06:48 michel__ joined #parrot
06:50 bacek joined #parrot
06:56 NotFound dukeleto: inverse hyp functions are not in math.h in all C libraries.
06:57 NotFound Better said, they are not in the library, math.h or not.
06:57 dukeleto NotFound: touché . they are in my OS X system headers
06:58 NotFound They are easily implememtable, but you need configure probes if you want to use the system ones where available.
07:04 user_5860 joined #parrot
07:04 kyle_l5l hmm, in Parrot_Object_thaw(), shouldn't we be calling some VTABLE_thaw() when extra_flags == EXTRA_IS_NULL?
07:05 dukeleto NotFound: good point. i am also noticing that the Complex PMC defines cot/csc etc, but they are not defined for real operands. seems a bit asymmetric
07:06 NotFound dukeleto: I think there is a ticket for that inconsistency
07:06 * dukeleto searches
07:07 kyle_l5l (background: I'm trying to get threads + HLLs working; in clone_interpreter(), when we try clone a class, we run into trouble after thawing an Object)
07:07 dukeleto NotFound: perhaps it is still in RT?
07:07 NotFound dukeleto: probably, is an old thing
07:11 NotFound dukeleto: linux man page says that asinh and family are C89, so maybe is safe to assume his availability.
07:11 NotFound "The variant returning double also conforms to SVr4,  4.3BSD, C89."
07:15 dukeleto NotFound: nice research
07:18 bacek o hai...
07:23 dukeleto NotFound: is this what you were thinking of ? http://rt.perl.org/rt3/Tic​ket/Display.html?id=38896
07:23 dukeleto bacek: hola
07:23 bacek dukeleto: G'Day
07:24 dukeleto so is the current policy to create a TT for any RT ticket that we still care about?
07:24 dukeleto i am seeing some RT's that are still relevant which don't have a TT
07:24 HG` joined #parrot
07:27 bacek It's not "policy", but "really good idea" afaik :)
07:28 dukeleto bacek: sounds good
07:30 NotFound dukeleto: not sure, maybe I was mixing Integer/Float/Complex in my head
07:30 dukeleto NotFound: it is similar enough
07:31 NotFound dukeleto: most probably they don't have a TT because nobody looked at them yet.
07:31 bacek msg Whiteknight If I read allison's mail about context_pmc3 branch correctly we have to unmerge pmc_sans_unionval branch... It breaks backward compatibility without deprecation cycle...
07:31 purl Message for whiteknight stored.
07:34 chromatic I don't like that.
07:34 chromatic If we don't maintain backwards compatibility for functions we didn't explicitly export, why do we maintain backwards compatibility for poking into private structs?
07:35 kjeldahl joined #parrot
07:36 bacek chromatic: I don't like either. But it's not my call.
07:38 chromatic You mentioned providing some macro abstractions to help with backwards compatibility.
07:39 bacek s:3rd/ / it /
07:41 cotto Either way, it seems like this should be spelled out more explicitly in the docs.
07:41 cotto The only thing that covers this currently is changes to PARROT_API functions.
07:42 NotFound ...wich desn't exists X-)
07:42 chromatic We've never promised that people can rely on things we haven't documented they can use.
07:44 chromatic If this falls under our backwards compatibility policy, we can't merge the PCC refactors until 2.0.
07:44 bacek_ joined #parrot
07:45 bacek_ me again.
07:45 chromatic <chromatic> If this falls under our backwards compatibility policy, we can't merge the PCC refactors until 2.0.
07:45 * bacek_ blame 3G network.
07:46 bacek_ chromatic: good point... Actually all refactors between 1.5 and 2.0 are focused on wrapping bare structs into API calls.
07:46 bacek_ almost all
07:46 purl hmmm... almost all is obvious but the checksum/verification/signature needs some discussion
07:47 bacek_ purl: forget almost all
07:47 purl bacek_: I forgot almost all
07:54 bacek Anyway, we need consistent policy about supported macros.
07:56 bacek E.g. I kept PMC_sub macro in Sub PMC refactoring branch, but I would like to get rid of it. Etc, etc, etc.
07:56 chromatic We need to figure out how to identify which macros are okay to use.
07:56 chromatic s/identify/document/
07:56 bacek Something like "All macros are for internal purpose only" will work :)
07:56 chromatic I responded to Allison's message on the list.
07:57 bacek chromatic: ok, thanks.
07:57 NotFound rakudo counts as "internal"? ;)
07:58 bacek I can offer patches for 3 "external" projects. Rakudo, partcl and Lua. They are on my check list for branch testing :)
07:58 chromatic Macros and structs aren't in the list of deprecation candidates.
07:58 chromatic $ perldoc docs/project/support_policy.pod
07:59 chromatic With that, I sleep.
07:59 dalek parrot: r40749 | dukeleto++ | trunk/t/op (2 files):
07:59 dalek parrot: [t] Many tests for Inf/NaN
07:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40749/
08:01 bacek chromatic: good night
08:14 dalek TT #943 created by dukeleto++: Real counterparts of some trig functions not defined
08:19 dukeleto joined #parrot
08:34 masak joined #parrot
08:39 dalek parrot: r40750 | bacek++ | trunk/src/pmc/exception.pmc:
08:39 dalek parrot: [cage] Don't deref NULL value in Exception.set_attr_str.
08:39 dalek parrot: This is partial fix for Rakudo's RT#68748 ticket.
08:39 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40750/
08:47 AndyA joined #parrot
08:53 TiMBuS joined #parrot
08:56 dalek parrot: r40751 | NotFound++ | trunk/src/oo.c:
08:56 dalek parrot: [core] better caching of class parent_overrides negative results
08:56 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40751/
09:03 dalek parrot: r40752 | bacek++ | branches/context_pmc3 (98 files):
09:03 dalek parrot: Bring branch up-to-date with trunk.
09:03 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40752/
09:03 cotto night
09:03 cotto happy hacking
09:05 NotFound Looks like r40751 improves speed of rakudo make test by 4%
09:06 mikehh I am just testing at r40750 - will try in a mo
09:07 moritz NotFound++
09:14 braceta joined #parrot
09:14 bacek Bah... Who is "exceptions master"?
09:15 dalek TT #944 created by bacek++: Exception handling in top-level main function is broken.
09:15 NotFound Not a master, but I've done several things inside them
09:16 NotFound Uh, that's ugly
09:17 bacek NotFound: indeed...
09:17 purl indubitably
09:18 * bacek have very strange feelings...
09:19 bacek Last time, in Hashes/Keys revamp branch I switched to cleanup GC bugs... Just because branch exposed a lot of them.
09:19 bacek Looks like time to check exceptions handling. *sigh*...
09:21 NotFound bacek: I think is better to wait for the pcc refactor
09:23 dalek parrot: r40753 | mikehh++ | trunk/MANIFEST:
09:23 dalek parrot: fix MANIFEST for addition of t/op/inf_nan.t
09:23 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40753/
09:26 * bacek is inpatient coding robot
09:26 bacek Bah! "impatient" :)
09:32 mikehh manifest_test FAIL, All others PASS (pre/post-config, smolder, nqp_test, rest of fulltest) at r40750 - Ubuntu 9.04 amd64 (g++) [fixed]
09:33 dalek parrot: r40754 | mikehh++ | trunk/t/op/inf_nan.t:
09:33 dalek parrot: set svn properties for t/op/inf_nan.t
09:33 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40754/
09:33 MoC joined #parrot
09:34 bacek NotFound: I don't think that pcc refactor will help... Some issues with Continuations handling
09:34 mikehh rakudo (c4c67da) builds on parrot r40750 - make test / make spectest (up to 28055) PASS - Ubuntu 9.04 amd64 (g++)
09:36 bacek Yay. TT#944 is about VTABLE invoke.
09:39 gaz joined #parrot
09:52 braceta left #parrot
09:58 mikehh All tests PASS (pre/post-config, smolder, nqp_test, fulltest) at r40754 - Ubuntu 9.04 amd64 (gcc)
10:12 mikehh rakudo (c4c67da) builds on parrot r40754 - make test / make spectest (up to 28055) PASS - Ubuntu 9.04 amd64 (gcc)
10:13 donaldh joined #parrot
10:44 riffraff joined #parrot
10:52 dalek parrot: r40755 | NotFound++ | trunk/src/pmc (4 files):
10:52 dalek parrot: [pmc] set auto_attrs on Hash and derived PMCs
10:52 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40755/
11:09 bacek joined #parrot
11:16 dalek parrot: r40756 | bacek++ | branches/context_pmc3 (10 files):
11:16 dalek parrot: [cage] Rename Parrot_ctx_ functions to Parrot_cx_ for consistency with other functions
11:16 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40756/
11:16 dalek parrot: r40757 | bacek++ | branches/context_pmc3 (2 files):
11:16 dalek parrot: [core] Fix codestd failures in context.c which causes headerizer to fail
11:16 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40757/
11:16 dalek parrot: r40758 | bacek++ | branches/context_pmc3 (2 files):
11:16 dalek parrot: [core] Add functions for manipulating Context recursion_depth
11:16 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40758/
11:19 dalek parrot: r40759 | bacek++ | branches/context_pmc3 (4 files):
11:19 dalek parrot: [core] Use Parrot_cx_inc|dec_recursion_depth instead of poking into
11:19 dalek parrot: struct directly.
11:19 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40759/
11:20 donaldh joined #parrot
11:24 mikehh partcl r817 builds on parrot r40754 - make test same as before (6 files FAIL but all subtests PASS) - running make spectest
11:33 dalek parrot: r40760 | NotFound++ | trunk/src (5 files):
11:33 dalek parrot: [pmc] set auto_attrs on Sub and derived PMCs - second attempt
11:33 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40760/
11:41 szabgab joined #parrot
11:42 mikehh partcl r817 on parrot r40754 - finished make spectest (results logged)
11:45 mikehh All tests PASS (pre/post-config, smolder, nqp_test, fulltest) at r40759 - Ubuntu 9.04 amd64 (g++)
11:53 dalek parrot: r40761 | bacek++ | branches/context_pmc3/src/context.c:
11:53 dalek parrot: [cage] Actually return value from Parrot_cx_dec_recursion_depth
11:53 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40761/
12:08 quek joined #parrot
12:08 mikehh joined #parrot
12:12 dalek parrot: r40762 | bacek++ | branches/context_pmc3 (23 files):
12:13 dalek parrot: Revert r40714 (manually). We can't preserve full backward compatibility
12:13 dalek parrot: for plain C structures anyway.
12:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40762/
12:13 dalek parrot: r40763 | bacek++ | branches/context_pmc3 (3 files):
12:13 dalek parrot: [core] Add caller_context accessors
12:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40763/
12:13 dalek parrot: r40764 | bacek++ | branches/context_pmc3/src (10 files):
12:13 dalek parrot: [core] Use Parrot_cx_get|set_caller_ctx instead of poking directly into
12:13 dalek parrot: struct.
12:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40764/
12:27 kjeldahl joined #parrot
12:29 bacek Hrm... There is no Parrot_interp_* at all, isn't it? Do everyone suppose to poke into Parrot_Interp structure directly?
12:30 NotFound bacek: Parrot_Interp is a pointer to some opaque thing.
12:31 bacek NotFound: I mean API functions.
12:31 bacek And Parrot_Interp isn't "opaque thing" from my point of view.
12:32 NotFound bacek: almost all API functions take an interpreter and use it some way, no point on prefixing all of them.
12:32 NotFound bacek: from the point of view of extend and embed, it is.
12:33 bacek What about functions to manipulate interpeter. Like set new Context, etc.
12:33 NotFound They are Context related, IMO.
12:34 bacek they are, indeed.
12:35 bacek But I want function to shortcut Parrot_cx_get_namespace(interp, CONTEXT(interp)).
12:35 bacek Something like Parrot_XXX_get_current_namespace(interp);
12:37 bacek Looks like XXX can be 'cv' or empty string.
12:38 NotFound cv?
12:38 purl cv is, like, curriculum vitae or a synonym for resume' or 105 in roman numerals or Code Value
12:38 bacek oops, cx
12:39 NotFound I don't like the idea of current namespace, but also don't like the idea of current context
12:40 bacek current namespace == namespace of current context
12:41 payload joined #parrot
12:41 NotFound Unfotunately, there is no other place to take the active context other than the interpreter.
12:41 bacek NotFound: not really.
12:41 bacek Sub, Continuation, etc can have own stored contexts
12:42 bacek ah, "active"
12:42 bacek You are right.
12:42 NotFound I use "active" as a way to diferentiate it from the Interpeter state
12:44 bacek hmm.. I don't quite understand difference
12:46 NotFound Ideally, IMO, the interpreter state will not have things that change just because of entering of exiting a sub, but more stable things.
12:47 NotFound We can't do that now, and maybe never, but we will try to avoid having too much 'current' things in the interpreter,
12:48 bacek All this things are encapsulated in Context.
12:48 bacek "current_namespace" is pointer inside Parrot_Context, not interpreter.
12:48 NotFound Yes, but the concept of "current namespace" is against that scheme.
12:49 NotFound So I'll keep the cx prefix, even if it doesn't take an explicit context argument.
12:49 bacek Deal :)
12:50 NotFound And make very clear that that function is for internal usage only
12:50 Coke ... by not adding PARROT_API to it?
12:50 Coke I wonder if we should add a NOOP PARROT_INTERNAL
12:50 moritz +1
12:50 purl 1
12:51 Coke (to differentiate between things we KNOW should be internal and things we just haven't classified)
12:51 NotFound Coke: provided that no one will add it by mistake...
12:51 NotFound Coke: I like that idea
12:54 bacek Coke: I like that idea too.
12:55 Coke "I'm just an idea Coke."
13:06 whiteknight joined #parrot
13:08 Coke whiteknight: hio.
13:08 whiteknight hello Coke
13:08 Coke whiteknight: note that my memory problems seem to have started after the commit that caused austin's problem.
13:09 Coke my *recent* memory problems.
13:09 whiteknight ah, okay
13:09 whiteknight probably still related, if what I have in my mind is right
13:09 Coke so there might be more than one thing going on.
13:09 Coke k.
13:27 Hunger joined #parrot
13:30 mikehh Coke: I ran make spectest on partcl and have the results logged - do you need anything from it
13:32 mikehh that was partcl r817 parrot r50754
13:32 mikehh r40754
13:47 mikehh bbl
14:10 slavorgn joined #parrot
14:14 MoC joined #parrot
14:16 Coke mikehh: sure, I'll take a copy.
14:24 Taulmarill joined #parrot
14:26 Psyche^ joined #parrot
14:36 ruoso joined #parrot
14:38 Alias joined #parrot
14:38 * Alias just noticed http://use.perl.org/~geoffrey/journal/39516
14:39 Alias #toolchain humbly suggests that if you are considering inventing a CPAN we might have some experience with that kind of thing, that may be of use
14:40 Alias (Nobody seems to have heard of it)
14:40 moritz japhb: feel highlighted
14:41 * szbalint feels stupid
14:41 szbalint I never made the metnal connection that japhb eq rgeoffrey
14:42 szbalint *mental
14:43 Hunger joined #parrot
14:45 payload joined #parrot
14:48 Alias japhb: Would you care to join us in #toolchain?
14:52 jonathan Alias: I'll mention this on #perl6 too, since there's been discussion about module tool chain stuff there too.
14:54 * Coke is all for not reinventing functional wheels
14:57 Alias jonathan: For the record, I can see 5-10 fatal design problems in the current spec
14:58 Alias About half of which already spent years getting rid of in the CPAN design
14:58 Alias That's from an initial 5 minute read
14:59 jonathan Alias: Current spec of...?
14:59 Alias The Parrot CPAN
14:59 jonathan Ah.
14:59 Alias It got designed in IRC over a weekend or something?
15:00 quek left #parrot
15:00 Coke or something.
15:01 moritz Alias: there was some mailing list discussions on parrot-dev before
15:01 Alias oh wait, I read the use.perl port wrong
15:01 Alias post
15:01 Alias Right, over a month
15:02 moritz Alias: but a huge problem was that that it's hard to get non-bikeshedding answers
15:02 Alias bikeshedding?
15:02 moritz discussing the color of the paint rather than real design issues
15:03 Alias Answers that don't discuss real design issues? To which questions?
15:04 moritz we had some threads with >200 postings on perl6-language back in June, with no substantial outcome
15:04 moritz I think the case of the parrot repo it was a bit better
15:04 Alias From experience, I'd suggest that language implementors are pretty bad at doing repositories
15:04 Alias p5p hasn't had a great track record either
15:05 moritz so let's wait for japhb to wake up and move to #toolchain
15:05 Alias ya
15:05 Coke Thanks for poking on this ; I'm sure whatever we end up with will be better for conversation.
15:05 Coke "poking us"
15:06 Alias Certainly the extra collective decades of experience should help :)
15:08 Alias Wonder if we can get someone to sponsor a cross-language source repository thing
15:08 Alias I tried to organise one a few years ago through Linux Australia, but it never got legs
15:09 moritz it would certainly help to meet face to face to discuss things
15:10 Alias Also did some preliminary work on an RDF grammar called PIG (for "Package Interdependency Grammar") from memory
15:10 Alias A kind of more rigourous META.yml that could act as a common point between all the repositories and all the distros
15:10 Eevee joined #parrot
15:11 Alias Which clarified some of the main problems in producing what you want
15:11 Alias (Although it didn't produce a useful language)
15:15 japhb Alias, moritz, jonathan, Coke, szbalint, et al. -- I will be available in a few minutes; need to take my daughter to school.
15:20 MoC joined #parrot
15:20 donaldh joined #parrot
15:26 dalek partcl: r620 | coke++ | wiki/SpecTestStatus.wiki:
15:26 dalek partcl: Add another 7 missed tests to the memory PANIC pile.
15:26 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=620
15:37 japhb OK, bak.  Moving discussion to #toolchain.
15:38 Coke joined #parrot
15:54 dukeleto mikehh++ for fixing my failing codingstd_tests. I forget that most of them don't get run in a git-svn clone, so I have to update an svn repo and then run them there
16:20 he_ Hm, is it just me or does the "pir_output_is( <<'CODE', <<OUTPUT, "Inf - ln" );" test in t/op/arithmetics.t in 1.5.0 fail for others as well?
16:21 Coke Infinoid: "git show git-svn" errors out. "git show -- git-svn" is silent.
16:21 kid51 joined #parrot
16:22 he_ I've written a small test program in C which gets the wrong result both on the platforms closest to "home" (NetBSD/i386, NetBSD/sparc64) and on Linux/i386.
16:24 he_ In short: log(-INFINITY) gives -INFINITY as result, not NaN.
16:26 he_ Hm, smolder appears to indicate this is my own platform problem...
16:31 davidfetter anybody here experienced in conference-wrangling?
16:33 kid51 define conference-wrangling
16:35 davidfetter well, i'm contemplating auctioning off some plum tasks
16:35 davidfetter lightning-talk-gonger, e.g.
16:36 Coke davidfetter: you have enough demand to help out to do that? awesome. =-)
16:37 davidfetter could well be
16:38 pmichaud what conference is this?
16:40 davidfetter pgday.eu, a postgresql conference in paris november 5-6
16:40 davidfetter france, not texas
16:42 soxet joined #parrot
16:49 dalek partcl: r621 | coke++ | trunk/tools/tcl_test.pl:
16:49 dalek partcl: Fix SKIPping for test files containing '-'
16:49 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=621
16:56 Limbic_Region joined #parrot
16:57 l3t0 he_: interesting
16:58 duk3leto he_: can i see your test program?
16:59 chromatic joined #parrot
17:00 mokurai joined #parrot
17:09 joeri joined #parrot
17:16 acajou joined #parrot
17:28 whiteknight joined #parrot
17:49 acajou Hi all.  BTW, topic states http://planet.parrot.org, but that's a DNS error for me...
17:50 Topic for #parrotis now http://www.parrot.org | http://planet.parrotcode.org | 1.5.0 "TEH PARROTZ!"  Released! | Feature freeze over, coders start your engines!
17:50 moritz acajou: thanks, fixed
17:51 acajou moritz: and thanks for fixing it! :)
17:54 acajou So I was wondering, is anyone working on ParrotQT? (Or Perl6QT, but I guess ParrotQT might make more sense?)
17:55 moritz not that I'm aware of
17:56 acajou Nokia has sponsored new Python QT bindings (Pyside) and it looks like they provide a kind of infrastructure that might make it easier to develop bindings for other languages...
17:56 acajou http://www.pyside.org/home-binding/
18:00 acajou It's probably a project that's much too big for little me, but since I'd like to write Qt/KDE apps in Perl6, I thought I throw the idea in here... :)
18:01 moritz acajou: if it's too big for you, you have to recruit slav^Wcontributors
18:02 acajou Indeed.
18:03 acajou I wonder if Nokia might be interested in funding that kind of work or if they just want to bet the farm on Python...
18:05 acajou For someone who's already familliar with make C/C++ libraries available in Parrot, it might be not too hard to get things off the ground I suppose...?
18:05 acajou s/make/making/
18:15 hercynium joined #parrot
18:15 japhb joined #parrot
18:22 Coke left #parrot
18:22 Coke joined #parrot
18:24 * Coke wonders if anyone will make him a pretty spec test chart out of http://code.google.com/p/partcl/source/​browse/trunk/docs/spectest-progress.csv
18:24 dalek rakudo: 7512462 | (Gianni Ceccarelli)++ | src/ (2 files):
18:24 dalek rakudo: move pick to setting
18:24 dalek rakudo: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
18:24 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​512462c718a2f4eb978d7110f4832a38946d25e
18:26 darbelo joined #parrot
18:27 dalek partcl: r622 | coke++ | trunk/docs/spectest- (2 files):
18:27 dalek partcl: update spectest info -
18:27 dalek partcl: - lost many files due to recent undiagnosed memory panics, but...
18:27 dalek partcl: - ... seem to be running the tests that are left slightly faster.
18:27 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=622
18:27 theory joined #parrot
18:27 jonathan Getting the wrong answer faster!
18:29 Coke jonathan: heh. thankfully, the tests that are still running seem to not be giving more wrong answers. =-)
18:30 Coke but the memory utilization has definitely spiked.
18:30 Coke er, shot up.
18:37 nopaste "in.c" at 158.38.62.77 pasted "Test program for duk3leto - testing log(-INFINITY) ++" (28 lines) at http://nopaste.snit.ch/17657
18:38 he_ duk3leto: sorry, had my gaze averted elsewhere; the nopaste above is the test program.
18:39 he_ At least one of our local Linux hosts gave the wrong results...
18:44 chromatic Anyone else getting SIGABORT when running HEAD Rakudo against HEAD Parrot?
18:45 * moritz tries
18:46 moritz latest rakudo with parrot r40737 worked fine
18:46 moritz (just in case we need to bisect)
18:46 chromatic Hm, also getting failures in Parrot's core test suite.
18:46 chromatic Let me rebuild.
18:49 chromatic Still seeing them.
18:50 chromatic Hm, it was the fixed size allocator.
18:50 moritz parrot core tests look fine so far; rakudo is still building
18:51 moritz (amd64)
18:59 whiteknight the fixed size allocator should be disabled in HEAD
18:59 whiteknight at least, that's the state I thought I left it in
18:59 whiteknight if it's not disabled, it should be
19:01 chromatic I have it enabled locally.
19:01 kid51 joined #parrot
19:07 mikehh_ joined #parrot
19:20 donaldh joined #parrot
19:24 davidfetter joined #parrot
19:30 iblechbot joined #parrot
19:34 whiteknight joined #parrot
19:36 kid51 I will be meeting with members of Buffalo.pm on Tuesday evening and Toronto.pm on Thursday evening.
19:37 kid51 Is there any particular message re Parrot you would wish me to convey?
19:37 Coke "Hey!"
19:37 whiteknight "Parrot is awesome, all you guys should use it and hack on it"
19:38 kid51 Umm, they've already had that talk!
19:38 particle wing night has new meaning!
19:43 davidfetter lol!
19:43 * davidfetter wonders where the nearest place with wings a buffalonian would accept are to his house
19:45 kid51 davidfetter:  Plug in your address to yahoo or google maps.  Enter Main & High Streets, Buffalo, NY.  Hit Enter.
19:45 davidfetter kid51, lol!
19:46 * davidfetter is in oakland, california, and would like to believe that somewhere a little closer would do
19:46 kid51 Kansas City is one time zone closer to you.
19:47 davidfetter thanks
20:03 Coke seen allison?
20:03 purl allison was last seen on #parrot 13 hours, 17 minutes and 44 seconds ago, saying: sleeps
20:03 Coke msg dukeleto ping
20:03 purl Message for dukeleto stored.
20:21 [1]Limbic_Region joined #parrot
20:22 dalek partcl: r623 | coke++ | wiki/SpecTestStatus.wiki:
20:22 dalek partcl: The only remaining failures here are trace related.
20:22 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=623
20:23 cotto joined #parrot
20:27 * chromatic will not be at the meeting tomorrow.
20:30 rg joined #parrot
20:30 darbelo cotto: ping
20:31 hercynium joined #parrot
20:37 cotto darbelo, pong
20:39 darbelo cotto: saw your msg. I'd like to keep working on decnum-dynpmcs too, I'll be arround most of this week at ~20:30 UTC (can't recall your tz), for some code-reviewing.
20:40 cotto clock?
20:40 purl cotto: LAX: Mon 1:40pm PDT / CHI: Mon 3:40pm CDT / NYC: Mon 4:40pm EDT / LON: Mon 9:40pm BST / BER: Mon 10:40pm CEST / IND: Tue 2:10am IST / TOK: Tue 5:40am JST / SYD: Tue 6:40am EST /
20:42 cotto sounds good.  How about Wed?
20:42 darbelo Sounds good to me.
20:46 particle speakers so loud...
20:46 bacek joined #parrot
20:49 cotto btw darbelo, the project could use some docs
20:51 darbelo Indeed, it could. I have some stuff laying about my home dir, I'll see if I can wikify it by tomorrow.
20:52 Limbic_Region joined #parrot
21:07 Whiteknight joined #parrot
21:08 Coke joined #parrot
21:19 bacek Good morning
21:19 purl Here I am, brain the size of a planet, and all they say is 'Good Morning'
21:19 davidfetter mornin' bacek
21:20 bacek G'Day davidfetter
21:21 davidfetter how goes?
21:22 he joined #parrot
21:24 japhb joined #parrot
21:25 bacek davidfetter: 7AM here... Trying to understand why I woke up so early...
21:25 * davidfetter passes the $beverage->{mood}
21:34 bobke joined #parrot
21:37 cotto bacek, it's your coding instinct telling you that there's crappy code that needs to be fixed.
21:37 bacek cotto: :)
21:38 cotto I'm surprised you can sleep at all.
21:39 bacek there is about 99% of crappy code in universe. Why to bother waking early?
21:39 mikehh All tests PASS (pre/post-config, smolder, nqp_test, fulltest) at r40764 - Ubuntu 9.04 amd64 (g++)
21:40 darbelo bacek: To fix the easy bugs and leave the hard one for the suckers that wake up later.
21:40 bacek darbelo: other way round. Fixing easy bugs it boring.
21:40 mikehh rakudo (7512462) builds on parrot r40764, make test / make spectest (up to 28056) PASS - Ubuntu 9.04 amd65
21:41 mikehh amd64 (g++)
21:43 mikehh tried for backspace - hit enter instead :-}
21:43 moritz mikehh: happens to me all the time. Very incovenient if you wrote 'rm *' and then hit the wrong key.
21:43 mikehh quite
21:51 ruoso joined #parrot
21:53 dalek rakudo: 9d9d416 | moritz++ | t/spectest.data:
21:53 dalek rakudo: [t/spectest.data] autoincrement-range.t needs the # icu flag, jrtayloriv++
21:53 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/9​d9d41675dab91fbeb6dcf29998210e00baecf97
22:01 cotto joined #parrot
22:13 dalek parrot: r40765 | bacek++ | branches/context_pmc3 (2 files):
22:14 dalek parrot: [core] Add current_namespace accessor to Context
22:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40765/
22:14 dalek parrot: r40766 | bacek++ | branches/context_pmc3/src (7 files):
22:14 dalek parrot: Use Parrot_cx_get|set_namespace accessors.
22:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40766/
22:14 dalek parrot: r40767 | bacek++ | branches/context_pmc3 (6 files):
22:14 dalek parrot: [cage] Fix some codetest failures.
22:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40767/
22:14 dalek parrot: r40768 | bacek++ | branches/context_pmc3/src/pmc/exception.pmc:
22:14 dalek parrot: [cage] Mark more ATTRs in Exception PMC.
22:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40768/
22:19 darbelo karma bacek
22:19 purl bacek has karma of 925
22:19 moritz wow, that's impressive.
22:19 moritz karma moritz
22:19 purl moritz has karma of 586
22:19 bacek o shi... looks like purl miscounted little but
22:20 moritz karma darbelo
22:20 purl darbelo has karma of 200
22:22 duk3leto msg he_ could you send along uname -a and gcc -v to go along with http://nopaste.snit.ch/17657 ?
22:22 purl Message for he_ stored.
22:22 duk3leto karma dukeleto
22:22 purl dukeleto has karma of 97
22:27 nopaste "darbelo" at 200.49.154.172 pasted "paste for duk3leto" (41 lines) at http://nopaste.snit.ch/17659
22:28 darbelo duk3leto: I'm not he_, but I figured you migh want to get the results of http://nopaste.snit.ch/17657 for other arches as well.
22:30 darbelo http://nopaste.snit.ch/17659 is for OpenBSD -current amd64
22:39 Whiteknight karma l3to
22:39 purl l3to has neutral karma
22:39 darbelo karma duk3leto
22:39 purl duk3leto has neutral karma
22:40 bacek Whiteknight: hi
22:40 duk3leto darbelo: dukeleto is where my karma lives
22:40 Whiteknight hello bacek
22:41 rg1 joined #parrot
22:41 bacek Will you have any chance to look at context_pmc3 branch? t/pmc/exceptionhandler.t failing test 7...
22:43 darbelo duk3leto: But that means I have to hit tab *twice* to ++ you.
22:46 darbelo That's a lot of work you're forcing people to do. And think of the strain on the poor keys! Getting pounded twice every time you do something good.
22:48 duk3leto darbelo: i am on my way to setting up irssi proxy :)
22:48 duk3leto darbelo++
22:49 darbelo My tab key appreciates it. dukeleto++ :)
22:50 dalek parrot: r40769 | bacek++ | branches/context_pmc3 (2 files):
22:50 dalek parrot: Relax checks for namespace in Parrot_cx_set_namespace.
22:50 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40769/
22:55 nopaste "dukeleto" at 70.102.219.22 pasted "Everyone loves a log -- Updated small C test file to find out if log ANSI-compiant on your system" (29 lines) at http://nopaste.snit.ch/17660
22:55 tokuhirom__ joined #parrot
22:56 duk3leto darbelo: if you try that one and paste your results, that would be super awesome super happy fun time :)
22:57 mokurai left #parrot
22:57 Andy joined #parrot
23:04 nopaste "darbelo" at 200.49.154.172 pasted "Another nopaste for dukeleto++" (43 lines) at http://nopaste.snit.ch/17661
23:04 darbelo duk3leto: Enjoy your super awesome super happy fun time :) with http://nopaste.snit.ch/17661
23:09 duk3leto darbelo++
23:19 patspam joined #parrot
23:20 donaldh joined #parrot
23:43 Whiteknight I'm looking to get an IDE, or something more sophisticated then medit. Anybody have any suggestions?
23:43 darbelo emacs? vim?
23:44 jonathan Visual Studio?
23:44 purl hmmm... Visual Studio is pretty nice
23:44 Whiteknight I'm on linux here
23:44 jonathan Ah.
23:44 jonathan Eclipse may be wroth a try.
23:44 Whiteknight but I'm inspired by visual studio, I use that at work and have grown accustomed to these things called "features"
23:45 jonathan Whiteknight: Aye, I do like Visual Studio too.
23:45 jonathan The C debugger is pretty decent.
23:45 darbelo Try Visual Studio on wine then :)
23:45 Whiteknight yeah, and all the .NET stuff is a breeze
23:45 dalek TT #549 closed by whiteknight++: Kill UnionVal
23:46 Whiteknight okay, I'll look into eclipse then
23:47 Whiteknight I had a bad experience with netbeans back in undergrad, and I haven't used any "real" IDEs since
23:48 darbelo Whiteknight: A Visual Studio junkie I know claims that MonoDevelop helps him cope with the withdrawal simptoms.
23:48 jonathan I never tried that, but I never heard anyone compliement it either.
23:49 Whiteknight well, I'm specifically looking for something to help with my parrot coding, so all C stuff
23:49 Whiteknight a little perl would be nice too, but I know that's like asking for a miracle
23:49 Austin joined #parrot
23:50 rg1 well i guess padre is always looking for help ;)
23:51 * acajou heard good things about EPIC - the Eclipse Perl plugin...
23:52 bacek_at_work hi again
23:52 purl oh, you're back!
23:52 Whiteknight rg1: I'm very very close to switching to Padre
23:52 Whiteknight but it's missing some features that I consider "deal breakers"
23:52 bacek_at_work Whiteknight: try vim. Very clever IDE :)
23:52 Whiteknight hell no, I tried vim a few times and I hated it
23:53 acajou xemacs perldb mode is pretty nice...
23:53 * rg is pretty happy with emacs too
23:53 Whiteknight When I first started using it, I had to leave my computer on for a week because I couldn't figure out how to save and close
23:53 Whiteknight and the online documentation was shittacular
23:54 rg but then people are expecting a lot more from an ide, like completion, help/autocorrection with errors, manpages and such things
23:54 * bacek_at_work wander about how my coding speed will be impacted without vim.
23:54 bacek_at_work 1/10 is about good estimation.
23:55 payload joined #parrot
23:56 Austin left #parrot
23:56 Austin joined #parrot
23:59 darbelo It looks like vim is an essential component of our magical coding robot!

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

Parrot | source cross referenced