Camelia, the Perl 6 bug

IRC log for #parrot, 2010-05-26

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:06 whiteknight joined #parrot
00:19 kyumoka joined #parrot
00:25 GeJ clock?
00:25 purl GeJ: LAX: Tue 5:25pm PDT / CHI: Tue 7:25pm CDT / NYC: Tue 8:25pm EDT / LON: Wed 1:25am BST / BER: Wed 2:25am CEST / IND: Wed 5:55am IST / TOK: Wed 9:25am JST / SYD: Wed 10:25am EST /
00:28 Mokurai1 joined #parrot
00:32 whiteknight I keep wishing we had a STRING_IS_NULL_OR_EMPTY() macro
00:41 Coke https://review.source.android.com/#patch,sidebyside,14699,1,libc/memset.c
00:42 sorear whiteknight: I think I proposed eliminating null strings in favor of empty strings a bit ago
00:42 whiteknight sorear: well, I can think of a few reasons why we would want a semantic difference between a null string and an empty string
00:42 tcurtis Coke: .... how recent was that change?
00:43 whiteknight I think of it analogously to how NaN is in relation to other floating point numbers
00:44 Coke tcurtis: iunno.
00:46 dalek parrot: r47000 | whiteknight++ | trunk (2 files):
00:46 dalek parrot: Add new PMC_IS_TYPE() and PMC_IS_TYPE_ENUM() macros on an experimental basis. Employ the first to prove that it works as expected
00:46 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47000/
00:46 dalek parrot: r47001 | whiteknight++ | trunk/src/pmc.c:
00:46 dalek parrot: [fix] oops. left in a C++ comment
00:46 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47001/
00:47 dalek plparrot: d1a8766 | dukeleto++ |  (2 files):
00:47 dalek plparrot: Link against parrot_config.o, and use Parrot_set_config_hash
00:47 dalek plparrot: Now we don't have to not jump through hoops loading P6object.pbc, Notfound++
00:47 dalek plparrot: review: http://github.com/leto/plparrot/commit/d​1a8766a207e0d808b8be91234ca104c8677508d
00:47 dalek plparrot: 6459ad5 | dukeleto++ | t/sql/test.sql:
00:47 dalek plparrot: Enable test which loads test_more.pir and add a test for load_bytecode
00:47 dalek plparrot: review: http://github.com/leto/plparrot/commit/6​459ad523931c46c65ce19630d37551133362210
00:47 Coke wow. This is the change that broke partcl: http://trac.parrot.org/parrot/changeset/45623
00:47 sorear wow?
00:48 * bacek_at_work hides under desk.
00:48 bacek_at_work Just in case if there is his fingerprints on commit
00:49 Coke ah. I just happened to end on a bad build.
00:56 abqar joined #parrot
00:57 abqar joined #parrot
00:58 tcurtis joined #parrot
00:58 dalek plparrot: 473d612 | dukeleto++ | README.plparrot:
00:58 dalek plparrot: Rename README to be more github-friendly
00:58 dalek plparrot: review: http://github.com/leto/plparrot/commit/4​73d61261eaed788f8fae4aa53362daa3122f83f
00:58 whiteknight src/global.c should be renamed to src/namespace.c, since every function in that file deals with namespaces
00:59 whiteknight and most of the functions therein should be renamed "Parrot_ns_*"
01:01 snarkyboojum joined #parrot
01:04 dalek plparrot: b736921 | dukeleto++ | html/i (2 files):
01:04 dalek plparrot: Add a website stub
01:04 dalek plparrot: review: http://github.com/leto/plparrot/commit/b​736921f6e90be62e7823b964317c8f2377b1bc0
01:26 cotto what's that smell?
01:27 cotto smells like ops.num clobbering time
01:29 whiteknight I smell what the cotto is cooking
01:30 Coke "the cott"
01:31 cotto coke--
01:31 cotto coke++
01:31 Coke ugh. I hate it when a bisect goes bad. these things take so long when they /work/
01:33 cotto maybe make test will finish faster if I check it every few seconds
01:34 cotto as opposed to something useless like parallel testing
01:35 sorear Andy++ gcc built using your instructions
01:35 Andy yay
01:36 sorear 176m
01:36 dalek parrot: r47002 | whiteknight++ | trunk/src/oo.c:
01:36 cotto done
01:36 dalek parrot: [oo] small refactor and TODO note for Parrot_oo_get_class. Only a fraction of the hurtin' that function requires
01:36 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47002/
01:36 dalek parrot: r47003 | cotto++ | trunk (9 files):
01:36 dalek parrot: [opsc] nuke ops.num and supporting code
01:36 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47003/
01:37 cotto That's less than 3 hours.
01:38 Coke hurm. that really is the build that broke partcl.
01:39 cotto the plot thickens
01:39 purl I know it! We are in a *novel*!
01:39 Coke r45623 introduced the "expected integer but got "0"" bug.
01:40 whiteknight src/oo.c: There is so much FAIL in this file
01:41 Coke msg plobsing r45623 introduced the "expected integer but got "0"" bug in partcl.
01:41 purl Message for plobsing stored.
01:43 Coke ah. he changed it from using an Undef to morph into the right type to always be a String.
01:51 bline joined #parrot
01:53 dalek parrot: r47004 | whiteknight++ | trunk/src/oo.c:
01:53 dalek parrot: [oo] some small improvments to func docs, for clarity. Much much more is needed
01:53 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47004/
02:09 dalek parrot: r47005 | petdance++ | trunk/src/pmc/coroutine.pmc:
02:09 dalek parrot: remove unused var
02:09 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47005/
02:09 dalek parrot: r47006 | petdance++ | trunk/src/string/encoding/ucs4.c:
02:09 dalek parrot: reran headerizer
02:09 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47006/
02:10 cotto btw, reconfig is recommended after my ops.num nuking
02:12 * cotto bumps PBC_COMPAT
02:19 kyumoka left #parrot
02:26 dalek parrot: r47007 | cotto++ | trunk/PBC_COMPAT:
02:26 dalek parrot: bump pbc_compat for ops.num removal
02:26 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47007/
02:28 dalek TT #1658 created by coke++: Change to config hash (45623)  broke partcl
02:28 dalek TT #1658: http://trac.parrot.org/parrot/ticket/1658
02:37 tcurtis joined #parrot
02:39 cotto hio tcurtis
02:39 tcurtis Hello, cotto.
02:42 dalek parrot: r47008 | darbelo++ | branches/ucs4_encoding:
02:42 dalek parrot: Branch has been merged into trunk and is no longer needed at HEAD.
02:42 purl i already had it that way, dalek.
02:42 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47008/
02:45 plobsing joined #parrot
02:48 janus joined #parrot
02:50 Coke Branch has been merged into trunk and?
02:50 purl i heard Branch has been merged into trunk and was no longer needed at HEAD.
02:50 LoganLK joined #parrot
02:55 snarkyboojum joined #parrot
02:56 plobsing Coke: (re: TT #1658) why do you have active HLL mappings if the functionality isn't ready to be used yet?
02:57 Coke plobsing: I think I will not answer that question until I get back the last 5 hours of parrot time that I wasted trying to pin down this commit. =-)
02:58 Coke I'm not sure I'll have a good answer anyway.
02:59 Coke but perhaps I'll have a less bitter one tomorrow. Night.
02:59 dalek parrot: r47009 | plobsing++ | branches/ops_massacre/t (17 files):
02:59 dalek parrot: fix some tests after moving i/o ops to dynops
02:59 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47009/
02:59 plobsing Coke: as the author on that commit, I'm sorry it cost you so much time. I'm willing to help remedy the problem. A superficial fix might be to use core PMCs explicitly.
02:59 plobsing However, IMHO, the underlying bug is the uninitialized but active HLL mappings
03:06 Coke left #parrot
03:18 tcurtis cotto, was that "hio" just being nice and friendly, or did you want to talk about something? I only ask because I'm about to leave #parrot for now(possibly for the night) unless it's the latter.
03:23 cotto just being friendly
03:23 cotto I ping if I want to talk.
03:23 cotto about something specific
03:28 bacek_at_work ~~
03:28 cotto ~~
03:28 bacek_at_work Does anyone know what soft used on cpantesters for continuous integration?
03:29 parthm joined #parrot
03:33 bacek_at_work cotto++ # nuking ops.num for good!
03:33 bacek_at_work and "experimental" ops are first class citizens now :)
03:35 mikehh packfile tests failing, needs sortin' out
03:36 mikehh darbelo: ping
03:39 bacek_at_work we have to rebuild packfiles after ops.num nuke
03:40 cotto no, for great!
03:41 cotto and "first class" wasn't ever especially well-defined on a technical level apart from "don't depend on these.  We might change them for no good reason whenever we feel like it."
03:44 nopaste "GeJ" at 192.168.1.3 pasted "[PATCH] reformat docs/pdds/pdd28_strings.pod to fix `make codetest`" (132 lines) at http://nopaste.snit.ch/20648
03:45 GeJ my POD-fu is weak. The patch would probably need some review.
03:45 GeJ last modification of the file was darbelo++ during the ucs4 merge, I believe.
03:47 GeJ Damn, I missed a <em>...</em> in the last paragraph.
03:49 dalek parrot: r47010 | mikehh++ | trunk/docs/pdds/pdd28_strings.pod:
03:49 dalek parrot: remove lines erroneously added in merge
03:49 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47010/
03:51 plobsing is there any way to get parrot's trace core to expand PMCs (eg: keys)?
03:52 mikehh Gej: I removed the lines in r47010 but kept them, similar to the patch, if they need adding later (they seem to be html of trak wiki or something - not pod)
03:55 GeJ mikehh: thank you.
03:57 mikehh GeJ: I am also getting a bunch of perlcritic failures (17 files) with TODO and/or XXX comments, since I upgraded to the latest Perl:Critic::Bangs
03:57 purl okay, mikehh.
03:57 plobsing msg Coke fix for config issue is at http://github.com/plobsing/partcl. next issue: [gs]et_.*_global seems to be a blackhole for _tcl;colons
03:57 purl Message for coke stored.
03:58 mikehh that of course should have :: in it
04:02 snarkyboojum joined #parrot
04:02 mikehh how do you get purl to forget things you said with I in them?
04:03 * plobsing wonders what mikehh is trying to hide
04:03 plobsing mikehh?
04:03 purl mikehh is giving make cover a try - let's see what happens - although I really should be able to pass make fulltest first or getting a bunch of perlcritic failures (17 files) with TODO and/or XXX comments, since I upgraded to the latest Perl:Critic::Bangs
04:04 plobsing purl forget mikehh
04:04 purl plobsing: I forgot mikehh
04:04 plobsing mikehh?
04:04 mikehh just when purl says okay, she has remembered sometrhing about you :-}
04:04 mikehh something
04:04 purl something is really wrong out there :)
04:04 plobsing the jig?
04:04 purl the jig is trying to write a sort of "expect" script in Perl with Win95.  Has anyone done this?  Jig need help.
04:05 plobsing the jig is also up
04:05 purl okay, plobsing.
04:07 * mikehh needs some sleep, or really am going to come across 6am
04:09 TiMBuS joined #parrot
04:26 GeJ "PackFile_unpack: This Parrot cannot read bytecode files with version 6.17." is there something that needs to be done somewhere? PBC_COMPAT points to a 6.18.
04:27 sorear packfiles do not have forward or backward compatibility
04:28 plobsing GeJ: run tools/dev/mk_native_pbc on i386
04:28 sorear if PBC_COMPAT is 6.18, all packfiles must be 6.18
04:28 sorear 6.17 packfiles are useless and must be regeneratated
04:28 GeJ plobsing: Ah, thank you.
04:29 GeJ sorear: thanks for the details.
04:29 plobsing btw, free karma for the 32 bit parrot dev who updates the packfiles
04:30 * GeJ is running amd64.
04:30 GeJ will that be a problem?
04:30 plobsing yes.
04:30 plobsing unless you have a virtualized 32 bit OS
04:30 cotto will do
04:31 cotto since I bumped pbc_compat in the first place
04:31 plobsing go cotto go!
04:31 cotto om nom nom
04:31 purl I eat your head!
04:32 bacek_at_work too late!
04:32 * bacek_at_work laughing evilly
04:32 sorear you shouldn't need a virtualized 32 bit OS
04:33 sorear just a multiarch one
04:33 sorear ie compile Parrot with -m32
04:33 plobsing you could in theory configure parrot to use 32-bit sized INTVALs/opcode_t/FLOATVALs, but the script doesn't do taht
04:33 plobsing no need for -m32 even
04:34 JimmyZ joined #parrot
04:38 dalek parrot: r47011 | bacek++ | trunk/t/native_pbc (4 files):
04:38 dalek parrot: Rebuild native PBCs
04:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47011/
04:40 plobsing note to self: do not stand between bacekbot and karma
04:41 bacek_at_work karma me
04:41 purl bacek_at_work has karma of 3
04:41 bacek_at_work erm...
04:41 bacek_at_work karma bacek
04:41 purl bacek has karma of 2647
04:44 GeJ karma me
04:44 purl gej has karma of 49
04:45 GeJ not too bad for mostly begging people to fix coding standards :)
04:46 bacek_at_work karma c
04:46 purl c has karma of 8613
04:47 bacek_at_work little bit more work and I can took over this cheater
04:47 cotto karma purl
04:47 purl purl has karma of 8967
04:47 cotto but purl cheats
04:47 plobsing karma me
04:47 purl plobsing has karma of 482
04:47 sorear karma me
04:47 purl sorear has karma of 126
04:47 cotto karma salami
04:47 purl salami has karma of 1
04:48 cotto karma bacon
04:48 purl bacon has karma of 120
04:48 plobsing karam (sliced bread)
04:48 cotto bacon >> sliced bread
04:48 plobsing karma (sliced bread)
04:48 purl (sliced bread) has neutral karma
04:48 cotto though if you put them together...
04:49 plobsing karma sandwich
04:49 purl sandwich has karma of 1
04:49 plobsing not greater than the sum of its parts
04:50 cotto you make the assumption that purl is an authoritative source on anything
04:59 snarkyboojum joined #parrot
05:11 plobsing filehandles opened with the open op can handle timely destroy but ones opened with new ['FileHandle'] cannot. why?
05:45 dalek parrot: r47012 | khairul++ | branches/gsoc_instrument (3 files):
05:45 dalek parrot: Removed duplicate functionality
05:45 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47012/
05:45 dalek parrot: r47013 | khairul++ | branches/gsoc_instrument/t/dynpmc/instrument.t:
05:45 dalek parrot: Added tests for instrument dynpmc
05:45 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47013/
05:45 dalek parrot: r47014 | pmichaud++ | trunk/src/pmc/sub.pmc:
05:45 dalek parrot: [core]:  Add .set_outer_ctx method to Sub PMC
05:45 dalek parrot: to explicitly set outer_ctx of next invocation (for
05:45 dalek parrot: times when the outer_sub doesn't have the needed context).
05:45 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47014/
05:49 dalek TT #1659 created by plobsing++: FileHandle objects don't honour timely destruction
05:49 dalek TT #1659: http://trac.parrot.org/parrot/ticket/1659
05:56 uniejo joined #parrot
06:01 dalek nqp-rx: 814727d | pmichaud++ |  (7 files):
06:01 dalek nqp-rx: Enable outer_ctx handling in NQP programs (e.g., via NQP::Compiler.eval).
06:01 dalek nqp-rx: Establish some infrastructure to make it easier for other compilers
06:01 dalek nqp-rx: to do the same.
06:01 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/8​14727d55f980f2d93efac9ef1a630e58a431147
06:01 dalek nqp-rx: d905d5b | pmichaud++ | src/cheats/parrot-callcontext.pir:
06:01 dalek nqp-rx: Skip any null lex_pad entries in the outer_ctx chain.
06:01 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/d​905d5b9f3e3f9f7bdb117d7fa38988311bcda98
06:01 dalek nqp-rx: 9125c3c | pmichaud++ | src/HLL/Actions.pm:
06:01 dalek nqp-rx: Add HLL::Actions.CTXSAVE to make it easy for compilers to save contexts.
06:01 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/9​125c3c4277abf6f1bb2f5ebfaf5863d8b2a3993
06:02 dalek nqp-rx: c5659bb | pmichaud++ | src/HLL/Compiler.pm:
06:02 dalek nqp-rx: Clean up exception printing when catching exceptions in interactive mode.
06:02 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/c​5659bb1223d697620f9dc3d591a493851c9999a
06:02 dalek nqp-rx: eaa4deb | pmichaud++ |  (4 files):
06:02 dalek nqp-rx: Add a more lexically-aware interactive REPL.  Still some tweaks wanted.
06:02 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/e​aa4deb2738645dfe7c1228027227c5d16117314
06:18 dalek parrot: r47015 | plobsing++ | branches/ops_massacre/t (3 files):
06:18 dalek parrot: fix many io tests after io ops moved to dynops
06:18 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47015/
06:21 GeJ as r47014 make fulltest PASS on FreeBSD 7 amd64
06:32 fperrad joined #parrot
06:35 fperrad_ joined #parrot
06:39 parthm joined #parrot
06:42 sorear How does "timely destruction" work?
06:58 cotto poorly atmp
06:58 cotto *atm
06:59 parthm joined #parrot
07:06 viklund joined #parrot
07:07 plobsing sorear: PMCs can register themselves as needing immediate cleanup. The GC can be run in "conditional" mode which only runs if some live PMCs are registered for timely destruction
07:08 parthm joined #parrot
07:08 sorear so, with a FH PMC around, the GC runs every insn?
07:09 plobsing no. there is a special "run the GC conditionally" opcode: "sweep 0"
07:10 plobsing the idea being that you stick one of those at the end of every scope to be able to support perlish autoclosing FHs
07:11 plobsing I'm not sure whether or not the "needs timely destruction" flag gets reset on GC run. Seems like things might get expensive otherwise.
07:12 plobsing (yes I'm aware that end-of-scope destruction is subtly weaker than refcounted destruction)
07:14 parthm joined #parrot
07:19 parthm left #parrot
07:40 snarkyboojum joined #parrot
07:57 dalek parrot: r47016 | gerd++ | trunk/examples/languages/squaa​k/doc/tutorial_episode_1.pod:
07:57 dalek parrot: remove some deprecated things and add the current use in the pct_tutorial_episode1
07:57 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47016/
08:11 bakkdoor joined #parrot
08:14 dalek parrot: r47017 | plobsing++ | branches/ops_massacre (3 files):
08:14 dalek parrot: fix remaining io tests
08:14 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47017/
08:19 bacek joined #parrot
08:31 dalek parrot: r47018 | plobsing++ | branches/ops_massacre/t/src/extend.t:
08:31 dalek parrot: fix src_tests
08:31 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47018/
08:31 dalek parrot: r47019 | chromatic++ | trunk/src/pmc/callcontext.pmc:
08:31 dalek parrot: [PMC] Avoided CallContext attribute refetching by extracting one static
08:31 dalek parrot: function from another and using the extracted version in two hot paths.  This
08:31 dalek parrot: produces a modest but measurable PCC performance improvement.
08:31 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47019/
08:36 iblechbot joined #parrot
08:41 bacek Aloha
08:46 JimmyZ hi~~
09:04 dalek parrot: r47020 | gerd++ | trunk/lib/Parrot/Configure/Options (3 files):
09:04 dalek parrot: Remove the languages option (perl Configure.pl --languages=...) which seems not to be in use any more. The option is deprecated.
09:04 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47020/
09:41 clinton joined #parrot
09:53 dalek parrot: r47021 | mikehh++ | trunk/src/pmc/callcontext.pmc:
09:53 dalek parrot: fix codetest failure - line length
09:53 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47021/
10:36 lucian joined #parrot
10:36 parthm joined #parrot
10:38 IllvilJa joined #parrot
10:51 aukjan joined #parrot
10:56 mikehh perlcritic fails on 17 files (out of 599) - all Flag comment TODO and/or XXX
10:56 mikehh all other tests PASS (pre/post-config, make corevm/make coretest, smoke (#34060), fulltest) at r47021 - Ubuntu 10.04 amd64 (g++)
11:01 IllvilJa joined #parrot
11:21 Coke joined #parrot
11:21 Coke gerd's mentor?
11:25 Coke message gerd - deprecated has a specific meaning - if that was deprecated, then someone should have opened a ticket, added the ticket to DEPRECATED.pod, and in general followed the support policy (r47020). if it is /obsolete/, that's different.
11:25 purl Message for gerd stored.
11:28 Coke message gerd - if that was driving 'gen::languages' then that config step should also be removed.
11:28 purl Message for gerd stored.
11:42 atrodo joined #parrot
11:49 JimmyZ joined #parrot
11:59 whiteknight joined #parrot
12:05 tetragon joined #parrot
12:08 bacek msg NotFound is GC_ATTRIB_POOLS_HEADROOM suppose to be sizeof(PMC_Attribute_Arena), not "8"?
12:08 purl Message for notfound stored.
12:09 whiteknight GC_ATTRIB_POOLS_HEADROOM is supposed to be a buffer value preventing us from making too many resizes on the pool aray
12:09 whiteknight array
12:09 whiteknight we allocate a little more than we need so we don't need to resize if we need a bigger object
12:11 bacek ah. I misunderstand it
12:12 bacek Anyway, I don't want to create PoolAllocator from GCMassacre anymore. I can just reuse "PMC_Attributes_Pool".
12:12 whiteknight make sure it has the right interface you need. Then we can go back and replace the guts of it later
12:12 bacek It will be little bit slower on checking that pointer contains in pool. But we use it only for tracing system stack.
12:13 bacek I need just one more function - "Allocator.is_owned" from GCMassacre page.
12:14 bacek And current PMC_Allocator should be faster than my proposal.
12:18 whiteknight The current allocator really isn't bad
12:21 whiteknight it's probably the fastest part of the whole GC
12:26 bacek whiteknight, it is fast. And I can't see how it can be improved for supposed usecases.
12:26 whiteknight bacek: I really like your GC ideas, by the way
12:27 bacek whiteknight, I spent months thinking how to refactor our GC :)
12:27 bacek Especially last few weeks...
12:29 dalek tracwiki: v9 | bacek++ | GCMassacre
12:29 dalek tracwiki: Add note about using "PMC_Allocator" as "Allocator"
12:29 dalek tracwiki: http://trac.parrot.org/parrot/wiki/G​CMassacre?version=9&amp;action=diff
12:29 dalek tracwiki: v10 | bacek++ | GCMassacre
12:29 dalek tracwiki: http://trac.parrot.org/parrot/wiki/G​CMassacre?version=10&amp;action=diff
12:34 bacek anyway, I need to sleep on this before going ahead with full speed
12:34 bacek Good night, folks
12:35 JimmyZ good night
12:36 elmex joined #parrot
12:39 dalek parrot: r47022 | pmichaud++ | trunk/ext/nqp-rx/src/stage0 (3 files):
12:39 dalek parrot: [nqp-rx]: Update bootstrap with latest REPL capabilities.
12:39 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47022/
12:52 dalek rakudo: 39c38ca | (Martin Berends)++ | src/Perl6/Compiler.pir:
12:52 dalek rakudo: [Perl6/Compiler.pir] change the order of directories in @*INC so that the
12:52 dalek rakudo: PERL6LIB ones come first
12:52 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/3​9c38caa2548d4d895d9034eedc4d9897b930e01
12:59 snarkyboojum joined #parrot
13:28 dalek rakudo: 1728e99 | (Solomon Foster)++ | src/core/Cool-str.pm:
13:28 dalek rakudo: Clarify logic at masak++'s suggestion.
13:28 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/1​728e99ffb00b811fbbc2b781a3ba44e504f90df
13:28 dalek rakudo: 51b61c5 | (Martin Berends)++ | src/Perl6/Compiler.pir:
13:28 dalek rakudo: [Perl6/Compiler.pir] re-order @*INC again, '.' goes to the end, pmichaud++
13:28 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/5​1b61c5969c5455a705eea01e7c1fda2fe026f57
13:28 dalek rakudo: 6f0e67d | (Martin Berends)++ | src/core/Cool-str.pm:
13:28 dalek rakudo: Merge branch 'master' of github.com:rakudo/rakudo
13:28 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/6​f0e67d7e72d0e45ebf8e3bf2199f173a7e00dbf
13:28 dalek rakudo: fd769c8 | pmichaud++ |  (2 files):
13:28 dalek rakudo: Update Perl6::Actions to grab dynamic outer compilation via HLL::Actions
13:28 dalek rakudo: instead of maintaining its own.
13:28 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/f​d769c80e75684d4ffc36a4d330eee453af1b360
13:28 dalek rakudo: 20068fa | pmichaud++ | src/ (2 files):
13:28 dalek rakudo: Merge branch 'master' of github.com:rakudo/rakudo
13:28 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/2​0068fa4ec19c0de7583af3036609385820450fa
13:49 patspam joined #parrot
13:59 dalek rakudo: 4569863 | masak++ | src/Perl6/Module/Loader.pm:
13:59 dalek rakudo: [Module::Loader] removed redundant period
13:59 dalek rakudo: The error message ends with a pair of parentheses. These stand on their own
13:59 dalek rakudo: without a full stop after them.
13:59 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/4​5698634de4135cf7b8c645887460ca7f3e5a715
13:59 gbacon joined #parrot
14:09 aukjan joined #parrot
14:20 ruoso_ joined #parrot
14:22 aukjan joined #parrot
15:03 theory joined #parrot
15:11 davidfetter joined #parrot
15:13 whiteknight joined #parrot
15:26 bakkdoor joined #parrot
15:27 dmalcolm joined #parrot
15:32 Mokurai joined #parrot
15:40 khairul joined #parrot
15:48 IllvilJa joined #parrot
15:54 integral joined #parrot
16:27 theory joined #parrot
16:32 darbelo mikehh: pong
17:30 mikehh darbelo: just wanted to ask about pdd28_strings.pod - what had you intended there
17:31 mikehh it looks like you commited some html/trac wiki code
17:37 mikehh darbelo: added in r46999 - removed in r47010
17:38 cotto_work anyone know when plobsing is planning on merging ops_massacre?
17:40 mikehh is he all done with the massa-cre
17:50 cotto_work seen plobsing
17:50 purl plobsing was last seen on #parrot 10 hours, 38 minutes and 0 seconds ago, saying: (yes I'm aware that end-of-scope destruction is subtly weaker than refcounted destruction)
17:50 darbelo mikehh: To be honest, I have no idea how that got there.
17:51 mikehh darbelo: looks like it got messed up as part of the merge
17:52 mikehh don't want to get too involved in svn/git wars but svn--
17:52 darbelo That text comes from a totally different file. I have no idea how svn could have done that.
17:52 dalek parrot: r47023 | NotFound++ | trunk (7 files):
17:52 dalek parrot: rename unroll to finalize
17:52 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47023/
18:21 iblechbot joined #parrot
18:25 dalek rakudo: 4e3d29d | (Hongwen Qiu)++ | src/core/Cool-str.pm:
18:25 dalek rakudo: add :samecase to Cool.subst
18:25 dalek rakudo: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
18:25 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/4​e3d29df4c64affddf6ff127652f92adce1077f7
18:25 dalek rakudo: bafb6d9 | moritz++ | src/Perl6/ (2 files):
18:25 dalek rakudo: remove rational number literal fossile
18:25 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/b​afb6d95ed74f6272fc91997722de560dd0ce81a
18:25 bluescreen joined #parrot
18:43 dalek parrot: r47024 | darbelo++ | branches/gsoc_nfg:
18:43 dalek parrot: Create a branch for my summer of code work on NFG.
18:43 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47024/
18:48 dalek winxed: r477 | julian.notfound++ | trunk/winxedst (2 files):
18:48 dalek winxed: rename unroll to finalize
18:48 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=477
18:49 tcurtis joined #parrot
19:18 bluescreen joined #parrot
19:33 dalek rakudo: 475d1c7 | (Solomon Foster)++ | src/core/Cool-str.pm:
19:33 dalek rakudo: Change .match to support the new list approach for :nth.
19:33 purl dalek: that doesn't look right
19:33 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/4​75d1c77fba6ff29c93c347336c51d203675b926
19:33 joeri joined #parrot
19:59 TimToady phone
20:03 dalek winxed: r478 | julian.notfound++ | trunk/examples/ajax.winxed:
20:03 dalek winxed: fix slurpy->flat in example ajax
20:03 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=478
20:17 theory joined #parrot
20:27 Mokurai1 joined #parrot
20:40 dalek parrot: r47025 | NotFound++ | trunk/src/pmc/string.pmc:
20:40 dalek parrot: change misuses of str_byte_length to str_length
20:40 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47025/
20:40 purl dalek: that doesn't look right
20:41 GeJ Good morning everyone.
20:54 tcurtis Good afternoon, GeJ.
21:00 GeJ heya tcurtis.
21:00 purl it has been said that tcurtis is Tyler Curtis <mailto:tyler.l.curtis@gmail.com>
21:01 patspam joined #parrot
21:09 bacek Good morning, humans.
21:14 cotto_work This is interesting.  It appears that PyPy doesn't have either its own set of ops or an ast that it operates on, but that it transforms an object-based representation of the input program.
21:16 cotto_work That gives it the ability to preserve a lot of semantic information through the transformation process so that the backend code emitters can dtrt with it.
21:16 cotto_work It's definitely a different approach than Parrot has been using.
21:17 tcurtis cotto_work: according to http://codespeak.net/pypy/trunk​/pypy/doc/interpreter.html#id5 they use CPython's bytecode.
21:19 GeJ G'Day bacek.
21:19 tcurtis Although it does seem to indicate that the actual implementation of the various ops are deferred to the object space.
21:19 bacek cotto_work, not really.
21:20 bacek PyPy is kind of PCT with various last step.
21:20 bacek e.g. PAST->(POST|C|Whatever)
21:20 whiteknight joined #parrot
21:21 tcurtis whiteknight: I have an idea of a possible use for non-destructive optimization. Remember that debugging annotation idea you suggested to me, where there would be an optimization pass to remove any nodes that were annotated as being for debugging purposes?
21:22 whiteknight yes
21:22 whiteknight but removing those things would be the very definition of "destructive"
21:22 whiteknight really, any change you make to the tree that causes some information to be lost is destructive
21:23 cotto_work clearly my brain isn't working as well as it thinks it is
21:24 whiteknight tcurtis: read your email, by the way. nice
21:27 tcurtis whiteknight: suppose a HLL allows you(possibly only at a REPL or with a special option turned on) to toggle removal of debugging statements on and off at runtime through some global variable or something. They'd need a copy of the un-optimized PAST for the $DEBUGGING_GLOBAL_VAR == true case. Or something similar.
21:29 whiteknight tcurtis: in that case, you wouldn't want the optmizer involved at all, I don't think. Just start off all your debugging functions with "if $DEBUGGING_GLOBL_VAR) ..."
21:30 whiteknight pruning the expression tree and then de-pruning it again at runtime doesn't have many benefits that I can think of, and adds far too much complexity to the optimizer and creates huge memory bloat
21:33 whiteknight plus, if we talk about chaining optimization stages, or even duplicating stages, are we going to copy the tree at each point, or just a single "before" and "after"?
21:33 particle that sounds like a configurable item
21:33 whiteknight in the later, if you turn off one optimization at runtime, you turn them all off
21:34 particle for example, some integrated development environments may want to keep the various stages for incredibly powerful debugging
21:35 particle something like visually diffing ast's before and after optimizations, with profile information
21:35 particle *profiling
21:36 particle some folks will want comments and whitespace carried along with the ASTs
21:36 tcurtis whiteknight: I'm not suggesting copy-by-default. I'm just suggesting that individual languages or optimizations might like to keep a copy for some purpose. Admittedly, I doubt that it would be anywhere near the best way to implement that particular usage.
21:37 particle one could implement a genetic optimizer
21:37 particle there are many reasons to allow highly-configurable copying
21:38 particle whether it's implemented or not with your soc project, it should not be made impossible
21:38 tcurtis particle++, I was about to mention something like that in a much less succinct way.
21:38 pmichaud tcurtis: I obviously have a lot of comments regarding past and optimization.  I haven't found a good place to chime in (or the right time)
21:39 pmichaud (and my head has been in other spaces lately)
21:39 particle pull it out of there.
21:39 pmichaud particle: well, some of those other spaces are non-maskable.  and some have deadlines associated with them (e.g., Rakudo *)
21:40 particle yes, my head has been in other spaces, too.  i haven't contributed anything but discussion for the last several months
21:41 pmichaud ("non-maskable" == "non-maskable interrupts", for those who missed the reference)
21:43 pmichaud tcurtis: perhaps we should set up a time to chat and I can brain-dump some thoughts?
21:44 Mokurai joined #parrot
21:44 tcurtis pmichaud: that would be excellent. When might be suitable for you?
21:44 whiteknight particle: a genetic optimizer could certainly use the feature, but I doubt that would be the only or even the best way to use it
21:45 whiteknight pmichaud: no huge rush we have tcurtis as our <s>slave labor</s> GSOC student for a few more months
21:47 pmichaud tcurtis: what tz are you in, and do you prefer evenings or mornings?
21:47 * bacek want a pon^W multi in nqp :)
21:49 bacek pmichaud,  ^^^^^^^ :)
21:49 whiteknight multi in nqp sounds quite nice. I wonder how hard it would be to implement
21:50 pmichaud whiteknight: not too difficult, if you're willing to accept parrot multi semantics :)
21:51 whiteknight pmichaud: haha, you know that begs the question "what do you wish Parrot's multi semantics were?"
21:51 whiteknight I am all about the satisfied customer
21:51 pmichaud anyway, I think we'll likely do multis relatively soon.  bacek++ has been putting together some good patches.
21:51 tcurtis pmichaud: CDT or UTC-5 from March to November, CST(UTC-6) for the rest of the year. I prefer evenings, generally, although since I don't have a non-GSoC job currently, I'm essentially free from when I wake up(between 9 and 11 CDT) to when I go to bed, barring specific plans.
21:52 pmichaud I'm CDT also.  evenings tend to work better for me too.
21:52 pmichaud (at least for planned events)
21:52 pmichaud let me check with $wife what is happening tomorrow and Fri evening  (tonight won't work for me)
21:53 pmichaud "evening" as in 6ish?  8ish?  10ish?
21:53 whiteknight pmichaud: how is $wife doing?
21:54 pmichaud whiteknight: the rollercoaster continues
21:54 pmichaud whiteknight: tomorrow is almost certain to be another downswing, one way or another.
21:55 pmichaud I did get some excellent coding in this week, but tomorrow (until at least the evening) may be very iffy
21:56 tcurtis Any of those would work for me tomorrow. 8 or 10 slightly preferable because 6 is normally is mealtime for me. I probably won't be free after 4:30 or so Friday, though.
21:58 pmichaud 8 or 10 is better for me too
21:58 pmichaud before 8 tends to be kid time
22:00 pmichaud tcurtis: I'll plan to be online tomorrow starting around 8 to 8:30ish
22:00 tcurtis pmichaud: okay. I will as well.
22:01 pmichaud (describing stuff on irc will be much easier than writing an email :-)
22:05 sorear pmichaud: Do you know what's up with the weekly meetings?  No new minutes have been posted since the tenth
22:05 pmichaud sorear: we're just behind a bit on the minutes, I think
22:05 pmichaud I missed last week's meeting altogether
22:05 pmichaud it happens from time-to-time
22:06 pmichaud (delays in meeting minutes)
22:06 pmichaud but I'll let chromatic know there's at least one person eager to see them each week.  that might get him to whip them out a little quicker :)
22:08 whiteknight I read them, but mostly to see if my name is mentioned :)
22:25 Khisanth joined #parrot
22:35 elmex_ joined #parrot
22:48 davidfetter joined #parrot
22:56 dalek parrot: r47026 | whiteknight++ | branches/ns_func_cleanup:
22:56 dalek parrot: creating a branch to start some cleanups in src/global.c
22:56 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47026/
22:59 Psyche^ joined #parrot
23:12 dalek parrot: r47027 | whiteknight++ | branches/ns_func_cleanup (12 files):
23:12 dalek parrot: Parrot_get_namespace_keyed* -> Parrot_ns_get_namespace_keyed*
23:12 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47027/
23:12 dalek parrot: r47028 | whiteknight++ | branches/ns_func_cleanup (11 files):
23:12 dalek parrot: Parrot_make_namespace_keyed* -> Parrot_ns_make_namespace_keyed*
23:12 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47028/
23:12 dalek parrot: r47029 | gerd++ | trunk/examples/languages/squaa​k/doc/tutorial_episode_1.pod:
23:12 dalek parrot: correct one little typo
23:12 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47029/
23:15 tetragon joined #parrot
23:29 dalek parrot: r47030 | whiteknight++ | branches/ns_func_cleanup (4 files):
23:29 dalek parrot: Parrot_make_namespace_autobase -> Parrot_ns_make_namespace_autobase
23:29 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47030/
23:29 dalek parrot: r47031 | whiteknight++ | branches/ns_func_cleanup (7 files):
23:29 dalek parrot: Parrot_[get|set]_global -> Parrot_ns_[get|set]_global
23:29 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47031/
23:42 dalek TT #1660 created by whiteknight++: Deprecate Parrot_find_global_s
23:42 dalek TT #1660: http://trac.parrot.org/parrot/ticket/1660
23:45 dalek parrot: r47032 | whiteknight++ | branches/ns_func_cleanup/DEPRECATED.pod:
23:45 dalek parrot: Add deprecation notice for a function in TT #1660
23:45 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47032/
23:45 dalek parrot: r47033 | whiteknight++ | branches/ns_func_cleanup (6 files):
23:45 dalek parrot: Parrot_find_global_curr -> Parrot_ns_find_current_namespace_global
23:45 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47033/
23:47 snarkyboojum joined #parrot
23:59 cotto_work whiteknight, your commit message for r47034 disagrees with your diff, fyi
23:59 whiteknight lolwut?
23:59 purl lolwut is probably the pic with that freaky giant pear with a mouth
23:59 whiteknight did I make a booboo?

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

Parrot | source cross referenced