Camelia, the Perl 6 bug

IRC log for #parrot, 2010-02-08

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:03 cotto Nice.  It does.
00:03 cotto Austin++
00:06 cotto Austin++
00:06 cotto Austin++
00:14 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32086), fulltest) at r43766 - Ubuntu 9.10 amd64 (g++ with --optimize)
00:15 cotto committed
00:15 cotto happy times
00:16 cotto even if I don't care about American football
00:16 Whiteknight I watch the game but don't really care about it too much
00:16 Whiteknight more a criticism of sunday-night television than it is a support of american football
00:22 dalek parrot: r43767 | cotto++ | trunk (3 files):
00:22 dalek parrot: [Austin++] Make pbc_merge also update argument name string constants, plus a test.
00:22 dalek parrot: This fixes TT #1425 and TT #1419.  Patch courtesy of Austin++.
00:22 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43767/
00:30 Whiteknight Urg, packfiles.
00:30 Whiteknight I really can't wait until they're all re-written as PMCs
00:31 Coke msg andy my PB login is apparently fubar, so I cannot just comment on the article; the bobby tables sample for coldfusion is better written as <cfqueryparam   > - no closing tag is necessary.
00:31 purl Message for andy stored.
00:33 Coke cotto: if we didn't install things that didn't work, would we install anything?
00:34 ash_ joined #parrot
00:36 cotto moot point now
00:36 ash_ NotFound: ping?
00:36 NotFound ash_: pong
00:37 ash_ i was just looking at winxed, and am curious, is there a reason you hand made your lexer and parser?
00:38 NotFound ash_: just personal preference for hand made recursive descent parsers.
00:39 dalek parrot: r43768 | cotto++ | trunk/compilers/pct/Rules.mak:
00:39 dalek parrot: [pct] move PCT to a two-step compilation process to simplify dependency
00:39 NotFound Also, that way is easier (to me) to have a C++ and a self hosted version.
00:39 dalek parrot: tracking and exercise pbc_merge
00:39 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43768/
00:40 ash_ I have been learning bison/flex recently, so i was just curious if there was a reason you didn't use a tool like that, but C++ compatibility is an issue with both of them
00:41 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32087), fulltest) at r43767 - Ubuntu 9.10 amd64 (gcc with --optimize)
00:41 NotFound ash_: some time ago I wrote a Z80 assembler with hand made parsing. Later, I wrote another version with bison. After that experience, I decide that I like better hand made ones.
00:42 NotFound If you are curious: http://pasmo.speccy.org/
00:42 dalek TT #1419 closed by cotto++: pbc_merge doesn't correctly merge
00:42 dalek TT #1425 closed by cotto++: pbc_merge does not correctly adjust constant offsets for parameter names
00:44 ash_ i have been reading the compilers principles techniques and tools book, and it talks about yacc and lex for parsing in its examples, its probably the main reason i am trying to learn bison and flex
00:44 ash_ NotFound: cool, i'll go look at that, just to see how you parse stuff in that
00:44 NotFound ash_: Have you readed "The Design and Evolution of C++" ?
00:45 ash_ not yet, but if you'd recommend it i'll add it to my list of things to read, i was also recommended "Crafting A Compiler"
00:46 NotFound Stroustrup says that people convinced him to write Cfront with yacc et al, and in retrospective he thinks that his initial idea of using recursive descent was better.
00:46 NotFound So I'm in a very godd company with my taste ;)
00:46 NotFound s/godd/good
00:48 NotFound Of course I recommend it, it has lots of recommendations for language designers.
00:49 NotFound And implementors.
00:50 ash_ hmm, okay, i'll definately look into it once i finish these two books
00:50 NotFound BTW, finally I've bought the Nokia N900 :)
00:51 NotFound Is a very nice toy^hol
00:53 NotFound And will be better when we have parrot running on it.
00:54 abqar joined #parrot
00:57 Whiteknight everytime i look into the packfile code my head hurts and i get bored and angry
00:57 Whiteknight horrible, horrible
00:58 Whiteknight plus, extremely light on the documentation, which complex code needs more of
00:59 kurahaupo1 joined #parrot
01:00 mikehh cotto: compilers/pct/src/ - POST/Node.pbc && Compiler.pbc, PAST/Node.pbc && Compiler.pbc, and PCT/Node.pbc are not under svn control or make clean/realclean
01:07 * Coke reminds the perl community about the 2010Q1 grant proposals:
01:07 Coke http://news.perlfoundation.org/20​10/02/2010q1_grant_proposals.html
01:07 Coke (the grants committee definitely reads your comments.)
01:10 Whiteknight rurban doesn't seem to have a good opinion of parrot
01:11 Coke ... given how he left the project, that doesn't surprise me at all.
01:12 Coke It does make me wonder if he knows who is on the grants committee.
01:13 Whiteknight no. but to mention it in an unrelated grant proposal...
01:15 plobsing pge people: why does data_json not work when invoked using "./parrot compilers/data_json/data_json.pbc --target=pir file.json", but does work when used from other PIR programs using compreg?
01:15 Coke The perl6 grant request might also be of interest to folks here.
01:16 Coke not every compiler uses PCT?
01:16 Coke me checks data_json; all the renames years ago confuse him still
01:19 Coke I don't see that it's using HLLCompiler, plobsing.
01:19 Whiteknight the David Mitchell proposal is impressive
01:19 Coke that one's not in the list. URL?
01:21 Whiteknight http://news.perlfoundation.org/2010/0​2/grant_proposal_fixing_perl5_co.html
01:21 Whiteknight i think its  a diffent type of grant
01:23 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32088), fulltest) at r43768 - Ubuntu 9.10 amd64 (g++ with --optimize)
01:23 * mikehh ok enough - need some sleep
01:30 kid51 joined #parrot
01:52 JimmyZ joined #parrot
02:15 TonyC joined #parrot
02:16 nopaste joined #parrot
02:24 eternaleye joined #parrot
02:54 dalek TT #1304 closed by coke++: Parrotcode.org docs show up before parrot.org on web searches
02:54 dalek TT #350 closed by coke++: Move wiki pages over from tpf to trac
02:54 dalek TT #776 closed by coke++: segfault in Parrot_str_equal...
02:55 cotto good day for closing tickets
03:02 Coke adding svn:ignore -> regen MANIFEST.SKIP
03:02 cotto I already did that. ;)
03:04 plobsing on how many pbcs would a nqp-rx based grammar depend?
03:07 dalek parrot: r43769 | cotto++ | trunk/compilers/pct (4 files):
03:07 dalek parrot: [build] add some svn:ignore properties and remove intermediate pbc files when cleaning
03:07 dalek parrot: mikehh++ for noticing
03:07 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43769/
03:07 dalek parrot: r43770 | cotto++ | trunk/MANIFEST.SKIP:
03:07 dalek parrot: [MANIFEST] manifest update
03:07 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43770/
03:09 cotto plobsing, depends on the grammar
03:09 plobsing maybe it would help if I describe my problem in more detail
03:10 plobsing I want to get tools/build/nativecall.pir to run from miniparrot (so that it can build parrot's nci interface)
03:10 plobsing It is table driven, currently using data_json.
03:11 plobsing but I don't want to have to generate all of TGE, PGE, and data_json from miniparrot
03:11 plobsing and I'd heard that nqp-rx output PIR that was mostly self-sufficient
03:11 cotto it is
03:12 plobsing so I'm considering writing a JSON parser with nqp-rx
03:13 plobsing of course, it'll only be worth the effort if I can use this parser without having to generate half of parrot's runtime libraries
03:27 cotto What's the objection to just checking in the generated src/nci.c and requiring a full parrot build to regenerate it?
03:29 plobsing src/nci.c can change depending on what features parrot is built with
03:29 cotto I didn't know that.
03:31 plobsing opengl, for example, enables several hundred signatures
03:31 cotto Why not just enable everything?
03:33 plobsing because my ultimate goal is to get rid of those hundreds of mostly-unused signatures
03:33 plobsing also because I'm not in the mood for spelunking through parrot's configure system ATM
03:36 plobsing but I am experiencing some severe head-desk right now
03:42 janus joined #parrot
03:56 Austin_away Cotto: You updated my patch to add another line. Is that in the "Loop over arguments" section directly above?
04:00 Austin Does PaFo have a similar granting system?
04:01 Austin Cotto: Never mind. There's only one similar switch statement (at least in my version) so I assume that's where you added it.
04:02 eternaleye joined #parrot
04:34 cotto Austin, yup
04:53 hercynium joined #parrot
05:00 kurahaupo joined #parrot
05:23 mikehh joined #parrot
05:32 dalek tracwiki: v15 | cotto++ | CottoTasklist
05:32 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Cot​toTasklist?version=15&amp;action=diff
05:48 mikehh cotto: I got a manifest_tests error with MANIFEST.SKIP - Coke also had this problem - I think it was a leftover from one_make - svn ignore stuff not in trunk
05:48 cotto I committed a regenerated manifest earlier today.
05:49 mikehh yeah I know - that is why i commented on it
05:49 cotto oic
05:50 cotto It's odd that my checkout has goofy properties.  svn--
05:50 cotto karma svn
05:50 purl svn has karma of -7
05:50 cotto good
05:50 cotto svn-- for being super slow to sync too
05:51 dalek parrot: r43771 | mikehh++ | trunk/MANIFEST.SKIP:
05:51 dalek parrot: Re-generate MANIFEST.SKIP
05:51 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43771/
05:51 mikehh I tend to do branch testing in different directories so I didn't get the problem
05:53 mikehh the problem being that svn ignore is ignored by svn so if that changes it still ignores it
05:53 mikehh I also tend to do a clean checkout fairly regularly
06:07 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32090), fulltest) at r43771 - Ubuntu 9.10 amd64 (gcc with --optimize)
06:45 Austin W00t!!! My class has parents.
06:46 Austin Go me.
06:59 Austin What the name for the kind of function passed to sort (e.g.) that compares items?
07:02 cotto comparator?
07:02 * cotto wonders if he's making up words
07:02 cotto again
07:02 Austin That's the conclusion I've been reaching.
07:03 Austin sort( comparator: &f )
07:03 Austin That works.
07:03 Austin sort( compare: &f )
07:03 Austin That also works.
07:04 * cotto really wants it to be git time.
07:07 cotto not that an hour to sync a branch with trunk is unreasonable or anything
07:08 uniejo joined #parrot
07:10 cotto epic sync commit incoming
07:11 kurahaupo ping svn.parrot.org => rtt min/avg/max/mdev = 211.832/212.957/216.780/1.703 ms
07:11 kurahaupo ping parrot.github.com => rtt min/avg/max/mdev = 259.936/260.792/262.497/1.024 ms
07:11 kurahaupo svn++
07:15 dalek parrot: r43772 | cotto++ | branches/ops_pct (1110 files):
07:15 dalek parrot: [opsc] sync branch with trunk, most likely breaking something
07:15 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43772/
07:35 eternaleye_ joined #parrot
07:55 fperrad joined #parrot
08:05 iblechbot joined #parrot
08:17 lucian joined #parrot
08:36 Austin Woot. Generated accessors.
08:37 dalek parrot: r43773 | cotto++ | branches/ops_pct (12 files):
08:37 dalek parrot: [opsc] convert opsc to nqp-rx and non-recursive makefiles
08:37 dalek parrot: opsc builds, tests still need work
08:37 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43773/
08:41 cotto sounds shiny
08:42 Austin It's pretty freaking sweet if you're using nqp.
08:42 Austin class X { use('P6object');   has( 'foo' );         method blah() { self.foo(1);      say(self.foo); } }
08:47 nopaste "cotto" at 96.26.227.153 pasted "I don't see any accessors" (62 lines) at http://nopaste.snit.ch/19514
08:48 cotto nm.  I suppose they could be generated by the 'has' method
08:48 Austin Heh. You have to be using Kakapo, which extends the P6object suite and compiles the accessors in memory at run time.
08:49 Austin You know anything about P6 "try" syntax?
08:49 cotto not at this point in the evening
08:49 Austin -rx has a try {...} syntax, but what are the rest of the keywords?
08:52 Austin looks like    try {       CATCH {...}    .... }
09:00 * kurahaupo wonders if he should give up converting tests to PIR and go for NQP instead?
09:01 Austin Only if you're sure you can write them in nqp.
09:02 Austin I wonder what's the difference between "set $P27['type'], 58" and "setattribute $P27, 'payload', $P28" ?
09:04 particle joined #parrot
09:05 Austin Hmm.. What's the right syntax for "throw" ?
09:06 kurahaupo Austin, who chose the name "Kakapo"? Is there someone else from NZ lurking around here?
09:06 Austin Nope, that was me.
09:06 Austin A buddy of mine sent me a video titled "You are being shagged by a rare parrot"
09:07 Austin (Which pretty much describes my experience as an HLL developer)
09:07 kurahaupo grep parrot /world/avians | sort --size | tail -1
09:07 Austin :)
09:07 kurahaupo s/size/weight/
09:08 Austin Part of the video - before the bird got dangerously frisky - included the comment that the kakapo looks old-fashioned.
09:08 Austin Since the Kakapo library is old-fashioned (compile, link, run) I figured it was a natural.
09:23 Austin Okay, http://www.youtube.com/watch?v=9T1vfsHYiKY
09:23 barney joined #parrot
09:25 bacek joined #parrot
09:26 cotto hio bacek.  Feel like getting back into opsc?
09:27 bacek cotto, I want to finish gc_encapsulate first (and properly plug Boehm after). It's crucial for Rakudo*
09:28 cotto sure
09:28 particle joined #parrot
09:33 bacek seen whiteknight
09:33 purl whiteknight was last seen on #parrot 8 hours, 12 minutes and 19 seconds ago, saying: i think its  a diffent type of grant
09:34 bacek time to wake up...
10:01 ruoso joined #parrot
10:02 eternaleye joined #parrot
10:58 bacek coverage?
10:58 purl coverage is probably http://cv.perl6.cz
11:04 dalek parrot: r43774 | bacek++ | trunk/src/ops (2 files):
11:04 dalek parrot: Revert r43414. RetContinuation doesn't kill itself anymore after invoke.
11:04 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43774/
11:11 dalek TT #1427 created by bacek++: [DEPRECATED] RetContinuation PMC
11:13 eternaleye joined #parrot
11:44 payload1 joined #parrot
11:50 bluescreen joined #parrot
12:26 dalek parrot: r43775 | bacek++ | branches/gc_encapsulate_part1:
12:26 dalek parrot: Split gc_encapsulate into smaller steps. Apparently I'm not able to find out why whole lot failing
12:26 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43775/
12:32 ruoso joined #parrot
12:42 kid51 joined #parrot
12:52 payload joined #parrot
12:56 bluescreen joined #parrot
12:59 dalek parrot: r43776 | bacek++ | branches/gc_encapsulate_part1 (5 files):
12:59 dalek parrot: Make callback functions accept explicit Memory_Pools*.
12:59 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43776/
13:13 kid51 Yippee:  Smolder just managed to accept two reports and properly report back results without a timeout
13:14 * kid51 wonders how long that will last :=)
13:15 dalek parrot: r43777 | bacek++ | branches/gc_encapsulate_part1/src/gc/mark_sweep.c:
13:15 dalek parrot: Update free_buffer to accept Memory_Pools*
13:15 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43777/
13:15 dalek parrot: r43778 | bacek++ | branches/gc_encapsulate_part1/src/gc (3 files):
13:15 dalek parrot: Use explicit Memory_Pools* in alloc_resources.c
13:15 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43778/
13:19 whiteknight joined #parrot
13:21 kid51 left #parrot
13:29 mj41 joined #parrot
13:31 payload joined #parrot
13:32 dalek parrot: r43779 | bacek++ | branches/gc_encapsulate_part1/src/gc (4 files):
13:32 dalek parrot: Switch mark_sweep functions to use explicit Memory_Pools*.
13:32 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43779/
13:32 dalek parrot: r43780 | bacek++ | branches/gc_encapsulate_part1/src/gc (2 files):
13:32 dalek parrot: Remove outdated function declarations. src/gc/pools.c was removed long
13:32 dalek parrot: time ago. Time to remove dangling foward declarations.
13:32 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43780/
13:42 tetragon joined #parrot
13:48 dalek parrot: r43781 | bacek++ | branches/gc_encapsulate_part1/src/gc (3 files):
13:48 dalek parrot: Switch src/gc/system.c to use explicit Memory_Pools*
13:48 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43781/
13:48 dalek parrot: r43782 | bacek++ | branches/gc_encapsulate_part1 (3 files):
13:48 dalek parrot: Remove unused, non-public Parrot_gc_ptr_in_memory_pool function
13:48 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43782/
13:50 mj41 joined #parrot
14:00 bacek whiteknight, aloha!
14:00 whiteknight hello bacek!
14:00 whiteknight I see you've been busy today
14:00 bacek tonight... It's 1am already
14:01 bacek do you have some box which is not linux/i386 to quick test gc_encapsulate_part1?
14:05 dalek parrot: r43783 | bacek++ | branches/gc_encapsulate_pa​rt1/src/interp/inter_cb.c:
14:05 dalek parrot: Copy logic of finding Interpeter for callback from gc_encapsulate branch
14:05 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43783/
14:05 dalek parrot: r43784 | bacek++ | branches/gc_encapsulate_part1 (2 files):
14:05 dalek parrot: Remove unused, non-public Parrot_gc_ptr_is_pmc function
14:06 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43784/
14:13 iblechbot joined #parrot
14:22 dalek parrot: r43785 | bacek++ | branches/gc_encapsulate_part1 (2 files):
14:22 dalek parrot: Remove unused, non-public Parrot_gc_get_pmc_index function
14:22 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43785/
14:22 dalek parrot: r43786 | bacek++ | branches/gc_encapsulate_part1/src/gc/api.c:
14:22 dalek parrot: Reorganize functions in src/gc/api.c to move attributes allocation close
14:22 dalek parrot: to other allocation functions. No functional changes.
14:22 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43786/
14:38 lucian joined #parrot
14:43 Coke I can answer the question about PaFo's grants if it's still on the table.
14:57 kurahaupo joined #parrot
15:06 PacoLinux joined #parrot
15:25 Psyche^ joined #parrot
15:28 bubaflub joined #parrot
15:46 Coke There is the possibility to do grants via the PaFo, but: 1) we do not have a steady stream of donations, corporate or otherwise at the moment. 2) we don't have a grants committee, nor a formal grant process; IIRC, it would basically be the board voting on disbursement.
15:54 Coke bacek: testing your part1 branch on strawberry now.
15:55 Coke bacek: all tests pass. (but strawberry using ming under the covers, so this may not be as interesting a platform as you might care about.)
16:03 leto joined #parrot
16:04 cotto_w0rk joined #parrot
16:04 NotFound pbc_merge is breaking c++ build
16:05 NotFound Maybe only in 64 bits
16:11 wagle_ joined #parrot
16:15 zostay joined #parrot
16:17 bacek joined #parrot
16:18 NotFound Who is working in pbc_merge?
16:19 Coke you can tell with the branch status tool!
16:19 Coke (but I think cotto)
16:20 NotFound It's comparing a char with 0x200
16:22 knewt joined #parrot
16:23 khisanth_ joined #parrot
16:23 purl joined #parrot
16:23 workbench joined #parrot
16:23 Coke trac--
16:23 Coke trac--
16:24 Coke NotFound: oh, this is on trunk.
16:24 Coke ?
16:24 Coke yah, prolly cotto.
16:24 Coke I have regen'd http://trac.parrot.org/parr​ot/wiki/BranchDescriptions though. =-)
16:25 jjore joined #parrot
16:26 khisanth_ joined #parrot
16:26 purl joined #parrot
16:26 workbench joined #parrot
16:30 NotFound Coke: NM, cretaing a ticket.
16:30 plobsing joined #parrot
16:31 Coke NotFound++
16:33 NotFound Why trac says that the ticket is owned by dukeleto?
16:35 dalek parrot: r43787 | mikehh++ | branches/gc_encapsulate_part1/src/gc/api.c:
16:35 dalek parrot: fix codetest failure - line length
16:35 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43787/
16:37 khisanth_ joined #parrot
16:38 workbench joined #parrot
16:39 dalek tracwiki: v25 | coke++ | BranchDescriptions
16:39 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Branc​hDescriptions?version=25&amp;action=diff
16:42 whiteknight bacek: I'll test first on OpenSolaris/x86, then Win32/x86. I have to do x64 testing at home tonight
16:44 dalek TT #1428 created by NotFound++: pbc_merge breaks c++ build
16:45 whiteknight fperrad: ping
16:47 mikehh gc_encapsulate_part1 branch:
16:47 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32106), fulltest) at r43787 - Ubuntu 9.10 amd64 (g++ with --optimize)
16:47 whiteknight purl msg fperrad How do I use setup.pir to create a dynops library? Do you have an example of that for me to see?
16:53 theory joined #parrot
17:02 whiteknight bacek: do we need to run all tests, or is coretest sufficient?
17:09 mikehh NotFound: it builds for me at r43787 on Ubuntu 9.10 amd64 g++ with --optimize
17:09 whiteknight coretest is good on opensolaris/x86  (minus the tests we fail there anyway)
17:10 mikehh joined #parrot
17:13 NotFound mikehh: fails for me with or without optimize
17:14 mikehh NotFound: let me try without --optimize
17:16 fperrad pong Whiteknight
17:16 fperrad see http://github.com/TiMBuS/fun/blob/master/setup.pir
17:16 fperrad or http://github.com/fperrad/wml​script/blob/master/setup.pir
17:16 whiteknight fperrad++
17:16 NotFound gcc version 4.3.2 (Debian 4.3.2-1.1)
17:18 whiteknight fperrad: is it possible to use multiple .ops files together?
17:24 fperrad Whiteknight, it's a hash, so you could create several dynops (but one .ops file produces one dynops shared lib)
17:24 whiteknight gotcha
17:26 kurahaupo1 joined #parrot
17:27 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32107), fulltest) at r43787 - Ubuntu 9.10 amd64 (g++ with --optimize)
17:29 mikehh NotFound: g++ (Ubuntu 4.4.1-4ubuntu9) 4.4.1
17:33 NotFound Bizarre
17:35 mikehh I get:
17:35 mikehh src/pbc_merge.c: In function ‘void pbc_merge_ctpointers(parrot_interp_t*, pbc_merge_input**, int, PackFile_ByteCode*)’:
17:35 mikehh src/pbc_merge.c:664: warning: case label value exceeds maximum value for type
17:35 mikehh but it doesn't generate an error
17:35 mikehh generate
17:36 NotFound Anyway, is a bug.
17:38 NotFound Don't know what is the correct result according the standard, but the code is wrong. If it can handle that value, having it is confusing. If it must handle it but can't, is a serious problem.
17:39 mikehh NotFound: yes - when fixed we need a test to catch it - All tests PASS at the moment
17:39 Coke NotFound: if you have any input on my just-now mail to -ev, I'd appreciate it.
17:39 Coke er, -dev
17:39 NotFound Coke: looking
17:40 Coke basically trying to avoid as much work as possible. =-)
17:42 Andy joined #parrot
17:42 Coke (if we need to support the full range of features we have no, I can cope.)
17:42 NotFound Last message I have is from fperrad
17:43 Coke "eliminating CFLAGS"
17:48 cotto_working joined #parrot
17:49 dukeleto 'ello
17:53 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32109), fulltest) at r43787 - Ubuntu 9.10 amd64 (g++)
17:57 whiteknight what does ** mean in P6 regex?
17:58 Coke http://feather.perl6.nl/syn/S05.html (there's a whole section starting with "the general repetition specifier is now **"...
17:59 Coke (it depends on what comes after it.)
17:59 Coke number or range == count of matches, literal or subrule, it's a separator.
17:59 whiteknight [ <statement> | <?> ] ** ';'
17:59 whiteknight I'm trying to figure out what that means
18:00 Coke there, it's a separator.
18:00 bacek joined #parrot
18:00 Coke so it's any number of statements, separated by a semicolon.
18:00 whiteknight okay
18:00 whiteknight so what is <?>
18:00 Coke whiteknight: also in S05 - match a null string
18:00 whiteknight gotcha
18:01 whiteknight Coke++
18:01 Coke whee!\
18:01 Coke (I keep mentioning S05 because that's wehere I'm looking this up right before I talk. =-)
18:01 whiteknight ok
18:02 Coke I still don't remember any of the new syntax for more than 10m. =-)
18:03 mikehh joined #parrot
18:08 Coke http://bobby-tables.com/, Andy++
18:08 NotFound Coke: don't get that message yet
18:08 Coke NotFound: odd.
18:08 Andy Thanks.
18:09 Coke ah well. I'm not going to write any code for a bit, so just reply when it shows up. =-)
18:09 NotFound Coke: maybe I'm too distracted playing with my new phone X-)
18:09 Coke Andy: I had to double check to make sure that it only said email when I emailed. =-)
18:10 NotFound Now I understand why they are called smartphones. It won me at chess.
18:11 * Coke ponders writing a fluxx game in some HLL targetting parrot.
18:11 payload joined #parrot
18:18 mikehh joined #parrot
18:20 Coke NotFound: the thrust was: do we really need all the exceptions in ./CFLAGS
18:20 Coke (because I can avoid having to recreate all this work at config time, awesome.)
18:21 Coke *if I..
18:25 mikehh_ joined #parrot
18:25 NotFound Coke: I'd like to kill all that shit. Unfortunately, bison and flex generated code will generate lots of warnings without it.
18:26 Coke right. one my questions was whether a clean build was worth it.
18:26 Coke whoa. tried to find out out who added the not about tsq never optimizing. it was brentax in r5583
18:27 Coke (when CFLAGS.in was created)
18:30 NotFound I like to have a clean build. And it avoids scaring new people at his first build.
18:31 NotFound Worth the price of a few seconds.
18:45 mikehh joined #parrot
18:46 Austin Coke: (re: grants) thanks
18:55 Coke That reminds me, I should consider hitting up my consulting company to donate a few grand.
18:56 Coke (might be easier once we are a not-for-profit)
19:03 Coke whiteknight: danke.
19:04 whiteknight we aren't not-for-profit?
19:06 Coke we are on the path. there's paperwork needs fillin'
19:06 Coke (plus it costs $$ to file)
19:24 ash_ joined #parrot
19:38 Khisanth joined #parrot
19:50 jan_ joined #parrot
19:51 kurahaupo joined #parrot
19:51 cotto_working includes some thoughts on llvm: http://blog.paulbiggar.com/archive/a-rant-about-p​hp-compilers-in-general-and-hiphop-in-particular/
20:05 joeri joined #parrot
20:07 ash_ cotto: interesting read, i mostly read the stuff linked to by the unladen swallow developers
20:10 mikehh joined #parrot
20:20 payload joined #parrot
20:39 kurahaupo1 joined #parrot
20:39 bubaflub left #parrot
20:39 purl joined #parrot
20:43 purl joined #parrot
20:44 purl joined #parrot
20:45 bacek o hai
20:52 mikehh hi bacek
20:52 dukeleto o hai
20:54 whiteknight bacek: I have a present for you: http://github.com/Whiteknight/parrot-hack​er-tools/blob/master/src/pmc/pmchack.pmc
20:54 whiteknight take a look at the method "force_gc_destroy" at the bottom, for extra fun
20:57 * whiteknight was looking for ways to be evil today
20:57 cotto_working Nice PMC there, but I notice it's missing the "segfault" method.
20:57 whiteknight cotto_working: I have plenty more methods to add still
20:57 cotto_working I'm sure you do.
20:57 whiteknight and I have a few more PMC types to add too
20:58 cotto_working Do you have an intended use for all that evil?
20:59 whiteknight cotto_working: yes and no.
20:59 cotto_working purl, yes and no
20:59 purl Go not to the mst for counsel, for he shall say both no and yes. also, breadsticks.
20:59 whiteknight I don't have a particular use for it, just idle curiosity
21:00 bluescreen joined #parrot
21:00 whiteknight would really help with some types of debugging to be able to access internal-only data from PIR programs
21:01 whiteknight plus in a Lorito world, we're going to need to provide much lower access to the API, so I want to explore that
21:02 whiteknight One thing I really wanted to explore was indirect op dispatch
21:02 whiteknight Imagine one op that took an opnumber as it's first parameter and a variadic argument list as it's second
21:03 whiteknight or, better yet, an ability to put in a op name and get an opnumber in return. Useful for PBC compilers written completely in PIR
21:05 dukeleto whiteknight: sounds cool
21:06 cotto_working You could hack up an OpInfo PMC to be an interface to interp->op_info
21:06 whiteknight cotto_working: exactly. I'm planning to prototype that
21:06 cotto_working That sounds shiny.
21:06 whiteknight Personally, I'm of the smalltalk school of thought: everything should be an object
21:06 cotto_working whiteknight++
21:07 bluescreen joined #parrot
21:07 whiteknight which has the benefit that if most of our ops become objects or methods on objects, out opset shrinks very naturally, and we can focus all optimization on the method dispatch hotpath
21:07 whiteknight but this is pie-in-the-sky rambling
21:10 cotto_working That's the second best kind of pie.
21:19 NotFound whiteknight: also, makes far easy to acess low-level parrot from object oriented HLLs. You just need to instantiate some object and call methods on it.
21:20 dukeleto cotto_working: key lime pie is first best, no?
21:20 cotto_working That is the subject of intense and divisive debate.  I think it's best not to get into it here.
21:20 mikehh_ joined #parrot
21:21 dukeleto cotto_working: you are correct. everyone's happiness pie is different
21:32 mikehh_ joined #parrot
21:34 kurahaupo1 joined #parrot
21:40 mikehh_ joined #parrot
21:44 Austin "I'm on a horse."  - This commercial should have been on the Superbowl.
21:44 Austin http://www.youtube.com/watch?v=owGykVbfgUE
21:46 mikehh_ joined #parrot
21:54 cognominal joined #parrot
21:55 Infinoid Austin: hah, awesome
21:55 Austin "I'm on a horse."
21:55 Infinoid That is a compelling argument.
21:55 cotto_working I don't want to smell girly.  Please tell me what I can buy to prevent that.
21:57 Austin "It's an oyster with two tickets to that thing you love"
22:17 mikehh__ joined #parrot
22:17 ash_ "I'am on a horse" is probably the best part IMO
22:25 mikehh joined #parrot
22:29 Austin whiteknight: You may appreciate this, being in the embedded biz: http://news.discovery.com/tech/t​oyota-recall-software-code.html
23:01 DrJoel joined #parrot
23:03 cotto_working joined #parrot
23:07 Whiteknight joined #parrot
23:11 patspam joined #parrot
23:19 cotto_w0rk joined #parrot
23:21 kid51 joined #parrot
23:38 cotto_w0rk joined #parrot
23:44 plobsing joined #parrot
23:53 zibri joined #parrot
23:57 tetragon joined #parrot
23:58 hercynium joined #parrot

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

Parrot | source cross referenced