Camelia, the Perl 6 bug

IRC log for #parrot, 2010-08-04

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 * cotto_work wonders how that got there in the first place
00:01 tcurtis cotto_work: src/namespace.c used to be src/global.c, I think.
00:02 tcurtis chromatic: I'll write up a blog post and then get to work on that.
00:04 ruoso joined #parrot
00:09 Austin How bizarre. Is the nopast-er not working?
00:09 cotto_work the bot is broken
00:09 Austin Anyway:  http://nopaste.snit.ch/22543
00:09 Austin Where's IMCC?
00:09 purl IMCC is the c of parrot or the problem or the intermediate code compiler or will be murdered soon or "fun" to hack on
00:09 sorear we kicked the bot after it became a spam vector
00:10 cotto_work Austin: that actually means that imcc couldn't find the file referred to by that line of code
00:10 cotto_work It's a pretty unhelpful message
00:10 cotto_work or misleading
00:11 cotto_work It's saying it can't find interpinfo.pasm
00:11 Austin So it probably means interpinfo.pasm, which is included at 198, right?
00:12 mikehh joined #parrot
00:12 Austin Okay.
00:13 Austin Hmm
00:13 Austin Given that parrot is not installed (this is on windows xp), how does imcc know where to search for stuff?
00:14 Tene does windows have an strace-like so you can see where it's looking?
00:15 Austin If it does, I don't know about it.
00:15 Austin The makefile does not seem to want to tell parrot -Ianything, so I'm assuming there's some built-in search logic. Is that true?
00:17 chromatic yes.
00:18 Austin Cool. Where's is that specified?
00:18 Austin grr
00:18 Austin *Where's is = Where is
00:20 Austin Adding -Iruntime/parrot/include in the makefile makes it work
00:21 Austin Hmm..until I try to compile dumper, and the whole thing starts again..
00:29 Austin Apparently, Configure.pl is not computing build_dir correctly.
00:29 Austin :(
00:30 Austin Coke: How do I specify a value for @build_dir@ in Configure?
00:37 bubaflub Austin: on windows machines in init/hints/mswin32.pm it's getting set
00:37 Austin bubaflub: Thanks, but my config_lib.pir has it set to null... :(
00:38 bubaflub Austin: hmmm.... lemme take a look. it *looks* like it's getting set there, but it could be overridden a number of different places
00:38 bubaflub i don't have a windows box to test it out, but lemme see what i can find
00:38 bubaflub Austin: do you have any errors configuring?
00:39 Austin I'm re-running configure with --ask now
00:41 Austin Have you ever built parrot on windows?
00:41 bubaflub Austin: no, i haven't had the pleasure
00:41 Austin Heh
00:41 bubaflub I'm working on getting out-of-directory building to work
00:41 Austin Yeah, I can't get shared library to work at all.
00:41 bubaflub and will eventually need to test it on Windows
00:42 Austin Your paste can be found at http://nopaste.snit.ch/22544
00:42 Austin There's a undef error
00:44 bubaflub hmmm, first of all that should cause the Configure script to bomb, not silently error
00:44 bubaflub can you run Configure.pl --verbose and no-paste the stuff?
00:44 bubaflub it'll be a lot, but it should be helpful
00:44 Austin What does that do?
00:45 Austin (IOW, do I have to hold its hand, or should I just let it run and pipe the output to nopaste?)
00:46 bubaflub --verbose should tell us more debug-esque information about what is being set during the configuration run
00:46 bubaflub and don't bother babysitting it; this step is just diagnostic so go ahead and pipe it to nopaste
00:51 Austin Your paste can be found at http://nopaste.snit.ch/22545
00:52 bubaflub Austin: looking now
00:53 Austin I don't see a single mention of build_dir until the error.
00:55 bubaflub Austin: yeah, me neither.  i thought it would have been set earlier.
00:55 bubaflub Austin: right towards the top the init::hints is loaded, that should set build_dir
00:56 Austin I wonder if the call to GetShortPathName is screwing things up.
00:56 bubaflub yeah, we can also go one level more verbose
00:56 bubaflub that'll give *tons* more output
00:57 bubaflub any item that is ever set in the config hash will be reported
00:57 Austin No, apparently build_dir is correct at that poiont.
00:58 bubaflub so then it's getting un-set somewhere else
00:58 Austin Well, it seems like (a) mswin32.pm is being run, and (b) build_dir is correct at that point.
00:59 bubaflub both init::defaults and init::hints::mswin32 should set 'build_dir'
00:59 bubaflub Austin: what revision are you running?
00:59 Austin Of?
00:59 bubaflub sorry, Parrot
00:59 Austin Trunk as of yesterday sometime.
00:59 bubaflub ok, just checking
00:59 Austin 48253
01:00 Austin No
01:00 Austin 48268
01:00 bubaflub oh, ok yeah that'll work
01:00 Austin So how do I get this config-setting-log to work?
01:01 bubaflub oh, lemme double check
01:01 Austin (As an aside: I would think that this sort of thing would be part of cotto_work's job requirements...)
01:01 bubaflub perl Configure.pl --verbose=2
01:02 Austin Whoa. That's pretty verbose.
01:02 bubaflub yeah, it'll output *every* time the config hash is updated
01:02 bubaflub and the steps that determine what headers you have and what warnings your compiler supports generate a lot of output
01:03 bubaflub cause they basically try to compile a test file and if it doesn't work, your compiler doesn't support it / you don't have that header
01:07 Austin That's weird.
01:07 Austin build_dir => undef
01:07 bubaflub ah! just as we suspected.  where is that getting set?
01:07 bubaflub init::defaults or init::hints?
01:08 Austin hints/win32
01:08 Austin I think it's the call to getshortpath
01:09 bubaflub yeah, try running that command on it's own
01:09 bubaflub i.e.
01:09 bubaflub perl -MWin32 -e "print Win32::GetShortPathName('C:/blahblahblah/');"
01:13 bubaflub Austin: and worst case scenario, try running Configure.pl with that line commented out and see if it's configures without errors
01:14 Austin Heh
01:14 Austin Heh
01:14 Austin Here's a funny:
01:14 Coke Austin: you don't specify /build_dir/, it's specified... when you build.
01:14 Austin If build_dir is set to "D:\My Documents\Sources\parrot-win" the function returns the same vaoue.
01:15 Austin But if build_dir is set to "D:/My\ Documents/Sources/parrot-win" then it fails.
01:15 Austin *vaoue=value
01:15 Austin Coke: getting there.
01:15 purl rumour has it getting there is the problem. or half the fun
01:15 Coke hokay.
01:20 cotto Austin, sadly my employer views PHP as a higher priority than Parrot.
01:20 Austin Heh
01:21 dalek parrot: r48304 | Chandon++ | branches/gsoc_threads (7 files):
01:21 dalek parrot: [gsoc_threads] Prep for working blocking I/O.
01:21 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48304/
01:25 bubaflub cotto: what about PHP on Parrot?
01:25 Coke cotto: ditto with me, except "Cold Fusion"
01:27 Austin So this mswin32 init code is getting build_dir like a/b/c, but it keeps using backslashes as directory separators. WTF?
01:27 Austin Is this an activestate vs strawberry perl-ism, or what?
01:28 Coke because backslash is the directory separator on windows?
01:28 Coke are you /setting/ build dir?
01:28 Austin That would be fine if it were consistent.
01:28 Austin Nope.
01:28 Austin I'm looking at the output of --verbose=2
01:30 Austin Your paste can be found at http://nopaste.snit.ch/22548
01:36 cotto currently just Zend PHP, though one coworker thinks we should roll our own.
01:38 kid51 joined #parrot
01:39 Austin If ever there was a language that MSFT *should* embrace and extend, PHP is it.
01:40 Austin Take it away from those idiots and make it useful.
01:41 cotto I'd love that.  I have no idea if it's got a chance though.  He's too busy to advocate it properly.
01:41 Austin Lol
01:41 Austin If you ever get a chance, take a look at the parser...
01:42 cotto why?
01:42 Austin Put's the ugh in ugly.
01:42 Austin *Puts
01:42 somebody_ joined #parrot
01:42 cotto That's par for the course.
01:55 somebody__ joined #parrot
01:58 Austin Is there a useful (& concise) TFM for me to R on building DLL's under Windows?
02:13 ash_ joined #parrot
02:23 * kid51 backscrolls
02:24 kid51 Austin:  If you are still having configuration problems, you can also call these options:
02:24 dalek website: tcurtis++ | PAST Optimization: plans for the final weeks of GSoC
02:24 dalek website: http://www.parrot.org/content/past-​optimization-plans-final-weeks-gsoc
02:24 kid51 --verbose-step=init::hints (and you can call multiple --verbose-step=some::step instances) ...
02:25 kid51 -- configure_trace (and then call perldoc lib/Parrot/Configure/Trace.pm to see how to get at that step-by-step data)
02:28 Austin kid51: Thanks, but I've figured out where the problem is. (Not to say I know how to fix it, though..)
02:48 kid51 Looks like Smolder is still down, just like last night:
02:48 kid51 Could not upload report to Smolder at http://smolder.plusthree.comHTTP CODE: 500 (Can't connect to smolder.plusthree.com:80 (connect: Connection refused))
02:58 bubaflub joined #parrot
03:00 janus joined #parrot
03:20 petdance joined #parrot
03:29 dafrito joined #parrot
03:43 Util purl msg tcurtis I just implemented your ticket report from #ps today. Feel free to tweak as needed, or to ask me to tweak it. Report {23} here: http://trac.parrot.org/parrot/report
03:43 purl Message for tcurtis stored.
03:48 jsut_ joined #parrot
03:57 tcurtis Util++
04:10 petdance In rakudo, the prime regex deosn't return for me.
04:10 petdance How can I try it in Parrot?  I don't know my Parrot at all.
04:22 darbelo joined #parrot
04:36 khairul joined #parrot
04:39 dalek parrot: r48305 | petdance++ | trunk/src/pmc/imageio.pmc:
04:39 dalek parrot: consting some pointers
04:39 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48305/
04:40 khairul cotto:ping
04:50 ash_ joined #parrot
04:55 darbelo joined #parrot
04:55 cotto khairul, pong
05:02 antivigilante joined #parrot
05:03 antivigilante where can I get the PIR/PASM source for a simple language
05:05 tcurtis antivigilante: I wouldn't recommend writing a compiler in pure PIR. Look at the Squaak compiler in examples/languages/squaak in the Parrot repository. It also includes a tutorial for implementing it.
05:05 cotto antivigilante, squaak might be what you're looking for.
05:06 cotto it's a simple language and includes a recently-updated 9-part tutorial.
05:07 tcurtis You can see the HTML version of the tutorial at http://docs.parrot.org/parrot/​latest/html/PCT_Tutorial.html , but there are a couple of typos there. They're fixed in the repository, and if you want to generate a HTML version of the latest revision of it, do "make html" in a Parrot svn checkout.
05:07 sorear Also, we don't recommend using PASM.  At all.  Ever.
05:10 TiMBuS joined #parrot
05:11 plobsing joined #parrot
05:13 dalek tracwiki: v10 | cotto++ | KhairulGSOC2010Schedule
05:13 dalek tracwiki: http://trac.parrot.org/parrot/wiki/KhairulG​SOC2010Schedule?version=10&action=diff
05:16 ruoso joined #parrot
05:29 dalek parrot: r48306 | khairul++ | branches/gsoc_instrument/runtime/parro​t/library/Instrument/EventLibrary.nqp:
05:29 dalek parrot: Added disable method of Instrument::Event::Object.
05:29 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48306/
05:44 bubaflub joined #parrot
05:46 dalek parrot: r48307 | khairul++ | branches/gsoc_instrument/src/d​ynpmc/instrumentinvokable.pmc:
05:46 dalek parrot: Expose call_sig and invocant to callback.
05:46 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48307/
06:26 davidfetter joined #parrot
06:33 bubaflub joined #parrot
06:42 fperrad joined #parrot
06:44 Casan joined #parrot
07:24 dalek parrot: r48308 | plobsing++ | trunk (4 files):
07:24 dalek parrot: Store PMC as a simpler "buf" type in PBC in stead of full-fledged strings.
07:24 dalek parrot: Slightly reduces PBC sizes and unpack consts.
07:24 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48308/
07:41 dalek parrot: r48309 | plobsing++ | trunk/t/native_pbc (7 files):
07:41 dalek parrot: native_pbc platform updates
07:41 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48309/
07:48 Psyche^ joined #parrot
07:50 robin-gvx joined #parrot
08:08 ambs joined #parrot
08:10 Coke joined #parrot
08:14 dalek parrot: r48310 | plobsing++ | trunk/src/packfile/pf_items.c:
08:14 dalek parrot: use arg assert macros
08:14 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48310/
08:36 perlite joined #parrot
08:50 perlite joined #parrot
09:02 ambs joined #parrot
09:04 dalek parrot: r48311 | cotto++ | trunk/compilers/pct/src/PCT/HLLCompiler.pir:
09:04 dalek parrot: make --stagestats a little less noisy
09:04 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48311/
09:07 dalek tracwiki: v28 | cotto++ | CottoTasklist
09:07 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Cot​toTasklist?version=28&action=diff
09:34 AndyA joined #parrot
09:50 TiMBuS joined #parrot
10:12 moritz good localtime()
10:12 moritz how do I control the umask of an mkdir?
10:16 sorear You can't.  mkdir always uses the current umask
10:16 moritz how do I set the current umask in parrot?
10:18 moritz or asked another way - why is the parrot umask not the one that I set in my .bashrc?
10:20 NotFound moritz: the value set depends on the umask and the flags used in the creation function.ç
10:21 NotFound METHOD mkdir(STRING *path, INTVAL mode)
10:22 moritz NotFound: thanks
10:22 NotFound mode, in that case
10:22 NotFound Ignored in windows
10:24 NotFound Maybe we must make that parameter optional and provide a useful default.
10:25 ambs joined #parrot
10:25 AndyA joined #parrot
10:28 moritz the problem was that rakudo had a default $mode = 0 :-/
10:28 moritz patching that now
10:32 muixirt joined #parrot
10:34 lucian joined #parrot
10:43 dalek parrot: r48312 | mikehh++ | branches/gc_massacre/src/call/context_accessors.c:
10:43 dalek parrot: [gc_massacre] add 3  missing const - still other problems
10:43 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48312/
11:03 AndyA joined #parrot
11:16 dalek parrot: r48313 | bacek++ | branches/gc_massacre/src/call/context_accessors.c:
11:16 dalek parrot: Fix context_accessors after merging with trunk
11:16 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48313/
11:20 jsut joined #parrot
11:42 mikehh bacek: much quicker than I am - builds with gcc, but not g++, make coretest failures, (same in make test)
11:46 bacek mikehh, :)
11:48 mikehh bacek: BTW codetest passes
11:48 bacek it's good.
11:49 bacek now time to fix other tests :)
11:50 dalek rakudo: 8f8ad30 | moritz++ | src/core/IO.pm:
11:50 dalek rakudo: fix RT #76978, mkdir() umask should default to 0o777
11:50 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/8​f8ad30c63be28b668fb99366426e1fadbe7406b
12:04 ruoso joined #parrot
12:19 macroron joined #parrot
12:22 dalek parrot: r48314 | mikehh++ | branches/gc_massacre/src/string/charset (2 files):
12:23 dalek parrot: [gc_massacre] fix g++ errors wrt charset - still other problems
12:23 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48314/
12:24 bluescreen joined #parrot
12:26 bluescreen joined #parrot
12:35 japhb joined #parrot
12:39 jhelwig joined #parrot
12:39 dalek parrot: r48315 | NotFound++ | trunk/src/hash.c:
12:39 dalek parrot: let the compiler know that the key compare function does not change during a search
12:39 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48315/
12:50 smash joined #parrot
12:52 mikehh bacek: still get the following g++ errors (nopaste not working)
12:52 mikehh src/gc/gc_ms2.c: In function ‘void Parrot_gc_ms2_init(parrot_interp_t*)’:
12:52 mikehh src/gc/gc_ms2.c:570: error: invalid conversion from ‘STRING* (*)(parrot_interp_t*, UINTVAL)’ to ‘Buffer* (*)(parrot_interp_t*, size_t)’
12:52 mikehh src/gc/gc_ms2.c:571: error: invalid conversion from ‘void (*)(parrot_interp_t*, STRING*)’ to ‘void (*)(parrot_interp_t*, Buffer*, size_t)’
12:52 mikehh src/gc/gc_ms2.c:592: error: invalid conversion from ‘void (*)(parrot_interp_t*, STRING*, size_t)’ to ‘void (*)(parrot_interp_t*, Buffer*, size_t)’
12:52 mikehh src/gc/gc_ms2.c:593: error: invalid conversion from ‘void (*)(parrot_interp_t*, STRING*, size_t)’ to ‘void (*)(parrot_interp_t*, Buffer*, size_t)’
12:53 mikehh bacek: tried to resolve them but it didn't work
12:54 Coke (writing a compiler in pure PIR) - for a comparison on that, see partcl vs. partcl-nqp ; doing it in PIR is muuuuch harder.
12:55 Coke mikehh: woof.
12:55 mikehh Coke: :-}
12:56 dalek parrot: r48316 | mikehh++ | branches/gc_massacre/src/gc/fixed_allocator.c:
12:56 dalek parrot: [gc_massacre] fix another g++ error - still problems with gc_ms2.c (Buffer *)/(STRING *)
12:56 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48316/
12:56 mikehh Coke: any idea what is happening with nopaste - apparently there was some spam or suchlike?
13:06 Coke no clue.
13:07 Coke purl, nopaste?
13:07 purl nopaste is http://nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl) or http://paste.scsys.co.uk or http://www.extpaste.com or http://gist.github.com or App::Nopaste or http://codepeek.com/paste/ or (: pastebot)
13:07 Coke purl, nopaste owner?
13:07 purl bugger all, i dunno, coke
13:38 ambs joined #parrot
13:38 [1]Casan joined #parrot
13:49 dalek partcl-nqp: c11d833 | Coke++ |  (3 files):
13:49 dalek partcl-nqp: after splitting up main.pm into its component chunks... remove main.pm.
13:49 dalek partcl-nqp: (whoops)
13:49 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/c11d83386b88e8fb10c85656eda4221606f369db
13:56 clinton joined #parrot
13:57 bubaflub joined #parrot
14:02 Coke how can one add more signatures to NQP's known opcodes?
14:03 Coke compilers/pct/src/PAST/Compiler.pir ?
14:03 bubaflub Austin: did you get your problem configuring on windows fixed?
14:05 dalek rakudo: e41c47b | moritz++ | src/ (2 files):
14:05 dalek rakudo: :ii is a synonym for :samecase
14:05 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​41c47b5db0ea07ec9ccdb47b7d967cc438418c1
14:06 Coke tried adding piropsig['new'] = 'Ps', but pir::new('Foo') still complains.
14:10 dalek rakudo: 0185653 | moritz++ | src/Perl6/Actions.pm:
14:12 dalek rakudo: :ii/:samecase now implies :i, as per spec
14:12 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/0​18565394a8e64af50941616eff59ec12062ed0e
14:16 ttbot joined #parrot
14:17 clinton joined #parrot
14:17 Casan joined #parrot
14:17 AndyA joined #parrot
14:17 TiMBuS joined #parrot
14:17 Coke joined #parrot
14:17 fperrad joined #parrot
14:17 darbelo joined #parrot
14:17 dafrito joined #parrot
14:17 aloha joined #parrot
14:17 bacek joined #parrot
14:17 dngor joined #parrot
14:17 estrabd joined #parrot
14:17 NotFound joined #parrot
14:17 s1n joined #parrot
14:17 snarkyboojum joined #parrot
14:17 simcop2387 joined #parrot
14:17 pmichaud joined #parrot
14:17 Util joined #parrot
14:17 NOTevil joined #parrot
14:17 kthakore joined #parrot
14:17 ingy joined #parrot
14:17 KatrinaTheLamia joined #parrot
14:17 Essobi joined #parrot
14:17 Ryan52 joined #parrot
14:17 sjn joined #parrot
14:22 Coke oh! it did work, just needed more of a rebuild.
14:22 bubaflub joined #parrot
14:22 jhelwig joined #parrot
14:22 japhb joined #parrot
14:22 lucian joined #parrot
14:22 muixirt joined #parrot
14:22 perlite joined #parrot
14:22 mikehh joined #parrot
14:22 Chandon joined #parrot
14:22 jdv79 joined #parrot
14:22 PacoLinux joined #parrot
14:22 Khisanth joined #parrot
14:22 particle joined #parrot
14:22 GeJ joined #parrot
14:22 wagle joined #parrot
14:22 mattp joined #parrot
14:22 p6eval joined #parrot
14:22 dzoe joined #parrot
14:22 sri joined #parrot
14:22 magnachef joined #parrot
14:22 pjcj joined #parrot
14:22 mikegrb joined #parrot
14:22 atrodo joined #parrot
14:22 Maddingue joined #parrot
14:22 frodwith joined #parrot
14:22 he joined #parrot
14:22 integral joined #parrot
14:22 tewk_ joined #parrot
14:22 szbalint joined #parrot
14:22 eirik joined #parrot
14:22 confound joined #parrot
14:22 sorear joined #parrot
14:22 Tene joined #parrot
14:22 Infinoid joined #parrot
14:24 davidfetter joined #parrot
14:29 Andy_ joined #parrot
14:30 plobsing joined #parrot
14:34 dalek rakudo: ecb4b0a | moritz++ | src/Perl6/Actions.pm:
14:34 dalek rakudo: remove left-over debugging output, moritz--
14:34 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​cb4b0a5c4f85dc76b7f2bb866b93ea244e42c4e
14:52 dalek parrot: r48317 | NotFound++ | trunk/src/hash.c:
14:52 dalek parrot: simplify parrot_hash_delete
14:52 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48317/
14:56 moritz rakudo: $_ = 'Foo'; s:ii/foo/bar/; .say
14:56 p6eval rakudo ecb4b0: OUTPUT«Bar␤»
14:56 moritz \o/
14:56 szbalint :)
15:02 ambs joined #parrot
15:04 erikh joined #parrot
15:05 erikh greets... kind of unrelated to parrot but I'm trying to find some wisdom if anyone wants to throw me a bone. I'm looking to bind to ICU and I am told at one point, parrot used this underneath the hood
15:05 erikh and then ripped it out. I was wondering what the rationale for that was?
15:06 moritz parrot still uses icu
15:06 erikh oh, nifty.
15:06 moritz at least optionally
15:07 erikh yeah, basically I'm trying to research the library a bit before I pour a week or two into making it play ball.
15:07 erikh seems like a dream come true, encoding-wise at least, on paper.
15:09 dukeleto erikh: parrot uses ICU for all unicodey things, what kind of wisdom are you looking for?
15:10 erikh actually, someone told me you dropped it, and I was curious what the reason was. I guess I/they were wrong about that.
15:11 erikh basically trying to expose (some, at least) traps before I start in on it.
15:11 erikh that's good to know, actually, because it tells me it's worth relying upon.
15:11 dukeleto erikh: yes, that is incorrect. Parrot  definitely still uses ICU (optionally) if you want a Parrot that understands unicode
15:11 erikh thanks guys.
15:11 dukeleto erikh: no worries. ICU is definitely good enough for Parrot, if that helps
15:12 erikh excellent.
15:13 chromatic joined #parrot
15:13 smash joined #parrot
15:14 ash_ joined #parrot
15:14 theory joined #parrot
15:15 Coke erikh: there is a kernel of truth in that rumor >>
15:16 Coke we used to bundle the icu source directly in our version control system. at some point we extracted it so we can just rely on the system ICU instead of lugging it around with us.
15:16 erikh aha. yeah, I was wondering how easy/hard it would be to bundle.
15:18 erikh kind of leaned on just this side of "hard" with the autotools requirement
15:23 dalek TT #1728 created by coke++: Add more signatures for for PAST op types.
15:23 dalek TT #1728: http://trac.parrot.org/parrot/ticket/1728
15:37 dukeleto 'ello
15:46 brianwisti joined #parrot
16:02 clinton joined #parrot
16:20 TonyC joined #parrot
16:26 theory joined #parrot
16:29 jevin joined #parrot
16:38 dalek parrot: r48318 | chromatic++ | trunk/src/hash.c:
16:38 dalek parrot: [hash] Doubled initial hash size to 8 buckets.
16:38 dalek parrot: This reduces resizing for hashes of 8 or fewer buckets: most of them.
16:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48318/
16:38 dalek parrot: r48319 | chromatic++ | trunk/src/hash.c:
16:38 dalek parrot: [hash] Removed small hash pseudo-optimization.
16:38 dalek parrot: It's not as useful as other optimizations, and less code is better anyway.
16:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48319/
16:38 dalek parrot: r48320 | chromatic++ | trunk/src/call/args.c:
16:38 dalek parrot: [PCC] Added positionals presizing to fill_args.
16:39 dalek parrot: Avoiding reallocation is useful here.  Perhaps this is a good use for
16:39 dalek parrot: init_pmc_int?
16:39 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48320/
16:42 ruoso joined #parrot
16:45 hercynium joined #parrot
17:01 clinton joined #parrot
17:10 * Coke nominates ...
17:12 jsut_ joined #parrot
17:28 donaldh joined #parrot
17:36 cotto_work ~~
18:10 donaldh joined #parrot
18:12 cotto_work http://collectiveidea.com/blog/arch​ives/2010/08/03/happy-git-commits/
18:17 Chandon joined #parrot
18:27 Paul_the_Greek joined #parrot
18:28 Paul_the_Greek Anyone available for a question?
18:29 sorear yes
18:29 Paul_the_Greek Howdy.
18:29 sorear (but even if noone was, you'd be better off just asking it)
18:29 Paul_the_Greek What is the difference between installing the latest developer release and using SVN to check out Parrot?
18:31 Paul_the_Greek I mean as far as working on Parrot is concerned.
18:31 chromatic SVN checkouts can be more experimental.
18:31 Paul_the_Greek What is the right thing to do if I want to work on Parrot?
18:32 chromatic That depends on your interest: building a language versus developing Parrot.
18:33 Coke if you're working on parrot guts, svn latest.
18:33 Coke if you're a HLL developer, latest release is typically ok. (but even then, I use latest all the time anyway)
18:33 Paul_the_Greek Oh, sorry, developing Parrot
18:34 Paul_the_Greek Okay, so SVN.
18:35 Coke note: if you prefer git (and aren't a committer) there is a git clone of the svn repo on github
18:35 Coke (if you prefer git and /are/ a committer, there's always git-svn)
18:35 Paul_the_Greek Do people lock the files they're working on?
18:36 cotto_work definitely not
18:36 sorear doesn't that defeat the whole point of using svn
18:36 cotto_work you might want to avoid working on a certain file if you know it's part of a current refactor, but there's definitely no manual locking
18:37 Paul_the_Greek Sorry, I'm new to svn.
18:37 particle svn uses an optimistic locking model
18:37 Chandon joined #parrot
18:37 Paul_the_Greek So I simply work away and then submit a patch.
18:37 particle aye
18:38 particle docs/submissions.pod will help
18:38 Paul_the_Greek Yes, that's what I'm looking at now.
18:42 Paul_the_Greek So svn looks at my copy of Parrot and the master copy and figures out what has changed?
18:42 cotto_work examples/c/nanoparrot.c is a self-contained minimalist version of Parrot.  It might make Parrot as a whole easier to grok.
18:42 cotto_work yup
18:43 cotto_work more or less
18:43 Paul_the_Greek Okay, time to read the Subversion docs, too.
18:43 particle svnbook.red-bean.com has the details on subversion
18:43 cotto_work svn book?
18:43 particle purl, svn book is http://svnbook.red-bean.com
18:44 particle svn book?
18:44 particle feh.
18:44 cotto_work unbotsnack
18:44 particle botslap
18:44 cotto_work Oh.  I guess we need purl to be here for her to work.
18:46 * darbelo fails to miss the bot.
18:46 cotto_work now we can say "+1" without her reply
18:46 atrodo +1
18:46 darbelo 1
18:47 particle darbeloslap
18:47 * cotto_work really should have seen that coming
18:47 particle :)
18:47 cotto_work darbelosnack
18:47 darbelo :)
18:52 mikehh one should note that purl keeps a record of karma :-}
18:52 preflex joined #parrot
18:52 particle purl++
18:52 particle ;)
18:53 cotto_work normally purl is a bot we keep around to record karma and various random factoids, fyi
18:55 mikehh she seems to have left in the great net split
18:55 mikehh but did not return
18:56 whiteknight joined #parrot
18:56 cotto_work yay for robust irc protocols
18:56 atrodo who/where is purl hosted?
18:57 sorear irc.perl.org itself
18:57 sorear go visit #purl, they have the tribunal atsmospher going *really well*
19:00 Paul_the_Greek Would a ticket like this be good for a newbie: http://trac.parrot.org/parrot/ticket/1605
19:01 chromatic If you're comfortable with memory management yes.
19:02 Paul_the_Greek Seems like a good way to learn a bit of internals without immediately hacking any actual code.
19:03 particle Paul_the_Greek++
19:04 Paul_the_Greek Fetch before the increment, you say?
19:06 Paul_the_Greek I'll start by reading the PDDs.
19:08 chromatic If you find anything confusing or wrong, please file tickets.
19:11 Paul_the_Greek One more question: Would documentation on the buffers belong in the source (as now), or in the Memory Internals documentation?
19:14 chromatic The docs are probably long out of date.
19:14 Paul_the_Greek Ah, so perhaps that's a project for later.
19:15 tcurtis joined #parrot
19:18 particle running svn log on the docs will give you some ideas as to their age and state of upkeep
19:19 Paul_the_Greek I'll do that, thanks.
19:19 cotto_work svn blame is nice for individual files
19:25 hercynium joined #parrot
19:32 Paul_the_Greek Thanks, folks.
19:48 theory joined #parrot
19:51 contingencyplan joined #parrot
19:53 davidfetter joined #parrot
19:57 muixirt joined #parrot
20:04 Chandon joined #parrot
20:39 tommyd joined #parrot
20:40 bluescreen joined #parrot
20:59 lucian joined #parrot
21:02 chromatic msg bacek http://www.coffeepowered.net/2009/06/1​3/fine-tuning-your-garbage-collector/
21:04 cotto_work joined #parrot
21:07 Chandon joined #parrot
21:37 dalek parrot: r48321 | chromatic++ | branches/gc_massacre/src/gc/gc_ms2.c:
21:37 dalek parrot: [GC] Added preliminary growth factor to ms2 GC.
21:37 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48321/
21:38 dalek tree-optimization: f7cd6ed | tcurtis++ | t/pastpattern.t:
21:38 dalek tree-optimization: Test that .transform does the right thing with deleting nodes.
21:38 dalek tree-optimization: review: http://github.com/ekiru/tree-optimization/com​mit/f7cd6ed5adcea6d64ea55d7bacd92d2a9be8b5e0
21:38 dalek tree-optimization: e3dd6b2 | tcurtis++ | src/Tree/Pattern/Transformer.nqp:
21:38 dalek tree-optimization: Fix deletion of nodes in .transform.
21:38 dalek tree-optimization: review: http://github.com/ekiru/tree-optimization/com​mit/e3dd6b24ff41593bd46a6c29104d7972768c6a1b
22:03 [1]Casan joined #parrot
22:24 theory joined #parrot
22:27 ascent joined #parrot
23:17 dalek parrot: r48322 | darbelo++ | branches/unshared_buffers:
23:17 dalek parrot: Create a branch for performance testing without shared buffers.
23:17 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48322/
23:24 theory joined #parrot
23:35 Paul_the_Greek joined #parrot
23:37 Paul_the_Greek The Memory Internals document is indeed sadly out-of-date.
23:39 dukeleto Paul_the_Greek: which document? the PDD?
23:40 Paul_the_Greek Yes.
23:40 Paul_the_Greek Oops, no, sorry, the developer doc.
23:59 Psyche^ joined #parrot

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

Parrot | source cross referenced