Camelia, the Perl 6 bug

IRC log for #parrot, 2009-07-15

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:04 * cotto puts on leisure suit
00:04 * treed hugs vim macros.
00:04 fdorothy Does anyone know what's been going on with Pirate?
00:05 cotto I think it died a while back.
00:05 cotto allison's working on a different Python implementation though
00:05 fdorothy :(  Are there any other python interpreters available for parrot?
00:05 cotto ha!
00:05 fdorothy Ok.
00:05 dalek TT #838 created by whiteknight++: Configure.pl mangles forwardslashed-flags for CL
00:06 cotto I think it's called "Pynie"
00:06 cotto If you're interested in helping, I'm sure she'd be glad to get help.
00:07 cotto http://code.google.com/p/pynie/
00:09 fdorothy I've tried it out.  It seems like it's broken against the trunk of parrot, but I may be wrong about that..
00:10 fdorothy Is she the only one working on it?
00:10 dalek parrot: r40090 | jkeenan++ | branches/tt761_keys_revamp/src (2 files):
00:10 dalek parrot: Make file conform to codingstd:  c_parens.
00:10 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40090/
00:13 dalek parrot: r40091 | jkeenan++ | branches/tt761_keys_revamp/src (2 files):
00:13 dalek parrot: Make file conform to codingstd:  linelength trailingspace.
00:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40091/
00:18 whiteknight parrot built!!!! Personal Best!
00:18 whiteknight of course, there were all sorts of errors and it doesn't pass many tests, but I do have an executable
00:18 hercynium_ joined #parrot
00:18 treed Is there anyone I should be communicating plans for cardinal to?
00:18 cotto it's progress
00:19 whiteknight treed: Tene I guess, otherwise you get to be the master and commander of that particular ship
00:20 whiteknight I'll stay on as your first mate and call you "dread pirate roberts" until the new crew believes you're really in charge
00:20 treed whiteknight: Tene has basically given me permission to do whatever I want.
00:20 whiteknight aye aye, capn' roberts, sir
00:20 treed Hoist the main sail, lads. We're goin' plunderin'.
00:22 cotto ARRRRRRR
00:25 * treed is in the midst of replacing the build system with rake.
00:25 fdorothy Ah, there we go..pynie worked.  I just had to copy over nqp into the Parrot install directory.
00:25 cotto fdorothy, did you run make install-dev?
00:25 * treed has it building cardinal, just needs to have it do the configuration stuff and then maybe add namespaces for the tests.
00:25 fdorothy Nope, is that what I'm suppose to do?
00:25 cotto That'll give you what you need to build a language.
00:26 fdorothy Good to know.  Thanks.
00:26 cotto s/give you/install/
00:26 cotto np
00:26 mokurai joined #parrot
00:33 whiteknight ah jeez, configure mangles the command line switches in all the makefiles through the repo
00:33 whiteknight now I have to manually fix all of them
00:34 dalek parrot: r40092 | jkeenan++ | branches/tt761_keys_revamp (7 files):
00:34 dalek parrot: Set SVN metadata properties.
00:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40092/
00:37 dalek parrot: r40093 | jkeenan++ | branches/tt761_keys_revamp/docs/book/pir (9 files):
00:37 dalek parrot: Set SVN metadata properties.
00:37 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40093/
00:38 whiteknight if make took any longer on this stupid platform, I'd miss the release
00:39 patspam joined #parrot
00:39 * cotto is beginning to see why we don't have any windows x64 reports
00:40 whiteknight I've got a commit I'm rolling up that will fix some issues, and then that TT that I filed earlier tonight and it *should* work
00:41 whiteknight ah damnit, fails when building PGE
00:42 treed What build system does Parrot use on Windows?
00:42 treed And I wonder if there might be something better.
00:42 treed SCons or whatever.
00:45 whiteknight no real build system, a custom-made Perl configure step and a makefile
00:45 treed Ah, so same as on Linux/OS X.
00:45 treed Why does it take so much longer?
00:46 whiteknight because Windows is TEH SUXXOR
00:46 whiteknight (and I mean that in the nicest possible way)
00:46 treed No offense here. I use OS X as my desktop with a lot of tasks done in a shell to my Linux server.
00:47 kid51 whiteknight:  I believe vi +102 config/inter/progs.pm is the first point in the config system where a CLI setting for --ccflags is pulled in.
00:47 treed I do have W7 installed on another partition, but I don't code in it.
00:47 kid51 But the config/init/hints/mswin32.pm is full of use of hyphens in this case.
00:48 kid51 treed:  Re build time:  I'm told (by particle, who should know) that the startup time of processes on Win is much greater than on *nix.
00:49 kid51 particle had me consolidate many of the test files in the preconfiguration tests so as to reduce the gross startup time.
00:49 kid51 ... even though my own preference was to keep the test files reasonably small to stay focused on particular source files.
00:50 kid51 treed:  "I use OS X as my desktop with a lot of tasks done in a shell to my Linux server."   -- like 80% of Parrot devs?
00:50 kid51 devs.
00:50 treed I have no idea about such a statistic.
00:50 treed I know Tene uses Linux almost exclusively.
00:50 kid51 The ? was accidental.  I was asserting it.  That's what I'm doing right now.
00:50 treed He's the only Parrot dev I know.
00:50 treed I figured you were asserting it.
00:50 kid51 Of course, Tene teaches RedHat.
00:50 treed I was responding that I have no idea whether or not your assertion holds water.
00:51 kid51 Based on F2F observation at hackathons, I'd say yes.
00:51 treed I'm prepared to accept that evidence.
01:14 dalek parrot: r40094 | jkeenan++ | branches/tt761_keys_revamp/confi​g/gen/config_h/has_header_h.in:
01:14 dalek parrot: Set SVN metadata properties.
01:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40094/
01:23 athomason1 joined #parrot
01:29 jdv79 linux and osx by far, win32 maybe third.
01:42 cotto kid51++ for understanding the configure system
01:50 TiMBuS joined #parrot
01:51 Theory joined #parrot
02:06 kid51_at_dinner cotto  I think I understand the *how* of the configure system -- but not the *why*
02:07 kid51 I think that there are only 4 or 5 people who understand the *why* -- and not all of them are currently active committers :-)
02:08 kid51 And since we don't have a PDD on configuration -- something that tells us what Parrot probes for and why -- nobody *fully* understands it.
02:18 patspam1 joined #parrot
02:19 cotto kid51, it's possible that it's better that way. ;)
02:24 patspam joined #parrot
02:26 dalek TT #678 closed by jkeenan++: Is tools/dev/parrot_8.supp still needed?
02:26 dalek TT #797 closed by jkeenan++: Eliminate need to expand Makefile variables in parrot_config and other ...
02:27 dalek parrot: r40095 | jkeenan++ | trunk (2 files):
02:27 dalek parrot: tools/dev/parrot_8.supp removed per TT #678.
02:27 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40095/
02:28 kid51 msg Coke Please review https://trac.parrot.org/parrot/ticket/727
02:28 purl Message for coke stored.
02:35 janus joined #parrot
02:38 Util In src/pmc.c, function `pmc_new` takes a parameter of `INTVAL base_type`.
02:38 Util While debugging, I see many different base types being created.
02:38 Util Is there a table of base_type => PMC_name anywhere?
02:38 jimmy joined #parrot
02:39 bacek_at_work Util: include/parrot/core_pmcs.h?
02:40 Infinoid Yeah, what he said :)
02:40 Util bacek_at_work: just what I was looking for. Thanks!
02:50 dalek parrot: r40096 | cotto++ | branches/ops_pct (15 files):
02:50 dalek parrot: [opsc] apply ops_format_update.patch to branch, simplifying opsc's job
02:51 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40096/
03:19 allison joined #parrot
03:20 donaldh joined #parrot
03:27 dalek parrot: r40097 | allison++ | trunk/config/gen/config_pm/config_pir.in:
03:27 dalek parrot: [install] Use configured path for library instead of prefix plus
03:27 dalek parrot: hardcoded 'lib/', which doesn't allow for platform variations. Applying
03:27 dalek parrot: TT #739. Thanks to Gerd Pokorra.
03:27 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40097/
04:21 dalek TT #476 closed by allison++: Binaries should not contain rpath
04:29 dalek close: r75 | Austin++ | trunk/ (3 files):
04:29 dalek close: Added find_caller_lex builtin
04:29 dalek close: review: http://code.google.com/p/close/source/detail?r=75
04:37 dalek parrot: r40098 | allison++ | trunk/config/init/defaults.pm:
04:37 dalek parrot: [install] Stripping -rpath from dynamic linking options that are copied
04:37 dalek parrot: from Perl when the --disable-rpath option is set. Resolves TT #830.
04:37 dalek parrot: Thanks to Gerd Pokorra.
04:37 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40098/
05:08 nnunley Hey.  I'm trying to make the bootstrap of nqp work, but have run into a minor bit of confusion.  The statement_block code does't seem to ever actually shift values into @?BLOCK
05:42 cotto nnunley, pmichaud would be the best person to talk to about nqp, but he's likely asleep atm
05:44 nnunley cotto:  Thanks.  I'll try to ask again later.
05:48 cotto Sure.  He's usually in Texas if that helps.
05:49 dalek close: r76 | Austin++ | trunk/ (6 files):
05:49 dalek close: Added tailcall builtin
05:49 dalek close: review: http://code.google.com/p/close/source/detail?r=76
05:50 bacek_at_work clock?
05:51 bacek_at_work ENOPURL???
05:51 purl couldn't get the headlines: http://rjbs.manxome.org/rss/clock.cgi wasn't successful
05:51 purl AM HERE TOO
05:52 cotto Wow.  An rss-based clock.
06:01 uniejo joined #parrot
06:27 chromatic joined #parrot
06:32 iblechbot joined #parrot
06:39 Tene Yay... I got my gui lib wrapper working from Perl 6.
06:39 moritz awesome!
06:39 purl well, awesome is a window manager or at http://awesome.naquadah.org or awesome!
06:40 MoC joined #parrot
07:20 donaldh joined #parrot
07:26 |MoC| joined #parrot
07:27 _MoC_ joined #parrot
08:21 moritz bacek_at_work: http://irclog.perlgeek.de/parr​otsketch/2009-07-14#i_1315689
08:40 mokurai left #parrot
08:52 TiMBuS joined #parrot
09:04 masak joined #parrot
09:14 mikehh All test PASS (pre/post config, smolder, fulltest) at r40098 - Ubuntu 9.04 amd64
09:33 TiMBuS joined #parrot
09:40 mikehh rakudo - d8d0640 builds on parrot r40098 make test PASS, make spectest (up to rev 27564) reports 1 failure - but actually passes
09:41 mikehh t/spec/S03-operators/hyper.rakudo - Parse errors: Bad plan.  You planned 91 tests but ran 93.
09:42 mikehh rakudo - d8d0640 builds on parrot r40098 make test PASS, make spectest (up to rev 27564) reports 1 failure - but actually passes
09:43 mikehh sorry meant to put the second in #perl6
10:40 MoC joined #parrot
10:46 mikehh rakudo - d8d0640 builds on parrot r40098 all tests PASS, make test/make spectest (up to rev 27565) - Ubuntu 9.04 amd64
10:48 jan joined #parrot
11:21 donaldh joined #parrot
11:25 rob joined #parrot
11:28 mikehh_ joined #parrot
11:54 Ehtyar joined #parrot
11:55 Whiteknight joined #parrot
11:56 Whiteknight good morning #parrot
11:59 masak good morning, Whiteknight.
12:01 bacek joined #parrot
12:01 bacek o hai
12:13 Whiteknight hello masak and bacek
12:13 moritz good localtime()
12:14 moritz bacek: did you see the link I posted to bacek_at_work?
12:14 bacek moritz: yes, thanks. Marking failing tests, smoking rakudo again, etc.
12:14 bacek BEWARE! HERE WILL BE DRAGONS! I'm going to merge keys_revamp branch!
12:20 masak ooh, dragons.
12:20 purl dragons are sexy
12:20 nopaste "bacek" at 114.73.169.200 pasted "Help wanted. How to rewrite test from pir_output_is to pir_output_like?" (16 lines) at http://nopaste.snit.ch/17280
12:20 dalek TT #739 closed by jkeenan++: path of installed config.fpmc file on a 64bit system
12:21 bacek Oh shi...
12:21 dalek parrot: r40099 | bacek++ | branches/tt761_keys_revamp/t/compilers/pct (2 files):
12:21 dalek parrot: [t] SKIP tests dependent on hash keys order.
12:21 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40099/
12:22 clinton joined #parrot
12:22 moritz bacek: sort on hash keys before printing seems easier than bilding a huge regex for that
12:23 bacek moritz: Output produced by "_dumper(past)". And I don't like idea to sort keys in "_dumper"...
12:23 dalek TT #830 closed by jkeenan++: [PATCH] Configure.pl disable-rpath option
12:24 eternaleye joined #parrot
12:25 MoC joined #parrot
12:26 * bacek tempted to add "BEWARE! HERE WILL BE A DRAGONS" into commit message.
12:29 mikehh joined #parrot
12:42 iblechbot joined #parrot
12:43 masak bacek: if you do, consider doing s/WILL // for extra faithfulness: http://en.wikipedia.org/wiki/Here_be_dragons
12:43 bacek masak: I WILL not :)
12:44 bacek ok, parrot's tests passed after merge, partcl passed, waiting for rakudo's spectest before dcommit it.
12:47 bacek ~1000 seconds to go. Time for coffee
12:54 * bacek wish to have rakudo's "make spectest" run in less than 1 minute...
12:58 Whiteknight oh, is that all you want? if somebody had just asked, I would have flipped the "fast" switch on the spectest
12:59 * Infinoid tapes the button down
13:00 * bacek put big brink on top of it
13:08 bacek My laptop going to melt... Whiteknight, where is this "fast" button?
13:08 Whiteknight I don't know, it's in there somewhere. You just have to spend the rest of your life looking for it
13:09 pmichaud the fast button is called "pcc refactor"
13:09 riffraff joined #parrot
13:10 Whiteknight I just heard from the IT guy that my work computer needs to be reformatted today
13:10 Whiteknight so /me is going to disappear from TEH INTERNETZ
13:13 * bacek opening cage with DRAGONS.
13:14 Whiteknight bacek: when is the merge happening?
13:14 bacek Whiteknight: now.
13:14 bacek right now
13:14 purl right now it's time to kick out the jams mother fuckers
13:14 Whiteknight purl++
13:14 bacek Files=415, Tests=14447, 2336 wallclock secs (12.88 usr  1.29 sys + 1985.08 cusr 54.76 csys = 2054.01 CPU)
13:15 Coke anyone able to run SDL that can look at #TT #798
13:15 Coke ?
13:15 bacek pmichaud: I don't feel that I can help with "pcc refactor"...
13:16 Coke src\hash.c: In function `parrot_hash_put':
13:16 Coke src\hash.c:1307: warning: suggest parentheses around && within ||
13:17 NotFound I suggest the same.
13:18 bacek Will do
13:18 dalek parrot: r40100 | bacek++ | trunk (42 files):
13:18 dalek parrot: Merge branch tt761_keys_revamp into trunk.
13:18 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40100/
13:19 NotFound BTW ( folowed by newline break codingstd.
13:21 bacek NotFound: imho it's less readable with newline in this case.
13:21 NotFound bacek: You can add: /* intentionally left blank */ X-)
13:22 bacek O! Good idea!
13:22 dalek parrot: r40101 | bacek++ | trunk/src/hash.c:
13:22 dalek parrot: [cage] Fix compiler warning in hash.c
13:22 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40101/
13:22 Whiteknight i sincerely do hope that pcc_rewiring does land shortly after 1.4
13:22 bacek NotFound: can I add /* Don't write there! */?
13:22 Whiteknight because it sets the stage for a lot of other cool optimization and refactoring work that we can do to Parrot
13:23 NotFound Or /* Put here an xkcd at your leisure */
13:24 * bacek have at least 3 xkcd's t-shirts :)
13:24 Whiteknight I don't have any!
13:24 * Whiteknight needs to catch up
13:25 bacek Whiteknight: It's 21st century! You can order them over TEH INTERWEB.
13:26 Whiteknight it's just a series of tubes
13:26 bacek How I can pacify compiler about cast INTVAL to (void*) in src/pmc/hash.pmc?
13:26 NotFound I'd like a shirt whit this: http://megatokyo.com/strip/1161
13:26 Whiteknight there was some kind of macro somewhere that allowed casting to/from INTVAL and pointers
13:26 bacek Whiteknight: not just tubes. But tubes with blackjack and hookers! :)
13:27 Whiteknight I would like this shirt: http://www.zazzle.com/rule_3​4_tshirt-235879767016272599
13:27 NotFound bacek: we have some horrid named macros for that type of things.
13:28 Whiteknight some horrid named macros that nobody can remember!
13:28 bacek And even grep can't help you!
13:28 NotFound grep '::horrid:'
13:28 Whiteknight I would try a grep for "2PTR"
13:28 NotFound Yes, something like that.
13:28 purl something like that. is, like, that true
13:29 Whiteknight although I would also warn that those things are bad because on some systems *cough*Win64*cough*
13:29 Coke purl, forget something like that.
13:29 purl Coke, I didn't have anything matching something like that
13:29 Whiteknight sizeof(intval) != sizeof(void*)
13:29 bacek INTVAL2PTR?
13:29 Whiteknight We need to sell Parrottshirts
13:30 Whiteknight I would buy one in a heartbeat
13:34 NotFound Did we have some vectorial format version of the parrot logo?
13:35 bacek Erm... INTVAL2PTR can't handle void*. Or gcc 4.4 is too pedantic...
13:39 NotFound bacek: nopaste it
13:39 purl nopaste it is See /who is nopaste for the URL.
13:39 bacek Bah! Found it. It brakes when I try to cast function call.
13:40 bacek If I assign it to variable than it cast silently...
13:41 NotFound Mmmm... I'm wondering if that macros lack some parens...
13:41 bacek NotFound: It's not.
13:48 bacek NotFound: r40102
13:49 dalek parrot: r40102 | bacek++ | trunk/src/pmc/hash.pmc:
13:49 dalek parrot: [cage] Pacify compiler on casting INTVAL to void*
13:49 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40102/
13:52 dalek TT #761 closed by bacek++: [RFC] Keys/Iterator deuglyfying.
13:52 dalek parrot: r40103 | bacek++ | trunk/MANIFEST:
13:52 dalek parrot: [cage]Regenerate MANIFEST
13:52 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40103/
13:53 NotFound bacek: yes, but anyway...
13:54 bacek NotFound: Some gcc shenanigans.
13:56 dalek parrot: r40104 | NotFound++ | trunk/include/parrot/parrot.h:
13:56 dalek parrot: [cage] add parens to U?INTVAL2PTR macros to prevent precedence problems in corner cases
13:56 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40104/
13:58 Andy joined #parrot
14:07 MoC joined #parrot
14:09 Whiteknight bacek++ # Very good work on the keys branch!
14:10 particle does it work with a c++ compiler?
14:13 mokurai joined #parrot
14:13 skids joined #parrot
14:15 bacek particle: it should be.
14:22 * Coke yawns
14:22 particle bacek: any time i see changes involving void * i think about c++ compilers
14:22 Andy mmm, void *
14:23 bacek particle: it was used in hashes before.
14:23 particle yes, i understand.
14:23 particle we're releasing a supported release on tuesday
14:23 particle i want to make sure it Just Works.
14:24 particle i haven't seen any failures from c++ smokers, but i only view failure reports, so i don't know if we have any c++ smokers
14:26 NotFound Negative, it fails with C++. I'm, going to fix it.
14:29 particle see!
14:29 particle i mean, see++!
14:29 * bacek printing facepalm.jpg
14:30 bacek How I can configure parrot to use C++ instead of plain C?
14:30 NotFound bacek: --cc=g++ --cxx=g++ --ld=g++ --link=g++
14:30 bacek NotFound++
14:31 Coke karma --cc=g
14:31 purl --cc=g has neutral karma
14:31 NotFound Is a trivial fix :)
14:31 Coke even if we had c++ smokers, broken builds make it hard to run 'make test'
14:31 Theory joined #parrot
14:32 NotFound Don't worry, I build almost daily with c++ and fix any problem.
14:33 NotFound Well, don't worry for g++ at least ;)
14:33 bacek :)
14:33 bacek NotFound: hash.pmc:544 is only failure?
14:33 moritz what's the advantage of building with g++ over gcc? when you want to embed it into a C++ program?
14:33 NotFound bacek: yes
14:34 bacek moritz: yes
14:34 particle taptinder?
14:34 purl taptinder is back, http://tt.perl6.cz/ .. need port watchdog to linux ... some test hang in r26917
14:34 particle moritz: sometimes the only good compiler on a platform is a c++ compiler
14:34 NotFound moritz: tha main advantage is that C++ is much more strict with casts, both implicit and explicit, and this has helped to avoid or find lots of problems.
14:35 moritz particle: example?
14:35 moritz NotFound: good point
14:36 bacek Can we rewrite parrot in C++??? Pleeease...
14:36 Whiteknight sometimes even I think that might be a decent idea
14:36 NotFound Regarding the embedded, if we don't make mistakes we can embed a c++ built parrot in a C program and viceversa. PARROT_EXPORT macro takes care of it.
14:36 Whiteknight actually no, I would just want C with some classes thrown in, not all sorts of templates and operator overloading and other creepy nonsense
14:37 bacek Whiteknight: templates are most powerful thing in C++.
14:37 Whiteknight KILL IT WITH A STICK!
14:37 Andy If we moved to C++ then all the splint wouldn't work and then what would I do?
14:38 moritz Andy: you'd have to write splint++ :-)
14:38 Andy Yeah, i'd get right on that.
14:38 bacek karma splint
14:38 purl splint has karma of 1
14:40 MoC karma C
14:40 purl c has karma of 7847
14:40 moritz karma C++
14:40 purl c++ has karma of -85
14:40 bacek bah! (C++)++
14:41 dalek parrot: r40105 | bacek++ | trunk/src/pmc/hash.pmc:
14:41 dalek parrot: [cage] Fix broken build with c++ compilers.
14:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40105/
14:41 mikehh getting some fulltest failures at r40104
14:41 bacek mikehh: "One test per pmc"?
14:42 mikehh codetest, examples_tests, distro_tests
14:44 mikehh distro_tests is a repeat of one of the failures in codetest
14:45 bacek mikehh: nopaste them. I'll check them tomor^W today's morning.
14:47 Alias joined #parrot
14:47 Alias Wow, people
14:47 Alias Are both the parrot and rakudo cliques here?
14:48 PerlJam pretty much.
14:48 PerlJam rakudo people also hang out on freenode #perl6
14:48 Andy We have the ID cards.
14:48 Alias So, Gabor wants me to build Strawberry Perl 6
14:48 Alias Which I'm totally fine with
14:48 Alias Should be great for PR
14:49 Alias I'm thinking Strawberry 5 + Padre + Rakudo + Padre::Plugin::Perl6
14:49 Andy I sense a butt
14:49 Andy i mean a but
14:49 Alias Butt
14:49 * bacek must sleep
14:49 purl $bacek->sleep(8 * 3600);
14:49 bacek see you soon.
14:49 Alias Currently I'm using a highly insane method to do that
14:50 Alias Involving getting Gabor to fetch directly from version control the current instantaneous state of both parrot and rakudo
14:50 Alias And him building it, zipping it and me just dropping it on top of the Strawberry build
14:50 Alias I'm so ashamed at this svn instant snapshot I'm too embarrased to call it Perl, and I've just labelled it as "Six"
14:51 Alias He mentioned something about not being able to build rakudo at all, except against svn parrot?
14:52 Alias Is there any stable way to build Perl 6 using stable releases?
14:52 szabgab not to be confused with http://en.wikipedia.org/wiki/Nu​mber_Six_(Battlestar_Galactica)
14:52 Alias Could Rakudo perhaps synchronise releases with parrot, so even if you hate me for using the tarball packages as hopefully days out of date, at least there's points in time where things are sane and I can show the world what Perl 6 looked like last week?
14:53 Alias s/hopefully/hopelessly/
14:53 Whiteknight Alias, they are basically synchronized
14:54 Whiteknight the Rakudo release comes out two days after the Parrot release
14:54 Alias For what definition of "basically"
14:54 Alias Is the Rakudo release guaranteed to work with the parrot release?
14:55 Whiteknight you'd have to ask pmichaud, he makes the guarantees
14:55 Whiteknight but I think so, yes
14:55 moritz Andy: fsvo "guaranteed", yes
14:55 moritz Andy: which means he does try to make each rakudo release work cleanly on the prior parrot release
14:55 Andy Similar to "fsvo 'pregnant'"
14:56 moritz I mean you can't sue him if it doesn't work
14:56 NotFound There is a file in rakudo that says wich release has been tested with, isn't it?
14:57 Whiteknight it has a PARROT_COMPAT marker somewhere that tells what revision of Parrot it requires
14:57 Alias And PARROT_BACKCOMPAT too I hope, given the deprecation policy
14:57 Alias :)
14:57 moritz build/PARROT_REVISION
14:57 purl i heard build/PARROT_REVISION was currently a svn rev
14:57 Whiteknight and PARROT_FUTURECOMPAT, in case we time travel
14:58 nopaste "mikehh" at 90.209.50.160 pasted "fulltest failures at r40104" (739 lines) at http://nopaste.snit.ch/17283
14:59 mikehh the itro_tests and one of the codetest failures require setting svn properties
15:00 mikehh distro
15:01 mikehh I haven't looked at the examples_tests failure - the BULK of the report
15:01 * Whiteknight doesn't know anything about svn properties
15:01 NotFound Doing...
15:01 Alias You can't build from an svn export?
15:02 mikehh NotFound got them
15:03 dalek parrot: r40106 | NotFound++ | trunk (6 files):
15:03 dalek parrot: [cage] svn properties
15:03 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40106/
15:05 PerlJam Alias: you can.
15:06 PerlJam Alias: you tell rakudo where to find parrot_config and it takes what it needs from there.
15:06 PerlJam so, you can theoretically build from an installed parrot (I've never done it though)
15:06 Alias Has anyone done it?
15:06 Alias :)
15:06 PerlJam Dunno.
15:06 * Alias tries to keep up the "Just make it fucking easy to install" line
15:07 Whiteknight we're working on it!
15:08 Alias I know
15:08 Alias And once it's smooth I shall be nothing but full of praise
15:08 NotFound I've doing it, but don't test with recent releases to do it with the parrot build directory renamed or deleted. But I think I heard allison had checked thatt
15:08 Andy BUT FOR NOW, SCORN!
15:08 Andy :-)
15:08 moritz Alias, PerlJam: there's a branch for that (building rakudo against installed parrot)
15:09 Alias moritz: For the first alpha I'm happy to just go with this "six" blob
15:09 Andy Alias: It's interesting to see the # of questions in freenode #perl becoming more Strawberry and less Activestate
15:09 Alias As long as it runs AT ALL I'll be happy
15:09 NotFound I just put the directory with the parrot executables in the PATH, configure, and make.ç
15:09 Alias Andy: I smell a crossroad approaching, yes
15:09 Alias Andy: Starting to encounter new types of problems I haven't seen before
15:09 Coke (scorn) (there's a branch for that.)
15:10 dalek parrot: r40107 | NotFound++ | trunk/src (2 files):
15:10 dalek parrot: [cage] codingstd parens
15:10 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40107/
15:10 dalek parrot: r40108 | petdance++ | trunk/src/pmc/hash.pmc:
15:10 dalek parrot: lots of consting.
15:10 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40108/
15:10 PerlJam Andy: given that many of us have been touting strawberry over activestate almost since strawberry has been in existence, it's not surprising that some momentum has been created :)
15:10 Andy PerlJam: Not surprised, either.  Just pleased.
15:10 Coke Andy: for the small places I've been able to insert perl at work, I'm using strawberry. very happy with it.
15:10 Coke er, s/Andy/Alias/
15:11 PerlJam Coke: strawberry on a stick?
15:11 Andy Coke, the man so lazy he could only type "A<tab>"
15:11 Coke Andy: it's a virtue, b<tab>
15:11 Andy Are we doing anything at OSCON?
15:11 Andy My OSCON week is very short this year.
15:12 * Coke tries the bitter route "oh, look, another con I cannot attend!"
15:12 * Coke grumbles into his coffee.
15:13 NotFound Time to make some smoke...
15:13 Alias And don't plan to learn it
15:13 Alias But they need to install it to use $some_other_program
15:13 Alias I never really meant Strawberry to be somebody's runtime
15:13 Alias Just a haven for Unix Perl people stuck on Windows
15:13 Alias Stick still has some boils
15:13 Alias Despite my "It's done" blog post
15:13 Alias (I was wrong)
15:13 Alias HOWEVER!
15:13 Alias Post July things get more interesting
15:13 purl Butt!
15:13 Alias Coke: Should have a nice ActiveDirectory-installable Strawberry + Padre IDE bundle shortly
15:13 Alias Perfect for handing to desktop support teams
15:13 Alias Padre still a teeensy bit immature of course, but it's clearly The Future
15:14 dalek parrot: r40109 | petdance++ | trunk/src/pmc/orderedhashiterator.pmc:
15:14 dalek parrot: local var consting
15:14 Andy Why clearly?
15:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40109/
15:14 szabgab Andy, it has embedded ack support
15:14 Coke OrderedHaShiterator
15:14 Coke <-- 12.
15:14 Andy szabgab: WELL FOR GOD'S SAKE THEN IT IS SO BLESSED.
15:15 szabgab I knew I am very convincing
15:15 Alias Andy: It feels like something close to all the commercial IDE products
15:15 Alias Or closer at least
15:15 Alias You install a single $thing, start it, write code, and never leave the interface
15:15 Andy Just wondering why that's a Good Thing.
15:15 NotFound You need a Close plugin to be even closer.
15:15 Andy You're aiming at people who want an IDE, then, not converting the CLI masses.
15:16 Alias Because it's what people are used to
15:16 PerlJam Alias: but ... but ... it's not vi!
15:16 Alias In that world
15:16 szabgab Andy, yes
15:16 Alias In the community of people who are not us
15:16 Andy I haven't played with it because it's not something that I'm personally likely to use.
15:16 Alias Right
15:16 szabgab about 99.37% of the programmer cannot handle command line
15:16 Alias It has that smell of The Thing That Might Bring People Back To Perl
15:16 Andy szabgab: Recap of the birthday party would be good Perlbuzz fodder.
15:16 Alias Hard to really break it down into specifics
15:17 PerlJam bring them back?  Have they really left?
15:17 NotFound szabgab: How do you calculate that? With a gui calculator?
15:17 Alias PerlJam: Sure
15:17 szabgab Andy, Alias we need someone who can actually right English
15:17 Alias hrm?
15:17 purl hrm is it possible to link up all my google/yahoo/pobox/whatever identities together?
15:17 Andy szabgab: I'm glad to apply my Powers Of Editing to whatever.
15:18 szabgab POE ?
15:18 purl somebody said POE was buu's last refuge from the Moose or perl -MAcme::POE::Acronym::Generator -le '$x = Acme::POE::Acronym::Generator->new();  print join " ", $x->generate for (0..100);'
15:18 Coke POE, POS, whatever it takes.
15:18 szabgab purl, POE is also Powers Of Editing
15:18 purl okay, szabgab.
15:18 NotFound You can add a vi mode to the editor, to also attract people on the other side ;)
15:18 Alias NotFound: Partially implemented
15:18 Andy szabgab: I'm saying that I will fix up whatever you write.
15:18 NotFound And an emcas mode, for completeness.
15:18 Alias That too
15:19 Andy I can indeed write English.
15:19 Alias And Ultraedit, and Notepad++
15:19 NotFound Impressive
15:19 Alias It's only bits and pieces of course, just a beginning
15:19 Alias Most are missing fairly large dependant features
15:19 Alias Like custom key bindings (which we don't have yet)
15:19 NotFound And a wordstar mode? ;)
15:19 Alias Or Find and Replace in files
15:20 Alias But we're working through them
15:20 Alias Plan to just keep adding a new committer per week until the problems go away
15:20 Alias So far the plan is working
15:20 Alias purl, 50 / 43
15:20 purl 1.16279069767442
15:21 Alias purl, 43 / 50
15:21 purl 0.86
15:21 NotFound I installed it yesterdat on my laptop, but haven't had time yet to play with it.
15:21 Alias So we're at 0.86 committers per week :)
15:21 donaldh joined #parrot
15:21 szabgab there is a vi mode
15:21 szabgab please feel being attracted :-)
15:21 szabgab partially, that is
15:21 Alias Emacs or Emacs Not, There is No Vi
15:22 NotFound Only pitfall I've seen is that it loads lots and lots of modules from cpan.
15:22 Coke the installer does?
15:22 Alias NotFound: Turns out you need to do a lot of different things to create an IDE
15:22 Coke (is the problem that they are not bundled, or that they are needed?)
15:22 NotFound And a lot more if you add perl6 and parrot modules.
15:22 Coke (because the latter isn't a problem. =-)
15:23 NotFound Not a problem for me, but for the targeted users, maybe.
15:23 Coke people who complain that we use CPAN need a talking to.
15:23 Alias NotFound: The targetted users get a single .msi installer that finishes in 30 seconds
15:23 szabgab we are going to have installers
15:23 szabgab and is being packaged my most distros
15:23 Alias ... aggressively
15:23 szabgab but that will take a few cycles to get to them
15:24 NotFound Nice, then.
15:24 PerlJam agression++
15:24 Alias Hopefully one day someone will make Perl easy to install on Mac
15:25 Coke Alias: does activestate not qualify?
15:25 NotFound Is easy, just drop Mac OS and install linux ;)
15:25 Alias Coke: How do you bundle Padre into ActivePerl?
15:25 Coke excellent question. =-)
15:25 Coke Alias: if you don't mind the usual nest of deps, you can use macports.
15:26 Alias The usual nest of deps is a problem
15:26 Alias Especially through the ugly region of Wx::
15:26 NotFound Smoke done for linux/amd64/g++: http://smolder.plusthree.com/app/pu​blic_projects/report_details/24974
15:26 Coke but generating something you can drop into /Applications that contains all of a pre-built perl shouldn't be horrific.
15:26 Alias hrm
15:26 NotFound A very good landing of the branch :)
15:26 Alias I suppose you could target it SPECIFICALLY to various cats
15:26 Alias Padre for Snow Leapord
15:26 Alias sp..
15:27 Coke Alias: yah. you might be able to do some sharing, but that's a good first pass.
15:27 Coke also a good idea to poke the camelbones people with a stick.
15:27 Alias Already did
15:27 Coke on list?
15:27 Alias They are, well, philosophically against the idea of cross-platform applications
15:28 Alias Seemed to think Padre was basically a stupid idea
15:28 szabgab I have started to build a strawberry -like Perl 5 distro for Linux/Unix
15:28 Alias So I didn't bother pushing after the initial conversation
15:28 szabgab with the aim to have padre in it that people can just unzip an go
15:28 Alias They can just steal code from us, like the vi people
15:28 Coke it's not in their niche, sure, but I figured they could help with the packaging issue. Ah well.
15:28 NotFound Alias: tell them that in Spain insulting Padre is a very very bad thing
15:29 Coke por que?
15:29 szabgab also in Lisbon :-)
15:30 Coke Alias: I can play with bundling something for 10.4/intel.
15:30 Alias Coke: That would be nice, although I'm not sure if it requires upgrades to core modules
15:30 Alias Which we might, we've caused a lot of dep cascades
15:31 Alias Packaging something with core upgrades could be iffy
15:31 szabgab anyway I'd need a lot of help with such a Perl distro
15:31 Alias I dunno how it works exactly on Mac
15:31 Alias szabgab: Since Mac is a relatively fixed target, might be worth looking at the overlay thing first
15:32 Alias Launcher puts /opt/padre/lib at the front of @INC, then boots from a mix of bundled and installed modules
15:32 Alias or whatever
15:32 szabgab I am planning to buy a mac but it is about 50% more expensive here than in the USA online store
15:32 Alias Ditto here
15:33 szabgab interesting as PCs are about the same price as in the US
15:33 Alias There's competition for PCs
15:33 Alias Jobs doesn't like Australia much, so splat
15:33 NotFound I've thinked about buying a Mac, but got a 42" TV instead
15:33 Alias 50%-100% tax
15:33 JC1 joined #parrot
15:34 Alias Ditto wherever else
15:34 Alias So I only buy Apple stuff for things like gifts, where the price shouldn't really get in the way of something the recipient would love
15:34 Theory joined #parrot
15:34 Alias Although I'm inches from caving and buying the Star Trek Data Pad... er... I mean iPhone
15:40 ruoso joined #parrot
15:54 davidfetter joined #parrot
15:55 cotto good morning
15:55 purl And good moroning to you, cotto.
15:57 Coke Alias: I have the old 4GB one. very happy even with that old 'clunker'
15:57 * Alias nods
15:58 Alias I've been pondering a second hand one, so I don't have to leave my mostly awesome mobile plan
15:58 Alias I discovered 365 day expiry prepay
15:58 * particle waits for the iPhone Shuffle
15:58 Alias Since I live across the road from work, if I'm making any really expensive calls I just duck into work :)
15:58 Alias So my phone bill now is negligable
16:00 particle how big is backpan in GB?
16:01 Psyche^ joined #parrot
16:02 Alias 11ish, compressed
16:02 Alias Something in that range
16:02 Coke particle: I wonder how much gear they'd need just to wire a phone into your head.
16:14 TimToady the actual phone part is easy, but putting the little touchscreen into your eye is a bit harder
16:15 Coke I imagine you'd put the touchscreen on teh back of your wrist.
16:20 Alias Or on your fingers
16:20 Alias Like a mouse, just invert it
16:20 Alias Drag across anything
16:22 TimToady your wrist and fingers are not part of your head
16:24 TimToady well, I suppose you can stick your fingers in your ears or your mouth, and then they're in your head, by some definition
16:25 mikehh examples_tests FAIL - All others PASS (pre/post config, smolder, fulltest) at r4109 - Ubuntu 9.04 amd64
16:25 Alias TimToady: So... I have to apologise
16:26 * TimToady sticks his fingers in his mouth
16:26 Alias TimToady: Since your trip to AU, I have managed to establish proof that the "first ever installable Perl 6 alpha" CD I gave you can never possibly actually install and work :)
16:26 Alias So I need to reset the clock on that one
16:26 mikehh examples_tests - Failed test 'examples/namespace/namespace_dump.pir' - at t/examples/namespace.t line 46.
16:26 TimToady well, I thought it was a bit premature, but it was a good stunt nonetheless
16:27 TimToady any PR is good PR, and all that...
16:27 Alias It was a great stunt :)
16:27 Alias And admittedly, it was BARELY at the limits of buildability
16:27 Alias This years stunt should be much better
16:27 TimToady so I refuse to accept your apology, since none was necessary
16:27 TimToady :)
16:28 Alias I'm hoping we can get this whole "Perl 6 IDE" thing ready in time for OSCON
16:28 TimToady yowser
16:28 Theory_ joined #parrot
16:28 Alias And distribute maybe some kind of "Perl 6 Beta + IDE" CD for that
16:28 Alias etc etc
16:28 Alias Maybe not beta...
16:28 Alias But whatever
16:28 Whiteknight quite ambitious!
16:29 Alias The basics all pretty much work
16:29 Alias http://padre.perlide.org/trac/raw-attachment/wi​ki/Screenshots/Padre-Plugin-Perl6-with-STD.png\
16:29 Alias http://padre.perlide.org/trac/raw-attachment/wi​ki/Screenshots/Padre-Plugin-Perl6-with-STD.png
16:29 PerlJam Alias++ for the publicity
16:30 Alias PerlJam: People keep forgetting Elaine's Law
16:30 Whiteknight Elaine's law is really less of a law and more like an order with an insult
16:31 PerlJam I don't think they forget as much as they are focused on other  things
16:31 Alias Whiteknight :)
16:31 Alias PerlJam: Potato, Potato
16:32 Alias Whiteknight: I need to note that joke down for later
16:32 masak Alias: it'd have been more impressive if you'd have written that using IPA. :)
16:33 TimToady the main thing that's misleading about Elaine's Law is the "just"  :)
16:34 Alias TimToady: Well, it's "just" in the sense that's all you need to do to be successful, really
16:34 Alias I ended up finding some actual research
16:35 Alias Ease of Procurement 2x the impact on market share than Product Quality
16:36 Alias So we all should be spending 2x as much time working on installers than we do on the actual making-it-work stuff :)
16:36 chromatic joined #parrot
16:39 NotFound Alias: unless that means we don't make it work ;)
16:39 Alias Make it Install, then Make it Work, then Make it Maintainable, then Make it Fast :)
16:39 jan joined #parrot
16:40 NotFound Respctfully disagree. If it installs but doesn't work, people get more frustrated.
16:40 Alias You just don't tell them it exists :)
16:40 Alias Till it works :)
16:41 NotFound Nice for a product company, but we need feedback to make it work.
16:42 Alias How can they give feedback if it doesn't install :)
16:43 NotFound Spending less time in foolproof installers ;)
16:44 cotto bacek_at_work, could you write up your changes from the merge?  Even if it's just quick and dirty, it'd be helpful.
16:44 cotto I'd be happy to flesh it out.
16:45 Theory joined #parrot
16:46 Coke Alias: as a language writer on top of parrot, I have trouble getting developers to work on bugs I raise because they are forced to do an install first.
16:47 Coke (though as a language writer, seems like I should be working against an install.)
16:48 Alias It's not easy enough for them?
16:54 Coke So I'm told.
16:54 chromatic It's true.
16:56 Alias Coke: Clearly more work needed on installer :)
16:57 cotto afk for much of the day
16:58 chromatic Installation doesn't lend itself to quick debugging.
16:59 Alias restate that?
16:59 Psyche^ joined #parrot
17:01 chromatic It's many, many times faster to modify and rebuild libparrot from a build tree for testing a language than to modify and rebuild and reinstall all of Parrot.
17:02 Alias Regardless, should THAT process be automated?
17:02 Alias Whatever is the easier way should also be trivialised
17:03 Coke part of the problem is that "building against installed parrot" and "building against built parrot" are two very different things atm.
17:03 Coke once that's resolved, chromatic will be happy, and I'll be ... as happy as I get.
17:04 * PerlJam surrounds Coke with cute bunnies and sunshine to see if that improves the happiness quotient
17:04 Coke Really?
17:04 chromatic I want to change the minimal amount of Parrot to test a change.
17:05 chromatic If I have to set up a new installation directory and reconfigure and rebuild all of Parrot to start testing a change, I'm less likely to do it.
17:05 * Coke roots around something blunt or sharp.
17:05 PerlJam Coke: it's your choice if the bunnies are the small furry kind or if they're more of the playboy variety  ;)
17:05 Coke ^_O
17:06 pmichaud Good afternoon, #perl6
17:06 Coke pmichaud: hello, patrick.
17:06 PerlJam pmichaud: this is #parrot  :)
17:06 Coke pmichaud: hey, any chance you'll have any tuits to fix the regex empty anchor issue before 1.4?
17:07 Coke (or failing that, put some hints on the ticket.)
17:07 pmichaud Coke: you mean the repeating zero-length match issue?  or am I confusing it with something else?
17:07 pmichaud Good afternoon, #PARROT
17:07 Coke that is what I probably meant, yes!
17:07 pmichaud Coke: very unlikely I'll fix it pre-1.4.
17:07 Coke (it's killing 2 partcl spec tests.)
17:07 pmichaud It will likely be fixed as part of the overall PGE refactor.
17:14 Coke Andy: Test::Pod is failing with a vanilla install via CPAN.
17:14 particle which is why it's on the FAIL100 iirc
17:14 Andy Coke: Need more details.
17:14 Alias It also broke back-compatibility for a long of things
17:15 Coke andy - opening a ticket on rt.cpan.org now.
17:15 Alias Because 600 modules depend on it that shouldn't have, because of stupid Kwalitee cargo culting
17:15 Alias Coke: Andy is allergic to rt
17:15 Andy left #parrot
17:15 Alias (If that wasn't a joke)
17:15 Alias ook
17:15 Alias Did we scare him off?
17:15 Andy joined #parrot
17:15 Alias nope
17:15 Andy ooops, back
17:16 Alias Andy: If you can without hurting anything, please restore support for 5.6.2
17:16 Alias Or thereabouts
17:16 Andy Alias: why?
17:16 purl Left field.
17:16 Alias Andy: It's not your fault, but you're a hard-dep for lots of things that shouldn't have made you a dep
17:16 Alias 600 modules I think
17:17 Alias You just accidentally broke 5.6.2 back-compat for all of them
17:17 Andy oK, but who cares about 5.6.2?
17:17 chromatic Not the channel #parrot.
17:17 Alias It's 5% of CPAN, I'm sure there's some in there
17:17 Andy chromatic: There there, Mr. Crabbypants. <smoochies>
17:18 Alias Getting rid of those cargo-culted deps is a separate issue
17:18 Alias I'm gunna make up a list for that I think
17:18 Andy But who's installed those modules on 5.6.2?
17:18 Andy Are there new installs going out in 5.6.2?
17:19 Alias There's module installs
17:19 Andy I'm pretty much in chromatic's camp on this.
17:19 Alias Usually on big iron
17:19 chromatic My camp on this is "Gentlemen, I don't see the relevance to #parrot".
17:19 Alias We've got some older AIX machines that have it
17:19 Alias true
17:19 Alias to #toolchain?
17:19 Andy There's a #toolchain?
17:19 chromatic Yes, please.  Thank you.
17:20 Alias Andy: We started it to resolve the lack of communications between the different areas
17:20 Coke Andy: https://rt.cpan.org/Ticket/Display.html?id=47913
17:20 Andy What is t/link.t?
17:20 Andy oops
17:20 Andy sorry
17:36 pmichaud Alias: (from backscroll)  Rakudo releases are synchronized with Parrot releases.  Each Rakudo release comes out two days after a Parrot release, and is intended to be able to build/run with the just-released Parrot.
17:36 pmichaud We're fairly careful to make sure that's the case.
17:36 Alias ok
17:37 Alias Wonder why gabor said he needed to use svn then...
17:37 pmichaud So Rakudo #18 was built from Parrot 1.3.0
17:37 pmichaud Rakudo #19 (July 2009)  will be released on July 22, based on Parrot 1.4.0 (released July 20)
17:37 * Alias nods
17:37 szabgab I just buildt the latest rakudo from git using --gen-parrot
17:38 szabgab I guess there are other ways as well
17:38 szabgab built
17:38 Alias No matter
17:38 Alias I've set up the zip file in the place it needs to be now
17:38 Alias Hopefully
17:38 pmichaud Currently Rakudo still cannot build an executable from an installed Parrot.
17:39 pmichaud (This is because Parrot doesn't yet provide all of the tools needed to do so.)
17:48 Coke you can, of course, get by against an installed parrot as partcl does. for me it's just a PITA to do both atm.
17:48 allison joined #parrot
17:48 pmichaud Coke: no, that's not true either, at least not for all platforms.
17:49 pmichaud I.e., there are some platforms on which Rakudo can build+execute now that doesn't work for the installed version of Parrot.
17:49 pmichaud so, switching over to completely using the installed version of Parrot (as partcl does) means cutting off those platforms from being able to use/run Rakudo.
17:50 Z3vil joined #parrot
18:01 pmichaud afk, lunch
18:05 davidfetter bon appetit
18:17 Coke pmichaud: I obviously have a smaller user base, but partcl is working against installed on win32, os x/intel, and at least one linux variant.
18:18 pmichaud Coke: well, I'll probably give working against an installed parrot a shot again today or tomorrow, but I know that the last time I tried it there were some outstanding issues that are (afaik) unresolved.
18:31 hercynium joined #parrot
18:33 sekimura joined #parrot
18:36 NotFound make: *** No rule to make target `/home/julian/parrot/runtime/parr​ot/library/PGE/Perl6Grammar.pbc', needed by `src/gen_grammar.pir'.  Stop.
18:36 NotFound This is what rakudo tell me if I rename the parrot build dir
18:37 NotFound So no, it doesn't use the installed dirs
18:37 Coke TimToady: hey, you're famous: http://www.groklaw.net/articl​e.php?story=20090714145746535
18:39 payload joined #parrot
18:40 payload i read so much "unkown reason" in   make test :(
18:54 pmichaud NotFound: correct, Rakudo doesn't use the installed Parrot yet.  (Because not all of the tools we use now are available in installed parrot.)
18:55 NotFound Perl6Grammar.pbc is
18:56 NotFound ./lib/1.3.0-devel/library/PGE/Perl6Grammar.pbc
18:57 pmichaud NotFound: I don't understand.
18:57 pmichaud NotFound: it's not sufficient for us to simply update the paths of the .pbc files, if that's what you're getting at.
18:57 mikehh joined #parrot
18:57 pmichaud the problem is in building the dynpmcs
18:58 HG` joined #parrot
18:58 NotFound Yes, but some of the problems can be solved, or at leaset clarified, in the meantime.
18:58 pmichaud no, they can't
18:59 pmichaud currently Rakudo uses dynpmc.pl and dynoplibs.pl from tools/build/ in order to generate the dynpmcs
18:59 pmichaud those two scripts aren't available in an installed parrot
19:00 NotFound Your value of 'some' if very strict ;)
19:00 pmichaud look, I'll put it another way.  I've already done most of the changes needed to get Rakudo to build from an installed parrot
19:00 pmichaud it's available from the ins branch in the github repository
19:00 pmichaud that version of rakudo uses the .pbc files from the installed parrot
19:01 pmichaud however, the pbc_to_exe that it was using didn't function without the build tree, and the dynpmcs won't compile on specific platforms (e.g., windows using certain compilers)
19:02 pmichaud if you're proposing that rakudo "solve" some problems by using some files from the build tree and other files from an installed parrot -- I *really* don't want to do that.
19:02 pmichaud I think we should either build completely from an installed parrot, or not at all.
19:03 pmichaud (because I don't want to troubleshoot issues of file incompatibilities between the two, or worrying about which files are being loaded from where.)
19:03 NotFound pmichaud: no, I'm just trying to understand what config values can be used (or are needed, if not present yet) to be able to build both from a parrot build tree or from an installed parrot.
19:03 pmichaud NotFound: okay.  Then you want to be looking at refactoring the result of Parrot's "make" so that it builds a replica of the install tree inside of the build directory.
19:04 pmichaud This is the current plan that has been discussed and tentatively approved by allison.
19:04 pmichaud then we don't have to worry about keeping two different sets of configuration variables.
19:04 pmichaud more precisely, Rakudo doesn't have to worry about managing to different sets of configuration variables.
19:05 pmichaud The message describing this is at http://lists.parrot.org/pipermail​/parrot-dev/2009-July/002482.html
19:06 allison pmichaud: are you still using the dynoplibs.pl and dynpmc.pl to build your ops and pmcs?
19:06 pmichaud allison: yes, because we don't have a workable alternative yet (unless you've resolved the RT ticket I pointed out yesterday)
19:07 jrtayloriv joined #parrot
19:07 allison pmichaud: okay, I'm just asking because those can never work installed, they're generated with hardcoded build paths
19:07 pmichaud agreed, I'm not arguing they should be present.
19:08 allison well, you could adopt them, if you want them
19:08 pmichaud I'm explaining why simply changing the locations of the .pbc files won't be sufficient to get Rakudo to build against an installed parrot.
19:08 allison pmichaud: which ticket? I've been looking at the three you mentioned in IRC yesterday, and working my way through the "install" tickets
19:08 pmichaud looking...
19:09 allison and, yup, definitely more than just file location
19:09 pmichaud RT #66558
19:10 allison oh, the spaces issue
19:10 allison yes, I'm fine with applying Andy's patch to parrot
19:10 pmichaud I haven't had a chance to look directly at Andy's patch to see what it does for Rakudo yet.
19:10 Andy Probably just a lot of consting.
19:10 allison all it does is direct substitution instead of setting up variables
19:11 allison (within the Makefiles)
19:11 pmichaud (Andy:  sorry, adougherty, not alester :-)
19:11 Andy pmichaud: I know.
19:11 Andy Or assumed.
19:11 purl assumed is a type of verb
19:11 pmichaud anyway, as mentioned above, I plan to re-attempt the rakudo from parrot install tomorrow or Fri.
19:11 allison pmichaud: so it's a work-around for incompatibilities in various Makefile implementations
19:12 pmichaud oh.  I thought it was more of a work-around for the fact that different compilers have different expecations/requirements
19:12 allison pmichaud: okay, cool, I'll get parrot patch applied today
19:12 pmichaud (some compilers require space, some forbid it, and parrot_config doesn't handle that decision logic for us)
19:12 allison pmichaud: yes, but the config variables have all the right spaces or lack of spaces already
19:13 pmichaud that seems.... wrong
19:13 allison it's just the Makefiles are inconsistent about keeping them when they're substituted from variables
19:13 allison well, that's why we have platform-specific config options
19:13 pmichaud if the config variables are already correct, it would seem the makefile would be straightforward.
19:13 allison so they can be exactly right
19:13 pmichaud anyway, I'll look at it tomorrow and report back on-list.
19:14 pmichaud right now I have to go get lunch if I'm to have any hope of eating before dinner.
19:14 allison ah, phone call, I've got to grab lunch
19:20 donaldh joined #parrot
19:24 Coke pbc_to_exe was broken for a while. it's fixed now.
19:25 Coke (and partcl is building dynpmcs, though that gets back to the 'use a tool' vs 'write a makefile')
19:33 jrtayloriv Sorry to be a bother, but I screwed up the first time I posted a patch to trac (didn't read the submissions.pod carefully enough), and wanted to know if I did everything right this time: https://trac.parrot.org/parrot/ticket/839 ... anything other information I should have added here?
19:34 dalek TT #839 created by jrtayloriv++: [PATCH]: Fix broken perldoc tags in docs/book/draft/ch01_introduction.pod
19:34 eternaleye joined #parrot
19:36 dalek parrot: r40110 | NotFound++ | trunk (4 files):
19:36 dalek parrot: [lib] Fix SDL font color, TT #798, richardh++
19:36 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40110/
19:41 dalek TT #798 closed by NotFound++: [BUG] and [PATCH] SDL Font color is wrong
19:41 Theory joined #parrot
19:42 eternaleye joined #parrot
19:43 Coke allison: ping.
19:44 allison Coke: pong
19:44 Coke allison: can you cut a new release of Pod::Simple with https://rt.cpan.org/Ticket/Display.html?id=39709 ??
19:45 Andy And ID who you're talking to here http://www.flickr.com/photos/kubina/892625446/
19:45 Andy http://twitter.com/Skud/statuses/2656169168
19:46 Andy And pick up a gallon of milk?
19:46 allison Coke: sure
19:46 chromatic Andy, that's Geneva Wall.
19:46 allison Andy: aside from Geneva and Aron Wall, and Tom Christiansen?
19:47 Coke allison: thanks. that'll help me get an osx distro of padre cut. =-)
19:47 Andy Geneva is who Skud was looking for.
19:47 Andy oops, I said it was Lewis, not Aron
19:47 allison Andy, looks like Gloria sort of hidden behind my sholder
19:48 allison Andy: oh, both the same photo
19:48 chromatic Aron's in the middle, Lewis is off to the left.
19:48 Andy Aha, so I was right.
19:48 allison yeah, Lewis has long hair, sort of standing out on the side looking on
19:49 allison Aron is in the conversation
19:49 Andy right, updated.
19:50 allison Andy: pretty certain the guy in the ballcap and printed shirt behind Aron is Larry too
19:51 Andy Skud was just askin' about Geneva.
19:53 chromatic will be ~5 late to the call
19:53 particle noted
19:58 sekimura_ joined #parrot
20:00 TimToady phone
20:04 darbelo joined #parrot
20:09 nnunley Weird.  ResizablePMCArrays are not clonable?
20:10 pmichaud they're clonable, but they don't clone what they refer to
20:10 nnunley That is, using the method call form.
20:13 nnunley pmichaud: Been hacking on the bootstrap version of nqp.  This is not a valid form:
20:13 nnunley $<name><ident>.clone();
20:15 nnunley At least, not yet.  What do I need to do?  Emit pir to work around the issue?  Create a new pmc?
20:18 mokurai joined #parrot
20:19 allison joined #parrot
20:28 hercynium joined #parrot
20:30 pmichaud do we need a clone at all?
20:40 nnunley When I eliminate the clone, it seems to corrupt a doubly linked list.  Of course, it could just be exposed by evaluation getting farther along.
20:45 dalek rakudo: dfe8518 | pmichaud++ | build/PARROT_REVISION:
20:45 dalek rakudo: Bump PARROT_REVISION to after the keys-branch-merge for more testing.
20:45 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/d​fe85183c5b9314152432b8e05a71fb3853e9d43
20:47 Theory joined #parrot
20:48 Theory joined #parrot
20:48 treed pmichaud: I think cardinal uses clone in some cases. Unsure if it's the same clone that you're talking about.
20:50 nnunley treed: It's more of an abstraction question.  The hand written pir code uses the clone op code, but the nqp bootstrap uses method dispatch on the ResizePMCArray pmc.
20:50 treed Ah.
20:54 allison joined #parrot
20:59 pmichaud seems like we should be able to avoid the clone somehow.
21:03 NotFound nnunley: do you have some pir example of that segfault?
21:04 nnunley NotFound:  I have a github repository with the version of nqp that I'm working with. git@github.com:nnunley/nqp.git
21:06 nnunley NotFound: I haven't narrowed it down yet in terms of the actual pir.  I'm using an x86-64 linux system.  It definitely exhibits the exception when I run t/26-method_ops.t
21:06 pmichaud afk for a short while
21:12 hercynium joined #parrot
21:15 nnunley I'm going to see if I can't play with a delta debugger to generate the error that I'm seeing.
21:21 payload joined #parrot
21:24 nnunley Ok.  Looks like it's something to do with the generated standalone executable, and not the pir itself.
21:25 hercynium joined #parrot
21:26 nnunley If I run the generated pir with parrot, I get error messages, but no stack corruption.  If I run it with the stand alone, things fly loose.
21:40 chromatic Does the backtrace show Parrot_really_destroy?
21:42 jrtayloriv joined #parrot
21:45 nnunley chromatic: It does.
21:45 purl if you say so...
21:45 chromatic Run parrot with --leak-test and you'll likely see the same thing.
21:47 nnunley chromatic: Nope. Slightly different: PackFile_Segment_destroy
21:47 nnunley Wasn't in the stack before.
21:47 nnunley It is now.
21:48 nnunley Still passing through Parrot_really_destroy, though.
21:49 chromatic The memory layout is slightly different, but it's not a problem localized to the fakecutable.  Good to know.
21:49 chromatic I think a segment gets double freed somehow.
21:50 jdv79 nnunley: some guy was in here a few days ago that said he may have bootstrapped nqp
21:50 nnunley jdv79: Cool.
21:50 jdv79 he wasn't ready to post the code though:(
21:50 jdv79 just fyi
21:51 nnunley My code is public.
21:51 jdv79 i know
21:51 nnunley If you see him, let him or me know.
21:51 nnunley :)
21:51 nnunley Back in a bit.
21:52 jdv79 his nick is einstein
21:54 nnunley jdv79: Thanks.  I'll watch for him.
21:55 nnunley chromatic: Is there anything I can do to help locate the leak?
21:55 chromatic Disable stack location randomization and set a breakpoint on malloc to see what wants that pointer.
21:55 chromatic Then it's standard leak checking.
21:56 chromatic If the PIR is small enough, Valgrind won't hurt.
21:56 chromatic Other than that, you already know how the process works.
21:57 nnunley Cool.  I'll attempt to find it later on tonight.
22:02 dalek rakudo: 10843c7 | moritz++ | src/setting/Operators.pm:
22:02 dalek rakudo: infix:<eqv> for hashes
22:02 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/1​0843c757f36fd2981fd88ad38b2985cbfb91627
22:06 bacek joined #parrot
22:17 dalek parrot: r40111 | NotFound++ | trunk (3 files):
22:17 dalek parrot: [config] generate pir constants from PARROT_LIB_ enums in library.h, TT #747
22:17 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40111/
22:21 athomason joined #parrot
22:25 bacek Good morning #parrot
22:29 mokurai joined #parrot
22:31 athomason joined #parrot
22:36 rg joined #parrot
22:37 Whiteknight joined #parrot
22:37 dalek rakudo: 9a7a1dc | moritz++ | src/setting/Operators.pm:
22:37 dalek rakudo: two identical objects are also equivalent
22:37 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/9​a7a1dce066c4136594693b44e1d220795c7cc88
22:38 clunker9__ joined #parrot
22:38 Limbic_Region joined #parrot
22:40 Whiteknight good afternoon again #parrot
22:41 Theory_ joined #parrot
22:43 dalek TT #839 closed by chromatic++: [PATCH]: Fix broken perldoc tags in docs/book/draft/ch01_introduction.pod
22:46 Zak joined #parrot
22:59 ruoso joined #parrot
22:59 Khisanth joined #parrot
23:18 jrtayloriv joined #parrot
23:20 bacek joined #parrot
23:20 donaldh joined #parrot
23:31 patspam joined #parrot
23:47 bacek_at_work hi again
23:47 purl oh, you're back!
23:47 bacek_at_work annoying girl...
23:47 bacek_at_work cotto: around?
23:48 bacek_at_work msg cotto Quick-and-dirty branch description in TT#761
23:48 purl Message for cotto stored.
23:53 Whiteknight we need to add test files for all the new PMC types
23:56 jrtayloriv joined #parrot

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

Parrot | source cross referenced