Camelia, the Perl 6 bug

IRC log for #parrot, 2009-07-21

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 mikehh All Tests PASS  (pre/post config, smolder, fulltest) at r40174 - Ubuntu 9.04 amd64
00:01 whiteknight joined #parrot
00:15 hercynium joined #parrot
00:17 TiMBuS joined #parrot
00:18 whiteknight Win64 is a mess. I don't know if this will ever work right
00:18 whiteknight it's failing t/op/64bit.t because "int" and "long" in MSVC64 are both 32-bit values, so IMCC isn't converting constants in PIR code correctly
00:19 whiteknight and MSVC doesn't implement strtoll like it should. It has _atoi64, but that doesn't let you specify a non-decimal base
00:19 whiteknight if we use "long" instead of "long long" to make things simple, sizeof(INTVAL) != sizeof(void*), and hillarity ensues
00:20 chromatic We shouldn't assume sizeof (INTVAL) is the same as sizeof (void *)
00:21 whiteknight we *shouldn't*, but we do in many places
00:21 whiteknight "we *shouldn::t*" :)
00:22 whiteknight And either way, the fact that 64-bit support in Windows is so poor is depressing
00:24 dalek decnum-dynpmcs: r114 | darbelo++ | trunk/aux/decTest/src/ (3 files):
00:24 dalek decnum-dynpmcs: [decTest] Made some changes to the grammar and altered the emitted code.
00:24 dalek decnum-dynpmcs: Context operations now work, as do the tests that involve operations implemented
00:24 dalek decnum-dynpmcs: in procs.pir.
00:24 dalek decnum-dynpmcs: All we need to do now is emit a valid plan() and we're generating valid TAP.
00:24 dalek decnum-dynpmcs: review: http://code.google.com/p/decnu​m-dynpmcs/source/detail?r=114
00:26 cotto whiteknight, how much work (and of what nature) is required to make win64 happy?
00:27 whiteknight cotto: lots of work and of an evil nature
00:27 whiteknight not before this release
00:27 whiteknight (maybe not before next release either)
00:27 chromatic we *{shouldn}t*
00:28 bacek joined #parrot
00:28 whiteknight this whole platform is so goddamn bonkers it's amazing
00:28 bacek hi again
00:28 purl oh, you're back!
00:28 mikehh All Tests PASS  (pre/post config, smolder, fulltest) at r40175 - Ubuntu 9.04 amd64
00:31 whiteknight doesn't help that nmake doesn't understand -j, and the build takes freaking forever
00:32 jrtayloriv joined #parrot
00:36 dalek parrot: r40176 | whiteknight++ | trunk/config/gen/platform/​generic/platform_limits.h:
00:36 dalek parrot: [win64] fix to a macro that wasn't quite right before.
00:36 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40176/
00:40 dalek parrot: r40177 | whiteknight++ | trunk/PLATFORMS:
00:40 dalek parrot: [PLATFORMS] update with information about the state of the build on Win64
00:40 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40177/
00:42 dalek TT #846 closed by coke++: Add 'testing' to Component pane in Trac
00:47 mikehh rakudo (0cdcc88) on parrot r40175 (spec at r27624) make test/make spectest PASS - Ubuntu 9.04 amd64
00:53 whiteknight smolder?
00:53 purl it has been said that smolder is http://sourceforge.net/projects/smolder or web-based smoke test aggregator used by developers and testers to upload (automated or manually) and view smoke/regression tests using the Test Anything Protocol (TAP). or http://smolder.plusthree.com/app​/public_projects/smoke_reports/8
00:54 cotto whiteknight, would a code audit for misuse of sizeof() be necessary and/or sufficient?
00:54 whiteknight cotto: I really don't know, I assume not
00:54 whiteknight there are plenty of places e.g. VTABLE_get_pointer that use INTVALs and void* interchangably
00:55 cotto sounds unpleasant
00:56 cotto epically so
00:56 whiteknight yes
00:56 * allison moving networks
00:58 cotto Is there a way we could use a macro to make such misuses obvious, i.e. POINTER_SIZE?
00:58 cotto (this obviously won't happen before the release, but we'll need to address it eventually)
01:04 whiteknight irclogs?
01:04 purl irclogs is http://irclog.perlgeek.de/parrot/today or see also: infrared clogs
01:22 cotto It seems like we could do some clever hacking (maybe with splint) to automatically detect misuses of void*.
01:22 whiteknight I dont' know. After the release I'm not working on this particular task anymore
01:22 whiteknight at least, not until next release
01:22 cotto heh
01:22 cotto I'm not surprised.
01:22 whiteknight you said you wanted testing on rare systems, and I did that
01:22 cotto Yup
01:23 cotto and I thank you for it
01:23 * whiteknight would love to submit yet another smolder report on Ubuntu9.04-amd64
01:23 whiteknight smolder reports on ubuntu are like potato chips: you can't have just one
01:23 whiteknight or just one dozen
01:24 whiteknight when is the release going out, before or after #ps tomorrow?
01:24 cotto om nom nom nom
01:24 purl om nom nom nom is http://apina.biz/4991.gif
01:26 cotto That's mesmerizing.
01:26 hydrapheetz ooh
01:27 Zak joined #parrot
01:37 ascent joined #parrot
01:41 dalek parrot: r40178 | jhorwitz++ | trunk (2 files):
01:41 dalek parrot: list candidates for embedding API so we can begin deprecating functions that
01:41 dalek parrot: don't belong
01:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40178/
01:48 dalek TT #588 closed by jhorwitz++: pdd10-extending/embedding/nci, define parrot api
01:54 kid51 'make fulltest' passes at r40174 on both Linux/i386 and Darwin/PPC
01:56 chromatic Alright, time for a couple of bugfixes!
01:57 ascent joined #parrot
02:01 Infinoid ugh.  I had plans to do a bunch of testing and parroty goodness, and then Monday happened
02:01 Infinoid how's 1.4 looking?
02:01 kid51 tests are passing
02:01 kid51 cotto is alive
02:02 kid51 what more could you ask for? ;-)
02:02 Infinoid gummy bears?
02:02 Infinoid ooh, a lava lamp nightlight for my bathroom!
02:02 * kid51 has been busy classifying Trac tickets by component
02:04 cotto Yup.  Definitely not dead.
02:05 darbelo A rainbow colored yak.
02:09 ascent joined #parrot
02:12 darbelo kid51++ # inundating our inboxes with reminders of our laziness.
02:16 dalek TT #265 closed by jkeenan++: patch to avoid linking with -lpthread on openbsd
02:17 Zak joined #parrot
02:30 cotto Who's Aldo Calpini?
02:30 chromatic The John Galt of Parrot.
02:30 cotto He's listed as the PocketPC platform porter in RESPONSIBLE_PARTIES, but I don't recognize the name.
02:31 cotto John Galt?
02:31 purl John Galt is probably a fictional character.
02:32 kid51 http://search.cpan.org/~acalpini/
02:32 cotto yup; just found that
02:34 cotto any reason not to remove him?
02:35 chromatic The most recent messages I've seen from him were in 2007.
02:35 cotto I'll take that as a yes.
02:35 cotto Now that most HLLs are out of svn, does it make sense to list their maintainers in RESPONSIBLE_PARTIES?
02:36 chromatic No.
02:38 cotto out they go
02:39 chromatic Coke's HLL map example doesn't get fixed this release.
02:39 chromatic Way, way too invasive.
02:39 chromatic Ugh, now I have to figure out a web site to report a bug.
02:39 chromatic trac--
02:39 chromatic trac--
02:39 chromatic trac--
02:43 dalek TT #867 created by chromatic++: Implicit NameSpace Creation Does Not Respect Compile-time HLL Map
02:44 dalek parrot: r40179 | cotto++ | trunk/RESPONSIBLE_PARTIES:
02:44 dalek parrot: [docs] various updates to RESPONSIBLE_PARTIES, including removing HLL devs since HLLs are externally hosted now
02:44 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40179/
02:44 dalek parrot: r40180 | chromatic++ | trunk/t/pmc/namespace.t:
02:44 dalek parrot: [t] Added TODO test for TT #867, where HLL maps in :init code doesn't change
02:44 dalek parrot: the NameSpace created in the rest of the file.
02:44 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40180/
02:44 tomyan joined #parrot
02:48 janus joined #parrot
02:55 dalek rakudo: d69cfe6 | pmichaud++ | build/PARROT_REVISION:
02:55 dalek rakudo: Bump PARROT_REVISION to get proper exception from unfound regex.
02:55 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/d​69cfe6514eb5689028332485298ee3dba173187
02:55 dalek rakudo: b756ac9 | pmichaud++ | src/parser/quote_expression.pir:
02:55 dalek rakudo: Allow comments in << ... >>  (qq:ww) lists.
02:55 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/b​756ac97b0a7761a0b19b43092e907365ae38a62
02:57 Andy joined #parrot
02:57 dalek TT #551 closed by chromatic++: t/pmc/nci.t:  2 test failures
02:57 dalek parrot: r40181 | chromatic++ | trunk (3 files):
02:57 dalek parrot: [NCI] Excluded NCI calls with 'V' in their signatures from JITting and JIT
02:57 dalek parrot: thunking at startup (TT #551, mostly a patch by donaldh).
02:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40181/
03:03 Andy I'm here now
03:28 Andy joined #parrot
03:38 tetragon joined #parrot
03:38 mokurai joined #parrot
03:52 ascent joined #parrot
03:58 Andy joined #parrot
04:17 allison joined #parrot
04:38 Andy joined #parrot
04:50 elmex joined #parrot
04:51 tomyan joined #parrot
04:58 dalek parrot: r40182 | pmichaud++ | trunk/DEPRECATED.pod:
04:58 dalek parrot: [pct]:  Add deprecation note about possible change in PAST::Val code generation.
04:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40182/
04:59 dalek TT #868 created by pmichaud++: DEPRECATE generation of PAST::Val constants
05:00 cotto It's a deprecation PARTY.
05:00 cotto and you're invited!
05:01 Andy joined #parrot
05:24 dalek rakudo: ae29fbf | pmichaud++ | src/parser/grammar.pg:
05:24 dalek rakudo: Better handle hyphens and apostrophes after certain keywords (RT #64656).
05:24 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/a​e29fbf88caa40bb04fb7c8aa255b0f8a63e868d
05:39 redicaps joined #parrot
05:42 dalek markdown: 9f9e594 | fperrad++ | src/Compiler.pir:
05:42 dalek markdown: use a root namespace for Math::Rand
05:42 dalek markdown: review: http://github.com/fperrad/markdown/commit​/9f9e5949c98022ac1b1fef83b2cb765b086fb3ad
05:46 redicaps left #parrot
06:00 cotto allison, the PIR book was sent to the publisher, right?
06:01 chromatic I reviewed the draft; it looks good.
06:03 cotto so it's in its final form but hasn't been sent off yet
06:04 chromatic They sent us a proof.  I meant proof.
06:12 allison on chromatic's vote of confidence, I approved the book
06:12 allison so, now it's just a matter of waiting for the data to reach the retailers
06:16 mj41 joined #parrot
06:16 allison last time we did this, 2 years ago, the whole process from submitting the manuscript to the publisher to having it in online stores took a little over a week
06:16 allison this time it's a little over a month, which is annoying
06:17 cotto I wonder how long it'll be until a Parrot book isn't obsolete before people can buy it. ;)
06:18 chromatic Parrot 3.6
06:18 cotto sounds about right
06:20 cotto chromatic, do you have a plan to turn L1 from an idea into a usable specification?
06:20 cotto I don't consider there to be an onus on you.  I just thought you might already have such a plan simmering.
06:21 chromatic I liked the PCT ops/PMC idea.
06:26 iblechbot joined #parrot
06:26 chromatic Once we have some idea of the operations we need to support, I figured I'd steal whatever Smalltalk's slang does.
06:28 flh joined #parrot
06:31 dukeleto joined #parrot
06:41 cotto Are Rakudo and Partlc building from an installed Parrot (without the build tree) now?
06:41 cotto *Partcl
06:42 moritz rakudo still uses the build tree; dunno if it would be possible
06:44 cotto thanks
06:45 cotto I remember hearing some noise about that but didn't know if the problem was resolved.
06:53 NotFound cotto: I checked a few days, and it still depends on the build tree for several things.
06:54 cotto NotFound, ok.  I'll make sure that NEWS doesn't claim otherwise.
06:55 * NotFound going to dayjob
07:13 dalek parrot: r40183 | cotto++ | trunk/NEWS:
07:13 dalek parrot: [NEWS] final (?) update to NEWS before tagging the release
07:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40183/
07:21 Topic for #parrotis now http://www.parrot.org/ | 277 RTs left | Next release: 2009-07-21 | Commit freeze in effect!
07:23 Eevee joined #parrot
07:49 moritz cotto++
07:49 cotto As I suspected, the native_pbc tests explode when I bump the PBC_COMPAT to a new major version number.  I hope mk_native_pbc takes care of that.
07:52 chromatic It should.
07:54 cotto It seems to.
08:07 Zak joined #parrot
08:09 dalek parrot: r40184 | cotto++ | trunk (14 files):
08:09 dalek parrot: [RELEASE] prepare for 1.4.0 release
08:09 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40184/
08:16 mikehh All Tests PASS  (pre/post config, smolder, fulltest) at r40183 - Ubuntu 9.04 amd64
08:29 GeJ cotto++
08:30 moritz is bumping PBC_COMPAT necessary? I mean we did that recently after bacek's branch merge
08:34 cotto It's part of the release process for a supported release.
08:35 cotto it's a major version bump rather than minor
08:38 Topic for #parrotis now http://www.parrot.org/ | 277 RTs left | Next release: 2009-07-21 | 1.4.0 is tagged.  Resume committing!
08:38 dalek parrot: r40185 | cotto++ | tags/RELEASE_1_4_0:
08:38 dalek parrot: tagging 1.4.0 release
08:38 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40185/
08:40 payload joined #parrot
08:52 bacek_at_work cotto++ # Yay!
08:53 riffraff joined #parrot
08:54 masak joined #parrot
09:06 cotto Is someone online who has admin powers on the Drupal site?
09:08 cotto I can post the announcement but not change the alias for release/current.  It'll look kinda silly if the release announcement gets sent out without that change.
09:08 mj41 joined #parrot
09:10 cotto The release announcement points to the right place, so I don't think this is a blocker.
09:10 cotto on with the show
09:12 chromatic Editing.
09:12 purl i heard editing was better on the director's cut
09:13 chromatic Edited.
09:14 chromatic Your name and quote confused me; it looked like something I would have chosen.
09:16 cotto heh.  As I was looking through old quotes I saw you had one by the same author.
09:16 chromatic One of my favorites.
09:17 chromatic Next release I do, I'll probably pull something from "Lord of Light".
09:23 dalek website: cotto++ | Parrot 1.4.0 "Mundo Cani" Released!
09:23 dalek website: http://www.parrot.org/news/2009/Parrot-1.4.0
09:28 donaldh joined #parrot
09:29 cotto use Perl doesn't seem to be a fan of <pre>
09:29 chromatic <ecode>
09:31 cotto Can I unsubmit?
09:39 dalek xml: 0304594 | fperrad++ | build/Makefile.in:
09:39 dalek xml: fix target 'clean'
09:39 dalek xml: review: http://github.com/fperrad/xml/commit/03​045941f2b203adec86f983c581f912a373af4a
09:41 elmex joined #parrot
09:43 Topic for #parrotis now http://www.parrot.org/ | Parrot 1.4.0 "Mundo Cani" Released!
09:49 GeJ anyone else witnessing erratic plan failures on t/pmc/packfileconstanttable.t ?
09:49 GeJ locally, it fails once out of 4 runs (on average).
09:51 nopaste "GeJ" at 61.5.215.29 pasted "failing test (on average once out of 4 consecutive runs)" (15 lines) at http://nopaste.snit.ch/17341
09:52 dalek tracwiki: v85 | cotto++ | WikiStart
09:52 dalek tracwiki: mention 1.4.0 release
09:52 dalek tracwiki: https://trac.parrot.org/parrot/wiki/​WikiStart?version=85&amp;action=diff
09:57 cotto That's no good.
10:00 cotto GeJ, I can't get that test to fail.  Are you all up-to-date?
10:00 cotto nm.  If you weren't it'd fail all the time.
10:06 cotto The only item I didn't get to was marking the Trac Milestone 1.3.0 completed, which I don't think I have permission for.
10:06 cotto That'll be easy enough to take care of at #ps.
10:06 cotto now it's time for sleep'
10:11 GeJ cotto: # parrot --version
10:11 GeJ This is Parrot version 1.4.0-devel built for amd64-freebsd.
10:12 GeJ I'll have a closer look at it tomorrow. I need to get back to Perl 5 for $job.
10:13 GeJ I wouldn't worry though, my boxes are known to never work as expected.
10:17 cotto Hmmm.  I checked that.
10:18 cotto Is that from the tarball?
10:18 bacek joined #parrot
10:18 cotto My understanding is that only the tarball doesn't say -devel
10:19 cotto clock?
10:19 purl cotto: LAX: Tue 3:19am PDT / CHI: Tue 5:19am CDT / NYC: Tue 6:19am EDT / LON: Tue 11:19am BST / BER: Tue 12:19pm CEST / IND: Tue 3:49pm IST / TOK: Tue 7:19pm JST / SYD: Tue 8:19pm EST /
10:19 cotto good evening, bacek
10:19 cotto ;)
10:19 bacek good morning, cotto :)
10:21 * bacek looking forward to brake Parrot badly :)
10:21 cotto I'm just going to double check that the tarball doesn't say "devel" and them I'm going to sleep.
10:21 cotto first remove some of those deprecations
10:22 cotto I think they've doubled in the past week
10:22 cotto ok.  A Parrot built from the tarball doesn't have -devel, so I'm going to sleep.
10:22 cotto good night
10:23 bacek good night
10:25 GeJ Good night cotto.
10:38 mikehh All Tests PASS  (pre/post config, smolder, fulltest) at r40185 - Ubuntu 9.04 amd64
10:39 dalek parrot: r40186 | bacek++ | trunk (2 files):
10:39 dalek parrot: [cage] Remove deprecated Parrot_get_namespace_autobase function
10:39 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40186/
10:41 * bacek picking low-hanging fruits
10:41 dalek TT #155 closed by bacek++: Remove unused function Parrot_get_namespace_autobase
10:42 bacek Yay. Cheap karma! :)
10:46 dalek parrot: r40187 | bacek++ | trunk (3 files):
10:46 dalek parrot: [cage] Remove unused Parrot_find_global_k function
10:46 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40187/
10:48 dalek TT #156 closed by bacek++: remove unused function Parrot_find_global_k
10:53 dalek parrot: r40188 | bacek++ | trunk (3 files):
10:53 dalek parrot: [cage] Remove pobject_lives compatibility macro. Also fix parrot.supp to use Parrot_gc_mark_PObj_lives
10:53 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40188/
11:09 * allison successfully built Ubuntu packages for a preview tarball built from cotto++'s pre-release commits
11:15 bacek allison: any thought/objections/etc about TT#866?
11:22 donaldh joined #parrot
11:44 kid51 joined #parrot
11:46 * kid51 wakes up and discovers that we've already released
11:46 kid51 cotto pulled an all-nighter
11:49 mikehh_ joined #parrot
11:57 ascent joined #parrot
12:10 whiteknight joined #parrot
12:16 GeJ clock ?
12:16 purl GeJ: LAX: Tue 5:16am PDT / CHI: Tue 7:16am CDT / NYC: Tue 8:16am EDT / LON: Tue 1:16pm BST / BER: Tue 2:16pm CEST / IND: Tue 5:46pm IST / TOK: Tue 9:16pm JST / SYD: Tue 10:16pm EST /
12:17 GeJ Good morning whiteknight
12:17 dalek parrot: r40189 | jkeenan++ | trunk (21 files):
12:17 dalek parrot: Merge darwin2hints branch into trunk.  Functionality for detecting Fink and Macports packages are transferred to config/init/hints/darwin.pm.   Tests revamped accordingly.
12:17 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40189/
12:17 dalek parrot: r40190 | jkeenan++ | branches/darwin2hints:
12:17 dalek parrot: Branch has been merged into trunk and is no longer needed at HEAD.
12:17 purl i already had it that way, dalek.
12:17 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40190/
12:17 whiteknight good morning GeJ
12:22 bacek good moroning whiteknight
12:22 whiteknight hello bacek
12:23 whiteknight excellent work today on the deprecations
12:23 whiteknight you move quickly!
12:23 bacek whiteknight: I steal your karma removing pobject_alive :)
12:24 dalek TT #727 closed by jkeenan++: remove config steps 'auto:macports' and 'auto:fink'
12:26 whiteknight go for it. The function is already gone, it's just a macro now
12:26 bacek s/now/was/ :)
12:36 donaldh1 joined #parrot
12:36 TiMBuS joined #parrot
12:44 * whiteknight is very busy at work today and will need to post his #ps report very very early
12:47 tetragon joined #parrot
12:47 clinton joined #parrot
12:59 dalek parrot: r40191 | bacek++ | trunk (46 files):
12:59 dalek parrot: [cage] Replace deprecated usage of 'new "Iterator"' with 'iter $P0'
12:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40191/
12:59 dalek parrot: r40192 | bacek++ | trunk (2 files):
12:59 dalek parrot: [cage] Remove Iterator.init
12:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40192/
13:02 dalek TT #869 created by bacek++: [cage] t/pmc/iterator.t should be removed
13:06 iblechbot joined #parrot
13:09 PerlJam Did the parrot get released a tad earlier than normal because of cotto's timezone?  ;)
13:11 bacek s/timezone// :)
13:13 dalek parrot: r40193 | bacek++ | trunk (7 files):
13:13 dalek parrot: [cage] Remove deprecate nextkey_keyed* VTABLE. Part of TT#866
13:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40193/
13:21 dalek tracwiki: v6 | bacek++ | Deprecation
13:21 dalek tracwiki: parrot_new_pmc_hash replacement
13:21 dalek tracwiki: https://trac.parrot.org/parrot/wiki/D​eprecation?version=6&amp;action=diff
13:23 MoC joined #parrot
13:24 dalek TT #494 closed by bacek++: [DEPRECATED] remove parrot_new_pmc_hash
13:24 dalek parrot: r40194 | bacek++ | trunk (2 files):
13:24 dalek parrot: [cage] Remove deprecated parrot_new_pmc_hash function
13:24 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40194/
13:38 dalek TT #861 closed by jkeenan++: mingw32-make fails upon tools\build\c2str.pl --all
13:38 dalek parrot: r40195 | bacek++ | trunk (7 files):
13:38 dalek parrot: [cage] Remove deprecated Parrot_MMD_method_name function
13:38 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40195/
13:41 * Coke is sad that the HLL build issues were not resolved.
13:41 Coke (in time for 1.40
13:41 Coke er, 1.4.0
13:42 Hinrik left #parrot
13:53 NotFound Mundo Cani? :o
13:53 mikehh joined #parrot
13:53 moritz what does it mean?
13:54 NotFound Hard to translate...
13:55 * NotFound searching help...
13:57 GeJ Wikipedia says that it's a character from the book used in cotto's quote.
13:58 GeJ �a depressed dog that is always crying out in anguish�
13:59 moritz a non-robot marvin? :-)
14:00 NotFound GeJ: yes, but I don't know if that meaning is related or not with the spanish slang meaning.
14:00 HG` joined #parrot
14:01 NotFound Anyway, users in Spain will laugh :)
14:02 bacek Oh shi... It's tomorrow again.
14:02 * bacek must sleep
14:02 purl $bacek->sleep(8 * 3600);
14:03 dalek parrot: r40196 | bacek++ | trunk/lib/Parrot/Vtable.pm:
14:03 dalek parrot: [cage] Don't generate deprecated Parrot_mmd_func_names
14:03 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40196/
14:03 dalek TT #625 closed by bacek++: Deprecate Parrot_MMD_method_name and related defines.
14:03 bacek Someone have to teach purl about "Constant folding" optimisation...
14:07 GeJ bacek: already tomorrow? I would have thought it was in one hour.
14:07 GeJ clock ?
14:07 purl GeJ: LAX: Tue 7:07am PDT / CHI: Tue 9:07am CDT / NYC: Tue 10:07am EDT / LON: Tue 3:07pm BST / BER: Tue 4:07pm CEST / IND: Tue 7:37pm IST / TOK: Tue 11:07pm JST / SYD: Wed 12:07am EST /
14:08 GeJ au temps pour moi.
14:13 mj41 joined #parrot
14:15 NotFound Suggested translation of 'Cani': youg gipsy raper.
14:15 NotFound s/youg/young
14:17 Coke !?
14:18 NotFound According to some sources, this meaning is deformation of the more known word: 'cañí'
14:19 masak sounds likely.
14:20 NotFound Don't care about that, just don't be surprised is people from Spain laugh when they hear the name.
14:21 jonathan lol
14:21 jonathan Languages are awesome. :-)
14:22 NotFound Some other day I'll explain why Nissan 'Pajero' is called 'Montero' in Spain X-)
14:24 GeJ NotFound: interesting reason indeed :)
14:28 jonathan And then there was the Nova ;-)
14:28 * jonathan forgets with car company that was
14:28 jonathan *which
14:29 moritz nova - no va?
14:29 jonathan Yeah
14:29 jonathan "doesn't go"
14:29 jonathan Not a good name for a car. ;-)
14:30 jonathan Of course, in various languages it means "new", which is rather better.
14:31 mikehh joined #parrot
14:32 TiMBuS joined #parrot
14:32 soxet joined #parrot
14:32 NotFound 'Pajero' is a lot more more amusing than 'Nova', and worse for the owner image X-)
14:33 NotFound Language interoperability is hard.
14:33 mj41 joined #parrot
14:34 GeJ NotFound: that's why we have Parrot, haven't we?
14:34 NotFound Yeah.
14:37 dukeleto joined #parrot
14:41 Coke chevy nova: http://www.snopes.com/business/misxlate/nova.asp
14:41 Coke (urban legend)
14:43 mj41 joined #parrot
14:45 NotFound Makes sense, nova in spanish is accented in the 'o', and pronounced as "no va" will be accented in the 'a'
14:46 NotFound But the Nissan is named 'Montero' in Spain, this is not a legend.
14:54 mj41 joined #parrot
15:01 TiMBuS joined #parrot
15:04 szabgab joined #parrot
15:12 whiteknight joined #parrot
15:12 mj41 joined #parrot
15:15 dukeleto r40196 : when I do "make fulltest" i get Failed test 'examples/benchmarks/overload.pir' : http://gist.github.com/151381
15:16 whiteknight cotto++ great release!
15:17 particle whiteknight: previous versions of msvc express (free version) did not have an x64 compiler
15:17 particle good to see the current version does
15:17 whiteknight particle: yeah, I think it's a recent inclusion
15:17 particle cl.exe is msvc
15:17 whiteknight not that it does us any good, Parrot + cl64 = b0rk
15:17 particle and always has been
15:17 whiteknight I know all about cl
15:18 particle i can get you a vm with msvc 2008
15:18 particle free and legal
15:18 whiteknight I have Vista64 with MSVC 2008 on another partition
15:18 iblechbot joined #parrot
15:18 particle i have an msdn subscription courtesy of microsoft to use for parrot development
15:18 whiteknight oh, well that is a nice perk
15:19 particle so, any ms tools you need, i can get you
15:19 whiteknight okay, that's a sweet deal
15:19 whiteknight The biggest steps right now are implementing a replacement for strtoll and strtoull, which MSVC doesn't have
15:19 whiteknight once we have a solid replacement for those, I think many problems will go away
15:20 donaldh joined #parrot
15:21 whiteknight if MinGW ever releases an x64 version, we should be able to work fine on that oo
15:21 whiteknight too
15:21 particle strtoll is not c89
15:22 particle long long didn't exist in c89
15:22 Andy joined #parrot
15:22 whiteknight you sure about that? I'm pretty sure I read somewhere that it was
15:22 particle i'm sure
15:22 whiteknight yeah, maybe you're right
15:23 particle parrot uses long long even though it's not c89, it's the one major exception we make
15:23 whiteknight the only place where I think we need that exception is in Win64
15:23 whiteknight I don't know of any other 64-bit platform where sizeof(long) != 8
15:23 particle yes, so we need parrot specific types
15:25 Coke ah, perhaps now we can drop our requirement on a hideously outdated stack!
15:28 mj41 joined #parrot
15:34 flh joined #parrot
15:34 tomyan joined #parrot
15:35 NotFound dukeleto: you are right, examples/benchmarks/overload.pir is failing
15:36 NotFound What are these .include in that file. They don't seem to be used.
15:39 NotFound And it doesn't look to be using the multi sub, BTW
15:45 soxet joined #parrot
15:46 NotFound whiteknight: I'm currently building with Debian amd64, not Ubuntu amd64, don't count me in that large number ;)
15:46 particle left #parrot
15:48 particle joined #parrot
16:01 darbelo joined #parrot
16:09 particle msg whiteknight you should read this: http://www.unix.org/version2/whatsnew/lp64_wp.html
16:09 purl Message for whiteknight stored.
16:17 darbelo joined #parrot
16:18 cotto bacek++ is a maniac
16:25 cotto mundo cani?
16:26 cotto mundo cani is the dog in the 1.4.0 release quote
16:26 cotto mundo cani?
16:26 purl mundo cani is, like, the dog in the 1.4.0 release quote
16:26 flh does some document explain PMC's memory structure (and maybe its subtleties), or should I just try to understand it from pobj.h?
16:26 theory joined #parrot
16:28 cotto flh, docs/pmc.pod is reasonably up-to-date and at least provides a good overview
16:33 flh thanks cotto, I now know that I shouldn't really try to understand UnionVal
16:33 cotto np.  I'm glad to know that the file will be useful to someone.
16:43 Coke Who broke "new Iterator" ?
16:43 Coke (they need to run examples_tests.
16:44 flh cotto: that's in r40192 (I won't name who did it, maybe someone who worked a lot to improve iterators :) )
16:44 flh Coke, not cotto, sorry...
16:45 Coke it's bacek. =-)
16:46 Coke I tagged TT #811.
16:46 Coke just a PITA to test other things when 'fulltest' is borked.
16:46 mokurai joined #parrot
16:48 dalek parrot: r40197 | coke++ | trunk (5 files):
16:48 dalek parrot: Remove the Pair dynamic PMC.
16:48 dalek parrot: TT #187, patch courtesy kraai++
16:48 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40197/
16:50 dalek TT #187 closed by coke++: deprecated: dynpmc Pair
16:52 Psyche^ joined #parrot
16:57 dalek rakudo: a74657c | pmichaud++ | docs/spectest-progress.csv:
16:57 dalek rakudo: spectest-progress.csv update: 416 files, 11854 passing, 0 failing
16:57 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/a​74657ca05633c605111cb6a72991c58c38e56ae
17:19 Andy joined #parrot
17:21 dalek TT #870 created by NotFound++: examples/benchmarks/overload.pir does not compile
17:23 dalek rakudo: 3b1bda7 | pmichaud++ | build/PARROT_REVISION:
17:23 dalek rakudo: Bump PARROT_REVISION to latest Parrot 1.4.0 release.
17:23 dalek rakudo: PARROT_REVISION is now frozen until Rakudo release.
17:23 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/3​b1bda7fdc360740d5c1ccc2945622d525571d74
17:33 hercynium joined #parrot
17:38 Tene joined #parrot
17:39 chromatic joined #parrot
17:39 nopaste "Util" at 68.191.99.24 pasted "Method not found - bug?" (26 lines) at http://nopaste.snit.ch/17345
17:39 Util When using ResizableIntegerArray, why can't 'push' be called as a method instead of just an opcode?
17:39 Util Both opcode and method work correctly for ResizablePMCArray.
17:39 Util See http://nopaste.snit.ch/17345
17:39 pmichaud we never defined it for RSA
17:39 pmichaud or RIA
17:40 pmichaud there was some debate about whether it belonged on RPA, even
17:40 Tene Why is push an op at all?
17:40 Util 3xTLA unknown. URLs?
17:42 Tene Util: ResizableStringArray, ResizableIntegerArray, ResizablePMCArray
17:42 Tene (took me way too long to remember what TLA meant ;)
17:43 Util doh!
17:43 treed Heh.
17:43 treed The context didn't help.
17:43 Util Tene: thanks!
17:43 NotFound Tycho Large Anomaly?
17:43 Tene I was thinking "TLA... is that like TLS, maybe?"
17:45 Util If an non-dynamic opcode exists, then I would think that every PMC would define it for themselves, unless it was nonsensical for that PMC.
17:45 particle or inherited
17:45 purl hmmm... inherited is for the classdata bit
17:45 Util (w.r.t. does it belong in RPA,RIA,RSA)
17:46 Util Tene: opcodes are cheaper/faster/easier to use?
17:46 Coke util: not if they delegate to something slow. =-)
17:47 Coke but in general, if there's a vtable, you can't access the vtable directly.
17:47 Coke so you use an opcode to invoke the vtable.
17:49 Coke hurm. a question for me is why RPA even has a push method.
17:49 Coke given that it has the push vtables.
17:50 Tene So that HLLs don't So that For HLLs, I'd suspect.
17:50 Coke what?
17:50 Tene Yay, lag eats my ^U
17:51 Coke if an HLL wants that, they can add it.
17:51 Tene Sure.
17:51 Coke I'd vote we kill it.
17:51 Coke we're already presenting a confusing picture in terms of op/vtable/method, no need to further mix and match.
17:52 Tene My preference would be killing the push op and replacing it with a methods and a virtual op.
17:52 Coke if you kill the push op, you kill the vtable entries, too.
17:52 Coke er, you 'might as well'
17:52 Coke and if you move that to a method... what's special about that over other vtable entries?
17:53 Coke (and also: pop, shift, and unshift)
17:53 Coke ... and don't you wish we had this discussion 3 weeks ago. =-)
17:54 payload joined #parrot
17:57 Tene Argh, connection fail. :(
17:57 Tene I'd rather they moved to methods, too.  I don't see what's so special about them that we have opcodes for them.  My understanding was that moving them to methods had been planned for a while.
17:58 Coke Tene: documentation?
17:58 purl documentation is => a core attribute of attributes
17:58 pmichaud The reason that RPA has a .push method is because NQP can't work without it.
17:58 Coke I mean, is there a ticket for that?
17:58 Coke NQP can't invoke an opcode?
17:58 pmichaud not directly.
17:58 Coke ...
17:58 pmichaud 3 + 4   invokes the 'add' opcode
17:59 pmichaud there's not an operator that means "push"
17:59 treed << ?
17:59 purl i think << is Anagram's friend, too.
17:59 Coke that seems leaky.
17:59 pmichaud << means something else entirely.
18:00 dalek cardinal: 3b2e0db | (Ted Reed)++ | src/parser/actions.pm:
18:00 dalek cardinal: Convert  to .ast() in actions.pm.
18:00 Tene Coke: you'd recommend that NQP look at *which* methods are being called, and rewrite just those methods calls to opcode invocations?
18:00 treed Depends on the context.
18:00 dalek cardinal: review: http://github.com/cardinal/cardinal/commit​/3b2e0db9a373675b906bf4a2771cc12c6df3d0b4
18:00 treed << is left shift in some cases, and basically like push when used with an array (at least in Ruby)
18:00 treed Heh.
18:01 pmichaud we had this discussion before, and that's how we ended up with push/pop/shift/unshift on RPA
18:01 treed It ate the $( ) from that commit message.
18:01 Coke Tene: that would be better than the current situation.
18:01 Coke let's either just use methods, or not use them.
18:01 pmichaud it was decided that providing the methods was better tahn forcing NQP to special-detect specific keywords
18:02 pmichaud especially since special-detecting keywords isn't dynamically possible in many cases
18:02 Coke pmichaud: isn't NQP written using the same tools that we expect language designers to be able to solve these sorts of problems with?
18:03 pmichaud Coke:  with the exception that NQP doesn't want a runtime beyond what Parrot provides.
18:04 Coke blah.
18:04 pmichaud whereas language designers typically provide a runtime library.
18:04 pmichaud there are *lots* of places where we have method forms that correspond to opcode forms now
18:04 Coke And I would argue that we shouldn't have any.
18:04 Util PMCs which have methods with opcode names:
18:04 Coke let's just do it one way.
18:04 Util resizablepmcarray and resizablestringarray : shift pop unshift push
18:04 Util filehandle and stringhandle : open close read print
18:05 Util .end
18:05 pmichaud classobj.'add_parent'(...)
18:05 pmichaud namespace.'make_namespace'(...)
18:05 Tene Didn't we have this discussion about 'read', too?
18:05 pmichaud if we have to get rid of either the opcode or method form, then I vote we eliminate the opcode form.
18:06 pmichaud The opcode form doesn't allow options to be passed, where the method forms do.
18:06 Tene cokewhat's so special about shift/pop/unshift/push that they should be opcodes instead?
18:06 Tene If we're doing it one way, doing it as methods seems to make a lot more sense to me.
18:06 Coke Tene: I don't /care/ which one they are, I just don't want BOTH of them.
18:06 pmichaud I suggest a list post.
18:06 Coke we already do not have a convincing story for what's a vtable/opcode/method.
18:06 Coke (er, more the why than the what.)
18:07 Tene The story as I understand it is just history.
18:07 pmichaud but method forms are _much_ easier to deal with than aopcode forms (from a compiler perspective)
18:08 Tene Isn't "dramatically reduce the number of opcodes" part of the bright shiny L1 future?
18:10 Coke (list post) I'll see if I can open a ticket. (it'll have to be for 2.0)
18:10 Util My thinking on the matter was shaped by Dan long ago: http://www.sidhe.org/~dan/​blog/archives/000404.html
18:10 dalek decnum-dynpmcs: r115 | darbelo++ | trunk/aux/decTest/src/pct/actions.pm:
18:10 dalek decnum-dynpmcs: Add plan() call to the generated pir. We are now feature-equivalent to the perl
18:10 dalek decnum-dynpmcs: script.
18:10 dalek decnum-dynpmcs: review: http://code.google.com/p/decnu​m-dynpmcs/source/detail?r=115
18:11 Tene I'm quite certain I remember "reduce the number of opcodes" being discussed at PDS.
18:11 Coke tene: <insert normal rant about not getting good documentation out of the PDS>
18:11 treed Is there any way to tell what kind of node a PAST node is? (NQP)
18:12 pmichaud treed:   .WHAT ?
18:12 purl .WHAT is like blessed
18:12 pmichaud or if you want to test it for type, use .isa(...)
18:12 treed k, thanks
18:13 treed And that'll look like "PAST::Op" or just "Op"?
18:13 Coke (list post) I lied. I'm not going to bother.
18:14 Coke (too many other parroty things to worry about.)
18:14 Andy joined #parrot
18:15 pmichaud treed:  "Op()"
18:15 pmichaud sorry, "PAST::Op()"
18:16 treed k, thanks
18:16 pmichaud if you want to know if a node is PAST::Op, though, you should use   $node.isa(PAST::Op)
18:16 treed Fair enough.
18:16 Coke seen chromatic?
18:16 purl chromatic was last seen on #parrot 8 hours, 46 minutes and 44 seconds ago, saying: <ecode>
18:16 treed I may end up solving this differently.
18:16 Coke seen cotto?
18:16 purl cotto was last seen on #parrot 1 hours, 42 minutes and 33 seconds ago, saying: np.  I'm glad to know that the file will be useful to someone.
18:16 treed So far the solution seems to be passing the test suite.
18:17 treed I'll have to see if it actually solves the problem that needs solving.
18:17 jhorwitz joined #parrot
18:23 dalek tracwiki: v2 | coke++ | Glossy-Brochure
18:23 dalek tracwiki: remove untrue marketing about perl and unicode, go with the facts. Clarify the mod_parrot statement (tcl, e.g., doesn't integrate so easily)
18:23 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Glo​ssy-Brochure?version=2&amp;action=diff
18:24 * Coke finally matches Util to his picture!
18:24 * Coke is an idiot.
18:24 Coke Util: ~~
18:24 Coke knew you as two separate people. =-)
18:26 Util Coke: Don't feel bad; I did the same thing with Cotto at YAPC::NA::2009.
18:29 Coke heyyyy, t/op/io.t was not fixed before the release, methinks.
18:33 dalek website: darbelo++ | This week on decnum-dynpmcs.
18:33 dalek website: http://www.parrot.org/cont​ent/week-decnum-dynpmcs.-0
18:38 mokurai joined #parrot
18:51 mikehh joined #parrot
19:02 dalek TT #871 created by coke++: add 'rand' as a dynop
19:03 dalek cardinal: da35935 | (Ted Reed)++ | src/parser/ (2 files):
19:03 dalek cardinal: Severely reduce what can go in lhs, and move the is_a_sub check farther up the chain so that it doesn't interfere with assignment.
19:03 dalek cardinal: review: http://github.com/cardinal/cardinal/commit​/da359352080ceb2b51d5d2f6a7bf0acf2b983b17
19:06 chromatic http://perso.ens-lyon.fr/fabrice.raste​llo/Biblio_Perso/Articles/CGO2009.pdf
19:08 MoC joined #parrot
19:08 Coke shorten that
19:08 purl That URL is at http://xrl.us/be4rk3 [perso.ens-lyon.fr]
19:10 Coke chromatic: is there a ticket/todo list/something regarding the profiler work?
19:10 Coke seen allison?
19:10 purl allison was last seen on #parrot 8 hours, 1 minutes and 45 seconds ago, saying: successfully built Ubuntu packages for a preview tarball built from cotto++'s pre-release commits
19:11 chromatic Not sure; cotto, did I post it somewhere for you?\
19:13 cotto lemme dig that up
19:15 * Coke thinks whiteknight is insane.
19:16 nopaste "cotto" at 74.61.2.46 pasted "my notes for getting a profiling runcore" (24 lines) at http://nopaste.snit.ch/17348
19:17 flh chromatic, is parrot using SSA?
19:18 Coke cotto, chromatic: can that go on the wiki?
19:18 chromatic Doing it now.
19:18 purl Doing it now. is, like, it paying the bills?  :)
19:18 cotto some editing may be required
19:19 dalek tracwiki: v86 | chromatic++ | WikiStart
19:19 dalek tracwiki: https://trac.parrot.org/parrot/wiki/​WikiStart?version=86&amp;action=diff
19:21 chromatic Huh.
19:21 chromatic That shouldn't be.
19:21 chromatic Oh, that's fine.
19:22 flh cotto, about TT#840, I can rewrite io.t in Perl, but do you understand why it is failing now? I can find a computer with windows on it, but I don't know how to reproduce this bug
19:22 flh bloody tab key, that was for Coke, sorry
19:23 NotFound flh: the output of the last test is lost somewhere.
19:23 cotto I'm much less confused now.
19:24 japhb cotto: just change your nick to Pepsi, and that will solve the tab completion problem right there.  ;-)
19:24 Coke flh: because the test wasn't converted properly?
19:24 Coke <shrug> if I knew why it was broken, I'd've already fixed it. =-)
19:25 Coke flh: and it doesn't need to be fully rewritten, so much as reverted, and then updated to track a few commits to the PIR version.
19:26 flh I'll attach a perl version to TT#840 tomorrow
19:26 dalek tracwiki: v1 | chromatic++ | ProfilingRuncore
19:26 dalek tracwiki: Added profiling runcore notes to the wiki
19:26 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Pro​filingRuncore?version=1&amp;action=diff
19:26 dalek tracwiki: v2 | chromatic++ | ProfilingRuncore
19:26 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Pro​filingRuncore?version=2&amp;action=diff
19:27 tomyan joined #parrot
19:27 Andy joined #parrot
19:27 flh I tried to build/run parrot under cygwin, but in this case, the test passes, so how can I build/run parrot on windows to see the failure?
19:27 NotFound Replacing the test with one that writes to a temporary file and checks that file might also be fine.
19:28 japhb Where should I post my thoughts regarding the HLL performance problem?  parrot-dev?  The wiki?
19:29 NotFound japhb: open a ssh connection to your brain ;)
19:29 moritz japhb: parrot-dev seems good to me
19:29 japhb NotFound: And where should the other end of the tunnel connect to?  :-)
19:29 Coke flh: install strawberry perl.
19:30 Coke it contains perl, a compiler, build tools...
19:30 Coke strawberry?
19:30 purl strawberry is a Perl for Windows that works just like Perl everywhere else. See http://win32.perl.org/wiki/in​dex.php?title=Strawberry_Perl
19:30 japhb moritz: OK, will do.
19:30 Coke japhb: do you have a work in progress?
19:30 Coke or something other than "DEAR GOD SO SLOW", which is pretty much what I have. =-)
19:30 NotFound japhb: let it open to any of us to connect.
19:31 flh ok, thanks (windows is really black magic to me, even for the simplest things)
19:31 japhb Coke: Actually, a few weeks ago I worked on trying to find some of the issues, and found a few general problem areas.
19:32 Coke excellent.
19:32 japhb I just don't have the Mad Skillz to fix said problems, nor the Copious Time to gain said Skillz.  :-(
19:33 japhb NotFound: No Necromancer episodes, please.  :-)
19:33 japhb Actually, hmm, wrong book reference I think, damn.
19:34 NotFound flh: strawberry is fine, I've been able to install it, build and test parrot, and implement the pipes win32 code in half a day.
19:34 NotFound japhb: I was thinking about searching for a related megatokyo episode, instead ;)
19:35 japhb heh
19:37 dalek TT #872 created by coke++: segfault with PASM1 compiler
19:39 NotFound Can't we just kill the PASM1 pseudo-compiler?
19:39 Coke NotFound: see my comment!
19:39 Coke yes, that'd be fine.
19:39 NotFound Aye, sir!
19:40 NotFound AFAIK is only used in the 'eval' instruction of parrot_debugger... wich doesn't work.
19:42 NotFound I mean kill, without even deprecating it, unless someone provide a reasonable scenario in which someone might be using it.
19:43 NotFound I can't think any good reason to do a deprecation cycle for a feature that doesn't work at all.
19:45 Coke NotFound: I'm assuming it might work.
19:45 Coke I have yet to prove my theory! =-)
19:47 Coke I agree that if it doesn't nothing but fail, our users are unlikely to be relying on a segfault.
19:47 Coke s/doesn't/does/
19:51 NotFound Coke: I took a look some days ago while reviewing TT #672, and were unable to find a way to make it work.
19:53 Coke wow, exactly 200 tickets apart.
19:53 NotFound That might explain why there are no test for it... or the contrary.
19:54 Coke seems reasonable to me, anyway - float the balloon on one of the tickets.
19:54 Coke ?
19:54 NotFound There are no test because it doesn't work, or it stopped working because it was never tested.
19:55 Coke Seems reasonable. chromatic, any opinion on removing something that can't demonstrably be shown to work?
19:57 chromatic If it doesn't work and never did, why keep it?
19:58 NotFound I suppose that when someone wrote the eval instruction in the debugger it must have been working.
20:01 NotFound Long, long time ago, in a galaxy far away...
20:13 eternaleye joined #parrot
20:20 Coke NotFound: seems like showing this also segfaulted in 1.0 would be sufficient for killing it.
20:22 NotFound Good idea, I'll try that tomorrow.
20:26 tomyan joined #parrot
20:27 Coke NotFound+=
20:27 Coke NotFound++
20:36 dalek cardinal: d5f39d2 | (Ted Reed)++ | src/parser/ (2 files):
20:36 dalek cardinal: Remove unused and misleading rules mlhs and rlhs.
20:36 dalek cardinal: review: http://github.com/cardinal/cardinal/commit​/d5f39d26a40bfea6d6cd798cd97573c78dfe61fc
20:38 hercynium joined #parrot
20:45 mberends joined #parrot
20:50 dalek parrot: r40198 | cotto++ | trunk (20 files):
20:50 dalek parrot: [ops2c] change ops file format to include explicit ops preamble (tt #836)
20:50 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40198/
20:50 nopaste "cotto" at 74.61.2.46 pasted "update to fix Rakudo with Parrot's r40198" (25 lines) at http://nopaste.snit.ch/17352
20:51 hudnix joined #parrot
20:51 cotto msg pmichaud There's a patch at http://nopaste.snit.ch/17352 to make Rakudo work with r40198.
20:51 purl Message for pmichaud stored.
20:52 particle (10:23:35 AM) dalek: rakudo: PARROT_REVISION is now frozen until Rakudo release.
20:52 pmichaud cotto: thanks!  Would that patch work with 1.4.0?
20:53 cotto no.  I just committed the deprecation.
20:53 cotto It's not needed to work with 1.4.0
20:53 whiteknight joined #parrot
20:53 whiteknight irclogs?
20:53 purl i guess irclogs is http://irclog.perlgeek.de/parrot/today or see also: infrared clogs
20:53 pmichaud that's not my question :-)
20:53 pmichaud I'm not asking if it's needed for 1.4.0, I'm asking if applying that patch would break Rakudo under 1.4.0
20:53 cotto sorry.  I suspect it'll break if you try to use it with 1.4.0
20:54 pmichaud okay.
20:54 pmichaud So we apply it after Rakudo's release.
20:54 cotto yes
20:54 pmichaud excellent, thanks.
20:54 cotto whenever you bump PARROT_REVISION
20:54 pmichaud could you forward that to rakudobug with [PATCH] so it doesn't get lost?
20:54 pmichaud As I'll undoubtedly lose it.
20:54 treed Easy way to tell how many items a PAST node has had pushed onto it? (NQP)
20:54 treed I presume there's some sort of sigil for that.
20:55 treed numify or whatever
20:55 treed (I don't know perl, for the curious.)
20:55 pmichaud treed:  +@($node)
20:55 treed Thanks.
20:57 cotto Coke: nopaste coming up for you...
20:57 nopaste "cotto" at 74.61.2.46 pasted "patch to make tcl work with Parrot's r40198" (21 lines) at http://nopaste.snit.ch/17353
20:58 cotto actually, lemme test that.  It's a very simple change but I've been known to get those wrong.
20:59 cotto it builds
21:00 darbelo_as__purl SHIP IT!
21:00 mokurai joined #parrot
21:00 cotto purl, it builds
21:00 purl cotto: huh?
21:00 cotto it compiles
21:01 purl Ship it!
21:01 cotto there we go
21:01 japhb pmichaud: What are the bus numbers for PGE and PCT?
21:01 pmichaud japhb: I dunno.  PGE probably 1.
21:01 pmichaud PCT >= 1
21:02 japhb That's what I was afraid of.
21:02 pmichaud sure, but what's the bus number for Perl 5 regex engine?
21:03 japhb pmichaud: I'd guess a handful.  Probably >1, <5.
21:03 pmichaud I'd be surprised if it's as high as 5.
21:03 pmichaud I suepect it's 2.
21:03 Tene I'm fairly comfortable with PCT.
21:03 pmichaud (where "bus number" means "availability of people to work with it")
21:04 pmichaud there may be 5 who have worked on the perl 5 regex engine, but I don't know that there are that many who *would* do so today
21:04 japhb Well, in any case, I'm more worried about PCT than PGE, except that it's never a bad idea to have at least 2 people at least aware of *how* to work on each block of code ...
21:04 japhb true.
21:04 pmichaud at any rate, I'm comfortable that PCT could be picked up w/o too much difficulty.  Several have done so.
21:05 pmichaud Tene++ being the most obvious candidate :-)
21:05 japhb :-)
21:05 pmichaud PGE would be a bit harder, but that's somewhat because of the nature of the beast.  But I'm certain that it would be easier for someone to pick up PGE than any other regex engine :-)
21:05 bacek joined #parrot
21:06 pmichaud and some major refactorings to PGE are imminent, and I think the result will be much more accessible.
21:06 cotto ltm?
21:06 purl i think ltm is longest token matching
21:06 Tene I've held off looking at PGE because of the pending refactors.
21:06 pmichaud protoregexes first.  Also I think I'll be redoing parts of PGE in NQP.
21:07 pmichaud after that will be ltm.
21:07 pmichaud protoregexes are far more important at this point than ltm.
21:07 Tene I plan to attentively watch the PGE refactors to start learning it.
21:07 Tene I was waiting for protoregexes to refactor Cardinal's parser.
21:07 japhb pmichaud: What has been blocking you on these?  I seem to recall at least protoregexes and LTM being "Real Soon Now" for a long time.
21:07 pmichaud Tene: there's a good possibility that "PGE refactor" will end up being "PGE rewrite"
21:07 Tene pmichaud: even better for me. :)
21:07 pmichaud japhb: main blocker is time.
21:08 japhb Ah.  Sucks, that.
21:08 pmichaud japhb: and energy, to some extent.  The time I had originally planned to spend working on PGE stuff got eaten up dealing with Parrot issues.
21:08 pmichaud er, "Parrot-related issues"
21:08 japhb That's very unfortunate.
21:09 pmichaud On the plus side, though, it's good that the rewrites were delayed, because a few Perl 6 regex details have changed since then
21:09 pmichaud (and may actually change again over the next couple of weeks)
21:09 japhb :-)
21:09 pmichaud so at least I'm only redoing it once instead of redoing it twice
21:10 pmichaud Tene: I'm thinking that there will be a new PGE that Rakudo will use and that hopefully other languages will be able to migrate to
21:10 pmichaud and the existing PGE will remain intact for backwards compatibility/deprecation purposes, and eventually be obsolete by around Parrot 2.6
21:10 cotto AGE - Awesome Grammar Engine
21:10 pmichaud I like that
21:10 pmichaud also gets rid of that silly "P"-acronym
21:11 cotto What kind of changes are you thinking about?
21:11 pmichaud well, fundamentally the operator precedence parser changes
21:11 pmichaud since everything in PGE regexes is based on that.... it's a fundamental change.
21:12 pmichaud the next fundamental change will be the heavy use of lexicals within regexes
21:12 cotto It seems like any given Parrot subsystem needs to be rewritten two or three times in a backwards-incompatible way before we know the right way to do it.
21:13 pmichaud at the time the current version of PGE was written, lexicals in Parrot were broken
21:13 pmichaud we also now have the advantage of having a working PCT
21:13 pmichaud so it's very possible that PGE ast nodes will turn into PGE nodes
21:13 cotto I look forward to stealing shamelessly from Squeak's VM to try to get that number down for L1.
21:13 pmichaud sorry, will turn into PCT nodes
21:14 pmichaud so if PGE can be written in terms of PCT... that's a big win.
21:14 pmichaud (another issue with PGE at the time it was written was that we didn't have -any- usable AST things for Parrot at the time)
21:14 * japhb adds the above to his mental list of reasons we really need more optimization work.
21:15 pmichaud the last fundamental change is that I'm thinking that NQP and its grammar will be the primary interface to PGE instead of having a separate interface (as is currently the case)
21:15 japhb pmichaud: Is the creation of new $_, $!, $/ in every Rakudo scope required by Perl 6, or an implementation quirk of Rakudo?
21:16 pmichaud japhb: every block has its own lexical $_, $!, and $/, iiuc
21:16 japhb Are block and scope interchangeable?
21:16 pmichaud every block defines a new scope
21:16 japhb (i.e. are there blocks that don't introduce a scope, or vice versa?)
21:17 pmichaud I'm not aware of any blocks that don't potentially define a new scope.
21:17 pmichaud it's possible that inlined blocks don't do that.
21:17 pmichaud but that's an optimization, which a compiler must be prepared to pessimize if the optimization preconditions turn out to be false
21:17 japhb nod
21:31 cotto seen austing
21:31 purl I haven't seen 'austing', cotto
21:31 cotto seen austin
21:31 purl austin was last seen on #parrot 9 days, 17 hours, 56 minutes and 52 seconds ago, saying: So does parray ever get reached?  [Jul 12 03:28:55 2009]
21:31 cotto seen austin_hastings
21:31 purl austin_hastings was last seen on #parrot 17 days, 19 hours, 50 minutes and 44 seconds ago, saying: Pmichaud: Thanks, but I'm just finishing the elsif chain... :(  [Jul  4 01:35:09 2009]
21:38 dalek rakudo: 2967a71 | moritz++ | docs/announce/2009-07:
21:38 dalek rakudo: [docs] stub 2009-07 release announcement with deprecation notices
21:38 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/2​967a7141ba4336747e330196a00c04952c75996
21:38 dalek rakudo: 5c1a814 | moritz++ | :
21:38 dalek rakudo: Merge branch 'master' of git@github.com:rakudo/rakudo
21:38 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/5​c1a81496bda4e1fd63a836f37d717c3eeb8bff2
21:43 theory joined #parrot
21:51 bacek good morning #parrot
21:52 dalek parrot: r40199 | bacek++ | trunk/examples/benchmarks/overload.pir:
21:52 dalek parrot: [cage] Remove .include 'mmd.pasm'. mmd.pasm doesn't exists anymore
21:52 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40199/
21:52 dalek parrot: r40200 | bacek++ | trunk/examples (11 files):
21:52 dalek parrot: [cage] Use iter opcode instead of new Iterator in examples
21:52 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40200/
21:56 cotto hi bacek
21:56 bacek hi cotto
21:59 dalek parrot: r40201 | coke++ | trunk/DEPRECATED.pod:
21:59 dalek parrot: According to TT #415, the trap opcode is staying -
21:59 dalek parrot: all other opcodes were already removed.
21:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40201/
22:04 Limbic_Region joined #parrot
22:05 dalek TT #836 closed by cotto++: [DEPRECATED] ops file preamble format
22:05 tomyan joined #parrot
22:19 kid51 joined #parrot
22:31 Coke anyone know if emit_jsr in the alpha jit needs to be removed if we're removing jsr?
22:31 chromatic I doubt it.
22:31 Coke ok. I'll leave it alone.
22:31 chromatic JSR there is likely the CPU instruction, not the Parrot opcode.
22:36 rg1 joined #parrot
22:37 Coke hey, we don't test the jsr opcode at all! whee.
22:37 Coke (makes ripping it out easy)
22:37 dalek TT #870 closed by bacek++: examples/benchmarks/overload.pir does not compile
22:38 Coke whoops, there is one test. my bad.
22:40 bacek Coke: kill it!
22:41 dalek parrot: r40202 | bacek++ | trunk (2 files):
22:41 dalek parrot: [cage] Remove references to mmd.pasm
22:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40202/
22:42 * bacek finished backlog
22:43 bacek cotto: I'm not "maniac"! I'm "THE MANIAC"!
22:51 bacek cotto: http://irclog.perlgeek.de/p​arrot/2009-06-14#i_1238572 Where is my title?! :)
22:54 japhb pmichaud: am I correct in assuming that PCT does essentially no optimization of the code between input PAST and output PIR?
22:55 japhb Which is to say, it's just following set rules for converting at each stage?
22:55 Tene japhb: yes.
22:55 japhb Tene: thx
22:56 Tene japhb: the plan is to allow additional stages that you can insert to do optimization at every step.
22:56 Tene (you could do that right now, but nobody does yet)
22:56 Tene so it's like... past->past_optimize->post->post_optimize
22:57 japhb right.
22:57 Tene and then ->pir for now, and just straight ->pbc eventually
22:59 cotto purl, bacek?
22:59 purl bacek is importing the last week's worth of patches from http://github.com/bacek/pa​rrot/tree/packfile_revamp, I imagine
23:00 cotto no, bacek is THE MANIAC
23:00 purl okay, cotto.
23:01 bacek :)
23:05 kid51 msg whiteknight Can you look at https://trac.parrot.org/parrot/ticket/490 ? Thanks.
23:05 purl Message for whiteknight stored.
23:09 bacek $dayjob time
23:09 dalek parrot: r40203 | bacek++ | trunk (3 files):
23:09 dalek parrot: [cage] Remove deprecated Parrot_str_free function
23:09 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40203/
23:13 dalek TT #826 closed by bacek++: Deprecate Parrot_str_free
23:13 payload joined #parrot
23:14 tetragon joined #parrot
23:28 dalek rakudo: 169fc1c | pmichaud++ | src/setting/ (2 files):
23:28 dalek rakudo: Add Int.Str and Num.Str.
23:28 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/1​69fc1c65b6a80d613c0c3e9fddc4c76aeb585d4
23:31 dalek parrot: r40204 | jkeenan++ | trunk (15 files):
23:31 dalek parrot: Merge pbc_info_remove branch into trunk.  Cf.:  https://trac.parrot.org/parrot/ticket/440.  This removes the 'pbc_info' executable and moves it to examples/c/pbc_info.c.
23:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40204/
23:34 dalek TT #440 closed by jkeenan++: move src/pbc_info.c to examples/c/pbc_sample.c
23:34 dalek parrot: r40205 | jkeenan++ | branches/pbc_info_remove:
23:34 dalek parrot: Branch has been merged into trunk and is no longer needed at HEAD.
23:34 purl i already had it that way, dalek.
23:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40205/
23:37 cotto < purl-- >
23:51 Coke parrot: Branch has been merged into trunk and?
23:51 purl it has been said that Branch has been merged into trunk and is no longer needed at HEAD.
23:52 Coke is 'make testb' failing for anyone?
23:52 Coke (it's failing in my work-in-progress to remove jsr, but it doesn't look JSR related)

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

Parrot | source cross referenced