Camelia, the Perl 6 bug

IRC log for #parrot, 2009-07-31

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:01 Coke segfault: http://smolder.plusthree.com/app/public_pr​ojects/report_details/25718#first_failure
00:02 Whiteknight in the name of the father, the son, and the holy particle, you are forgiven
00:02 NotFound I'd like also the Higgs Boson
00:03 NotFound Uh, some codetest...
00:03 Whiteknight Coke: that segfault is old
00:03 Coke is there a ticket?
00:03 Whiteknight v1.2.0, about 1000 revisions ago
00:04 Whiteknight i dont know if there is or not, but a bug from two months ago that hasn't been mentioned since is hardly a prime candidate
00:04 Coke (if there's a ticket, the test should be todo'd with the ticket #)
00:04 Coke if there's no ticket, there should be.
00:04 Coke having failures in smolder is bad.
00:04 Whiteknight I don't deny that, but it's an old failure and there's no evidence that it still exists
00:05 Coke Whiteknight: ah, I see what you're saying.
00:05 Coke one wonders why someone is generating a smolder report from such an old version.
00:06 Coke (the report is from a few minutes ago.)
00:06 NotFound The alarm problems?
00:06 NotFound They are gone.
00:06 Coke I'll email the guy and ask him to update his smolder.
00:06 NotFound Well, alarm is gone
00:07 chromatic Hm, ~30 new passing Rakudo tests per day on average in the past week.
00:07 Whiteknight chromatic: wow, that's impressive
00:07 Coke meanwhile, partcl is heading in the wrong direction, keep losing tests to segfaults. :|
00:08 Coke I will try to generate tickets. :|
00:08 Whiteknight ...and we will try to ignore them :)
00:12 dalek parrot: r40350 | NotFound++ | trunk (5 files):
00:12 dalek parrot: [cage] use vtable slot numbers instead of names in null pmc access and defaults
00:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40350/
00:15 Whiteknight particle: ping
00:15 Coke Whiteknight: you make baby coke cry. :|
00:15 Whiteknight sorry (i kid!)
00:18 particle Whiteknight: wazzup?
00:29 davidfetter joined #parrot
00:30 Whiteknight particle: we need parrot t-shirts and other buyables
00:31 Whiteknight we should be able to set up a cafepress shop or something
00:31 NotFound I think I asked several times, without answer: Did we have a scalable version of the parrot logo?
00:32 darbelo NotFound: It's not hard to make one from the raster image on the web site.
00:33 NotFound darbelo: for someone with talent for that type of things, I think so, but I'm not such someone ;)
00:38 darbelo NotFound: I'll see if I can manage an svg version of the image later.
00:42 NotFound Coke: tcl.ops:40: error: ‘ParTcl_binary_scan’ was not declared in this scope
00:48 NotFound Coke: I think tcl.ops needs a BEGIN_OPS_PREAMBLE / END_OPS_PREAMBLE around the #include block
00:58 Eevee joined #parrot
00:58 TiMBuS joined #parrot
01:04 Maddingue joined #parrot
01:05 Coke NotFound: yup, I haven't fixed that because of the bsr/jsr/ret stuff.
01:05 tetragon joined #parrot
01:05 jdv79 joined #parrot
01:07 NotFound Coke: ah, well. I was just checking c++ buildability. It build with that change, but fails with the pir files because of that stuff.
01:07 Coke I'm not failing to build tcl.ops. wonder why that is.
01:08 NotFound Without warnings?
01:08 Coke ah. got the warnings once, and then it was biult. =-)
01:08 NotFound C++ strictly requires prototypes.
01:10 NotFound Enough for today, in any case, going to bed :O
01:13 kid51 joined #parrot
01:13 Coke NotFound: try now?
01:13 Coke or tomororw. =-)
01:21 Coke NotFound++
01:31 japhb joined #parrot
01:32 dalek decnum-dynpmcs: r138 | darbelo++ | trunk/src/pmc/decint.pmc:
01:32 dalek decnum-dynpmcs: Make DecInt divide_* and i_divide_* VTABLEs forward to floor_divide_* and
01:32 dalek decnum-dynpmcs: i_floor_divide_*.
01:32 dalek decnum-dynpmcs: Fixes division by constants.
01:32 dalek decnum-dynpmcs: review: http://code.google.com/p/decnu​m-dynpmcs/source/detail?r=138
01:47 kid51 Whiteknight didn't blog today
01:47 * kid51 slits his wrists out of terminal boredom
01:48 chromatic You could fix a memory leak with SUPER in dynpmcs.
01:48 chromatic INTVAL i = SUPER();
01:48 chromatic becomes
01:49 chromatic INTVAL i = Parrot_Foo_get_vtable(interp)->get_integer(interp, pmc);
01:49 chromatic and leaks the return value of Parrot_Foo_get_vtable.
01:49 chromatic Not that that code should work very well anyway....
01:59 jdv79 kid51: whatsup?
02:01 kid51 yo
02:02 kid51 Did the Pittsburgh workshop stoke your appetite for parrot?
02:02 dalek partcl: r549 | coke++ | trunk/src/ops/tcl.ops:
02:02 dalek partcl: attempt to update to new OPS syntax.
02:02 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=549
02:02 jdv79 kid51: haha, are you taking a survey?
02:03 kid51 Well, I'm trying to identify people in NYC area to get together with, invite to hackathons, etc.
02:03 kid51 And, of course, to get to present at perlsemny
02:05 wayland76 Well, I guess I've no chance of finding P6 people in country Australia then :)
02:13 GeJ wayland76: There's damian of course. But I think bacek is in Australia. You should check with him.
02:14 kid51 bacek is in Sydney
02:14 kid51 GeJ is in New Caledonia
02:14 kid51 same hemisphere, at least :-)
02:14 sjn joined #parrot
02:15 kid51 And of course lots of other P5 people in Oz:  alias; Paul Fenwick; gnat in New Zealand; sisyphus somewhere in the outback
02:15 Casan joined #parrot
02:15 GeJ 2 hours and some by plane. Way longer by pirogue, with a lower chance of reaching your destination.
02:16 wayland76 I've met Paul Fenwick (to speak to) and Damian (just sort of "hi"/handshake; naturally he doesn't remember me :) )
02:18 davidfetter damian reviewed my talk at oscon
02:19 wayland76 But I met both Paul and Damian in Melbourne, which is the big smoke around here :)
02:23 Zak joined #parrot
02:26 * davidfetter thinks TimToady might recognize him from geekcruises
02:28 davidfetter who's all going to osdc this year?
02:30 dalek cardinal: 3508b5e | (Danius Michaelides)++ |  (3 files):
02:30 dalek cardinal: Implement Array#replace() with tests.
02:30 dalek cardinal: Signed-off-by: Ted Reed <ted.reed@gmail.com>
02:31 dalek cardinal: review: http://github.com/cardinal/cardinal/commit​/3508b5e94cbbda141a7dfba5a3ee9484bb181fe1
02:32 jimmyZ joined #parrot
02:36 dalek cardinal: eca2172 | (Danius Michaelides)++ |  (3 files):
02:36 dalek cardinal: Partial implementation of Array#insert().
02:36 dalek cardinal: Signed-off-by: Ted Reed <ted.reed@gmail.com>
02:36 dalek cardinal: review: http://github.com/cardinal/cardinal/commit​/eca2172e1166ee951276538797298f3db3263467
02:41 janus joined #parrot
02:57 dukeleto joined #parrot
03:20 donaldh joined #parrot
03:44 eternaleye joined #parrot
04:03 tetragon_ joined #parrot
04:07 tetragon joined #parrot
04:08 Andy joined #parrot
04:16 dalek parrot: r40351 | petdance++ | trunk (3 files):
04:16 dalek parrot: gratuitous consting to test my commit rights
04:16 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40351/
04:19 jimmyZ IIRC, the semicolon was removed by sameone
04:20 jimmyZ Any reaso
04:20 jimmyZ Any reason to add it once again?
04:42 GeJ JimmyZ: I think that the file is generated by yacc (or bison, or whatever). And for what I've been told, depending on the version you use you may experience extra-but-meaningless semicolons.
05:42 dukeleto joined #parrot
06:03 hiroyuki_y joined #parrot
06:07 uniejo joined #parrot
06:32 mikehh All tests PASS at r40351 (pre/post-config, smolder, nqp_test, fulltest) - Ubuntu 9.04 amd64
06:49 cotto chromatic, thanks for noticing that memory leak.
06:50 mikehh rakudo (a53a1cd) builds on parrot r40351 - make test PASS/ make spectest (up to 27822) same two tests FAIL - Ubuntu 9.04 amd64
07:13 dukeleto joined #parrot
07:13 iblechbot joined #parrot
07:21 donaldh joined #parrot
07:38 mj41 Coke: myops.t num 7 is weird one ... it start/stop failing randomly on different machines/OS ... e.g. http://tt.ro.vutbr.cz/report/pr-Parr​ot/do?trun-1831=on&amp;trun-1839=on http://tt.ro.vutbr.cz/report/pr-Parrot/do?tr​un-3856=on&amp;trun-3857=on&amp;trun-3851=on​&amp;trun-3854=on&amp;trun-3830=on&amp;trun-​3834=on&amp;trun-3831=on&amp;trun-3835=on
07:41 payload joined #parrot
07:42 mj41 http://tt.ro.vutbr.cz/report/pr-Pa​rrot/do?trun-2373=on&amp;trun-2368
07:45 dukeleto joined #parrot
07:55 flh joined #parrot
07:59 JimmyZ joined #parrot
08:03 cotto time for sleeps
08:19 JimmyZ good night cotto ;)
08:21 japhb joined #parrot
08:26 tetragon_ joined #parrot
08:41 dukeleto joined #parrot
08:51 jan joined #parrot
09:28 Whiteknight joined #parrot
09:37 wayland76 joined #parrot
09:42 particle joined #parrot
09:47 particle joined #parrot
09:50 bacek joined #parrot
09:50 bacek o hai
09:51 JimmyZ hello bacek
09:52 bacek hi JimmyZ
09:53 JimmyZ but who is bacek_at_work?
09:53 purl bacek_at_work is probably suggesting derangement
09:53 gaz joined #parrot
09:53 bacek JimmyZ: it's me also.
09:53 bacek But it's Friday night already. So I've moved to my other alter ego.
09:54 JimmyZ ;)
09:55 bacek_with_scotch purl: forget bacek_at_work
09:55 purl bacek_with_scotch: I forgot bacek_at_work
09:55 bacek_with_scotch purl: bacek_at_work is bacek_without_scotch
09:55 purl OK, bacek_with_scotch.
10:00 Whiteknight ah scotch: the solution to all of our problems
10:01 wayland76 I've got scotch blood, I don't need that alcohol stuff :)
10:01 dalek parrot: r40352 | whiteknight++ | trunk/docs/project/release_manager_guide.pod:
10:01 dalek parrot: [Release] Replace rgrjg as the release manager for August
10:01 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40352/
10:16 hiroyuk__ joined #parrot
10:25 * bacek can drink it later
10:26 wayland76 Hmm, I hope you're not planning to drink my blood later :)
10:26 bacek wayland76: I've got release manger at $work with scotch blood :)
10:27 preflex joined #parrot
10:27 bacek But I'm not sure how much blood left in his veins. There is definitely a lot of scotch in.
10:28 wayland76 Well, my scotch blood was in the USA for about 300 years, so I don't know how much there is either :)
10:28 * bacek passing bottle of scotch to wayland76 to refresh his blood
10:32 wayland76 wayland76 passes it back -- listening to Celtic music seems to have made it sufficiently scotch, and wouldn't want to deprive bacek in favour of someone who doesn't enjoy scotch.  But thanks for the offer :)
10:35 mikehh don't think I got much scotch blood, even if I do currently reside in Scotland
10:40 mikehh as for ancesrty two of my grandparents were English, one Lithuanian and one was born in South Africa of Estonian ancestry
10:46 mokurai Mikehh: My grandfather was born in a part of
10:46 mokurai Poland that was in Lithuania at the time.
10:47 mokurai left #parrot
11:04 wayland76 I'm about 1/2 scots/Irish, 1/4 Swedish, 1/8 German, and 1/8 bits & pieces :)
11:05 UltraDM joined #parrot
11:20 donaldh joined #parrot
11:33 bacek joined #parrot
11:54 wayland76 joined #parrot
11:57 MoC joined #parrot
11:59 kj joined #parrot
12:01 quek joined #parrot
12:16 ruoso joined #parrot
12:28 payload joined #parrot
12:51 HG` joined #parrot
13:11 JimmyZ joined #parrot
13:13 JimmyZ blogspot is held up by China Great Firewall :(
13:14 payload joined #parrot
13:15 donaldh joined #parrot
13:19 wayland76 purl: seen allison
13:19 purl allison was last seen on #parrot 2 days, 18 hours, 31 minutes and 14 seconds ago, saying: (and expanding from there)  [Jul 28 18:42:29 2009]
13:59 mberends joined #parrot
14:11 iblechbot joined #parrot
14:13 NotFound mj41: Is yours the systema that is reporting that smoke?
14:16 mj41 no
14:16 NotFound mj41: then just forget it, the opcode that is shown the failure no longer exists.
14:47 clinton joined #parrot
14:49 Coke NotFound: is this the smoke from the really old revision?
14:51 NotFound Coke: is the repeated alarm failure, it must be an old revision.
14:52 Coke NotFound: I just fixed partcl. can you try C++ again?
14:52 Coke NotFound: yup.
14:52 Coke see my mail to the list.
14:52 Coke (and my freaking out in here yesterday)
14:52 NotFound Coke: The bsr thing also?
14:52 Coke yup.
14:53 Coke my previous attempts to fix the bsr/ret thing were masked by the ops failure.
14:53 NotFound Just one moment, I'm filling a ticket...
14:53 Coke building dyn pmcs is MUCH more verbose than core pmcs, so you tend to miss those warnings.
14:55 Andy joined #parrot
14:55 Coke Andy: welcome back to the land of the committed.
14:55 Andy I feel so whole.
14:55 Coke <mrfl mrfl mrfl>
14:56 Psyche^ joined #parrot
14:56 NotFound Coke: that's the main reason to build sometimes with C++: it doesn't allow you to ignore that problems.
14:58 dalek TT #892 created by NotFound++: PARROT_IN_OBJECTS_C used only in one file
15:00 NotFound Coke: partcl builds fine
15:01 NotFound It fails 2 array tests
15:03 Coke NotFound: I bet if you run the test by hand, it works.
15:03 Coke parrot tcl.pbc t/cmd_array.t
15:05 NotFound Yes, it works.
15:05 NotFound t/cmd_lsort.t fails at exit, after passing all tests.
15:06 NotFound attempt to access code outside of current code segment
15:06 NotFound Other than these 2, all test pass.
15:07 dalek rakudo: e02bc06 | jnthn++ | src/builtins/guts.pir:
15:07 dalek rakudo: Improve error output when we have a junction of types.
15:07 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​02bc06ee0bb9aabe3c38c171c86739fb0994794
15:08 Coke whoever installed parrot on feather, please "install-dev", not just "install"
15:08 Coke NotFound: yah. that lsort bug is long known, inferior runloop problem.
15:08 Coke danke.
15:08 Coke Not sure how to diagnose the array failures. Any help appreciated.
15:10 Coke (in fact, it'll probably work if you just rerun 'make test')
15:11 NotFound Yes :D
15:11 Coke a ticket for that would probably be good.
15:11 Coke partcl?
15:11 purl partcl is tcl on parrot or http://code.google.com/p/partcl
15:17 Coke Anyone looking for a PIR only task? =-)
15:17 Coke http://code.google.com/p/p​artcl/issues/detail?id=83
15:20 donaldh joined #parrot
15:21 quek left #parrot
15:25 Coke seems that partcl commits are no longer coming through.
15:29 theory joined #parrot
15:33 Coke msg Infinoid seems that partcl commits are no longer coming to #parrot.
15:34 purl Message for infinoid stored.
15:44 iblechbot joined #parrot
15:56 payload joined #parrot
16:28 dukeleto joined #parrot
16:29 donaldh left #parrot
16:43 dalek partcl: r550 | coke++ | trunk/runtime/builtin/uplevel.pir:
16:43 dalek partcl: Remove bsr/ret usage from this file
16:43 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=550
16:43 dalek partcl: r552 | coke++ | trunk/Configure.pl:
16:43 dalek partcl: Issue 87 - add a warning if ICU is not present.
16:43 dalek partcl: re-order list of probed config vars
16:43 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=552
16:48 dalek partcl: r555 | coke++ | wiki/ParrotIssues.wiki:
16:48 dalek partcl: This issue was resolved.
16:48 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=555
16:50 MoC joined #parrot
17:06 kj joined #parrot
17:09 Coke msg Infinoid no. there it is. whee. =-)
17:09 purl Message for infinoid stored.
17:10 sekimura_ joined #parrot
17:13 bacek joined #parrot
17:17 mj41 new TapTinder testing machine added - cygwin 32bit VMWare Win XP machine (id 12, name ttcl-win-32)
17:21 chromatic joined #parrot
17:22 abesapien joined #parrot
17:33 Coke seen on the wx-perl list: "progress is slow and distraction is cheap. :("
17:33 Coke amen.
17:47 NotFound What do you said? I was distracted X-)
17:57 darbelo joined #parrot
17:57 chromatic http://gbracha.blogspot.com/2​009/07/miracle-of-become.html
18:00 mokurai joined #parrot
18:03 * treed got weirded out that his laptop's fans were suddenly going fullblast for no apparent reason.
18:03 * treed then remembered that he was building Parrot in the background.
18:04 sekimura joined #parrot
18:19 Coke chromatic: any suggestions for tracking down partcl core tests that fail occasionally with 'make test' but not when run by hand?
18:20 chromatic Are there special CLI switches in make test, like there are in Parrot's test harness?
18:20 Coke If you say "create a small PIR test case", I'll just cry on you.
18:22 Coke I don't think so; 'make test' runs t/harness, which just calls "prove". lemme try to run the test with perl instead of tcl. (should auto-exec.)
18:23 Coke *sigh*
18:23 Coke perl t/cmd_array.t does fail consistently. Thanks chromatic. :|
18:23 * Coke hates it when he misses obvious diagnostics.
18:28 Coke so, when running with perl, it uses Tcl::Test.pm which just invokes exec with the parrot used to build partcl, the built tcl.pbc, and the path to the test. (only difference between what I did by hand is that it used full paths.
18:28 Coke now if I print out the exec line before it's executed, and run it by hand, it passes all tests.
18:29 chromatic We've had bugs related to full paths before.
18:29 Coke http://code.google.com/p/partcl/so​urce/browse/trunk/lib/Tcl/Test.pm
18:30 Coke (that's what does the re-exec). Test file at http://code.google.com/p/partcl/s​ource/browse/trunk/t/cmd_array.t
18:30 Coke chromatic: if I run full paths by hand, it passes.
18:30 chromatic I'm sorry, I don't read Urdu.
18:30 Coke freaky.
18:30 Coke chromatic: ?
18:31 chromatic I'm trying out new catchphrases.  Please be honest.  It's for posterity.
18:31 PerlJam (Princess Bride)++
18:31 Coke "that's your new catchphrase?"
18:31 Coke (/dr. horrible)
18:32 * cotto starts a land war in Asia
18:32 chromatic There's Dr. Horrible fanfic?  Wow.
18:33 * chromatic starts his own Internet
18:33 Coke GAH. not SLASH horrible. no no no
18:33 Coke ... the evil thing is, it'd be bad horse/dr. horrible.
18:33 * Coke shuddlers.
18:33 * Coke finds an extra L.
18:35 chromatic Like with pie.
18:36 Coke chromatic: my kids, with no prompting from me, love the soundtrack.
18:37 Coke so, the full path thing has some merit.
18:37 Coke any idea why it only seems to affect that one .t?
18:38 Coke ah. if I remove the rel2abs call, still fails via perl.
18:42 joeri joined #parrot
18:44 Coke ... I think this might be another sorting issue.
18:45 Coke looks like [array names] isn't guaranteed to return things in order, and I think parrot is shuffling things every N invocations.
18:45 Coke so I should just explicitly sort hte results.
18:46 NotFound You can't just relax the check?
18:48 Coke I'm sorting in the test.
18:48 bacek joined #parrot
18:48 Coke (not in the result of [array names])
18:49 mokurai joined #parrot
18:52 Coke NotFound: there, shoudl be fixed.
18:53 dalek partcl: r556 | coke++ | trunk/t/cmd_array.t:
18:53 dalek partcl: [array names] doesn't return results sorted, so to test, sort explicitly.
18:53 dalek partcl: Should fix an occasional failure where the underlying parrot hash returned
18:53 dalek partcl: things not in the order they were inserted.
18:53 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=556
18:57 NotFound Coke: doesn't fail now.
18:58 Coke whee.
18:58 NotFound Going to build an optimized parrot and recheck
19:03 Coke if I wanted to convert, say,
19:03 Coke http://code.google.com/p/partcl/source/​browse/trunk/runtime/builtin/after.pir
19:04 Coke to NQP, would I have to prefix it with the .HLL/.namespace stuff?
19:08 Coke also: can NQP handle slurpy args?
19:11 pmichaud NQP doesn't know how to put things into an HLL yet.  It probably should do that.  I'm not sure what syntax to use.
19:11 pmichaud If NQP cannot handle slurpy args yet then it can be made to do so fairly quickly.
19:14 * pmichaud decides to take a short nap.
19:16 * Coke finds a post that suggest they do, and finds that *@argv is probably the right syntax.
19:17 PerlJam Coke: are you converting partcl to use PCT?
19:17 Coke PerlJam: no, that's hard.
19:17 Coke I'm seeing if I can rewrite one builtin using nqp.
19:18 Coke (instead of PIR)
19:18 Coke hurm. can I have typed variables? (basically, get at the int registers?)
19:19 pmichaud That's a planned update, yes.
19:19 pmichaud Might even happen this weekend.
19:19 pmichaud It doesn't exist yet.  When it does, it'll be    my int $foo
19:20 Coke hurm. I think I'd need that before converting quite a few things.
19:21 pmichaud well, it's a high priority for me at the moment, as I expect the PGE refactors will want it.
19:22 PerlJam Coke: you're always just a little ahead of the technology curve  :)
19:23 Coke does nqp have 'die' ?
19:23 Coke (or something spelled differently that does the same thing? =-)
19:23 pmichaud Coke: nope.  One would have to use Q:PIR { die "..." }
19:23 pmichaud nqp doesn't have any builtin functions
19:24 pmichaud we might come up with a syntax that allows nqp to directly generate opcodes (other than the Q:PIR approach).  I'll have to think about that one a bit.
19:25 pmichaud I'd prefer not to try to map function call syntax directly to opcodes, though, because it would interfere with the calling of actual functions
19:25 nopaste "Coke" at 72.228.52.192 pasted "attempt to translate http://code.google.com/p/partcl/source/​browse/trunk/runtime/builtin/after.pir into NQP" (11 lines) at http://nopaste.snit.ch/17447
19:26 pmichaud that's reasonable, except I'm not sure about the $sec in the Q:PIR
19:26 Coke that's why it has the ??? - it fails.
19:27 NotFound pmichaud: inlined functions might be a useful mechanism.
19:27 pmichaud NotFound: example?
19:27 pmichaud Coke:
19:27 pmichaud Q:PIR {{
19:27 pmichaud $P0 = find_lex '$sec'
19:27 pmichaud $I0 = $P0
19:27 purl $I0 = $P0 is, like, VTABLE_get_integer_native
19:27 pmichaud sleep $I0
19:27 pmichaud }};
19:27 PerlJam pm: you could map "functions" that are prefixed by "pir:" directly to opcodes.
19:28 Coke pmichaud: so when I say my $sec, I'm creating a visible lexical?
19:28 pmichaud Coke: yes.
19:28 Coke is there a way to just have a local PMC?
19:28 pmichaud Coke: not yet.  Eventually perhaps    "my pmc $sec"
19:28 Coke (what i really want is a local int and float, but I understand that's not there yet."
19:28 NotFound pmichaud: if nqp had an inline mechanism, it can have a predefined die function with the Q:PIR die opcode
19:29 pmichaud PerlJam: I'm not sure that "pir:" would be valid Perl 6 syntax
19:29 Coke ok. Then I don't think this is the right hammer for this nail (writing guts of partcl in something a little higher level than PIR.)
19:29 pmichaud eventually I would probably expect
19:29 pmichaud my int $ms := @argv[0];
19:29 PerlJam pm: I was thinking of it similarly to syntactic categories, but I guess that only works well for the defs and not for the calls so much :)
19:29 pmichaud my int $sec := $ms / 1000;
19:30 pmichaud Q:PIR { sleep $sec };   # or some alternate syntax
19:30 PerlJam pir-sleep($sec);
19:30 PerlJam (that's certainly valid perl 6 :)
19:30 NotFound A possible perlish way: use PIR::die;
19:30 pmichaud oh, another perlish way might be just   use PIR;
19:30 pmichaud which imports the PIR opcodes as functions
19:31 * Coke ponders HLL_macros
19:32 PerlJam NQP doesn't do multis or overrides does it?
19:32 pmichaud it will likely do multis at some point
19:32 einstein joined #parrot
19:32 pmichaud basically I'd like NQP to be able to do most everything that PIR lets us do
19:33 Coke pmichaud: well, based on this simple example, I think having "my (regtype) $var" would enable me to convert quite a bit over.
19:33 pmichaud Coke: right
19:33 Coke and I think the nqp version of that does look nicer than the PIR version.
19:33 pmichaud Coke: My goal is to be able to write things like PCT itself in NQP
19:33 pmichaud (and NQP, for bootstrapping)
19:33 Coke (just having the conditionals indented there makes me much happier.)
19:33 NotFound And a pir compiler.
19:34 pmichaud the other thing that NQP will be getting soon is that it won't created nested subs for { ... } blocks
19:34 pmichaud (if there aren't any lexicals declared, it'll inline the block)
19:35 einstein I would like to tell you that I have made nqp in nqp instead of pir, so it becomes easyer to work on nqp
19:35 pmichaud einstein: yes, I've been following your work :-)
19:36 pmichaud and the other thing that NQP will be getting is the ability to declare grammars and regexes
19:36 einstein but now I busy with the typeof,getclass and class ops, so i can make a good working typeof operator in nqp
19:36 PerlJam einstein: as long as you aren't still hung up on god playing dice with the universe :)
19:37 pmichaud when NQP can do grammars and regexes, then we'll be able to embed NQP blocks into regexes :-)
19:37 Coke pmichaud: ah, yah, the blocks as subs is kind of scary. I already have too many sub calls.
19:37 pmichaud and Perl6Grammar.pir can be deprecated/disappear
19:37 PerlJam pm: at some point nqp will lose the not quite moniker if you keep that up
19:38 pmichaud PerlJam: it'll always be "not quite" because we won't have the runtime library
19:38 pmichaud (although if NQP does regexes I guess that means it'll have PGE loaded.  Oh well.)
19:39 Coke I think perhaps a short term code-improvement would be to use hll_macros
19:43 Coke (just for partcl, sorry =-)
19:44 PerlJam buoyancy for partcl helps all language-boats float
19:44 PerlJam (like I said before, you always seem to be a little ahead of the curve)
19:51 einst31n joined #parrot
19:51 pmichaud partcl++ Coke++
19:53 pmichaud afk # nap, switching locations
19:57 iblechbot joined #parrot
20:10 Coke hurm. can anyone think of a way to macroize ".local pmc $foo\n$foo = get_root_global ['bar'], '$foo' ?
20:10 Coke the foo in quotes ends up being a literal '.foo', not '<the string I want>'
20:47 treed Anyone around who can tell me what <[!@&`'+~=/\\,;.<>_*$?:"] > means?
20:47 treed I thought <> was for references to named rules.
20:48 PerlJam <[]> is a character class
20:49 treed Ah.
20:49 treed So <> and <[]> are two completely different things
20:50 treed Why the space at the end though?
20:50 PerlJam AFAIK, the space doesn't have any semantic.
20:50 treed k
20:50 treed Thanks a lot.
20:50 PerlJam what <> means is determined by the character immediately following the <
20:51 treed Ahhhh.
20:51 treed Is there a good doc for this?
20:51 treed Googling for rules usually takes me to a broad overview.
20:51 PerlJam I'm working on a quick ref that'll describe such things but it's not  there yet.
20:54 PerlJam Were I you, I'd just read S05
20:55 PerlJam http://perlcabal.org/syn/S05​.html#Extensible_metasyntax_(%3C...%3E)
21:36 dalek cardinal: 275c208 | (Joeri Samson)++ | src/parser/grammar.pg:
21:36 dalek cardinal: Parse several special variable (like $, $! etc...)
21:36 dalek cardinal: Signed-off-by: Ted Reed <ted.reed@gmail.com>
21:36 dalek cardinal: review: http://github.com/cardinal/cardinal/commit​/275c2080aad9f3e21b97ef80e1ad9acd2b0ea1cb
21:36 dalek cardinal: 4ccae36 | (Joeri Samson)++ | src/parser/ (2 files):
21:36 dalek cardinal: Make self and nil literals instead of variables. Add literals true and false.
21:36 dalek cardinal: Signed-off-by: Ted Reed <ted.reed@gmail.com>
21:36 dalek cardinal: review: http://github.com/cardinal/cardinal/commit​/4ccae3655d172bad5bd8aaaf57b661eca9b9d79d
21:36 dalek cardinal: dd12a49 | (Joeri Samson)++ | src/parser/grammar.pg:
21:37 dalek cardinal: Whitespace fix
21:37 dalek cardinal: Signed-off-by: Ted Reed <ted.reed@gmail.com>
21:37 dalek cardinal: review: http://github.com/cardinal/cardinal/commit​/dd12a49ca797f6e6e96cc6baff121d773d0e052a
21:51 mokurai joined #parrot
21:53 theory joined #parrot
21:54 Limbic_Region joined #parrot
21:58 Andy am I the only one with packfile*.t failures?
21:58 chromatic Have you run 'make realclean' recently?
21:59 Andy dunno
22:00 Andy do you have a specific reason for that, or jsut that's a general thing you do
22:00 chromatic I believe we have some missing dependencies.
22:00 chromatic ... in the Makefile rules, that is.
22:00 Andy from .c to .h?
22:01 chromatic Probably.
22:01 Andy isn't there a makedeps something?
22:01 chromatic I don't know.
22:02 Andy makedepend.
22:02 purl well, makedepend is known. you either need an old gcc (2.95 or so) or patch it. see perlbrowse
22:02 Andy I'll try to look at it tonight
22:02 chromatic That would be fantastic.
22:03 Andy so lemme see....
22:03 Andy is there a task tracker in trac?
22:03 Whiteknight joined #parrot
22:03 Andy oh my, many
22:05 dalek TT #893 created by petdance++: Rebuild dependencies in makefile
22:12 Coke someone should create a petdance to andy karma link.
22:12 Coke karma petdance?
22:12 purl petdance has karma of 1159
22:12 Andy heh
22:12 Coke karma andy?
22:12 purl andy has karma of 601
22:12 Andy That would requiring me giving a shit. :-)
22:12 Coke Andy: I think you can do it with an update to CREDITS. (shit) true.
22:13 Coke but we have enough OCD people with us that it'll probably happen without any further effort from either of us.
22:13 * Coke tries to write a simple .Try (catch/finally) that could get sucked into hllmacros if anyone cared.
22:14 einst31n Can it be true that parrot strings get never released due to a bug in the gc?
22:14 Andy Anything is possible.
22:15 einst31n It seems that the  pool->gc_object does not get set in the string pool
22:15 Whiteknight the we've fixed a lot of GC bugs over the last few months, surely there must be none left :)
22:16 Andy hey, src/stacks.c went away recently.  Anyone can tell me why?
22:16 * Coke wonders who azawazi is.
22:16 Coke azawawi, whoops.
22:16 Coke andy: there was a ticket.
22:16 Whiteknight Andy: was deprecated for a while and not used
22:16 * Whiteknight did the removing
22:16 Andy i had so much splinty goodness. :-)
22:17 Tene Whiteknight: I still have a GC bug.
22:17 Whiteknight Tene: no, it's not a bug it's a feature
22:17 Whiteknight you have a GC feature
22:18 Tene Whiteknight: parrot crashing is a feature?
22:18 MoC joined #parrot
22:19 Whiteknight it's not "crashing", it's being more efficient by terminating early and saving extra processing cycles
22:20 Tene Ah.
22:20 Tene But why does it only fail when I load steme before rakudo and not rakudo before steme?
22:20 Andy hey look, make realclean fixed my test bummers, chromatic
22:21 Andy and that ancient t/library/pcre.t failure is still there
22:34 NotFound I think is time to fix that pcre thing. Just by dropping any configure checks for it, doing them at pbc loading time, and providing a way to report the failure other than dying.
22:36 Coke I don't think we want to drop the config checks.
22:36 Andy is that wy it's failing?
22:36 dalek partcl: r557 | coke++ | trunk/src/macros.pir:
22:36 dalek partcl: Add some macros to make it easier to initialize .locals
22:36 Andy For not having pcre installed?
22:37 dalek partcl: Add a Try() that just eats any thrown exceptions.
22:37 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=557
22:37 dalek partcl: r558 | coke++ | trunk/runtime/tcllib.pir:
22:37 dalek partcl: Make hllmacros available for use in our code.
22:37 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=558
22:37 Coke IIRC, ther's a config check that the tests ignore.
22:37 Coke we probe for it 3 different ways instead of just one.
22:37 NotFound Coke: they are completely unuseful
22:37 Andy Pioint me to a ticket on it.
22:37 Andy or assign it
22:37 Andy I wish Trac could let you have multiple owners
22:37 Andy or watchers
22:37 purl i guess watchers is fixed in 3.0.
22:37 Andy oh well, gotta run home
22:38 NotFound Andy: I think is failing just because in most cases the libpcre.so link to the real lib isn't provided by the regular packages, but the dev ones.
22:38 Coke msg andy closest you can get to watchers is cc.
22:38 purl Message for andy stored.
22:38 Coke NotFound: he's gone.
22:39 NotFound I type fast, but not enough ;)
22:39 rg joined #parrot
22:39 Coke assigned #406 to andy
22:40 nopaste "coke" at 72.228.52.192 pasted "wonders if this is cleaner. (.int => .local int foo\nfoo=value)" (21 lines) at http://nopaste.snit.ch/17449
22:46 Whiteknight if Padre had syntax highlighting for C, I would use it as my primary editor
22:46 Whiteknight well, I would also like to be able to define my own colorscheme, but I won't be choosy
22:47 japhb I wonder if tewk's parser could be repurposed to act as a syntax highlighting backend for Padre?
22:49 japhb My big Padre wish is a screen(1)-friendly text mode.  I don't want to switch until I can use it everywhere (I hate the mental context switch overhead when I have to switch between editor syntaxes every time I switch windows).
22:54 pmichaud cla?
22:54 purl somebody said cla was Contributor License Agreement or http://www.perlfoundation.org/​contributor_license_agreement or http://www.parrot.org/foundation/legal or http://www.parrot.org/files/parrot_cla.pdf
23:03 kid51 joined #parrot
23:05 mokurai joined #parrot
23:08 bacek joined #parrot
23:31 bacek Good morning
23:31 purl And good moroning to you, bacek.
23:34 dalek parrot: r40353 | NotFound++ | trunk/t/library/pcre.t:
23:34 dalek parrot: [t] check and diagnose pcre library loading
23:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/40353/
23:44 payload joined #parrot

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

Parrot | source cross referenced