Camelia, the Perl 6 bug

IRC log for #parrot, 2011-02-22

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 dalek winxed: r817 | NotFound++ | trunk/winxedst (2 files):
00:01 dalek winxed: new predef 'push', use it instead of the push method in a lot of places
00:01 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=817
00:02 dalek winxed: r818 | NotFound++ | trunk/pir/winxed_ (2 files):
00:02 dalek winxed: update installable files
00:02 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=818
00:07 dalek parrot/opsc_full_parse: f884bb0 | bacek++ | / (2 files):
00:07 dalek parrot/opsc_full_parse: Handle type casting as term. Fix parsing of PTR2INTVAL, etc.
00:07 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/f884bb0acf
00:07 dalek parrot/opsc_full_parse: fafa6a1 | bacek++ | t/compilers/opsc/01-parse-body.t:
00:07 dalek parrot/opsc_full_parse: Add more tests.
00:07 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/fafa6a1f4f
00:07 dalek parrot/opsc_full_parse: 5ed4d8e | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
00:07 dalek parrot/opsc_full_parse: Add more math ops.
00:07 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/5ed4d8ecad
00:07 dalek parrot/opsc_full_parse: 4453ad3 | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
00:07 dalek parrot/opsc_full_parse: Mark all statements with 'endstmt'
00:07 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/4453ad3a0e
00:07 dalek parrot/opsc_full_parse: 840561c | bacek++ | t/compilers/opsc/01-parse-body.t:
00:07 dalek parrot/opsc_full_parse: Add test for if-else without {}
00:07 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/840561c0c7
00:12 kid51 bacek:  I'm getting build failures in the opsc_full_parse branch.  Are they to be expected at this point?
00:12 bacek_at_work kid51, yes. It's broken. Heavily :)
00:13 kid51 Do you want a paste, or shouldn't I bother yet?
00:14 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#10575) fulltest) at 3_1_0-213-g40c0d36 - Ubuntu 10.10 i386 (g++-4.5)
00:14 dalek parrot/opsc_full_parse: 487a341 | bacek++ | / (2 files):
00:14 dalek parrot/opsc_full_parse: Opname can be keyword. E.g. 'if'
00:14 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/487a3419bb
00:14 mikehh again had to run make -j world twice to complete build
00:15 bacek_at_work kid51, not yet. It will not be ready in next few days.
00:15 kid51 k
00:19 dalek parrot/opsc_full_parse: 13b01ea | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
00:19 dalek parrot/opsc_full_parse: Add prefix:~
00:19 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/13b01ea018
00:22 lucian left #parrot
00:36 snarkyboojum left #parrot
00:37 cotto ~~
00:42 bacek_at_work aloha, cotto
00:47 cotto hi bacek_at_work.  looks like you've got something interesting to hack on
00:47 bacek_at_work cotto, parsing of op bodies is mostly done.
00:47 cotto what's the battle plan?
00:50 bacek_at_work finish parsing
00:50 bacek_at_work implement pasting
00:51 bacek_at_work implement emitting of C back
00:51 bacek_at_work (optional implement JIT for current ops :)
00:51 cotto wheee
00:51 cotto bacek++
00:51 cotto I hope to have some tuits later this evening for that.  it sounds like fun
00:52 bacek_at_work cotto, you welcome :)
00:54 cotto ]'m worried that opsc is too fast.  This should help. ;)
00:56 bacek_at_work cotto, it will be faster actually.
00:56 cotto orly?
00:56 bacek_at_work yes. Less backtracking
00:57 cotto great
00:58 nwellnhof left #parrot
00:59 dalek parrot/opsc_full_parse: c81d95e | bacek++ | / (2 files):
00:59 dalek parrot/opsc_full_parse: Add (broken) control switch
00:59 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/c81d95ec09
00:59 bacek_at_work cotto, feel free to fix control<switch> :)
01:00 cotto wfm
01:03 cotto can't do it right now, but it'll be great in 60m +/-
01:11 bacek_at_work cotto, we just need enable backtracking over <statement>. Problem is: second "case:" consumed by <statement> up to ":" (afaiu)
01:13 cotto ok
01:13 dalek Rosella: c5ecd69 | Whiteknight++ | / (3 files):
01:13 dalek Rosella: fix the new query files so they build, and add them to setup.nqp
01:13 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/c5ecd6950a
01:13 dalek Rosella: 21b3173 | Whiteknight++ | query/ (2 files):
01:13 dalek Rosella: fix providers so we can run some basic tests
01:13 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/21b31735c6
01:13 dalek Rosella: 8e111dc | Whiteknight++ | / (4 files):
01:13 dalek Rosella: fix some bugs in queries. Add a stub test file to show that mapping arrays and hashes works
01:13 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/8e111dc944
01:13 dalek Rosella: 97732a8 | Whiteknight++ | / (3 files):
01:13 dalek Rosella: fixes for 'filter' and some tests for the behavior
01:13 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/97732a8a17
01:13 hudnix left #parrot
01:14 hudnix joined #parrot
01:15 bacek_at_work too late
01:15 dalek parrot/opsc_full_parse: 53fccb0 | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
01:15 dalek parrot/opsc_full_parse: Add case and default to keywords. Fix switch statement.
01:15 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/53fccb0d56
01:16 cotto oh noes
01:21 bacek_at_work cotto, io.ops can't be parsed atm. Due "string concatenation". E.g. INT_FMT "\n"
01:21 cotto i'll look at it if you don't first
01:23 whiteknight we need 6model
01:23 whiteknight we need it yesterday
01:23 cotto whiteknight, ?
01:23 whiteknight just saying
01:23 whiteknight nothing to see here. Continue with your normal conversations
01:23 cotto wfm
01:26 cotto whiteknight, are you looking at our oo code or something?
01:26 whiteknight maybe
01:26 whiteknight I've been reading through the 6model code all day
01:27 cotto it's a bit odd, but I really like it
01:27 cotto it'll be great in its final form
01:29 Tene final form, eh?  makes me thing of final fantasy.
01:31 mtk left #parrot
01:37 kid51 cotto Any comment on http://trac.parrot.org/parrot/ticket/683 ?
01:38 mtk joined #parrot
01:41 dalek TT #692 closed by jkeenan++: PBC to PIR fallback fails.
01:41 dalek TT #692: http://trac.parrot.org/parrot/ticket/692
01:54 sigue joined #parrot
02:04 whiteknight Tene: you a final fantasy fan?
02:05 whiteknight cotto: Rakudo is moving to 6model. tewk is talking about moving Cardinal to 6model. Those are currently our two biggest HLLs.
02:05 whiteknight I can't imagine a decent JS compiler using our current metamodel
02:05 whiteknight moving 6model into Parrot core is the best way to reduce impedance between Parrot and these compilers
02:07 cotto whiteknight, that's the plan
02:07 whiteknight right. I know that's the plan. I want to facilitate and accelerate
02:09 whiteknight I really need to get this damn IMCC work out of the way so I can move on to the next project
02:10 plobsing sure we want it. but lets not rush it and get a half baked implementation.
02:12 nopaste "Whiteknight" at 192.168.1.3 pasted "Can anybody spot the problem?" (9 lines) at http://nopaste.snit.ch/33556
02:13 whiteknight if anybody can figure out what the problem is with that test failure there, they win a prize
02:15 tadzik whhitespace?
02:15 tadzik yes. If you look slocesly, you'll notice two newlines in the first one
02:16 tadzik ...damh lag
02:16 whiteknight oh shit
02:16 whiteknight tadzik wins the prize
02:17 tadzik yay, shake-a-hand with whiteknight awaits!
02:17 dalek parrot/ptrbuf: af78853 | plobsing++ | src/pmc/structview.pmc:
02:17 dalek parrot/ptrbuf: add documentation for StructView
02:17 dalek parrot/ptrbuf: review: https://github.com/parrot/parrot/commit/af7885367a
02:17 dalek parrot/ptrbuf: 572a1e4 | plobsing++ | src/pmc/structview.pmc:
02:17 dalek parrot/ptrbuf: [codingstd] c_struct
02:17 dalek parrot/ptrbuf: review: https://github.com/parrot/parrot/commit/572a1e4d03
02:17 dalek parrot/ptrbuf: f42c67d | plobsing++ | src/pmc/ptr (2 files):
02:17 dalek parrot/ptrbuf: [codingstd] tabs
02:17 dalek parrot/ptrbuf: review: https://github.com/parrot/parrot/commit/f42c67d7cf
02:17 dalek parrot/ptrbuf: a4d4bdd | plobsing++ | src/pmc/ (2 files):
02:17 dalek parrot/ptrbuf: [codingstd] c_parens
02:17 dalek parrot/ptrbuf: review: https://github.com/parrot/parrot/commit/a4d4bdd726
02:17 dalek parrot/ptrbuf: ceb741c | plobsing++ | src/pmc/structview.pmc:
02:17 dalek parrot/ptrbuf: [codingstd] trailing_space
02:17 dalek parrot/ptrbuf: review: https://github.com/parrot/parrot/commit/ceb741c40b
02:17 dalek parrot/ptrbuf: fbf595a | plobsing++ | lib/Parrot/Headerizer.pm:
02:17 dalek parrot/ptrbuf: ignore begin/end pmc header preamble statements
02:17 dalek parrot/ptrbuf: review: https://github.com/parrot/parrot/commit/fbf595a86d
02:17 dalek parrot/ptrbuf: ab9ce37 | plobsing++ | src/pmc/structview.pmc:
02:17 dalek parrot/ptrbuf: [codingstd] pmc_docs
02:17 dalek parrot/ptrbuf: review: https://github.com/parrot/parrot/commit/ab9ce3779e
02:18 tadzik it may be quite useful to substitute newlines with literal \n in TAP output
02:28 sigue left #parrot
02:30 sorear alternatively, with \n#
02:30 cotto kid51, we should tag that tt as "6model" or something similar and close it when 6model lands in master.  Yes, it'll be a while.
02:31 cotto I don't like closing tickets with a reference to something nyi
02:38 dalek parrot/whiteknight/imcc_compreg_pmc: a74e161 | Whiteknight++ | compilers/imcc/debug.c:
02:38 dalek parrot/whiteknight/imcc_compreg_pmc: fix whitespace errors to fix a failing test. tadzik++
02:38 dalek parrot/whiteknight/imcc_compreg_pmc: review: https://github.com/parrot/parrot/commit/a74e161994
02:38 dalek parrot/whiteknight/imcc_compreg_pmc: 795f437 | Whiteknight++ | src/pmc/imccompiler.pmc:
02:38 dalek parrot/whiteknight/imcc_compreg_pmc: actually throw an exception from IMCCompiler.compile_file method. This fixes behavior in the API and reclaims another test
02:38 dalek parrot/whiteknight/imcc_compreg_pmc: review: https://github.com/parrot/parrot/commit/795f437096
02:38 whiteknight left #parrot
02:39 Tene aloha: tell whiteknight Tewk is talking about adding async IO to Parrot.  Tene is talking about moving Cardinal to 6model.
02:39 aloha Tene: Okay.
02:50 ShaneC left #parrot
02:54 kid51 left #parrot
03:07 ShaneC joined #parrot
03:08 bubaflub joined #parrot
03:09 davidfetter joined #parrot
03:14 cotto seen japhb
03:14 aloha Sorry, I haven't seen japhb.
03:15 davidfetter left #parrot
03:18 dalek parrot: 1b11ce1 | cotto++ | config/gen/opengl.pm:
03:18 dalek parrot: make the gen::opengl step less noisy on unknown function prototypes
03:18 dalek parrot: review: https://github.com/parrot/parrot/commit/1b11ce1948
03:19 dalek parrot/opsc_full_parse: c1a030a | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
03:19 dalek parrot/opsc_full_parse: Add hack to support string concatenations.
03:19 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/c1a030aaa9
03:19 dalek parrot/opsc_full_parse: 821dc1d | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
03:19 dalek parrot/opsc_full_parse: Add optional type letter to integer constants.
03:19 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/821dc1d4b5
03:19 dalek parrot/opsc_full_parse: e4249f0 | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
03:19 dalek parrot/opsc_full_parse: Add more prefix and infix tokens.
03:19 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/e4249f0579
03:19 dalek parrot/opsc_full_parse: 6f41bf3 | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
03:19 dalek parrot/opsc_full_parse: Allow multiple * for type_definition
03:19 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/6f41bf3f57
03:20 bacek_at_work cotto, you are way too slow!
03:20 bacek_at_work bacek@illusion:~/src/parrot (opsc_full_parse) $ prove t/compilers/opsc/02-parse-all-ops.t
03:20 bacek_at_work t/compilers/opsc/02-parse-all-ops.t .. ok
03:20 bacek_at_work All tests successful.
03:20 bacek_at_work Files=1, Tests=19, 10 wallclock secs ( 0.02 usr  0.00 sys + 10.60 cusr  0.02 csys = 10.64 CPU)
03:20 bacek_at_work Result: PASS
03:21 ttbot Parrot 1b11ce19 i386-linux-thread-multi make error http://tt.taptinder.org/cmdinfo/26306
03:21 cotto bacek_at_work, my fleshy fingers and frail brain can't keep up with your robot-like speed.
03:21 cotto I'm not even half-way through the stuff I decided I needed to plow through before looking at opsc.
03:23 cotto bacek++
03:23 cotto it's nice that they let you hack on Parrot at work (or that you can get away with it)
03:24 ttbot Parrot 1b11ce19 i386-linux-thread-multi make error http://tt.taptinder.org/cmdinfo/26317
03:24 cotto can we fix that damn json bug?
03:25 bacek_at_work cotto, just revert commit
03:25 bacek_at_work much easy than unmerge whole branch
03:26 cotto It should be an easy fix.
03:26 dukeleto cotto: yes, please revert the commit
03:26 cotto I'd have done it from work but I'm a little sleep-deprived.
03:26 cotto is the branch still there?
03:26 dukeleto cotto: i am talking about reverting the commit in master. I don't know which branch it came from.
03:27 bacek_at_work cotto, just whinging
03:28 * cotto needs less being crabby and more clear thinking
03:29 atrodo don't we all
03:31 bacek_at_work I don't. I'm computing not thinking.
03:32 atrodo correction, don't all humans
03:32 * cotto wants to rewrite himself
03:33 cotto though debugging could get ugly
03:35 dalek TT #1947 closed by cotto++: excessively noisy output from Configure.pl
03:35 dalek TT #1947: http://trac.parrot.org/parrot/ticket/1947
03:41 ttbot Parrot 1b11ce19 i386-linux-thread-multi make error http://tt.taptinder.org/cmdinfo/26339
03:50 cotto dukeleto, ping
03:53 dalek parrot: 2ad0afc | cotto++ | tools/dev/nci_thunk_gen.pir:
03:53 dalek parrot: Revert "change to new JSON language API"
03:53 dalek parrot:
03:53 dalek parrot: This reverts commit adafa9f7d885c1265a590f439caa9eb531a12115.
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/2ad0afc371
03:59 cotto Some days I wish Parrot were more active.  Some days I wish the flood would slow down.
04:01 bubaflub cotto: i hear ya; i'm stuck in an Airport with free wi-fi till tomorrow morning
04:02 cotto bubaflub, been there
04:04 plobsing cotto: what kind of day is today?
04:04 cotto flood
04:11 dalek parrot: e18da5e | (Gerd Pokorra)++ | config/gen/makefiles/root.in:
04:11 dalek parrot: add dependency to fix JSON error
04:11 dalek parrot: review: https://github.com/parrot/parrot/commit/e18da5eb10
04:13 * cotto tries to summon gerd
04:19 * atrodo hands cotto his staff of +2 summoning
04:24 cotto It's not working.
04:42 plobsing config/auto/pmc.pm generates dependancies of PMCs on all ancestors. Pmc2c only creates an include for direct parents. which, if either of these is right?
04:46 cotto direct parents is all that's necessary
04:48 cotto the code that pmc2c generates for a PMC only depends directly on that PMC's immediate parent(s)
04:48 plobsing I was leaning that way.
04:50 dalek cardinal/nqprx: 2befbd4 | tadzik++ | / (5 files):
04:50 dalek cardinal/nqprx: Moved from PGE to NQP-rx. Doesn't yet work, there's a TODO in a grammar but it compiles
04:50 dalek cardinal/nqprx: review: https://github.com/parrot/​cardinal/commit/2befbd4bf3
04:51 plobsing what do PMCs need to include from even their immediate parents? the only thing I ever use those headers for are attribute accessors, which are regenerated in the child's header anyways.
04:51 bacek_at_work cotto, stop slacking! We do need new "pasting" for opsc :)
04:51 bacek_at_work plobsing, "update_vtable" functions.
04:51 cotto It's a little more annoying when trying to figure out which VTABLE functions come from where, but not having all the symbols sped up libparrot startup measruably
04:51 cotto what bacek_at_work said
04:51 cotto Parrot_x_get_vtable et al
04:52 Tene aloha: tell dukeleto to talk to Tene about admin privs on the cardinal repo
04:52 cotto bacek_at_work, not slacking; responding to the keys deprecation tt
04:52 bacek_at_work hooray! Finally :)
04:52 plobsing cotto: not having vtables static *sped up* libparrot?
04:53 gbacon left #parrot
04:53 cotto plobsing, making them static, i.e. not exporting the symbols
04:53 cotto sorry, misread
04:54 plobsing when I'm thinking about static, I'm thinking constant, shared memory. statically defined in the file, not created at runtime.
04:54 cotto yeah, that's what I misread
04:55 cotto the VTABLE functions in the post-processed C files are no startic now, but the VTABLE structs are built at startup
04:55 cotto s/no startic/static/
04:55 cotto *sigh*
04:56 cotto bacek_at_work, one question about that ticket is how to deal with ["foo";3;"wqrwqew"]
04:56 cotto fpa sounds like it might be expensive
04:58 bacek_at_work I want to kill such syntax
04:58 bacek_at_work leave only ["foo";"bar";"baz"] with FSA
04:58 plobsing yes, and I'd been thinking for a little while that those vtable structs could be defined such that they resided in read-only, shared memory with the implicit assumption that that would be cached by the OS, giving better load time and better memory utilization
05:00 cotto bacek_at_work, do you know how often that sytax is used?
05:00 cotto *syntax
05:00 bacek_at_work cotto, not much. In PLA only.
05:00 bacek_at_work afaik
05:01 cotto pipp used to use it, but it'll have bitrotted pretty severely by now
05:01 plobsing keyed-access is a useful idiom. We can "kill Keys", but something else that's basically equivalent will be needed and eventually come into being.
05:02 cotto plobsing, +1
05:02 cotto The plan (afaiu) is to use the same syntax to produce non-key PMCs.
05:02 plobsing I do agree that the scope is too large (eg: use as namespace identifiers), and that their implimentation and interface sucks (linked-list with many gcables, fragile custom functions that can only handle Key PMCs)
05:03 cotto It's going to be wonderful to see that code go away.
05:03 bacek_at_work plobsing, what about "Keyed access" vs "Nested aggregates Keyed access"?
05:03 bacek_at_work plobsing, I don't think that we do need "nested" version
05:04 plobsing just because it exists doesn't mean your agregate has to support it
05:04 plobsing just because it exists doesn't mean you have to use it
05:04 bacek_at_work but we have to maintain it
05:04 plobsing but it is useful in some situations to expose some interfaces
05:05 cotto How expensive would it be to use an fpa and boxing for what keys are used for now?
05:05 bacek_at_work plobsing, yes. For namespaces. http://trac.parrot.org/parrot/ticket/903
05:05 bacek_at_work With special "NamespaceKey"
05:05 plobsing cotto: FPA doesn't act as a reference to registers. That is a very useful thing
05:05 bacek_at_work not generic Key-of-something-for-something
05:06 bacek_at_work we can have NamespaceKey to support register references. I just want it specialiazed for Namespace only.
05:06 bacek_at_work afk
05:07 plobsing NamespaceKey shouldn't support register references. That does not make sense.
05:07 bacek_at_work plobsing, why?
05:07 plobsing .namespace [$P0, $I0] ??
05:07 plobsing really? what does that mean?
05:07 bacek_at_work [$S0, $S1]
05:08 bacek_at_work $P0 = get_root_ns; $P1 = $P0[$S0, $S1]
05:08 plobsing ah, you're using "namespace key" to mean that which is used in namespace lookups.
05:08 plobsing I thought we were only talking about namespace defn's
05:09 plobsing what you are proposing has the potential to greatly increase the number of specialized key types
05:10 plobsing the great thing about Keys, was that they were the only type you had to understand to interpret PBC completely (this was before PCC ruined that)
05:12 dalek parrot: 5c3a01e | (Gerd Pokorra)++ | config/gen/makefiles/root.in:
05:12 dalek parrot: remove dependency; move if to correct place
05:12 dalek parrot: review: https://github.com/parrot/parrot/commit/5c3a01ed65
05:12 cotto What if fxa could hold register references?
05:13 plobsing Are you proposing a new type, or ammending the existing types with a third leg?
05:13 cotto amending existing types, but it's just a thought experiment atm
05:15 plobsing I don't think that's a great idea. The existing fxa types are one thing - a simple array of x.
05:16 plobsing We need *different* types to do register references because they fundamentally act differently.
05:17 plobsing register-reference objects should only ever be used from constants static to the PBC. A register allocator will cause any other types of usages to fail mysteriously.
05:18 cotto Then what if you could use registers to build FxA, but needed to copy the values?
05:18 plobsing That's one alternative to Keys. It involves a lot of GC churn.
05:18 cotto yeah
05:19 cotto I'm starting to wonder if the right thing is to get rid of keys.
05:20 plobsing My opinion: reduce their scope, change their interface to be flexible and allow other types to be supported, and then leave well enough alone.
05:20 plobsing actually, once you have the flexible interface, the guts could get some much-needed tweaking.
05:21 cotto bacek_at_work, how does that strike you?
05:22 dalek parrot: 2010eaa | cotto++ | docs/pdds/ (2 files):
05:22 dalek parrot: Move pbc pdd into drafts.
05:22 dalek parrot:
05:22 dalek parrot: It's helpful for understanding a slightly out-of-date version of our
05:22 dalek parrot: bytecode format, but it's going to get much more dated in the near
05:22 dalek parrot: future.
05:22 dalek parrot: review: https://github.com/parrot/parrot/commit/2010eaaa60
05:24 Coke btw, -1 to having a "draft/" directory as opposed to just marking the file with DRAFT in the title.
05:24 Coke (because I'm pretty sure that the index of PMCs has not been following which are drafts and which are not.)
05:24 Coke s/PMC/PDD/
05:25 cotto Coke, I'm not attached to any convention.  I just did it that way because I didn't feel like inventing something better.
05:27 cotto whiteknight++ for beating me to many of these tickets
05:27 Coke ah, looks like someone (could have been me) fixed it so you don't have to manually track it anymore.
05:27 Coke of course, now docs/html/pdds.html looks funny because you did not also change the title to draft.
05:27 cotto me--
05:28 cotto Coke, do you mind fixing that?
05:28 Coke (oh, right, it was probably fixed with the html rewrite I started and mikehh finished.
05:28 Coke cotto: aye. ;)
05:28 cotto coke++
05:28 Coke "yes, I mind.", sorry, wasn't clear.
05:29 Coke heading to bed anyway. ~~
05:29 cotto 'night
05:44 cotto it makes me happy every time I see a commit sha1 in a trac ticket.
05:46 dalek TT #683 closed by cotto++: implement method lookup iterators
05:46 dalek TT #683: http://trac.parrot.org/parrot/ticket/683
05:46 dalek TT #611 closed by cotto++: pdd06-pasm (complete syntax/semantics)
05:46 dalek TT #611: http://trac.parrot.org/parrot/ticket/611
05:57 cotto Apart from it occasionally overwhelming my browser, I'm really liking github for commit review.
05:57 benabik joined #parrot
05:58 benabik left #parrot
06:06 bubaflub left #parrot
06:15 rurban_ joined #parrot
06:17 rurban left #parrot
06:17 rurban_ is now known as rurban
06:29 dalek parrot: ee403d8 | mikehh++ | MANIFEST:
06:29 dalek parrot: re-generate MANIFEST
06:29 dalek parrot: review: https://github.com/parrot/parrot/commit/ee403d8c5e
06:40 cotto tewk, ping
06:56 cotto seen dukeleto
06:56 aloha dukeleto was last seen in #parrot 3 hours 29 mins ago saying "cotto: i am talking about reverting the commit in master. I don't know which branch it came from.".
06:57 dukeleto cotto: pong
07:00 cotto dukeleto, see privmsg
07:05 JimmyZ joined #parrot
07:08 cotto How hard is it to get ttbot to email a list when a build breaks?
07:09 dukeleto cotto: may want to msg that to mj41
07:09 dukeleto cotto: i assume if the feature isn't there, it shouldn't be hard to add to TapTinder. All the code is on github
07:09 cotto seen mj41
07:09 aloha mj41 was last seen in #parrot 9 hours 58 mins ago saying "Fixed (html escaped) TapTinder error output http://tt.taptinder.org/cmdinfo/25763".
07:10 dukeleto cotto: i plan on setting up jitterbug for parrot, which has email functionality
07:10 * dukeleto has a linode that he is still trying to put to use
07:11 cotto dukeleto, we need a version of you that's not crazy busy
07:14 cosimo left #parrot
07:15 dukeleto cotto: yeah, i keep trying to work on that
07:17 cotto You should get involved with one of those secret unethical companies working on human cloning that I keep not hearing about.
07:18 he__ joined #parrot
07:25 JimmyZ left #parrot
07:25 dalek parrot: 8682534 | dukeleto++ | docs/pdds/ (30 files):
07:25 dalek parrot: [cage] Remove more svn-- remnants from our PDDs
07:25 dalek parrot: review: https://github.com/parrot/parrot/commit/8682534f7c
07:26 cotto 30 files?  blarg
07:27 TiMBuS hello parroteers. yesterday i noticed i can't use sockets in rakudo, and it seems to be a parrot error. was the socket pmc changed recently? what is the way you are meant to open a socket?
07:29 TiMBuS more specifically: the connect method checks to see if the socket is open before it connects. what does it want me to do to open it?
07:29 cotto TiMBuS, does t/pmc/socket.t tell you what you need?
07:30 TiMBuS didnt even think of checking the tests!
07:30 cotto they're handy
07:31 dukeleto TiMBuS: we have tried to add ipv6 support recently, but that might have broken something for you
07:32 TiMBuS hmmm maybe
07:32 TiMBuS this test example might have helped
07:32 TiMBuS ah yep thats the problem
07:33 TiMBuS you need to call .socket on a socket object now
07:33 dukeleto TiMBuS: t/pmc/sockaddr.t and t/pmc/socket_ipv6.t may also be helpful
07:33 dukeleto TiMBuS: when is the last time you know that your code worked?
07:34 TiMBuS hmmmmm last time i updated rakudo.. about 3 weeks?
07:34 TiMBuS ago
07:35 TiMBuS thats ok, im pretty sure i can get it working now
07:35 cotto TiMBuS, great
07:36 baest joined #parrot
07:36 he__ left #parrot
07:37 baest left #parrot
07:37 baest joined #parrot
07:38 baest left #parrot
07:41 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#10575) fulltest) at 3_1_0-213-g40c0d36 - Ubuntu 10.10 i386 (g++-4.5)
07:46 dalek nqp: a007136 | moritz++ | src/HLL/Compiler.pm:
07:46 dalek nqp: rewrite parse_name in nqp
07:46 dalek nqp: review: https://github.com/perl6/nqp/commit/a00713648f
07:59 lucian joined #parrot
08:02 fperrad joined #parrot
08:08 contingencyplan left #parrot
08:09 he__ joined #parrot
08:24 lucian left #parrot
08:44 theory left #parrot
08:54 dalek parrot/m0-spec: c759e28 | dukeleto++ | docs/pdds/draft/pdd32_m0.pod:
08:54 dalek parrot/m0-spec: Add chromatic++'s description of ops that are too magical for M0
08:54 dalek parrot/m0-spec: review: https://github.com/parrot/parrot/commit/c759e2830d
08:57 dalek parrot/m0-spec: 765a66e | dukeleto++ | docs/pdds/draft/pdd32_m0.pod:
08:57 dalek parrot/m0-spec: Clarify our stance on minimization of opcodes
08:57 dalek parrot/m0-spec: review: https://github.com/parrot/parrot/commit/765a66e7b1
09:04 ShaneC left #parrot
09:19 mtk left #parrot
09:24 bacek ~~
09:25 cosimo joined #parrot
09:26 mtk joined #parrot
09:42 tadzik hello parrots
10:33 dalek parrot: 60b9483 | fperrad++ | runtime/parrot/library/Archive/Tar.pir:
10:33 dalek parrot: [Archive/Tar] fix on Win32
10:33 dalek parrot: review: https://github.com/parrot/parrot/commit/60b948337e
10:43 cosimo left #parrot
10:49 dalek lua: 44b666d | fperrad++ | t/lua-TestMore:
10:49 dalek lua: update submodule lua-TestMore
10:49 dalek lua: review: https://github.com/fperrad/lua/commit/44b666dce4
11:32 ligne joined #parrot
12:04 lbr joined #parrot
12:16 bluescreen joined #parrot
13:19 JimmyZ joined #parrot
13:34 Coke left #parrot
13:35 whiteknight joined #parrot
13:35 whiteknight good morning, #parrot
13:36 JimmyZ good morning whiteknight, how are you
13:37 whiteknight I am doing well, JimmyZ. How are you today?
13:38 JimmyZ I'm fine, but a bit bored again :)
13:39 Coke joined #parrot
13:39 tadzik oh, did you have a Java lecture today? :)
13:39 plobsing left #parrot
13:40 moritz JimmyZ: when you are bored, write tests for those tickets here: http://rakudo.de/testneeded :-)
13:40 plobsing joined #parrot
13:40 tadzik I stayed awake for the entire night hacking on Cardinal mostly. When I got up, I wasn't tired at all. But as soon as I reached my lecture, the sleep felt so tempting :>
13:43 whiteknight tadzik++ The all-nighters are key
13:44 whiteknight I'm very happy to see renewed interest in Cardinal
13:44 whiteknight One of my coworkers came up to me yesterday and said "Do you know anything about this Cardinal project on github?". One of his friends started following it, and he knew I did Parrot stuff
13:45 tadzik whiteknight: do you know it was almost converted to nqp-x?
13:45 tadzik *nqp-rx
13:47 whiteknight what is it written in now?
13:48 moritz old nqp, iirc
13:48 whiteknight tewk was talking about using 6model, so I assume he's talking about building it with the new NQP
13:48 tadzik pge
13:48 whiteknight PGE? damn
13:48 moritz ah right, old NQP used PGE as the grammar engine
13:48 tadzik yeah. Tonight I almost translated lolcode to nqprx (: I'll need some lolcode hackers to gimme a hand
13:49 tadzik or maybe just ordinary Parrot hacker :)
13:50 tadzik lolcode uses some magic trick to inject a custom method written in PIR into the grammar written in perl6 rules. That kinda worked on PGE, but acts quite funny on NQP-RX
13:50 moritz it should Just Work I think
13:50 moritz oh wait
13:50 moritz PGE and NQP-rx use a different argument passing style
13:51 moritz I think you need pmichaud++ for this one
13:51 moritz or maybe jnthn++
13:51 sigue joined #parrot
13:52 nopaste "tadzik" at 192.168.1.3 pasted "pir-written methods and nqp grammars don't mix very well" (17 lines) at http://nopaste.snit.ch/33587
13:52 tadzik hmm, I can actually push it to a branch
13:55 Coke is there any way to determine at runtime what HLL is in effect?
13:56 Coke (preferably PIR, but a dynop would work too.)
13:57 Coke would help with my "why is this an Integer instead of a TclInt" problem.
14:06 JimmyZ moritz: I can do it, but I don't have a commit bit .
14:06 moritz JimmyZ: what's your github ID?
14:07 JimmyZ zhuomingliang
14:07 Coke anyone know of a tool that goes through all the commits in a repo and changes author email A to B ?
14:07 moritz JimmyZ: now you have a commit bit. Welcome :-)
14:07 JimmyZ moritz: thanks.
14:07 atrodo Coke> http://www.kernel.org/pub/software/​scm/git/docs/git-filter-branch.html
14:09 whiteknight Coke: "$P0 = get_hll_namespace"
14:09 whiteknight Coke: I think I have TT #2022 figured out, but I haven't been able to reproduce locally
14:10 whiteknight when I run the example code you gave me, I get an "lrepeat not implemented" kind of error
14:10 Coke atrodo: danke.
14:10 Coke whiteknight: which version of partcl are you using?
14:10 atrodo Coke> No problem.  Had to do it a few weeks ago myself
14:10 whiteknight Coke: partcl, not partcl-nqp
14:10 Coke atrodo: trying to do a svn-to-git conversion, and preferred email addresses are still trickling in. pita.
14:11 Coke whiteknight: odd error, then.
14:11 bluescreen left #parrot
14:11 atrodo Coke> Oh, that sounds like a load of unpredictable fun
14:13 rurban_ joined #parrot
14:14 Coke whiteknight: weird. that says that the namespace /is/ tcl.
14:14 Coke (and thank you, I have no idea why it didn't occur to me to use that opcode!)
14:15 whiteknight Coke: I suspect it's a problem inside IMCC. In some place inside IMCC it forces the HLL to "0", which is the parrot HLL
14:15 whiteknight I suspect that my recent IMCC cleanups may have exposed that code in a way that it was never used before
14:15 Coke yah, when I box, in the 'tcl' namespace, it's generating a String, not a TclString.
14:16 Coke but it doesn't happen all the time - if I write a small test case, it works fine.
14:16 Coke I haven't had the time to strip this down from "all of partcl" to "small test case".
14:17 rurban left #parrot
14:17 rurban_ is now known as rurban
14:17 Coke perhaps put a break on that box in src/grammar/expr/past2pir.tg and see where it goes inside IMCC.
14:17 whiteknight if it goes inside IMCC
14:17 Coke (line 91)
14:18 whiteknight line 91?
14:20 whiteknight set a break on Parrot_pcc_set_HLL or Parrot_pcc_set_HLL_func
14:20 JimmyZ moritz: most teams can 'delete' Repositories?
14:20 Coke line 91 of src/grammar/expr/past2pir.tg is the box that gets it wrong.
14:20 moritz JimmyZ: what exactly are you talking about?
14:21 lbr left #parrot
14:22 Tene tadzik: I'd be glad to help you with lolcode.  I wrote that too.  :)
14:22 JimmyZ moritz: the power of the teasm :)
14:22 JimmyZ *team
14:22 Tene tadzik: It'll have to wait until after I'm done with work, though. (currently woken up early by the pager)
14:26 whiteknight pagers are the worst
14:26 moritz JimmyZ: thanks, that was not intended
14:26 moritz JimmyZ: I've changed the permissions to Push & Pull now
14:26 moritz (was : Push, Pull & Administrative)
14:26 moritz which is a bit exaggerated
14:26 contingencyplan joined #parrot
14:26 Tene whiteknight: Especially when the problem is that a third party is failing; so the only action I can take is watching my inbox to hear that it's fixed.
14:27 JimmyZ moritz: hehe
14:27 whiteknight double-sucks
14:30 JimmyZ moritz: I got Permission denied (publickey) error .
14:31 moritz JimmyZ: then double-check the ssh key you uploaded to github
14:31 moritz JimmyZ: you're in a github group grants push access to perl6/roast - there's nothing more I can do
14:31 JimmyZ moritz: I see
14:32 whiteknight_ joined #parrot
14:32 dalek winxed: r819 | NotFound++ | trunk/winxedst1.winxed:
14:32 dalek winxed: cleanups and style fixes
14:32 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=819
14:34 whiteknight__ joined #parrot
14:37 whiteknight left #parrot
14:38 whiteknight__ is now known as whiteknight
14:39 whiteknight_ left #parrot
14:49 Coke so, box seems pretty straightforward. unless something is messing with the current HLL in the context...
14:49 PacoLinux left #parrot
14:54 lucian joined #parrot
15:00 dalek winxed: r820 | NotFound++ | trunk/winxedst0.cpp:
15:00 dalek winxed: some refactor in stage 0
15:00 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=820
15:03 JimmyZ sigh, I can't do git pull parrot too.
15:05 cotto wfm
15:06 moritz GCF?
15:07 cotto pdq
15:11 whiteknight Coke: at the point of the box the HLL namespace is "tcl"?
15:11 whiteknight before and after the box?
15:14 JimmyZ E:\parrot>git pull Permission denied (publickey). fatal: The remote end hung up unexpectedly
15:14 Coke just before the box, yes.
15:14 JimmyZ oh
15:14 Coke er, s/just/
15:14 moritz JimmyZ: did you check those ssh keys?
15:15 JimmyZ I didn't modify it, it worked before.
15:15 Coke whiteknight: just before and just after the box, tcl both times.
15:16 whiteknight weird
15:16 dalek TT #1951 closed by doughera++: t/dynpmc/os.t incorrectly uses system-specific error strings
15:16 dalek TT #1951: http://trac.parrot.org/parrot/ticket/1951
15:18 he__ left #parrot
15:20 lucian JimmyZ: that sort of thing used to happen to me with gitorious, occasionally
15:23 * JimmyZ doesn't how to fix it
15:23 JimmyZ doesn't know
15:28 JimmyZ shit, I use cmd for git, it's wrong way.
15:39 cotto_work ~~
15:41 gbacon joined #parrot
15:45 bluescreen joined #parrot
15:55 theory joined #parrot
16:02 cotto_work http://blog.ksplice.com/2011/01/8-gdb-tricks/
16:13 dalek Rosella: cca93a9 | Whiteknight++ | / (2 files):
16:13 dalek Rosella: Add in a path-query class. It's similar to xpath, but for Parrot objects.
16:13 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/cca93a95e4
16:13 dalek Rosella: 65b389a | Whiteknight++ | tap_harness/ (2 files):
16:13 dalek Rosella: Fixes to the harness, I was getting the premature exit logic backwards
16:13 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/65b389ae75
16:13 dalek Rosella: 3aa2160 | Whiteknight++ | query/Path.nqp:
16:13 dalek Rosella: fixes to Path.nqp
16:13 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/3aa21607e2
16:13 dalek Rosella: c822d18 | Whiteknight++ | t/ (2 files):
16:13 dalek Rosella: Add t/query to the harness. Add some basic tests for new path functionality
16:13 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/c822d187ea
16:17 Patterner left #parrot
16:17 Psyche^ joined #parrot
16:17 cotto_work tewk: ping
16:18 Psyche^ is now known as Patterner
16:18 bluescreen left #parrot
16:22 dalek winxed: r821 | NotFound++ | trunk/winxedst1.winxed:
16:22 dalek winxed: some little refactors for clarity and a bit of speed
16:22 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=821
16:38 cotto_work I thought our tests griped when a PMC function wasn't documented.
16:38 tadzik aloha: clock
16:38 tadzik aloha: clock?
16:38 aloha tadzik: LAX: Tue, 08:38 PST / CHI: Tue, 10:38 CST / NYC: Tue, 11:38 EST / UTC: Tue, 16:38 UTC / LON: Tue, 16:38 GMT / BER: Tue, 17:38 CET / TOK: Wed, 01:38 JST / SYD: Wed, 03:38 EST
16:38 Coke is this the doc test where people started putting in empty docs?
16:39 * Coke hated that and hoped he got it removed from wherever it was used.
16:42 JimmyZ left #parrot
16:42 cotto_work Empty docs are worthless, but the test is valuable if we can get people to contribute something substantial.  The empty docs come from a mentality of making the tests shut up rather than actually fixing anything.
16:42 dalek parrot: 948e3ee | (Gerd Pokorra)++ | config/gen/makefiles/root.in:
16:42 dalek parrot: add more dependencies
16:42 dalek parrot: review: https://github.com/parrot/parrot/commit/948e3ee810
16:42 dalek parrot: cfc6507 | (Gerd Pokorra)++ | tools/dev/nci_thunk_gen.pir:
16:42 dalek parrot: try to switch to JSON.nqp again
16:42 dalek parrot: review: https://github.com/parrot/parrot/commit/cfc6507f85
16:42 cotto_work gah
16:42 cotto_work that's what branches are for
16:42 cotto_work gerd--
16:43 gerd joined #parrot
16:43 Coke msg dukeleto 48933675c362e09ace9dc1f86cd4010532593002 -- you need to tack on the EXE there to work on windows, me thinks.
16:43 aloha OK. I'll deliver the message.
16:43 cotto_work gerd: hi.  Could you move that into a branch?
16:44 gerd Do you mean the last changes I have just made?
16:46 ttbot Parrot cfc6507f i386-linux-thread-multi make error http://tt.taptinder.org/cmdinfo/26868
16:46 cotto_work gerd: yes
16:46 cotto_work ^ that's why
16:46 Coke at least now we know they don't work. ;)
16:46 bluescreen joined #parrot
16:46 JimmyZ joined #parrot
16:46 Coke cotto_work: well, it's hard to tell if they don't work if they're not tested. putting them on master /does/ test them pretty quickly these days.
16:46 Coke (but, knowing that we already did this dance at least once does make me gun shy)
16:47 cotto_work Coke: exactly
16:47 gerd I hope that they work with the dependencies they now have
16:48 cotto_work gerd: they don't
16:48 cotto_work that's why I'd like a branch
16:49 gerd I make new dependencies for 2 minutes, are you sure they do not work
16:49 cotto_work ttbot just reported a failure
16:51 Coke gerd: 11:46 <+ttbot> Parrot cfc6507f i386-linux-thread-multi make error http://tt.taptinder.org/cmdinfo/26868
16:51 cotto_work I'm happy to help iron out the wrinkles
16:51 gerd okay then I will start with a branch
16:51 cotto_work gerd: great.  Thank you.
16:51 JimmyZ left #parrot
16:51 Coke looks like it failed on 50% of the builds at that commit.
16:52 gerd cotto_work: Should I switch back?
16:52 cotto_work gerd: please do
16:52 Coke you need to at least revert cfc6507f
16:52 cotto_work (assuming you mean "revert")
16:53 cotto_work right.  cfc65 is the one to revert
16:53 ttbot Parrot cfc6507f i386-linux-thread-multi make error http://tt.taptinder.org/cmdinfo/26914
16:55 dalek parrot/ptrbuf: a42364c | plobsing++ | t/src/checkdepend.t:
16:55 dalek parrot/ptrbuf: skip checkdepend on pmc directories
16:55 dalek parrot/ptrbuf:
16:55 dalek parrot/ptrbuf: checkdepend doesn't handle inheritance properly, but pmc2c and config/auto/pmc.pm do.
16:55 dalek parrot/ptrbuf: review: https://github.com/parrot/parrot/commit/a42364c459
16:55 Andy joined #parrot
16:56 tewk cotto pong
16:56 tewk cotto_work: pong
16:56 cotto_work tewk: in tewk/select, why did you add int keyed access VTABLE functions to Hash?
16:57 cotto_work (I blink on "cotto" and "cotto_work".  no need to ping both)
16:58 dalek parrot: c1dc008 | (Gerd Pokorra)++ | / (2 files):
16:58 dalek parrot: switch back to data_json
16:58 dalek parrot: review: https://github.com/parrot/parrot/commit/c1dc00846b
16:58 cotto_work gerd++
16:59 tewk cotto_work: convenience, are they provided by a superclass?
16:59 gerd cotto_work: no gerd++; I sorry that it do not work
16:59 dalek parrot: 61ad091 | cotto++ | / (2 files):
16:59 dalek parrot: add gerd alias to CREDITS
16:59 dalek parrot: review: https://github.com/parrot/parrot/commit/61ad091f6d
17:00 cotto_work gerd: have you read docs/project/git_workflow.pod?
17:00 cotto_work tewk: I don't think they're appropriate for a Hash.
17:01 gerd cotto_work: I know that docs
17:01 cotto_work gerd: ok
17:01 tewk cotto_work: Why not? hash keys should be able to be INTVL
17:01 tewk ALS.
17:02 cotto_work gerd: you can easily revert with "git revert c1dc008"
17:03 gerd cotto_work: okay; I will send you an email if I have created a branch
17:03 ttbot Parrot cfc6507f i386-linux-thread-multi make error http://tt.taptinder.org/cmdinfo/26972
17:04 cotto_work tewk: using INTVAL keys implies that the PMC is array-like, which it isn't
17:06 cotto_work It's a significant change in the semantics of an important PMC that seems to be unrelated to the the purpose of the branch.
17:06 tewk The PMC name is Hash, pmclass Hash provides hash
17:06 PacoLinux joined #parrot
17:06 cotto_work hard to disagree with that
17:07 tewk It doesn't change the semantics at all, it just saves you from having to box ints when you want to use them as keys
17:07 tewk In Perl you can do $myhash{1}
17:07 mtk left #parrot
17:07 tewk Hash already has set{integer,string}_keyed_int, we should be complete and add set_pmc_keyed_int and set_number_keyed_int
17:08 cotto_work cotto--
17:08 cotto_work indeed it does
17:09 cotto_work you're just making it more consistent
17:09 cotto_work wfm
17:10 cotto_work The Select PMC's functions need documenting.
17:10 tewk To tell the truth I originally boxed them myself, reached inside the Hash PMC to pull out the HASH* struct and do the add, but that breaks encapsilation, terribly.
17:10 cotto_work quite
17:12 tewk Yeah I missed those, will fix, timeout should probably be a NUM not an INTVAL.
17:12 whiteknight tewk++ on the Select PMC
17:13 tewk patches and commits welcome.
17:13 tewk Feel free to commit to my branch
17:13 cotto_work tewk: why do you have separate ATTRs for (rwe)b and (rwe)_array?
17:13 whiteknight I have too many branches of my own that need working on
17:14 jsut joined #parrot
17:15 mtk joined #parrot
17:15 cotto_work handling of those strikes me as odd
17:16 tewk cotto_work: select mutates the passed in bit vectors
17:17 Coke I thought lolcode was already converted to nqp-rx
17:18 cotto_work wfm
17:18 tewk Hmm, let me remember why I did that.
17:18 Coke ... and why not just convert it to nqp at this point?
17:18 tadzik Coke: what do you mean?
17:18 cotto_work Coke: it wouldn't be a bad learning experience for someone with the tuits.
17:18 tadzik It runs on nqp, the old nqp
17:19 Coke tadzik: what do you mean, what do I mean?
17:19 cotto_work recursion depth exceeded
17:19 Coke ok. there is a NEW nqp, called "nqp", that is replacing nqp-rx.
17:19 tadzik ah, this one
17:19 jsut_ left #parrot
17:19 Coke the old "nqp" should probably be called "nqp-old" at this point.
17:19 cotto_work +1
17:20 tadzik right. But it's not there yet. And the transition from nqprx to "new shiny nqp" is supposed to be less painful than the last time
17:20 cotto_work it'll be a fertile source of confusion otherwise
17:20 tadzik or we can stick to PGE describing "what's old and bad"
17:20 bluescreen left #parrot
17:21 Coke pge does not necessarily imply old nqp.
17:21 Coke partcl (old) uses pge/tge, e.g.
17:21 * cotto_work tries to remember what he was doing.  It seems like there was something about a dayjob.
17:24 gerd left #parrot
17:35 bluescreen joined #parrot
17:49 dalek parrot/tewk/select: 6b83661 | tewk++ | src/pmc/hash.pmc:
17:49 dalek parrot/tewk/select: Hash pmc gets set_pmc_keyed_int and delete_keyed_int
17:49 dalek parrot/tewk/select: review: https://github.com/parrot/parrot/commit/6b836618b7
17:49 dalek parrot/tewk/select: 7f96dd7 | tewk++ | / (9 files):
17:49 dalek parrot/tewk/select: Select PMC
17:49 dalek parrot/tewk/select: review: https://github.com/parrot/parrot/commit/7f96dd7227
17:53 * cotto_work is confused
18:11 whiteknight what are you confused about?
18:15 ShaneC joined #parrot
18:18 cotto_work why tewk's commits seem to have been applied twice
18:18 cotto_work er, pushed
18:19 dukeleto he may have done a force push
18:21 rurban left #parrot
18:23 dalek winxed: r822 | NotFound++ | trunk/winxedst1.winxed:
18:23 dalek winxed: change handling of volatile lexicals
18:23 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=822
18:31 ColloquyUser joined #parrot
18:35 dalek parrot: 9a2e3a7 | fperrad++ | runtime/parrot/library/Archive/Tar.pir:
18:35 dalek parrot: [Tar] specify encoding
18:35 dalek parrot: review: https://github.com/parrot/parrot/commit/9a2e3a758d
18:35 dalek parrot: c51cc9f | fperrad++ | t/dynpmc/ (2 files):
18:35 dalek parrot: [t] typo
18:35 dalek parrot: review: https://github.com/parrot/parrot/commit/c51cc9f63a
18:40 NotFound left #parrot
18:40 NotFound joined #parrot
18:41 cotto_work tewk: what's no_ro in Select for?
18:44 tewk GLOBAL_PARAM_W_ARITY("thread"                        , sch_thread            , "thread"                        , 1 , 1)  , env);
18:44 tewk opps
18:45 tewk What is no_ro?  no read-only?
18:45 tewk It may be a cut and paste error
18:50 tewk looks like its from the old STM impl
18:51 tewk gone!
18:55 dukeleto tewk: i have some non-linux machines to try your Select PMC branch on, would you like that?
18:56 dukeleto tewk: i can try it on NetBSD and FreeBSD. I have an openbsd box to test on as well, but it doesn't seem to have enough memory to compile parrot.
18:56 tewk dukeleto: sounds great.
18:57 cotto_work tewk: instead of having separate ATTRs for the fd_set and the fd_set*, why not just get the pointer as needed?
18:57 ColloquyUser left #parrot
18:58 tewk cotto_work: Yeah, I had the same, thought, I'm not sure why I did it that way.  This code is probably over a year old.  I resurrected it.
18:58 cotto_work tewk: ok.  I thought you'd just written it.
18:58 dalek parrot/tewk/select: 7e45688 | tewk++ | src/pmc/select.pmc:
18:58 dalek parrot/tewk/select: Removed no_ro, whitespace fixes
18:58 dalek parrot/tewk/select: review: https://github.com/parrot/parrot/commit/7e456880c1
19:02 estrabd_ joined #parrot
19:02 estrabd left #parrot
19:04 cotto_work github++
19:05 ColloquyUser joined #parrot
19:08 dalek parrot/tewk/select: 62d9b6d | tewk++ | src/pmc/select.pmc:
19:08 dalek parrot/tewk/select: Removed unneeded fd_set *r_ attributes from Select PMC
19:08 dalek parrot/tewk/select: review: https://github.com/parrot/parrot/commit/62d9b6d48d
19:08 bluescreen left #parrot
19:09 cotto_work tewk++
19:09 cotto_work much cleaner
19:14 dalek parrot/whiteknight/imcc_compreg_pmc: 4b647be | Whiteknight++ | / (4 files):
19:14 dalek parrot/whiteknight/imcc_compreg_pmc: clean up a mismatch in the IMCC API that created some harmless, but humorously wrong code in the parrot frontend. Also, fix t/src/embed/api.t
19:14 dalek parrot/whiteknight/imcc_compreg_pmc: review: https://github.com/parrot/parrot/commit/4b647be298
19:14 whiteknight down to three failing tests
19:15 dukeleto "humorously wrong code" is the best kind of wrong code
19:15 whiteknight t/pmc/io_stdin.t segfaults, but I don't know why and I don't understand the perl-side of the test enough to debug it. t/src/checkdepend.t hates me and I hate it, and t/perl/Parrot-Test.t, which I also don't understand
19:16 Andy left #parrot
19:19 Andy joined #parrot
19:20 dukeleto whiteknight: interesting. If you show me the outputs of the failing/coredumping tests, I will try to help you debug
19:20 dukeleto whiteknight: which perl bits of those tests have you confused?
19:20 nopaste "Whiteknight" at 192.168.1.3 pasted "test summary for dukeleto++" (8 lines) at http://nopaste.snit.ch/33637
19:21 whiteknight The three tests in t/pmc/io_stdin.t all segfault. When I take the PIR code out into a separate file and run it directly with the same inputs, no segfault
19:21 mikehh left #parrot
19:21 mikehh joined #parrot
19:22 whiteknight so I have no idea what the discrepancy there is, but I suspect some weird command-line parameters are being passed from the perl-based test functions to the parrot executable
19:23 dukeleto whiteknight: if you set POSTMORTEM=1 in your shell, the tests will leave behind any binaries/junk they create as part of the test
19:23 dukeleto whiteknight: that way, you can gdb/strace the binaries/coredumps that come out
19:23 dukeleto whiteknight: prove -v output would be more useful
19:24 whiteknight dukeleto: t/pmc/io_stdin.t doesn't seem to respond to POSTMORTEM. It doesn't output any binaries or other junk
19:24 whiteknight doesn't even output the .pir files to a findable location
19:25 whiteknight prove -v doesn't output anything that "perl t/pmc/io_stdin.t" doesn't
19:26 dukeleto whiteknight: interesting, i haven't seen this "pir_stdin_output_is" test function before
19:27 whiteknight exactly. And it does not appear to be as well-behaved as other test functions
19:27 dukeleto whiteknight: yes, POSTMORTEM is only for c_output_is and friendly
19:27 dukeleto s/friendly/friends/
19:27 cotto_work tewk: let me know when you take a break from your branch.  I'd like to do some cleanups.
19:30 dukeleto whiteknight: i am checking out your compreg branch and building it now, so I can run the test and attempt to give you some help
19:31 cotto_work #ps in 57
19:34 dukeleto whiteknight: running gdb on the core file gives us a clue
19:35 dukeleto whiteknight: https://gist.github.com/839219
19:35 tewk cotto_work: I'm done working on it.
19:35 dukeleto whiteknight: the badness is happening in parseflags
19:36 dukeleto whiteknight: hope that helps
19:36 dalek parrot: 6bfd0d1 | NotFound++ | t/pmc/fixedpmcarray.t:
19:36 dalek parrot: one more test for FPA, reaching 100% coverage
19:36 dalek parrot: review: https://github.com/parrot/parrot/commit/6bfd0d1505
19:36 cotto_work tewk: cool
19:36 dukeleto whiteknight: also, here is the full output for prove -v on my system : https://gist.github.com/839221
19:37 dukeleto whiteknight: let me know if you need other testing
19:38 lucian_ joined #parrot
19:39 whiteknight dukeleto: how did you run gdb on that?
19:40 dukeleto whiteknight: gdb parrot core
19:40 dukeleto whiteknight: the core file gets left in the root of the parrot.git repo
19:40 dukeleto whiteknight: and only one is left behind, so that is only the core file for the last test
19:41 dukeleto whiteknight: since each test's core file gets overwritten by the next coredump
19:41 ColloquyUser_ joined #parrot
19:41 ColloquyUser left #parrot
19:41 ColloquyUser_ is now known as ColloquyUser
19:41 whiteknight dukeleto: "gdb parrot core" doesn't do anything
19:42 whiteknight it just opens up gdb like normal and doesn't create a core file
19:42 lucian left #parrot
19:45 ColloquyUser_ joined #parrot
19:46 ColloquyUser left #parrot
19:46 ColloquyUser_ is now known as ColloquyUser
19:46 whiteknight ah nevermind, I wasn't dumping core
19:50 dalek parrot/whiteknight/imcc_compreg_pmc: 4e6b125 | Whiteknight++ | frontend/parrot/main.c:
19:50 dalek parrot/whiteknight/imcc_compreg_pmc: test a pointer before dereferencing it. this fixes t/pmc/io_stdin.t. dukeleto++
19:50 dalek parrot/whiteknight/imcc_compreg_pmc: review: https://github.com/parrot/parrot/commit/4e6b1252f5
19:53 whiteknight two test failures. ah ah ah
20:01 cognominal joined #parrot
20:13 nopaste "Whiteknight" at 192.168.1.3 pasted "checkdepend.t output on imcc for....(anybody who is interested)++" (172 lines) at http://nopaste.snit.ch/33643
20:16 cotto_work aloha: whiteknight?
20:16 aloha cotto_work: No clue. Sorry.
20:16 cotto_work aloha: whiteknight is The Count
20:16 aloha cotto_work: Okay.
20:17 Coke whiteknight: is that in your imcc branch?
20:18 bacek ~~
20:18 bacek Good morning, humans.
20:18 cotto_work HELO BACEK
20:18 dalek parrot/opsc_full_parse: eea3dcc | bacek++ | compilers/opsc/src/Ops/Compiler/Actions.pm:
20:18 dalek parrot/opsc_full_parse: Start building new PAST tree:
20:18 dalek parrot/opsc_full_parse:
20:18 dalek parrot/opsc_full_parse: 1. C<blockoid>
20:18 dalek parrot/opsc_full_parse: 2. C<declarator>
20:18 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/eea3dcc818
20:18 dalek parrot/opsc_full_parse: 4cfe233 | bacek++ | t/compilers/opsc/03-past-declarator.t:
20:18 dalek parrot/opsc_full_parse: Add more test for PAST generation.
20:18 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/4cfe233a9f
20:18 bacek EHLO cotto_work
20:19 donaldh joined #parrot
20:19 Coke so why do the gengc changes show no improvement on isparrotfastyet?
20:20 atrodo Coke> Initial looking seems to indicate bad test choices
20:21 Coke what is "rakudo" there mean. the build?
20:22 atrodo Startup.  rakudo -e ''
20:23 * jnthn is sure he heard somebody on #perl6 mention startup was faster with gengc.
20:23 Coke if someone comes up with a benchmark that does show improvment, how hard would it be to get the old data for those graphs?
20:23 Coke also, why are are we only showing a week's worth of data on that page?
20:23 dukeleto whiteknight: i think checkdepend.t may just be incorrect for your branch
20:23 dukeleto Coke: "we" is atrodo++
20:23 atrodo Not hard
20:24 Coke dukeleto: yes, I'm using the royal we.
20:24 atrodo And it's only a week, since I just chose 20 as a number to use
20:24 Coke dukeleto: how can checkdepend.t be "wrong" for the branch?
20:24 Coke Far more likely that whiteknight moved things about and didn't update dependencies.
20:25 dukeleto Coke: because it has a bunch of names of things hard-coded into it, and if the structure of files and deps have changed, then checkdepend probably needs to be updated
20:26 tcurtis joined #parrot
20:26 Coke dukeleto: what names are hardcoded?
20:26 Coke I only see:       qw/src compilers include frontend/ );
20:27 Coke (for directories in which to check for .c and .h files)
20:29 Coke given his nopaste, it's far more likely that, assuming he updated the .c/.h files, we just need to add the rules to the makefile.
20:31 mikehh #ps time
20:32 Coke I'll have that branch checked out; I'll see if I can fix the deps up this evening if no one beats me to it.
20:37 jan left #parrot
20:43 whiteknight I moved a bunch of files and changed a bunch of dependencies
20:43 dalek parrot/opsc_full_parse: 5209840 | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
20:43 dalek parrot/opsc_full_parse: Tweak grammar a bit to avoid splitting function call into
20:43 dalek parrot/opsc_full_parse: type_declarator and postcircumfix().
20:43 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/5209840b70
20:43 dalek parrot/opsc_full_parse: f34ceb1 | bacek++ | compilers/opsc/src/Ops/Compiler/Actions.pm:
20:43 dalek parrot/opsc_full_parse: Add more PAST.
20:43 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/f34ceb13cf
20:43 dalek parrot/opsc_full_parse: 3ce3c61 | bacek++ | t/compilers/opsc/03-past.t:
20:43 whiteknight now I need to slog through and fix the test
20:43 dalek parrot/opsc_full_parse: Remove misleading comment in test
20:43 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/3ce3c61bc5
20:43 dalek parrot/opsc_full_parse: 342b755 | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
20:43 dalek parrot/opsc_full_parse: Add infix:,
20:43 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/342b755c91
20:43 dalek parrot/opsc_full_parse: a010f2f | bacek++ | compilers/opsc/src/Ops/Compiler/ (2 files):
20:43 dalek parrot/opsc_full_parse: Rename term:sym<name> to sym<call> and implement pasting
20:43 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/a010f2f3f2
20:45 whiteknight Are we going to turn opsc into a full C parser?
20:45 cotto_work whiteknight: only a subset of C
20:45 whiteknight do ops only use a subset of C?
20:46 bacek whiteknight, yes. And parser already implemented :)
20:46 * cotto_work boggles
20:46 jan joined #parrot
20:47 Tene cotto_work: my understanding is that we're going to migrate to a subset of C that maps to things implementable in the M0 ops, approximately.
20:48 cotto_work Tene: sure.  I'm just boggling at bacek++'s eerie speed.
20:49 cognominal left #parrot
20:50 Tene I've always felt just a little bit that "magical coding robot" is a bit dismissive, and doesn't give enough credit to the hard work bacek puts in now, and has put in in the past to develop those skills.
20:50 Tene Could be I'm just a bit oversensitive or something.  He's certainly never objected that I've seen.
20:51 whiteknight cotto_work: three words for you: Magical. Coding. Robot
20:51 donaldh there was a C grammar written for parsing headers into NCI sigs. What happened to that?
20:51 NotFound whiteknight: we miss you at #ps
20:51 bacek Tene, "coding robot" is just a joke. I don't mind at all :)
20:51 whiteknight damnit.
20:51 whiteknight Tene; it's a metaphor
20:51 tewk donaldh: I think its in the parrot github c99 repo
20:52 Tene Okay.  That makes sense.  Thanks for the explanation.  :)
20:52 donaldh tewk: ah thanks.
20:52 whiteknight Tene: Calling him "our very special, super-speedy, fantastic coder guy" is too long-winded
20:53 Tene whiteknight: I'm not surprised that I was missing some social context.  That's very reasonable.
20:53 dalek parrot/opsc_full_parse: fe1f209 | bacek++ | compilers/opsc/src/Ops/Compiler/Grammar.pm:
20:53 dalek parrot/opsc_full_parse: Remove infix:,. It broke parser.
20:53 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/fe1f209a5c
20:59 bluescreen joined #parrot
21:04 dalek TT #2028 created by NotFound++: pcre is deprecated
21:04 dalek TT #2028: http://trac.parrot.org/parrot/ticket/2028
21:05 bacek afk # $dayjob
21:05 bluescreen left #parrot
21:13 dalek parrot/opsc_full_parse: bba2c2b | bacek++ | compilers/opsc/src/Ops/Compiler/Actions.pm:
21:13 dalek parrot/opsc_full_parse: Handle constants
21:13 dalek parrot/opsc_full_parse: review: https://github.com/parrot/parrot/commit/bba2c2bd2c
21:16 lucian_ is now known as lucian
21:26 kid51 joined #parrot
21:30 whiteknight left #parrot
21:34 tadzik bacek: poke
21:41 cotto_work tewk: one test fails on the Select PMC.  Is the test broken or is the PMC?
21:42 dalek parrot/tewk/select: 71cb482 | cotto++ | src/pmc/select.pmc:
21:42 dalek parrot/tewk/select: code cleanup; no functional changes
21:42 dalek parrot/tewk/select: review: https://github.com/parrot/parrot/commit/71cb48245e
21:42 dalek parrot/tewk/select: 6535c49 | cotto++ | src/pmc/select.pmc:
21:42 dalek parrot/tewk/select: more code cleanup; still no functional changes
21:42 dalek parrot/tewk/select: review: https://github.com/parrot/parrot/commit/6535c497cb
21:42 dalek parrot/tewk/select: 24aff85 | cotto++ | / (2 files):
21:42 dalek parrot/tewk/select: rename data_hash to fd_map, update METHODs and tests
21:42 dalek parrot/tewk/select: review: https://github.com/parrot/parrot/commit/24aff85d00
21:42 dalek parrot/tewk/select: c373c86 | cotto++ | src/pmc/select.pmc:
21:42 dalek parrot/tewk/select: ATTR alloc/free isn't needed with auto_attrs
21:42 dalek parrot/tewk/select: review: https://github.com/parrot/parrot/commit/c373c86408
21:42 cotto_work (I confirmed that the test failed before those changes.)
21:47 NotFound Could not upload report to Smolder at http://smolder.parrot.org
21:47 NotFound HTTP CODE: 500 (Internal Server Error)
21:52 bacek_at_work ~~
21:52 bacek_at_work tadzik, pong
21:52 bacek_at_work cotto_work, we don't need manual WB in Select PMC
21:54 tadzik bacek_at_work: might looking at that YAML::Tiny problem of mine?
21:54 cotto_work bacek_at_work: ok
21:54 tadzik https://github.com/tadzik/p​arrot-deprecation-detector
21:55 bacek_at_work tadzik, can you give me commit bit?
21:56 kid51 left #parrot
21:56 ColloquyUser left #parrot
21:57 tadzik bacek_at_work: sure thing
21:58 tadzik bacek_at_work: you've got it
21:58 bacek_at_work tadzik, ok.
21:58 bacek_at_work tadzik, are you fighting with Null PMC access in find_method('new')?
21:59 tadzik bacek_at_work: aye
21:59 tadzik bacek_at_work: the error has recently moved to YAML::Tiny, hence I'm asking you
22:00 * dukeleto hacked the Github Organization Gibson
22:00 dukeleto i figured out how to make more teams and give admin bits to specific repos
22:00 dukeleto so I created  "HLL developers" and "HLL owners" teams
22:01 NotFound It's a Gibson Les Paul?
22:01 dukeleto "HLL Developers" have a commit bit to everything but parrot.git (can be tweaked).
22:01 bacek_at_work tadzik, It couldn't create Regex. Looks like something isn't loading properly.
22:02 dukeleto i may have to rethink "HLL Owners". I think we will need a team for each HLL
22:02 NotFound dukeleto++
22:02 cotto_work dukeleto: no sense in adding complexity prematurely
22:02 tadzik bacek_at_work: what bytecode should I load? Or is YAML's fault?
22:02 dukeleto cotto_work: well, the thing is, that "HLL owners" will have admin bits to every HLL in the parrot org, instead of, for instance, only Cardinal owners having an admin bit to Cardinal
22:03 bacek_at_work tadzik, -load_bytecode P6object; + load_bytecode "nqp-rx.pbc"
22:03 dukeleto cotto_work: currently Cardinal is the only thing in HLL owners, so it doesn't matter
22:03 tewk cotto_work: They all pass for me, which one?
22:03 dukeleto cotto_work: but once we add another, we can fiddle with the structure
22:03 tewk cotto_work: opps I need to pull
22:03 dukeleto cotto_work: if i understand correctly, we can give commit bits to the "HLL dev" team without a CLA, correct ?
22:04 NotFound It may make sense only for languages with very active groups of developers.
22:04 cotto_work bacek_at_work: what do you mean that wb isn't needed?
22:04 tadzik bacek_at_work: alright, now the errors are on my side. Thanks :)
22:04 Tene dukeleto: I'm still a member of the parrot devs team though, right?  You haven't finally kicked me out?  ;)
22:04 dukeleto Tene: yes, of course :)
22:04 cotto_work dukeleto: If I'm the person to ask my answer is "yes", but I'm not sure I am.
22:05 bacek_at_work tadzik, now figure out regexes problem :)
22:05 bacek_at_work tadzik, loading is fixed
22:05 nopaste "cotto_work" at 192.168.1.3 pasted "test failure on select pmc" (26 lines) at http://nopaste.snit.ch/33647
22:05 cotto_work tewk: ^
22:06 ColloquyUser joined #parrot
22:07 tadzik bacek_at_work: oh, was just going to ask you about 21d6178d :)
22:17 dalek parrot/tewk/select: cd8ab92 | tewk++ | / (2 files):
22:17 dalek parrot/tewk/select: Fix getmaxfd test
22:17 dalek parrot/tewk/select: review: https://github.com/parrot/parrot/commit/cd8ab9292d
22:19 cosimo joined #parrot
22:21 bacek_at_work tadzik, all fixed. Time to fix regexes in api.yaml. They are Perl6, not Perl5.
22:24 tadzik hey, I wanted to fix that :)
22:24 tadzik bacek_at_work++
22:26 Coke NotFound: if a language has a very active group of developers, it probably doesn't to be in /parrot/
22:36 gbacon left #parrot
22:38 bacek_at_work tadzik, / $regex / doesn't work
22:38 tadzik yeah, seems so
22:39 tadzik plus there's a newline at the end of a string
22:39 dalek parrot: 9405706 | (Gerd Pokorra)++ | NEWS:
22:39 dalek parrot: add a news
22:39 dalek parrot: review: https://github.com/parrot/parrot/commit/940570612e
22:39 dalek parrot/whiteknight/imcc_compreg_pmc: 0abb532 | Coke++ | co (2 files):
22:39 dalek parrot/whiteknight/imcc_compreg_pmc: Fix several build dependencies
22:39 dalek parrot/whiteknight/imcc_compreg_pmc: review: https://github.com/parrot/parrot/commit/0abb532781
22:48 lucian left #parrot
22:52 dalek parrot/whiteknight/imcc_compreg_pmc: 84de9ec | Coke++ | MANIFEST:
22:53 dalek parrot/whiteknight/imcc_compreg_pmc: Regenerate the MANIFEST
22:53 dalek parrot/whiteknight/imcc_compreg_pmc: review: https://github.com/parrot/parrot/commit/84de9ece69
22:54 Coke msg whiteknight I think the remaining failure in t/src/checkdepend.t on your branch is due to a bug in config/auto/pmc.c that does not properly track includes of includes to get all the dependencies for a PMC - you just happened to be the first PMC to include something that included something else.
22:54 aloha OK. I'll deliver the message.
23:00 bacek_at_work tadzik, git pull; make run
23:08 Andy left #parrot
23:20 whiteknight joined #parrot
23:23 fperrad left #parrot
23:24 kid51 joined #parrot
23:32 whiteknight hello, #parrot
23:34 whiteknight Coke++ # Thanks for looking at that test
23:38 dalek parrot: 1a92ce7 | (Gerd Pokorra)++ | NEWS:
23:38 dalek parrot: add more news
23:38 dalek parrot: review: https://github.com/parrot/parrot/commit/1a92ce7eef
23:39 ColloquyUser left #parrot
23:40 ColloquyUser joined #parrot
23:41 ColloquyUser left #parrot
23:47 donaldh left #parrot

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

Parrot | source cross referenced