Camelia, the Perl 6 bug

IRC log for #parrot, 2011-04-15

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 bbatha joined #parrot
00:00 bacek_at_work aloha, msg bacek_at_work ping
00:00 aloha joined #parrot
00:00 bacek_at_work aloha, msg bacek_at_work ping
00:00 aloha bacek_at_work: OK. I'll deliver the message.
00:00 bacek_at_work ~~
00:00 bacek_at_work aloha, msg whiteknight test
00:00 aloha bacek_at_work: OK. I'll deliver the message.
00:01 whiteknight ...?
00:01 whiteknight got it
00:01 aloha left #parrot
00:01 whiteknight ...lolwat?
00:01 aloha joined #parrot
00:03 bbatha_ joined #parrot
00:03 bbatha left #parrot
00:03 bbatha_ is now known as bbatha
00:04 whiteknight has aloha undergone any major upgrades lately
00:04 whiteknight ?
00:05 whiteknight no commits to parrot master. I take that to mean that nobody has magically resolved our gc heisenbug
00:09 bbatha left #parrot
00:15 cosimo joined #parrot
00:18 bbatha joined #parrot
00:32 bbatha left #parrot
00:32 whiteknight this bug is killing me
00:34 mikehh whiteknight: I re-ran teh rakudo test with my g++ build and got exactly the same result
00:35 whiteknight mikehh: Yeah, there's definitely some kind of algorithmic problem
00:36 whiteknight plobsing: ping
00:37 mikehh whiteknight: I don't know if it helps - rakudo smoke test #14656 and now #14672 - parrot smoke test #14655
00:37 whiteknight my current theory is that we aren't setting write barriers on the PtrObj PMC enough. the contents of the packfile are not getting promoted to the older generation, and are therefore getting marked/swept more often
00:39 mikehh whiteknight: my tests with gcc --gc=gms seem a bit random, but g++ was consistent
00:40 bbatha joined #parrot
00:43 mikehh anyway need to do some testing on i386 - bbiab
00:49 whiteknight the core.pm build takes FOREVER
00:50 whiteknight even if it took the same amount of time, compiling all the .pm files individually and using pbcmerge would at least have more visible progress movement
00:50 Coke can't do that due to limitations in parrot, as I recall.
00:51 Coke the setting needs to refer to other parts of itself when compiling - compile in chunks, you don't end up with the same logic.
00:53 ShaneC left #parrot
00:54 ShaneC joined #parrot
00:55 whiteknight blah, I can dream
00:57 whiteknight ....okay, my next attemped fix didn't work either
01:05 bbatha left #parrot
01:05 bbatha_ joined #parrot
01:17 whiteknight damnit. Still no love
01:18 whiteknight all the easy-looking places to stick writebarriers don't fix it
01:19 whiteknight and it's going to require some non-trivial changes to jam in more write barriers in other places
01:20 whiteknight this wouldn't be so frustrating if the rakudo build didn't take ~10 minutes
01:22 bbatha joined #parrot
01:22 bacek_at_work whiteknight, how you build parrot? --optimize --gc=gms?
01:22 whiteknight no --optimize. Just --gc=gms
01:22 bacek_at_work ookey
01:22 whiteknight that seems to make the failures happen most consistently, for me at least
01:22 benabik whiteknight: Where does it fail?  Build or spectest?
01:23 whiteknight benabik: for me, spectest
01:23 bbatha_ left #parrot
01:23 whiteknight I haven't seen any build failures in a while
01:23 benabik whiteknight: Blech.  That takes literally hours on my machine.  :-(
01:23 benabik Or seems to.
01:23 mikehh left #parrot
01:23 bacek_at_work whiteknight, try to compile with -DPARROT_GC_VALIDATE
01:23 whiteknight benabik: I don't run the entire spectest. I only need to run the first one or two. All spectests segfault
01:24 benabik whiteknight: Ah.  That's simple enough to test for.
01:24 whiteknight bacek_at_work: okay, how to specify that on the configure? --ccflags="-DPARROT_GC_VALIDATE"?
01:24 bacek_at_work whiteknight, yes
01:25 bacek_at_work hmm. Is it gcc or g++ build?
01:25 bbatha left #parrot
01:25 whiteknight mikehh is using g++. I am using clang
01:25 woosley joined #parrot
01:30 rurban joined #parrot
01:30 benabik How do you compile with g++?  Just --cc=g++?
01:31 bacek_at_work benabik, --link=g++
01:35 JimmyZ left #parrot
01:37 JimmyZ joined #parrot
01:40 bacek_at_work whiteknight, Parrot_pf_get_packfile_pmc
01:40 whiteknight what about it?
01:40 bacek_at_work looks like it doesn't work properly...
01:40 whiteknight I added a write barrier there on my machine, but that didn't fix the problem
01:40 whiteknight why not?
01:41 bacek_at_work hmm
01:41 bacek_at_work let me dig bit more.
01:42 whiteknight PARROT_GC_VALIDATE didn't give me any information
01:43 whiteknight still fails the same assertion at the same place
01:44 whiteknight okay, I'm going to bed. I'll look at this again tomorrow and add in some more write barriers
01:45 whiteknight thanks for looking at it bacek++
01:45 whiteknight left #parrot
01:48 S_Arms left #parrot
01:48 S_Arms joined #parrot
01:50 dmalcolm left #parrot
02:00 bubaflub left #parrot
02:02 Coke left #parrot
02:02 Coke joined #parrot
02:04 S_Arms left #parrot
02:04 S_Arms joined #parrot
02:05 mikehh joined #parrot
02:12 benabik Well, no segfaults with g++, gms, non-opt.  Ah, well.
02:12 JimmyZ left #parrot
02:31 Coke left #parrot
02:31 Coke joined #parrot
02:35 ShaneC left #parrot
02:47 ShaneC joined #parrot
02:49 mtk left #parrot
02:56 mtk joined #parrot
02:59 mikehh rakudo - got a whole bunch of segfaults spectest_smolder #14694 running with same g++ build on i386 as amd64
03:00 mikehh g++ --optimize --gc=gms --buildframes --inline
03:01 * mikehh had enough for the moment - must sleep (after 4am here)
03:11 hudnix left #parrot
03:55 petdance joined #parrot
03:57 cotto ~
03:57 Eduardow left #parrot
04:37 petdance left #parrot
04:38 bacek_at_work ~~
04:38 bacek_at_work I can reproduce assert failure in t/compilers/pge
04:38 bacek_at_work Some PackFile vs GC shenanigans
04:38 cotto great
04:39 cotto seen plobsing
04:39 aloha plobsing was last seen in #parrot 7 hours 56 mins ago saying "oh no. self-awareness.".
05:10 sorear_ is now known as sorear
05:25 cotto Has anyone here done international travel for a conference?
05:28 sorear I know jnthn has
05:29 sorear also TimToady
05:30 sorear moritz and tadzik will soon if they haven't already
05:31 sorear cotto: ^^
05:31 cotto ok
05:43 cotto allison, ping
05:49 dukeleto cotto: yes, i have
05:49 dukeleto cotto: what do you want to know?
05:51 cotto dukeleto, is there anything surprising about traveling to a conference on another continent?
05:53 cotto how much should I have between arrival and the start of the conf, and end of conf and departure?
05:58 ShaneC left #parrot
06:00 TimToady we usually prefer to arrive a day early and walk around in the sun, which seems to help the jet lag
06:00 TimToady also gives you a cushion if the flight is delayed
06:01 TimToady coming back usually depends more on whether there's a hackathon or we want to do some sightseeing
06:04 cotto TimToady, thanks.  That's helpful.
06:10 dukeleto cotto: where are you going?
06:17 dalek parrot/leto/embed_grant: 84871f5 | dukeleto++ | t/src/extend_vtable.t:
06:17 dalek parrot/leto/embed_grant: [t] Parrot_PMC_delete_keyed(_int)
06:17 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/84871f57f4
06:18 theory left #parrot
06:21 fperrad joined #parrot
06:24 cotto dukeleto, yapc::eu
06:24 dukeleto cotto: very awesome
06:25 cotto Yeah.  I'm looking forward to it.
06:25 dukeleto cotto: arriving at least 2 hours before international flights is usually a good idea. Sometimes 3. Depends on how busy the airport will be
06:26 cotto I suppose I should get my passport.
06:26 dukeleto cotto: that would be a good idea
06:29 cotto apparently my old one only expired a year ago
06:29 dalek parrot/leto/embed_grant: 4d52021 | dukeleto++ | t/src/extend_vtable.t:
06:29 dalek parrot/leto/embed_grant: [t] Parrot_PMC_delete_keyed_str
06:29 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/4d52021be8
06:32 dalek parrot/leto/embed_grant: 67ffbc8 | dukeleto++ | t/src/extend_vtable.t:
06:32 dalek parrot/leto/embed_grant: [t] Parrot_PMC_delprop
06:32 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/67ffbc8ae5
06:33 birdwindupbird joined #parrot
06:35 dukeleto cotto: # Exception is: type 36 severity 2 message 'exists_keyed_str() not implemented in class 'ResizablePMCArray''
06:35 dukeleto cotto: bug or feature?
06:35 dukeleto cotto: i guess arrays should only allow integer indexing
06:36 * dukeleto votes features
06:36 cotto feature
06:37 dukeleto cotto: is there going to be hackathoning at yapc::eu ?
06:37 cotto dukeleto, yup
06:37 dalek parrot/leto/embed_grant: eab67bd | dukeleto++ | t/src/extend_vtable.t:
06:37 dalek parrot/leto/embed_grant: [t] Parrot_PMC_exists_keyed_str
06:37 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/eab67bd42e
06:37 cotto http://yapceurope.lv/ye2011/
06:38 cotto It might be fun to stick around and get schooled in Perl 6 by jnthn.
06:38 dukeleto cotto: sounds like good times
06:39 dalek parrot/leto/embed_grant: de2f736 | dukeleto++ | t/src/extend_vtable.t:
06:39 dalek parrot/leto/embed_grant: [t] Parrot_PMC_get_integer_keyed_str
06:39 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/de2f736680
06:40 dalek parrot/leto/embed_grant: c6f1fe7 | dukeleto++ | t/src/extend_vtable.t:
06:40 dalek parrot/leto/embed_grant: [t] Parrot_PMC_get_number_keyed_str
06:40 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/c6f1fe741f
06:42 cotto I'm surprised by how not completely insane the site for renewing my passport is.
06:44 dukeleto cotto: things are looking up :
06:44 dukeleto :), rather
06:45 dukeleto # Exception is: type 5 severity 2 message 'Null PMC access in destroy()'
06:46 * dukeleto is getting pretty good at debugging odd vtable error messages. Not sure if that reflects well on my sanity.
06:47 UltraDM joined #parrot
06:47 cotto null pmc access in destroy
06:47 cotto it sounds like a Zen koan
06:48 dalek parrot/leto/embed_grant: 986d760 | dukeleto++ | t/src/extend_vtable.t:
06:48 dalek parrot/leto/embed_grant: [t] Parrot_PMC_get_pmc_keyed_int
06:48 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/986d7601bc
06:50 dalek parrot/leto/embed_grant: ce3f9ec | dukeleto++ | t/src/extend_vtable.t:
06:50 dalek parrot/leto/embed_grant: [t] Parrot_PMC_get_pmc_keyed
06:50 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/ce3f9ecdf3
06:54 dalek parrot/leto/embed_grant: d7afbd7 | dukeleto++ | t/src/extend_vtable.t:
06:54 dalek parrot/leto/embed_grant: [t] Parrot_PMC_get_repr
06:54 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/d7afbd7a60
06:55 cotto If a null fails to be destroyed, has it failed to be destroyed?
07:05 cotto dukeleto, any idea what a decent price would be and where I could find it?
07:12 ShaneC joined #parrot
07:18 Coke left #parrot
07:19 Coke joined #parrot
07:22 rurban_ joined #parrot
07:22 rurban left #parrot
07:22 rurban_ is now known as rurban
07:24 UltraDM left #parrot
07:31 UltraDM joined #parrot
07:44 SHODAN joined #parrot
08:04 jrtayloriv left #parrot
08:15 cosimo left #parrot
08:36 aloha left #parrot
08:37 aloha joined #parrot
08:50 UltraDM left #parrot
08:58 woosley left #parrot
08:59 UltraDM joined #parrot
09:15 ShaneC left #parrot
09:22 bacek joined #parrot
09:28 bacek ~~
09:29 bacek we need some bloody "GC guy" to fix rakudo's segfault
09:29 bacek oh wait...
09:32 birdwindupbird left #parrot
09:32 birdwindupbird joined #parrot
09:54 contingencyplan left #parrot
10:00 ShaneC joined #parrot
10:01 bacek msg whiteknight Sorry mate, you have to unmerge your branch. Main reason - PtrObj's packfile gets updated behind the scene by Parrot_load_bytecode which pokes inside interp->initial_pf. Without switching guts to use Packfile PMCs it's almost impossible to fix it.
10:01 aloha OK. I'll deliver the message.
10:02 bacek msg whiteknight I'll try to comeup with some workaround but don't hold your breath.
10:02 aloha OK. I'll deliver the message.
10:05 davidfetter left #parrot
10:08 davidfetter joined #parrot
10:20 ShaneC left #parrot
10:23 dalek parrot/packfile_wrap: 2d1a1fe | bacek++ | include/parrot/interpreter.h:
10:23 dalek parrot/packfile_wrap: Brake the build - replace "PackFile *initial_pf" with "PMC *current_pf"
10:23 dalek parrot/packfile_wrap: review: https://github.com/parrot/parrot/commit/2d1a1fe09a
10:37 mtk left #parrot
10:44 mtk joined #parrot
10:54 dalek parrot/packfile_wrap: af010fa | bacek++ | src/gc/mark_sweep.c:
10:54 dalek parrot/packfile_wrap: Fix mark in trace_roots
10:54 dalek parrot/packfile_wrap: review: https://github.com/parrot/parrot/commit/af010fa3fa
10:54 dalek parrot/packfile_wrap: 62b5f53 | bacek++ | / (8 files):
10:54 dalek parrot/packfile_wrap: Mass replace of interp_pf with current_pf.
10:54 dalek parrot/packfile_wrap: review: https://github.com/parrot/parrot/commit/62b5f53d99
10:56 bacek moritz, ping
11:06 moritz bacek: pong
11:06 bacek moritz, can you test packfile_wrap branch with rakudo?
11:06 bacek GMS, optimize, etc
11:07 bacek moritz, and side question - how is your parenting going? :)
11:08 moritz bacek: I'm first testing with a commit that whiteknight++ has asked me to test, then will test the branch
11:09 bacek moritz, which commit?
11:09 moritz bacek: the past two nights have been quite exhausting, I'm staying at home today. Apart from that: very rewarding :-)
11:09 moritz bacek: 364ad6bc22
11:10 bacek moritz, this commit will not help.
11:11 bacek moritz, I still remember that time. First few month with my first kid was very tough. Second was much easier :)
11:13 moritz bacek: Ronja started to smile a lot about two weeks ago, that's just *SO* cute :-)
11:14 bacek moritz, wow!
11:14 moritz oh, and she's started "talking" ("goo", "brrr") :-)
11:15 bacek moritz, did you ever read Lindgren? "Ronja Rövardotter" (in Russian translation) was one of my favourite book in childhood :)
11:18 moritz bacek: I did :-)
11:18 moritz bacek: my wife is half Norwegian, so we looked for names that are known both in Scandinavia an Germany. That one won :-)
11:19 bacek moritz, should I call you Matias now? :)
11:19 moritz bacek: no, we don't disclose who the robber is :-)
11:20 bacek moritz, ok-ok. I'll keep this secret for ya! :)
11:24 tadzik hello #parrot
11:26 bacek aloha, tadzik
11:28 tadzik hello bacek
11:28 tadzik how are things?
11:30 bacek so-so
11:30 bacek Quite Friday night.
11:31 tadzik thank you
11:31 tadzik oh, ww
11:38 mikehh left #parrot
11:51 mikehh joined #parrot
11:56 Patterner left #parrot
11:56 Psyche^ joined #parrot
11:56 Psyche^ is now known as Patterner
12:35 whiteknight joined #parrot
12:36 whiteknight good morning, #parrot
12:36 whiteknight msg bacek I figured that was the problem. I've got an idea in mind to fix it. If I can't fix it today I will unmerge
12:36 aloha OK. I'll deliver the message.
12:37 bacek whiteknight, morning.
12:37 whiteknight hello bacek
12:37 whiteknight thanks for looking at the segfault last night. bacek++
12:37 bacek whiteknight, I did initial "workaround" in packfile_wrap branch.
12:37 whiteknight okay, I have to look at that branch
12:38 bacek It's kinda "what we should have at the end of the day"
12:38 whiteknight one way, we will have this issue fixed
12:38 whiteknight I just hope we can do it before the release
12:38 bacek whiteknight, unlikely.
12:38 whiteknight what functions modify the packfile?
12:39 bacek https://github.com/parrot/parrot/commit/62b5f53d99
12:39 bacek pf_set_packfile
12:39 bacek Called from quite few places
12:40 whiteknight okay, will the problem be fixed if we can add a write barrier to that function?
12:40 bacek and append_packfile
12:40 whiteknight I have an idea to add write barriers to those places
12:40 bacek whiteknight, yes. It will be fixed.
12:40 whiteknight okay, I think I can do that today
12:40 whiteknight I am going to add a ->parent_pmc to the PackFile* structure, or one of those structures. If that field is not null, we can write barrier
12:41 bacek Problem is - atm interp doesn't have pointer to PtrObj-holding-Packfile-ptr
12:41 ambs joined #parrot
12:41 bacek whiteknight, (parent_pmc) this is what I basically done in packfile_wrap branch
12:42 whiteknight bacek: in Parrot_pf_get_packfile_pmc, I register all PtrObj pmcs, so they should always be safe
12:42 whiteknight that's a temporary hack to mimic current behavior
12:42 bacek whiteknight, yes
12:42 whiteknight so all we should need are write barriers, and we should have this fixed. Unless I am missing something
12:42 bacek but PackFile* was updated behind the scene
12:43 whiteknight right, that's where the write barriers are needed
12:43 whiteknight unless there are other updates I am not aware of
12:43 bacek yes
12:43 bacek but in master there is no way to WB PMC-holding-PF
12:44 bacek because interp store raw pointer only
12:44 whiteknight even if we add a ->parent_pmc field to the packfile itself?
12:44 whiteknight the pmc will still be available, just one more indirection
12:45 bacek it's even more awful hack than packfile_wrap branch...
12:45 whiteknight Any fix would be temporary. We just need to get to Packfile PMCs everywhere
12:46 bacek I don't like it. It will probably work. But I really don't like it.
12:47 whiteknight i don't like it either, but we only need it to work until tuesday
12:47 whiteknight then we can start breaking things again
12:47 whiteknight if it turns into a really big mess we can unmerge. That's always plan B
12:50 bacek Yes...
12:50 whiteknight bacek: We can play today. Tomorrow we make the hard decision for the release. Wednesday we throw it all out and break things
12:51 bacek whiteknight, today is almost over for me
12:51 whiteknight bacek: my day is only just beginning. We'll give you an extra day to play if you want
12:51 whiteknight :)
12:51 bacek And I will not have time over weekend to hack parrot
12:51 bacek There is "plan C"
12:51 bacek Keep it as is.
12:52 whiteknight that's okay. You've been very helpful, and I will make sure it gets fixed
12:52 whiteknight bacek: plan C is not a good plan. Segfaults are not good.
12:52 bacek We'll have supported release with GC MS2
12:52 whiteknight this is true
12:52 whiteknight we can do plan C if everything else is too hard
12:52 bacek Put big deprecation notice like "everything is going to change"
12:53 bacek Rewrite/refactor/fix/whatever right after release
12:55 bacek At least we can deprecate exposure of "PackFile*" to embedders
12:55 bacek (src/embed.c)
12:55 bacek or whole "old embed API"
13:01 bacek anyway, it's bed time. Please hold release until-after #ps.
13:04 whiteknight goodnight bacek
13:07 moritz \o/ all rakudo spectests pass on packfile_wrap parrot branch
13:13 whiteknight see? that's the power of bacek
13:14 moritz I'm doing a second run, just to be sure
13:14 moritz and if that's clean, I'll try again without --optimize
13:14 whiteknight awesome
13:14 whiteknight I know the branch isn't complete yet
13:15 tadzik hmm, I can try it on this pc
13:15 tadzik what's desirable to test? gms, --optimize? It's 32bit linux
13:15 moritz gms
13:16 moritz with and without --optimize
13:16 bubaflub joined #parrot
13:16 tadzik --gen-parrot won't do, aye?
13:17 hudnix joined #parrot
13:17 moritz only if you bump PARROT_REVISION first
13:17 moritz (locally please :-)
13:17 tadzik wha should I put there?
13:17 tadzik (sure thing :))
13:17 Coke left #parrot
13:17 Coke joined #parrot
13:18 whiteknight the packfile_wrap branch looks very promising
13:18 whiteknight and it is the kind of change we want to make in the long run
13:18 moritz tadzik: try RELEASE_3_2_0-377-g62b5f53
13:18 tadzik that's this packfilw-wrap?
13:20 moritz yes
13:20 atrodo btw, the spectest choked my linode
13:20 atrodo w/gms and optimize
13:21 tadzik ok. Fetching parrot, this will take some time...
13:22 kaneplusplus joined #parrot
13:46 Eduardow joined #parrot
14:04 Andy_ joined #parrot
14:14 tadzik 1 failed test so far
14:15 tadzik bbs
14:34 UltraDM left #parrot
14:38 davidfetter left #parrot
15:04 S_Arms left #parrot
15:11 M_o_C joined #parrot
15:13 hercynium joined #parrot
15:20 rurban_ joined #parrot
15:22 rurban left #parrot
15:22 rurban_ is now known as rurban
15:24 benabik left #parrot
15:24 theory joined #parrot
15:25 benabik joined #parrot
15:27 benabik left #parrot
15:53 benabik joined #parrot
16:07 rurban left #parrot
16:08 benabik left #parrot
16:12 benabik joined #parrot
16:17 birdwindupbird left #parrot
16:20 tadzik rakudo spectests pass for me, 32bit linux, gms, --optimize
16:25 Eduardow left #parrot
16:36 cotto_work ~
16:37 tadzik ~~
16:39 benabik ~~~
16:41 M_o_C left #parrot
17:00 Eduardow joined #parrot
17:16 moritz bacek, whiteknight: all my rakudo tests on the packfile-wrap branch have been successful.
17:20 tadzik mine too
17:23 cotto_work bacek++
17:23 cotto_work moritz++
17:24 cotto_work whiteknight++
17:24 cotto_work this sounds promising
17:26 shell joined #parrot
17:27 davidfetter joined #parrot
17:28 whiteknight what sounds promising?
17:29 moritz the packfile-wrap branch
17:29 moritz passes all tests here
17:37 dmalcolm joined #parrot
17:37 whiteknight bacek seemed to think that the branch was still lacking or incomplete in some way
17:37 whiteknight I don't know what problems there may be.
17:39 shell left #parrot
17:39 whiteknight it definitely is a step in the right direction
17:40 whiteknight I was hoping that the imcc_compreg_pmc branch could be workable without having to take that next step, but apparently not
17:40 whiteknight it's a row of dominos that hasn't stopped falling yet
17:42 whiteknight msg bacek the packfile_wrap branch appears to be working and passing tests (including Rakudo spectests). What is it missing? Is there any reason why we couldn't merge that soon?
17:42 aloha OK. I'll deliver the message.
17:46 dodathome joined #parrot
17:57 whiteknight I suspect strongly that the problems with that branch involve that godforsaken Eval PMC
18:02 moritz oh, make fulltest says
18:02 dukeleto we still have an Eval PMC?
18:02 moritz t/src/extend.t       (Wstat: 1536 Tests: 18 Failed: 6) Failed tests:  10-13, 16-17 Non-zero exit status: 6
18:02 dukeleto Seems like a bad idea waiting to punch us in the face.
18:03 dukeleto moritz: can you gist the failing output? I might be able to help debug
18:04 whiteknight dukeleto: Yes. the imcc_compreg_pmc branch did a lot of work to make Eval unnecessary, but we still aren't at a point yet where we can excise it
18:05 moritz dukeleto: https://gist.github.com/922155
18:05 ShaneC joined #parrot
18:15 benabik left #parrot
18:16 dukeleto moritz: wow, those assertions have useless error messages
18:16 moritz dukeleto: not my fault :-)
18:17 dukeleto moritz: agreed. I am lamenting at our code :)
18:17 dukeleto moritz: which branch is that error output from?
18:18 dukeleto moritz: i don't see that assertion in the master branch anywhere
18:24 benabik joined #parrot
18:25 hudnix left #parrot
18:25 moritz dukeleto: packfile_wrap
18:29 dukeleto moritz: ah
18:31 cxreg left #parrot
18:33 * davidfetter waves to dukeleto
18:34 dukeleto davidfetter: wazzup
18:34 davidfetter does pl/parrot still work?
18:36 dukeleto davidfetter: define "work"
18:36 dukeleto davidfetter: PL/Parrot, last time I checked, "worked"
18:37 davidfetter as in still passes "make check" (or "make test" ,  whichever it was)
18:37 dukeleto davidfetter: PL/Perl6, "worked", except that in newer releases of Parrot+Rakudo (let's say the last 6ish months) certain tests cause postgres to coredump
18:37 dukeleto davidfetter: PL/Parrot should still pass "make test" (make check might need to be updated)
18:38 davidfetter cool
18:38 dukeleto davidfetter: PL/Perl6 will not. I needs to fix a bug that I haven't had the tuits to fix
18:38 davidfetter what's the bus number on that bug?
18:39 dukeleto davidfetter: 1 (me), although I described it best in the GCI task that I attempted to make an unlucky student work on, but they all were too smart to work on it :)
18:40 davidfetter dukeleto, apart from fooling some poor GSoC student into it, how might the bus number increase?
18:43 dukeleto davidfetter: i suppose I should make a github issue describing exactly what happens and what I suspect the bug is
18:43 davidfetter k
18:43 dukeleto davidfetter: do you want to try to fix that bug?
18:43 dukeleto davidfetter: i can be your cheerleader :)
18:44 * davidfetter 's coding & debugging skills may be up to it, but thinks his pointy-hairing skills may be a better fit
18:44 contingencyplan joined #parrot
18:46 dukeleto davidfetter: what pointy-haired things do you have up your sleeve?
18:47 davidfetter i am pretty decent at finding suc^Wvict^H^H^Holunteers
18:48 davidfetter ...then getting them what they need and keeping Others™ out of their hair
18:49 rohit_nsit08 joined #parrot
18:49 rohit_nsit08 hello #parrot
18:49 tadzik hello rohit_nsit08
18:53 dukeleto rohit_nsit08: howdy
18:53 dukeleto davidfetter: so you want to wave the PL/Parrot flag? Where do you see the future of PL/Parrot being?
18:53 whiteknight hello rohit_nsit08
18:54 rohit_nsit08 dukeleto: hi, perfectly fine :-) reading design docs of parrot
18:54 rohit_nsit08 whiteknight: hi
18:54 rohit_nsit08 tadzik: hello
18:54 davidfetter dukeleto, are you sure you won't get all pissed off at me again?
18:54 dukeleto rohit_nsit08: sounds lovely
18:55 dukeleto davidfetter: I fully recognize that I was not in a good mood when we last talked about that stuff and apologize for being so negative.
18:55 rohit_nsit08 whiteknight: checked that last post on my blog? any comments?
18:56 whiteknight rohit_nsit08: I did look at it, I haven't had time to really dig into it. I've been busy looking for a bug in parrot
18:57 rohit_nsit08 whiteknight:  ya i know, read it on the mailing list
18:57 rohit_nsit08 whiteknight: last week i did a complete test while installing Parrot on my system and all tests passed successfully
18:58 whiteknight rohit_nsit08: awesome. passing tests are good to see
18:58 rohit_nsit08 whiteknight: agree
19:00 davidfetter dukeleto, well, so, i'd like to see pl/parrot as the basis of all PLs, apart from C
19:01 davidfetter well, maybe apart from C and C++, should C++ get to be a viable alternative
19:06 dukeleto davidfetter: how do you feel about putting PL/Parrot on PGXN ?
19:26 benabik left #parrot
19:27 GodFather joined #parrot
19:41 whiteknight left #parrot
19:43 particle1 joined #parrot
19:46 particle left #parrot
19:48 rohit_nsit08 left #parrot
19:58 benabik joined #parrot
19:59 jevin left #parrot
20:00 [hercynium] joined #parrot
20:02 davidfetter dukeleto, that'd be great :)
20:05 hercynium left #parrot
20:05 [hercynium] is now known as hercynium
20:06 dukeleto davidfetter: are you volunteering? ;)
20:07 davidfetter dukeleto, heh. i'll probably pester some other guy from portland about this
20:07 davidfetter david something
20:10 dukeleto davidfetter: i am sure theory will love to hear that you are volunteering him :)
20:11 theory wtf?
20:11 theory dukeleto: I can sit down with you at a hackathon to work on it if you like.
20:11 theory But otherwise I have no tuits.
20:11 dukeleto theory: that would be cool
20:11 * theory is doing a talk on PGXN at next week's PDXPUG.
20:11 davidfetter :)
20:11 dukeleto theory: i can trade you a few beers for a few tuits
20:11 dukeleto theory: i will do my best to attend that
20:11 theory Mmmm…beer
20:11 theory dukeleto: Right on. And tell all your peeps.
20:12 theory In fact, I should adversize it on pgx.pm list.
20:12 dukeleto theory: i see it is on Calagator http://calagator.org/events/1250460492
20:13 dukeleto theory: just added it to my calendar
20:14 theory right on
20:14 davidfetter o/` i'm not an alky i just drink a lot o/`
20:15 davidfetter (with apologies to Big Pun)
20:19 dodathome left #parrot
20:32 Andy_ bacek_at_work: amything I can do to help on your branch?
20:34 [hercynium] joined #parrot
20:38 hercynium left #parrot
20:38 [hercynium] is now known as hercynium
20:47 mtk left #parrot
20:50 Eduardow left #parrot
20:55 Eduardow joined #parrot
20:59 benabik left #parrot
21:08 benabik joined #parrot
21:11 cotto_work coke++
21:12 hercynium left #parrot
21:18 ambs left #parrot
21:26 fperrad left #parrot
21:34 jevin joined #parrot
21:35 GodFather left #parrot
21:46 SHODAN left #parrot
21:51 bacek Good morning, humans.
21:54 dalek parrot/packfile_wrap: d7afc1f | bacek++ | config/gen/config_h/config_h.in:
21:54 dalek parrot/packfile_wrap: Brake the build again. Replace typedef of Parrot_PackFile with PMC version.
21:54 dalek parrot/packfile_wrap: review: https://github.com/parrot/parrot/commit/d7afc1f369
21:55 bacek msg whiteknight After fixing build of packfile_wrap (d7afc1f369) branch will be mergeable back to master.
21:55 aloha OK. I'll deliver the message.
21:55 bacek Andy_, ^^^
21:56 Andy_ ok, what do you want me to do
21:57 cotto_work bacek++
21:58 Andy_ and yes, it does break the build now
21:59 Andy_ I can't build the branch
22:00 Andy_ ./include/parrot/config.h:73:9: error: ‘Parrot_PMC’ does not name a type
22:03 jrtayloriv joined #parrot
22:04 bacek left #parrot
22:08 Andy_ left #parrot
22:09 dalek parrot/packfile_wrap: f65e92a | bacek++ | src/embed/api.c:
22:09 dalek parrot/packfile_wrap: Update code to new Parrot_PackFile typedef.
22:09 dalek parrot/packfile_wrap: review: https://github.com/parrot/parrot/commit/f65e92a536
22:10 bacek joined #parrot
22:11 bacek ~~
22:11 bacek my adsl router just died...
22:16 bacek Andy, look at f65e92a5369526a3fdf343cfde55660ad9f1a403. This is direction what have to be done.
22:17 * davidfetter wonders whether bacek has replaced the aforementioned router with TCS technology
22:18 bacek no. reboot helps. But I do want to replace it...
22:19 soh_cah_toa joined #parrot
22:27 bubaflub left #parrot
22:40 whiteknight joined #parrot
23:01 whiteknight good evening, #parrot
23:02 soh_cah_toa hey
23:04 whiteknight hello soh_cah_toa
23:05 soh_cah_toa looks like aloha finally decided to leave me alone
23:06 whiteknight yeah, bacek probably fixed it
23:10 bubaflub joined #parrot
23:38 dmalcolm left #parrot

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

Parrot | source cross referenced