Camelia, the Perl 6 bug

IRC log for #parrot, 2010-12-23

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 Yuki`N left #parrot
00:01 whiteknight I have a few in
00:02 whiteknight we need dukeleto to approve them
00:12 whiteknight Yuki'N: There are a few tasks in the queue still that you could do. Some of them should be pretty quick and easy, even if they aren't super fun
00:12 sorear gci isn't over yet?
00:13 cotto_work It ends on the 10th
00:14 whiteknight GCI is like the gift that keeps giving
00:15 whiteknight We have 76 tasks completed already
00:15 whiteknight Next year I think we need to have at least 100 tasks set up
00:15 whiteknight our slowness at creating and managing tasks has been the biggest limiting reagent here
00:16 cotto_work Bah.  Getting make cover to fail quicky with nmake is not easy.
00:19 whiteknight nmake sucks at failing
00:19 cotto_work Ah.  windows' echo treats everything as its argument, not just the quoted stuff
00:19 whiteknight s/ at failing//
00:21 cotto_work If I just put "exit 1", it stops correctly
00:24 whiteknight Who's on the January release, cotto?
00:24 cotto_work yup
00:29 whiteknight Okay
00:29 whiteknight I would definitely like to see PLATFORMS be updated this month
00:29 whiteknight and lots of smoke reports submitted
00:38 whiteknight aloha: coverage?
00:38 aloha whiteknight: coverage is http://cv.perl6.cz or http://tapir2.ro.vutbr.cz/cover/cover-results/
00:40 whiteknight how is it that Parrot_pcc_build_sig_object_from_varargs has zero test coverage?
00:41 whiteknight Parrot_pcc_parse_signature_string too. There are a few functions in src/call/args.c that I feel like either should be covered, or are probably dead code
00:42 whiteknight Parrot_pcc_allocate_empty_context is probably dead if it isn't covered
00:44 whiteknight Parrot_pcc_invoke_sub_from_c_args. That one has probably been completely replaced by Parrot_ext_call
00:44 whiteknight so many of these functions SHOULD NOT be labeled PARROT_EXPORT
00:45 bluescreen whiteknight: ping
00:45 whiteknight pong
00:45 bluescreen one question
00:45 whiteknight sure
00:45 bluescreen all example in t/src/embed do this: "GET_INIT_STRUCT(initargs);" thats calls calloc
00:46 whiteknight right
00:46 bluescreen should we give user a FREE_INIT_STRUCT macro?
00:46 whiteknight what do you mean?
00:46 bluescreen i mean, if that nevers get freed we're in troubles
00:46 cotto_work bluescreen: ping
00:46 bluescreen the other option is let the user call free himself
00:46 bluescreen cotto_work: pong
00:47 cotto_work bluescreen: can you make a note in #1899 as to which commit fixed the issue?
00:47 bluescreen sure... that was mi comment in the commit,
00:48 bluescreen cotto_work: do you mean like a comment in track saying "Commit XXXXX fixed the issue"?
00:48 cotto_work bluescreen: yes.
00:49 bluescreen cotto_work: done.. I'm used to do the opposite put the ticket in the commit message
00:49 whiteknight src/dynext.c:make_string_pmc is basically just a box operation. It's the same as the code in the box_p_s op, but it doesn't support HLL mapped types
00:52 bluescreen whiteknight: In the same way we gave users a GET_INIT_STRUCT we should give them a FREE_INIT_STRUCT.. I'd wouldn't let them use "free" directly to get rid of "initargs"
00:53 whiteknight oh, right. I see what you are saying
00:53 whiteknight yes, that probably causes a leak, doesn't it?
00:53 bluescreen yes.. well not a big problem in the case you exit immediately
00:53 bluescreen but if you are to be creating and destructing interps it can be a problem
00:53 whiteknight actually better yet, we should probably change GET_INIT_STRUCT to take a structure, not a structure pointer
00:54 whiteknight allocate it on the stack, don't use calloc
00:54 bluescreen yeah much better, and that would be the only struct we ever publish outside parrot
00:55 kennym left #parrot
00:57 whiteknight we should probably rename that macro to something better like PARROT_INIT_ARGS() or something
00:57 whiteknight i didn't put a lot of thought into that macro, obviously
00:59 bluescreen he he... I like much better PARROT_INIT_ARGS()
01:01 dalek parrot: f7512cd | cotto++ | lib/Parrot/Pmc2c/PCCMETHOD.pm:
01:01 dalek parrot: [pmc2c] remove some unused generated code
01:01 dalek parrot: review: https://github.com/parrot/parrot/commit/f7512cd690
01:04 perlite joined #parrot
01:09 macroz left #parrot
01:11 bluescreen one question: what prevents gabarge collector from destroying string, hashes you create in an embedded application ? for example in main.c (of my app ) pmc = Parrot_api_pmc_new(...);
01:12 bluescreen it won't be referenced from any other pmc... so in theory mark and sweep should sweep it, isn't it?
01:12 cotto_work either they're on the stack where the stack-walking code will find them or they're manually added to the root set
01:15 whiteknight bluescreen: the init args define a "stacktop" pointer. GC traces the system stack up until this location
01:15 whiteknight if the embedding application sets their stack top high enough, the GC will trace into embedding code
01:16 whiteknight also, there is a function somewhere that allows us to register a PMC for GC
01:17 whiteknight src/pmc.c:Parrot_pmc_gc_register and Parrot_pmc_gc_unregister
01:17 whiteknight Those might make good routines to expose through the new API
01:17 cotto_work +1
01:18 whiteknight Although I feel like it' ahuge waste to add two new functions to the API for such an arcane kind of thing
01:18 whiteknight cotto: You took a look at the new draft of PDD10. Besides the spelling/grammar fixes, what did you think about it?
01:19 cotto_work whiteknight: I'd like to read through it a couple times and mull it over a bit.  So far I've only skimmed through it.
01:19 whiteknight okay, what do you think of the style of it?
01:19 whiteknight I was trying not to descend to function-level documentation
01:22 whiteknight It mostly documents what we have in the branch now. There are very few points in there which are TODO
01:24 cotto_work I do like the overall organization.  It's approachable.
01:28 cotto_work registration is important.  Most users may not end up needing it, but for the ones that do the only alternative is to repeatedly call VTABLE_mark on the PMC manually or accept semi-random failures.
01:29 cotto_work that's my first thought on it, at any rate
01:29 sorear the alternative is to have a global Hash PMC of "tracked C globals", like Lua does
01:30 cotto_work sorear: that's basically what the root set is now
01:31 sorear is there a Parrot API that I can use to save and restore PMC* values by name?
01:31 cotto_work sorear: how would that be used?
01:31 sorear Parrot_pmc_gc_register is going to lead to problems if we want the collector to be able to move stuff
01:32 sorear cotto_work: void run_script_button() { Parrot_call_sub(interp, Parrot_get_cglobal(interp, "script")); }
01:34 Yuki`N joined #parrot
01:36 cotto_work by "save and restore", do you mean full freeze/thaw or just referencing?
01:38 cotto_work whiteknight: do you plan on keeping the limitation that interps after the first have to have a parent?
01:43 plobsing sorear: Parrot_pmc_gc_register is only a problem if the handles we provide embedders are guarranteed to be pointers. Nothing I have read makes that guarrantee.
01:44 cotto_work plobsing: interesting
01:45 sorear plobsing: for some reason I thought you were going to typedef PMC *Parrot_PMC;
01:45 sorear ha.
01:45 sorear ah.
01:45 sorear cotto_work: referencing.
01:45 cotto_work whiteknight: the docs don't specify how :immediate and friends are handled
01:47 plobsing sorear: that's the way it works now (and I agree that the current implementation won't work with a collector that moves stuff indiscriminantly), I'm just saying it doesn't have to work that way.
01:52 cotto_work whiteknight: I'm not sure how I feel about Parrot_api_disassemble_bytecode.
02:00 whiteknight cotto_work: I don't understand the limitation that interps after the first must have a parent, and I am going to ignore it
02:00 whiteknight we'll treat that like a bug if we have to
02:01 whiteknight cotto_work: :immediate is the province of the compiler. :load and :init should probably be covered by docs. I haven't wanted to discuss them because plobsing and I are trying to change them
02:01 cotto_work ok
02:01 whiteknight Parrot_api_disassemble_bytecode only exists to implement pbc_disassemble.exe
02:02 whiteknight I don't like it at all, but until we separate out that functionality into the packfiles or whatever other tool has providence over it, it has to be there
02:03 Kristaba Hi whiteknight
02:03 whiteknight hello Kristaba
02:04 Kristaba I would do some tasks for Parrot, but I'll go to see my family for a week at the other end of France (more than ten hours of car)
02:04 whiteknight ouch. 10 hours is hard
02:04 Kristaba So, have you something to do that doesn't require an internet access? :p
02:05 Kristaba Yeah, France with its small country road :D
02:05 Kristaba (never mind if it's not really a GCI task)
02:06 Kristaba Or, if you have an idea of something that can allow me to prepare a future task
02:06 whiteknight Kristaba: I don't have any off the top of my head. When do you leave?
02:07 Kristaba In 5 hours :x
02:08 Kristaba Oh, some days ago, you tell me about a C# API, or something like
02:09 whiteknight Kristaba: oh, right. http://github.com/Whiteknight/parrotsharp
02:09 whiteknight I do it in MonoDevelop, which is free
02:10 whiteknight if you want to create wrappers for a few more API functions in that, I'll make tasks for whatever you do and count it for GCI
02:12 whiteknight a lot of it is still a prototype, so you can play and experiment with things if you want
02:12 Kristaba Perfect, that will keep me busy for the road ;)
02:14 whiteknight awesome. I don't have any unit tests for that either. I need to set those up, or you can play with that if you want
02:14 whiteknight whatever you want to do. The project is young and needs everything :)
02:15 Yuki`N whiteknight, ping
02:15 Yuki`N I'm still up for some API work, if you found anything.
02:15 * sorear is curious about this "the far side of France has no internet access" thing
02:16 Yuki`N As repetitive as it sounds.
02:16 Kristaba whiteknight: Yeah, I don't promise anything, but I'll play with for sure :p
02:16 Kristaba Well, about 3h a.m here., I go bed :)
02:16 whiteknight goodnight
02:17 mtk joined #parrot
02:17 Kristaba So, good night and good Christmas everyone
02:17 Kristaba left #parrot
02:20 cotto_work whiteknight: I like the pdd10 update so far.
02:20 cotto_work nice work
02:20 whiteknight thanks. That's my christmas present to you
02:21 cotto_work whiteknight: did you see that pauseless gc article?
02:21 whiteknight cotto_work: yeah, read it enthusiastically
02:21 Yuki`N Hmmm :/
02:22 cotto_work does it sound like something Parrot could do post-Lorito?
02:22 whiteknight makes some extremely aggressive use of virtual memory, I would be *very* surprised if we could implement that in a way that works on *nix, solaris, and windows
02:22 whiteknight and I highly doubt we could use it on an embedded device like android
02:23 whiteknight (which is where we want pauseless GC in the first place)
02:23 whiteknight cotto_work: in short, I guess I really can't tell you how hard it would be, I would need to take a look at some of their code to see what all their techniques were
02:24 sorear "pauseless GC" is actually pretty easy if you don't care about constant factors
02:25 * sorear goes to find an article
02:25 plobsing Azul, the people who made the pauseless GC in question, require either a custom hypervized OS or a custom Linux kernel on which to run their VM.
02:25 whiteknight right, I highly doubt we could use it
02:26 whiteknight and even if linux got a permanent kernel improvement and that got shipped to all major distros, we still wouldn't have it on Windows or solaris
02:27 mtk left #parrot
02:28 cotto_work That's unfortunate.  Maybe if the patch makes it into Linux we could have a specialized gc to take advantage of it, but it doesn't sound like a portable solution.
02:30 cotto_work time to go home
02:31 whiteknight time for me to go to bed. See you tomorrow
02:31 whiteknight left #parrot
02:32 mtk joined #parrot
02:35 mtk left #parrot
02:41 mtk joined #parrot
02:45 kid51 joined #parrot
03:04 Yuki`N left #parrot
03:22 AzureSto_ left #parrot
03:23 cotto ~.~
03:25 AzureStone joined #parrot
03:27 AzureStone left #parrot
03:29 AzureStone joined #parrot
03:39 Coke aloha: seen mj41?
03:39 aloha Coke: mj41 was last seen in #parrot 23 days 18 hours ago saying "dukeleto: Ok. Will look at it too as fast temporary solution.".
03:40 Coke aloha: tt?
03:40 Coke aloha: taptinder?
03:40 aloha Coke: taptinder is continues integration tool - http://taptinder.org . For Parrot project running on http://tt.taptinder.org/ and reporting build failures to #parrot channel as ttbot.
03:41 Coke msg mj41 - anything I can do to get http://tt.taptinder.org/ working again?
03:41 aloha OK. I'll deliver the message.
03:50 particle left #parrot
03:52 particle joined #parrot
03:55 bluescreen left #parrot
04:06 bluescreen joined #parrot
04:09 Coke left #parrot
04:09 dalek parrot: d519bc8 | cotto++ | config/auto/coverage.pm:
04:09 dalek parrot: [configure] use $(PERL) instead of trying to figure out portable shell code for windows and *nix
04:09 dalek parrot: review: https://github.com/parrot/parrot/commit/d519bc8602
04:09 Coke joined #parrot
04:15 cotto seen fbrito
04:15 aloha fbrito was last seen in #parrot 1 days 8 hours ago leaving the channel.
04:26 * cotto accidentally a word
04:35 kid51 left #parrot
05:12 ilbot2 joined #parrot
05:12 Topic for #parrot is now Parrot 2.11.0 Released | http://parrot.org | Log: irclog.perlgeek.de/parrot/today | Onward and upward with Google Code-In | Please test rakudo with bleeding edge parrot! | merge html_cleanup and embed_api2
05:14 snarkyboojum joined #parrot
05:15 Infinoid joined #parrot
05:15 estrabd joined #parrot
05:34 rurban_ joined #parrot
05:36 rurban left #parrot
05:37 rurban_ is now known as rurban
05:43 dalek tracwiki: v7 | cotto++ | Platforms/Windows
05:43 dalek tracwiki: fix up the msvc parts a bit
05:43 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Plat​forms/Windows?version=7&action=diff
06:09 confound joined #parrot
06:29 dalek parrot: 8d253b8 | Coke++ | t/codingstd/perlcritic.t:
06:29 dalek parrot: Mention env var in POD
06:29 dalek parrot: (not just buried in docs)
06:29 dalek parrot: review: https://github.com/parrot/parrot/commit/8d253b8b33
06:29 dalek parrot: 4b96417 | Coke++ | t (2 files):
06:29 dalek parrot: Make checkdepend a real codingstd test.
06:29 dalek parrot:
06:29 dalek parrot: * eliminate Test::Differences dependency
06:29 dalek parrot: * eliminate App::Ack dependency
06:29 dalek parrot: * Rip out any code that dealt with PIR -
06:29 dalek parrot:   Results of this were always ignored anyway, and we can put it back when
06:29 dalek parrot: we have code that will pass.
06:29 dalek parrot: review: https://github.com/parrot/parrot/commit/4b9641742d
06:29 dalek parrot: d70323e | Coke++ | src/dynpmc/ (2 files):
06:29 dalek parrot: Fix last failure on t/codingstd/checkdepend.pl
06:29 dalek parrot:
06:29 dalek parrot: (I think I did this differently in a branch. This way is better.)
06:29 dalek parrot: review: https://github.com/parrot/parrot/commit/d70323eecb
06:30 dalek parrot: 08e9d21 | Coke++ | MANIFEST:
06:30 dalek parrot: track file rename
06:30 dalek parrot: review: https://github.com/parrot/parrot/commit/08e9d21e09
06:32 cotto coke++
06:41 NotFound exit 0 \
06:41 NotFound /usr/bin/perl Configure.pl --ccflags=" -fprofile-arcs -ftest-coverage" \
06:41 NotFound --linkflags="-fprofile-arcs -ftest-coverage" --ldflags="-fprofile-arcs -ftest-coverage"
06:41 NotFound /bin/sh: line 0: exit: too many arguments
06:41 NotFound http://tapir2.ro.vutbr.cz/cover/logs/last-run.txt
06:42 dalek parrot: f50b101 | Coke++ | t/codingstd/checkdepend.t:
06:42 dalek parrot: remove trailing whitespace.
06:42 dalek parrot: review: https://github.com/parrot/parrot/commit/f50b101dcf
06:42 Coke cotto: I just got sick of having to run it by hand once a week. ;)
06:43 Coke next time I will try to remember to reconfig & run codetest before committing.
06:43 Coke s/committing/pushing/
06:44 cotto NotFound, I thought I tested that.
06:44 cotto my mistake
06:46 NotFound Is a shame, I was expecting to see exporter.pmc at 100%
06:49 cotto It's an easy fix.  I'm running make cover now to make sure there isn't anything else.
06:58 bacek joined #parrot
07:01 cotto still broken
07:02 cotto hi bacek
07:02 cotto thanks for getting aloha back up
07:03 cognominal left #parrot
07:07 bacek aloha, cotto
07:34 theory left #parrot
08:23 dalek parrot: 14aa06c | cotto++ | config/gen/makefiles/root.in:
08:23 dalek parrot: [cover] get make cover working again
08:23 dalek parrot: review: https://github.com/parrot/parrot/commit/14aa06c095
08:35 mj41 Coke: Please join irc://irc.freenode.org/taptinder and look at https://github.com/mj41/TapTinder/issues
08:45 mikehh_ is now known as mikehh
08:45 mikehh opbots, names
08:57 slavorgn left #parrot
09:26 fperrad joined #parrot
09:32 rfw left #parrot
10:01 dalek parrot: a6f2157 | (Gerd Pokorra)++ | docs/project/release_manager_guide.pod:
10:01 dalek parrot: add me to take release 3.2
10:01 dalek parrot: review: https://github.com/parrot/parrot/commit/a6f215732b
11:14 bluescreen left #parrot
11:23 cognominal joined #parrot
12:07 contingencyplan left #parrot
12:19 bluescreen joined #parrot
12:25 slavorgn joined #parrot
12:38 mikehh Coke: checkdepend.t does not work before make, and has a failure after
12:42 mikehh Coke: or out it this way - it has multiple failures - Tests: 97 Failed: 95 before make, after configure and Tests: 121 Failed: 1 after make
12:42 mikehh s/out/put/
12:45 mikehh all files in include/pmc/* seem to be generated and are required by the test - it needs to go somewhere else - not make codetest which can be run after configure
13:34 rurban_ joined #parrot
13:34 whiteknight joined #parrot
13:36 rurban left #parrot
13:36 rurban_ is now known as rurban
13:38 whiteknight good morning, #parrot
13:45 smash joined #parrot
13:45 smash hello everyone
13:47 whiteknight good morning, smash
13:48 plobsing_ joined #parrot
13:50 whiteknight msg cotto: I'm ready to start the embed_api4 merge any time. Let me know if you have any questions or concerns before I push the button.
13:50 aloha OK. I'll deliver the message.
13:51 kennym joined #parrot
13:52 plobsing left #parrot
13:52 whiteknight when we get that branch merged and the coverage bot starts giving us statistics for it, I will start setting up GCI tasks to increase coverage for it
14:11 Coke mikehh: the test doesn't fail if Makefile isn't present, it skips everything.
14:11 Coke That said, feel free to move it to another directory.
14:12 Coke (I don't see any targets that are appropriate.)
14:13 Coke any suggestions?
14:15 mikehh Coke: don't know - have to think about that - Configure generates the makefile, but you have to make to generate some of the dependencies
14:17 mikehh in testing I usually run Configure with --test, make manifest_tests, make codetests, make corevm/make coretest, make world/make test, then fulltest
14:19 mikehh all after a make realclean and git pull --rebase
14:20 mikehh maybe in the t/src directory
14:22 JimmyZ joined #parrot
14:25 JimmyZ good evening, parrot
14:27 mikehh hi JimmyZ
14:27 JimmyZ hello mikehh
14:28 mikehh JimmyZ: how's things in your part of the world?
14:30 JimmyZ everything goes well
14:32 JimmyZ except looking for a job
14:33 JimmyZ GCI work for chinese students may be unpractical.
14:47 JimmyZ left #parrot
15:25 Matt_ joined #parrot
15:41 jan joined #parrot
15:50 dalek TT #1901 created by Matt221++: Parrot_io_get_last_file_position should be removed in the next deprecation ...
15:50 dalek TT #1901: http://trac.parrot.org/parrot/ticket/1901
15:51 whiteknight I strongly suspected that function was not being used
15:52 Matt_ whiteknight: Submitted the task for review :)
15:53 Matt_ whiteknight: https://github.com/parrot/parrot/pull/38
15:54 whiteknight Matt_: awesome. Let me pull it and I'll close it
15:54 Matt_ thanks!
16:06 bluescreen left #parrot
16:13 plobsing_ aloha: msg NotFound assign (=:) is an unacceptable substitute for real closure variables. Any chance I could use unary * for fetch/push update? Or get 'volatile' for the generation of pessimistic code?
16:13 aloha plobsing_: OK. I'll deliver the message.
16:13 aloha plobsing_: Okay.
16:14 NotFound plobsing_: unary * ?
16:14 plobsing_ *x; // pointer-ish syntax
16:14 NotFound I see... let me think about it...
16:21 Matt- joined #parrot
16:26 Matt_ left #parrot
16:33 Matt- left #parrot
16:33 Matt_ joined #parrot
16:34 NotFound plobsing_: volatile seems to be a cleaner approach, I'll look at that way.
16:35 plobsing_ thanks
16:36 dalek parrot/gci_deprecate_io_position: 8b88b11 | (Matt Rajca)++ | / (3 files):
16:36 dalek parrot/gci_deprecate_io_position: Removed Parrot_io_get_last_file_position for deprecation
16:36 dalek parrot/gci_deprecate_io_position: review: https://github.com/parrot/parrot/commit/8b88b11fa7
16:39 whiteknight Matt_: pulled. I've closed the task. Thanks
16:44 Matt_ whiteknight: Ill take this too: http://www.google-melange.com/gci/task/show/google​/gci2010/parrot_perl_foundations/t129295777080#c1
16:45 dmalcolm joined #parrot
16:53 whiteknight Matt_: approved. Good luck!
17:09 whiteknight left #parrot
17:14 cotto msg whiteknight Merge whenever you're ready.
17:14 aloha OK. I'll deliver the message.
17:32 theory joined #parrot
17:33 cotto_work ~~
17:37 dalek ohm-eta-wink-kzd: 2839fce | plobsing++ | src/Makefile:
17:37 dalek ohm-eta-wink-kzd: add rules for generating pbc
17:37 dalek ohm-eta-wink-kzd: review: https://github.com/plobsing/ohm​-eta-wink-kzd/commit/2839fce39e
17:37 dalek ohm-eta-wink-kzd: a773923 | plobsing++ | src/ometac.winxed:
17:37 dalek ohm-eta-wink-kzd: add stage1 winxed compiler
17:37 dalek ohm-eta-wink-kzd: review: https://github.com/plobsing/ohm​-eta-wink-kzd/commit/a7739234ee
17:37 dalek ohm-eta-wink-kzd: 9087d9f | plobsing++ | src/ (2 files):
17:37 dalek ohm-eta-wink-kzd: change build to accomodate stage1 compiler
17:37 dalek ohm-eta-wink-kzd: review: https://github.com/plobsing/ohm​-eta-wink-kzd/commit/9087d9feb5
17:39 ambs joined #parrot
17:48 dalek TT #1902 created by Matt221++: Parrot_freeze_size should be removed for deprecation
17:48 dalek TT #1902: http://trac.parrot.org/parrot/ticket/1902
17:57 zaur joined #parrot
18:04 whiteknight joined #parrot
18:05 cotto_work hio whiteknight
18:05 whiteknight hey cotto_work
18:06 whiteknight you're happy with the embed_api4 branch?
18:06 cotto_work yes
18:06 whiteknight okay then. I'll merge soon
18:07 cotto_work whiteknight: do you see any value in including Visual Studio instructions in Platforms/Windows?  All that's necessary is the Windows SDK.
18:07 particle1 left #parrot
18:08 cotto_work http://trac.parrot.org/par​rot/wiki/Platforms/Windows
18:08 whiteknight cotto_work: I was looking at that page yesterday too. No, VisualStudio isn't necessary but it is a way to get the SDK
18:08 cotto_work are the instructions worth keeping or can we assume that anyone who has visual studio will know how to use it?
18:09 cotto_work I'd like to remove any unnecessary instructions from the page.  There's more there now than I think is necessary.
18:10 whiteknight are you on windows?
18:10 whiteknight for any of your developing?
18:10 particle joined #parrot
18:10 cotto_work I have a windows machine set up with the msvc toolchain and I'm going to get the mingw toolchain working later today.
18:10 cotto_work It's not as painful as I was hoping.
18:10 cotto_work er, afraid of
18:11 whiteknight oh, okay. I have the msvc toolchain here at work
18:11 whiteknight I don't see any reason to talk about VisualStudio except to mention that it's a vector to obtain the SDK
18:11 cotto_work wfm
18:14 whiteknight Matt_: I'm running a merge and tests here locally. I'll get your pull request after that
18:15 whiteknight msg plobsing Can you take a quick look at #1902? A GCI student has identified it as dead code, and I want a second set of eyes on it before we move to deprecate/delete
18:15 aloha OK. I'll deliver the message.
18:16 whiteknight we wouldn't need to worry about deprecation at all if functions like these weren't marked PARROT_EXPORT
18:17 whiteknight stupid PARROT_EXPORT
18:18 whiteknight I would like to put in a blanket deprecation notice for a cull of PARROT_EXPORT functions that are not used, not usable, or that simply should not be exported
18:18 whiteknight We have too damn many functions being exported
18:24 plobsing_ whiteknight: done
18:32 plobsing_ as I understood it, PARROT_EXPORT didn't mark something as "part of the api", it just made it visible for parrot core tools (which were special at the time).
18:34 fperrad left #parrot
18:36 fperrad joined #parrot
18:38 whiteknight plobsing_: I like that interpretation of things much more than the alternative
18:39 cotto_work Note that "PARROT_API" is spelled differently from "PARROT_EXPORT".
18:41 plobsing_ also, IIRC, PARROT_EXPORT is sometimes needed to make the windows build work for some reason.
18:42 plobsing_ and then there's the issue of ops libraries and C++ forcing me to extern a symbol
18:42 cotto_work I have a working window build environment now, so we can test what breaks it pretty easily.
18:44 Matt- joined #parrot
18:46 cotto_work whiteknight: are you set up to build on Windows?
18:50 Matt_ left #parrot
18:50 Matt- left #parrot
18:50 Matt_ joined #parrot
18:53 dd070 joined #parrot
18:57 dalek tracwiki: v8 | cotto++ | Platforms/Windows
18:57 dalek tracwiki: rework instructions on getting the mscv and mingw environments set up
18:57 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Plat​forms/Windows?version=8&action=diff
19:04 lucian joined #parrot
19:06 nwellnhof joined #parrot
19:07 nwellnhof cotto: why did you remove the instructions for Microsoft Visual C++ 2010 Express from the wiki page?
19:08 cotto_work nwellnhof: they're not necessary.  The Windows SDK includes the compiler.
19:09 cotto_work I want it to be as straightforward as possible to start hacking on Parrot on windows.
19:10 cotto_work nwellnhof: does that strike you as sensible?
19:11 nwellnhof yes, if the SDK contains everything needed. the Visual C++ Express install might be a bit leaner, though.
19:11 cotto_work I suspect that the SDK is lighter.
19:11 cotto_work not certain though
19:12 nwellnhof the SDK seems to contain the .net stuff, too.
19:12 cotto_work I had to install it separately.
19:13 nwellnhof but Visual C++ Express installs all kinds of other stuff, too.
19:15 ambs left #parrot
19:17 whiteknight Visual C++ Express definitely contains the .net stuff
19:19 kennym hi, who's in charge of this task? http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129031964617
19:20 cotto_work kennym: any Parrot mentor can deal with any task
19:27 Patterner left #parrot
19:27 theory left #parrot
19:37 Psyche^ joined #parrot
19:37 Psyche^ is now known as Patterner
19:38 Khisanth left #parrot
19:56 lucian https://github.com/appcelerator/kroll similar to mozilla's XPCOM i think
19:56 * lucian has to go
19:59 contingencyplan joined #parrot
19:59 lucian left #parrot
20:05 sorear left #parrot
20:12 sorear joined #parrot
20:13 he_ joined #parrot
20:14 he__ left #parrot
20:14 Kapace_ left #parrot
20:22 simcop2387_ joined #parrot
20:22 whiteknight has somebody played with dependencies or the makefile recently or anything?
20:22 whiteknight I'm getting a lot of failures in t/codingstd/checkdepend.t here locally where I'm merging embed_api4 into master
20:23 whiteknight and a lot of these are files I that have not been changed in the branch
20:23 cotto_work that test was only added yesterday
20:24 simcop2387_ left #parrot
20:24 whiteknight okay, can I ignore it for now?
20:25 cotto_work probably good to fix what can be fixed
20:32 dalek ohm-eta-wink-kzd: 24b594b | plobsing++ | src/Ωη.coda.winxed:
20:32 dalek ohm-eta-wink-kzd: re-open classes the hard way (winxed does not support this with syntactic sugar yet)
20:32 dalek ohm-eta-wink-kzd: review: https://github.com/plobsing/ohm​-eta-wink-kzd/commit/24b594b17b
20:32 dalek ohm-eta-wink-kzd: f135bf8 | plobsing++ | src/Makefile:
20:32 dalek ohm-eta-wink-kzd: track more generated files
20:32 dalek ohm-eta-wink-kzd: review: https://github.com/plobsing/ohm​-eta-wink-kzd/commit/f135bf87b0
20:32 dalek ohm-eta-wink-kzd: d50db48 | plobsing++ | src/ometa-base.winxed:
20:32 dalek ohm-eta-wink-kzd: use quotes to avoid pseudo-constructor
20:32 dalek ohm-eta-wink-kzd: review: https://github.com/plobsing/ohm​-eta-wink-kzd/commit/d50db483d5
20:33 dalek ohm-eta-wink-kzd: 2ec1b38 | plobsing++ | src/ometa-base.winxed:
20:33 dalek ohm-eta-wink-kzd: lookup class for instantiation - String.new doesn't work
20:33 dalek ohm-eta-wink-kzd: review: https://github.com/plobsing/ohm​-eta-wink-kzd/commit/2ec1b38531
20:35 whiteknight cotto_work: is that test passing in master?
20:36 cotto_work whiteknight: no.  there's one failure
20:38 bluescreen joined #parrot
20:38 rfw joined #parrot
20:42 whiteknight blah.
20:43 whiteknight well, I have no idea why this test is failing, or how to fix it
20:44 whiteknight merge is going to have to wait till tomorrow, at least
20:49 dalek parrot/embed_api5: 4e02ead | Whiteknight++ | / (53 files):
20:49 dalek parrot/embed_api5: fix some conflicts in branch merge
20:49 dalek parrot/embed_api5: review: https://github.com/parrot/parrot/commit/4e02ead2d0
20:49 dalek parrot/embed_api5: c92d340 | Whiteknight++ | / (3 files):
20:49 dalek parrot/embed_api5: fix a few codetest failures
20:49 dalek parrot/embed_api5: review: https://github.com/parrot/parrot/commit/c92d3403ae
20:49 dalek parrot/embed_api5: 95a4a78 | Whiteknight++ | src/embed/pmc.c:
20:49 dalek parrot/embed_api5: fix more codestd problems
20:49 dalek parrot/embed_api5: review: https://github.com/parrot/parrot/commit/95a4a785ac
20:55 cotto_work I'll take a look at it.
20:56 whiteknight thanks. I can't make heads nor tails of it
20:57 Matt_ whiteknight: When you get a chance to, please review: https://github.com/parrot/parrot/pull/39 plobsing_ already approved the change on the trac ticket page
20:58 bluescreen left #parrot
20:59 dalek parrot: 8780440 | plobsing++ | config/gen/makefiles/root.in:
20:59 dalek parrot: fix checkdepend.t
20:59 dalek parrot: review: https://github.com/parrot/parrot/commit/8780440eb0
21:00 plobsing_ why does embed_api keep incrementing?
21:01 whiteknight plobsing: because I keep merging from master, and I don't want to merge in-place
21:02 plobsing_ why not?
21:02 plobsing_ is now known as plobsing
21:02 dalek parrot/gci_removed_freeze_size: 0aee73a | (Matt Rajca)++ | / (2 files):
21:02 dalek parrot/gci_removed_freeze_size: Removed Parrot_freeze_size for deprecation
21:02 dalek parrot/gci_removed_freeze_size: review: https://github.com/parrot/parrot/commit/0aee73a125
21:02 whiteknight Matt_: done
21:03 plobsing it's not like you are rewriting history or anything.
21:03 whiteknight plobsing_: is it a problem that I'm incrementing it?
21:03 cotto_work It's a curious git workflow.
21:03 plobsing not really a technical problem, no
21:04 plobsing it does make it harder to keep track of what the latest embed_api branch is
21:04 whiteknight I'm experimenting with it
21:04 plobsing that and you're not cleaning up your defunct branches
21:05 whiteknight I've cleaned up some. I wasn't deleting branches that people said they still had outstanding commits on
21:05 dalek parrot: 1819260 | cotto++ | / (7 files):
21:05 dalek parrot: various minor fixes to make msvc less sad
21:05 dalek parrot: review: https://github.com/parrot/parrot/commit/1819260aeb
21:05 dalek parrot: 658e577 | cotto++ | src/pmc/orderedhash.pmc:
21:05 dalek parrot: [pmc] remove some unneeded code from OrderedHash
21:05 dalek parrot: review: https://github.com/parrot/parrot/commit/658e57727e
21:07 atrodo "less sad"  Brilliant
21:07 plobsing whiteknight: you can delete branches on which people have unpushed changes. they just have to merge those into the latest branch (which they have to do anyways).
21:08 plobsing and it gives a good indication of "hey, you're working on a defunct branch. stop doing that"
21:12 rfw hey whiteknight
21:12 rfw or whoever
21:12 rfw http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129295757244 <-- looking to claim
21:12 plobsing is now known as whoever
21:12 whoever hi rfw!
21:12 whoever is now known as plobsing
21:12 rfw hi whoever, formerly known as plobsing!
21:13 rfw where are the namepsace tests, by the way
21:14 cotto_work t/pmc/namespace.t
21:14 rfw ah got it
21:14 rfw thanks
21:14 plobsing > find t | grep namespace # there's a couple of places
21:21 Matt_ whiteknight: can you close the task: http://www.google-melange.com/gci/​task/show/google/gci2010/parrot_pe​rl_foundations/t129295777080#c3002
21:21 cotto_work I wish melange didn't have one url that worked and another that didn't.
21:22 whiteknight rfw: Accepted. Good luck
21:22 rfw whiteknight: thanks
21:22 whiteknight Matt_: Accepted and closed. Thanks
21:22 cotto_work http://notanumber.net/archives/54/​underhanded-c-the-leaky-redaction
21:23 simcop2387 left #parrot
21:23 rfw whiteknight: i don't quite understand what make_namespace_autobase does
21:23 rfw i don't think it's deprecated though, since i can see it in the code
21:25 whiteknight rfw: you see it used somewhere? that's disappointing
21:25 rfw whiteknight: yeah
21:25 whiteknight rfw: Where is it used from?
21:25 rfw src/pmc/role.pmc:            _namespace = Parrot_ns_make_namespace_autobase(interp, _namespace);
21:25 rfw src/pmc/role.pmc:            _namespace = Parrot_ns_make_namespace_autobase(interp, _namespace);
21:25 whiteknight stupid role PMC
21:25 whiteknight rfw: Add tests to t/pmc/role.t to try and execute that function
21:26 rfw ah okay
21:26 whiteknight doesn't matter what it does, just get it covered for now
21:26 whiteknight (understanding it would be good too, but for now I want it to be covered at least a little)
21:26 whiteknight once we know that we are properly exercising it, we can fix it, replace it, delete it, whatever and know that we aren't breaking anything
21:27 rfw okay
21:28 dalek parrot/embed_api5: 8780440 | plobsing++ | config/gen/makefiles/root.in:
21:28 dalek parrot/embed_api5: fix checkdepend.t
21:28 dalek parrot/embed_api5: review: https://github.com/parrot/parrot/commit/8780440eb0
21:28 dalek parrot/embed_api5: 1819260 | cotto++ | / (7 files):
21:28 dalek parrot/embed_api5: various minor fixes to make msvc less sad
21:28 dalek parrot/embed_api5: review: https://github.com/parrot/parrot/commit/1819260aeb
21:28 dalek parrot/embed_api5: 658e577 | cotto++ | src/pmc/orderedhash.pmc:
21:28 dalek parrot/embed_api5: [pmc] remove some unneeded code from OrderedHash
21:28 dalek parrot/embed_api5: review: https://github.com/parrot/parrot/commit/658e57727e
21:28 dalek parrot/embed_api5: 1a76c9b | Whiteknight++ | / (9 files):
21:28 dalek parrot/embed_api5: Merge branch 'master' of github.com:parrot/parrot
21:28 dalek parrot/embed_api5: review: https://github.com/parrot/parrot/commit/1a76c9b907
21:29 whiteknight checkdepend.t is still broken on that branch
21:29 rfw aahhhhh oh god what
21:29 rfw parrot didn't compile
21:29 rfw error: Parrot_DynOp_core_2_10_1 undeclared
21:30 rfw whiteknight: could you take a look at that
21:30 plobsing rfw: you need to reconfigure after parrot revision number changes
21:30 rfw plobsing: i did
21:30 plobsing make realclean?
21:31 rfw i did make clean before
21:31 cotto_work whiteknight: I'd say go ahead and merge.  I'm getting inconsistent results with checkdepends and I'm not entirely convinced that your code is wrong.
21:31 cotto_work or just use plobsing's fix
21:31 rfw oh works now
21:31 rfw probably not clean enough
21:32 whiteknight cotto_work: too late now. I'm heading home. Will do it tomorrow
21:32 whiteknight later
21:32 whiteknight left #parrot
21:32 cotto_work wfm
21:33 dd070 left #parrot
21:34 simcop2387 joined #parrot
21:34 rurban_ joined #parrot
21:34 dmalcolm left #parrot
21:36 rurban left #parrot
21:37 rurban_ is now known as rurban
21:37 rfw wh
21:37 rfw argh
21:37 cotto_work whargh
21:42 cotto_work anyone think it'd be a bad idea for me to merge embed_api5?
21:45 rfw yay
21:45 rfw managed to cover autobase
21:47 smash left #parrot
21:49 rfw anyone available: https://github.com/parrot/parrot/pull/40 and http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129295757244
21:50 plobsing cotto_work: I was going to attempt to fix all the checkdepends tests to allow for just that. If you are willing to let those tests slide, all the better.
21:53 jan_ joined #parrot
21:54 plobsing vtable_set_pmc on FPA can cause it to change size. that seems wrong to me.
21:55 cotto_work merginated
21:55 dalek parrot: 1a76c9b | Whiteknight++ | / (9 files):
21:55 dalek parrot: Merge branch 'master' of github.com:parrot/parrot
21:55 dalek parrot: review: https://github.com/parrot/parrot/commit/1a76c9b907
21:56 jan left #parrot
21:56 jan_ is now known as jan
22:07 cotto_work msg whiteknight I went ahead and merged embed_api5.  If you're happy, please delete the old embed_api* branches.
22:07 aloha OK. I'll deliver the message.
22:08 Matt_ I'm getting a couldn't read 'src/extend_vtable.c': No such file or directory at lib/Parrot/Headerizer/Functions.pm line 124. error when running make headerizer. Is this a known issue?
22:08 dalek parrot: 99ba321 | plobsing++ | lib/Parrot/Distribution.pm:
22:08 dalek parrot: remove some references to files that no longer exist
22:08 dalek parrot: review: https://github.com/parrot/parrot/commit/99ba321a6e
22:14 plobsing Matt_: (1) I don't have that issue. (2) that file *should* exist.
22:14 plobsing you should run 'make' before running the headerizer?
22:14 plobsing s/you/maybe you/
22:15 rfw could someone look over my role.t test case to cover make_namespace_autobase?
22:23 cotto_work rfw: merging
22:23 rfw cotto_work: thanks
22:23 rfw could you approve the task too
22:24 cotto_work sure
22:26 dalek parrot: c8e0332 | (Tony Young)++ | t/pmc/role.t:
22:26 dalek parrot: Added a test in role.t to cover make_namespace_autobase.
22:26 dalek parrot: review: https://github.com/parrot/parrot/commit/c8e033205a
22:28 rfw got it, thanks
22:37 Khisanth joined #parrot
22:55 nwellnhof msg moritz thanks for applying my Rakudo patches
22:55 aloha OK. I'll deliver the message.
22:57 kid51 joined #parrot
22:57 fperrad left #parrot
23:00 contingencyplan left #parrot
23:05 kid51 If I "nice" a command-line program, and if that program calls several other programs (e.g., 'make' 'make test') are those other programs "nice-d" as well?
23:06 kid51 cotto_work: ping
23:06 Khisanth left #parrot
23:06 cotto_work kid51: pong
23:06 kid51 The conditions you msg-ed me about -- they're still broken?
23:07 cotto_work kid51: can you remind me what I msg'd you about?
23:07 kid51 make cover not failing gracefully when Devel::Cover not installed
23:07 cotto_work ah
23:07 cotto_work I think that's fixed now.
23:08 kid51 Yeah, I saw some commits to that today.  I will review.
23:08 cotto_work It should also be portable.
23:08 kid51 (I expected problems like that.)
23:09 cotto_work It's a bit hacky.  If you can think of a more elegant way to fail quickly I'm open to it.
23:09 nwellnhof kid51: regarading your "nice" question: yes, i tihnk so.
23:09 kid51 nwellnhof:  Thanks.
23:10 kid51 cotto_work:  Yes, I didn't know how to do an IF block in a Makefile.   (and don't, really)
23:12 Matt_ cotto_work: Does a Parrot_api_pmc_box_string exist in src/embed/api.c ? I grep'ed all the embed_api* branches and got no results yet:
23:12 Matt_ "Update the function Parrot_api_pmc_box_string in src/embed/api.c to use your new function"
23:14 cotto_work Matt_: src/embed/pmc.c
23:16 Matt_ cotto_work: thanks!
23:27 Khisanth joined #parrot
23:29 Matt_ cotto_work: Can you do a quick review of https://github.com/parrot/parrot/pull/41
23:30 Matt_ and https://github.com/parrot/parrot/pull/42 (on the embed_api2 branch)
23:31 cotto_work Matt_: I just merged embed_api5
23:33 Khisanth left #parrot
23:41 Yuki`N joined #parrot
23:41 Matt_ cotto_work: what do you want me to do?
23:48 cotto_work Matt_: you don't need to do anything.  I
23:48 cotto_work 'll just apply the diffs manually.
23:50 Yuki`N dukeleto, ping
23:57 nopaste "kid51" at 192.168.1.3 pasted "[codingstd] checkdepends: Possible solution" (112 lines) at http://nopaste.snit.ch/27430
23:59 NotFound plobsing: ping
23:59 kid51 Coke, whiteknight:  That patch may solve that codingstd failure.  With it I get a PASS on 'make test' and 'make codetest'
23:59 cotto_work kid51: go for it if it doesn't break anything
23:59 dalek winxed: r722 | NotFound++ | trunk/winxedst1.winxed:
23:59 dalek winxed: initial support for volatile in lexical var
23:59 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=722
23:59 kid51 Cf:  http://smolder.parrot.org/app​/projects/report_details/1833

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

Parrot | source cross referenced